Drinking Water Regulation and Health

DRINKING WATER REGULATION AND HEALTH

FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

A JOHN WILEY & SONS, INC. PUBLICATION

DRINKING WATER REGULATION AND HEALTH

DRINKING WATER REGULATION AND HEALTH

FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

A JOHN WILEY & SONS, INC. PUBLICATION

The reader should not rely on this publication to address specific questions that apply to a particular set of facts. The authors and publisher make no representation or warranty, express or implied, as to the completeness, correctness or utility of the information in this publication. In addition, the authors and publisher assume no liability of any kind whatsoever resulting from the use of or reliance upon the contents of this book.

1 This book is printed on acid-free paper.


Copyright # 2003 by John Wiley & Sons, Inc. All rights reserved. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4744. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 605 Third Avenue, New York, NY 10158-0012, (212) 850-6011, fax (212) 850-6008, E-Mail: [email protected]. For ordering and customer service information please call 1-800-CALL-WILEY. Library of Congress Cataloging-in-Publication Data: Pontius, Frederick W. Drinking water regulation and health = Frederick W. Pontius. p. cm. Includes bibliographical references and index. ISBN 0-471-41554-5 (cloth) 1. Drinking water—Law and legislation—United States. 2. United States. Safe Drinking Water Act, I. Title. KF3794.P658 2003 346.73040 69122–dc21 2003006645 Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

CONTENTS PREFACE ACKNOWLEDGMENTS CONTRIBUTORS ACRONYMS

PART I

1

THE SAFE DRINKING WATER ACT AND PUBLIC HEALTH

Drinking Water and Public Health Protection

xix xxi xxiii xxvii

1 3

Daniel A. Okun

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11

2

Introduction, 3 Water Supply for the City of Rome, 4 The Middle Ages and the Industrial Revolution, 5 The Great Sanitary Awakening, 6 The Emergence of Water as a Public Health Issue, 9 The Beginning of Water Treatment, 11 The Chemical Revolution, 13 The Introduction of Regulations, 14 Prelude to the 1974 Safe Drinking Water Act, 17 Drinking Water in Developing Countries, 19 The Future of Public Water Supply, 21

Improving Waterborne Disease Surveillance

25

Floyd J. Frost, Rebecca L. Calderon and Gunther F. Craun

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8

Introduction, 25 Background, 26 Limitations of the Current Disease Surveillance Systems, 28 Early Detection of Outbreaks, 31 Endemic Disease, 32 Applicability of Outbreak Investigations, 34 Monitoring Infection Versus Disease, 36 Improving Disease Surveillance, 38 v

vi

3

CONTENTS

Waterborne Outbreaks in the United States, 1971–2000

45

Gunther F. Craun, Rebecca L. Calderon, and Michael F. Craun

3.1 3.2 3.3

3.4

3.5

3.6

4

Introduction, 45 Waterborne Disease Outbreak Surveillance System, 46 Waterborne Outbreak Statistics, 48 3.3.1 Type of Water System, 48 3.3.2 Type of Water Source, 51 3.3.3 Outbreak Etiologies, 53 3.3.4 Severity of Illness, 55 Causes of Outbreaks in Drinking Water Systems, 55 3.4.1 Etiology of Drinking Water Outbreaks, 55 3.4.2 Water System Deficiencies, 58 3.4.3 Water Quality During Outbreaks, 59 Outbreaks Associated with Recreational Waters, 61 3.5.1 Lakes, 61 3.5.2 Pools, 61 3.5.3 Recreational Outbreaks Reported Since 1991, 63 Outbreak Trends, 65

History of the Safe Drinking Water Act (SDWA) Frederick W. Pontius

4.1 4.2 4.3

4.4 4.5 4.6

4.7 4.8

Introduction, 71 Early Development of Drinking Water Standards, 72 The Safe Drinking Water Act of 1974, 73 4.3.1 The National Interim Primary Drinking Water Regulations, 75 4.3.2 National Academy of Sciences (NAS) Study, 77 4.3.3 1977–1980 SDWA Amendments, 77 1986 SDWA Amendments, 79 1988 Lead Contamination Control Act, 80 1996 SDWA Amendments, 81 4.6.1 Reauthorization Issues Emerge, 81 4.6.2 GAO Studies Note Deficiencies, 82 4.6.3 102nd Congress, 83 4.6.4 103rd Congress, 84 4.6.5 USEPA Redirection of Regulatory Priorities, 88 4.6.6 104th Congress Activity, 90 Public Health Security and Bioterrorism Preparedness and Response Act, 91 Future Outlook, 95

71

CONTENTS

5

SDWA: Looking to the Future

vii

105

Diane VanDe Hei and Thomas Schaeffer

5.1 5.2

5.3

5.4

5.5

5.6

PART II 6

Introduction, 105 U.S. Governmental Structure, 105 5.2.1 The Executive Branch, 106 5.2.2 The Legislative Branch, 106 5.2.3 The Judicial Branch, 106 How Laws Are Made, 107 5.3.1 How Legislation Originates, 107 5.3.2 The Committee–Subcommittee Process, 108 5.3.3 Floor Action on Bills, 109 5.3.4 The Conference Committee Process, 109 5.3.5 Final Passage, Approval, and Publication, 110 5.3.6 Authorization and Appropriation Measures, 110 Forces Shaping the SDWA and Amendments, 111 5.4.1 The Setting for the 1974 SDWA, 111 5.4.2 The Setting for the 1986 Amendments, 114 5.4.3 The Setting for the 1996 Amendments, 116 5.4.4 The Setting for the 2002 Amendments, 121 Future Amendments to the SDWA, 121 5.5.1 Political Dimension, 121 5.5.2 Social Dimension, 122 5.5.3 Scientific Dimension, 123 5.5.4 Unresolved Issues, 124 5.5.5 Emerging Issues, 125 Outlook for Major Change, 127

REGULATION DEVELOPMENT

Toxicological Basis for Drinking Water Risk Assessment Joyce Morrissey Donohue and Jennifer Orme-Zavaleta

6.1 6.2

6.3

6.4 6.5

Introduction, 133 Toxicological Evaluation of Drinking Water Contaminants, 133 6.2.1 Human Studies, 136 6.2.2 Animal Studies, 137 Use of Toxicity Information in Risk Assessment, 137 6.3.1 Cancer Risk Guidelines, 138 6.3.2 Effects Other than Cancer, 139 6.3.3 Maximum Contaminant Level Goal (MCLG), 141 Health Advisories, 143 Future Outlook, 145

131 133

viii

7

CONTENTS

Epidemiologic Concepts for Interpreting Findings in Studies of Drinking Water Exposures

147

Gunther F. Craun, Rebecca L. Calderon and Floyd J. Frost

7.1 7.2 7.3 7.4 7.5 7.6

7.7

7.8 8

Introduction, 147 What Is Epidemiology?, 149 Historical Origins, 149 Disease Models, 150 Basic Measures of Disease Frequency, 152 Types of Epidemiologic Studies, 156 7.6.1 Ecological Studies, 158 7.6.2 Time-Series Analyses, 161 7.6.3 Random and Systematic Error, 162 7.6.4 Measures of Association, 167 7.6.5 Strength of Association, 168 7.6.6 Causality of an Association, 168 7.6.7 Meta analysis, 169 Examples: Experimental, Cohort, and Case–Control Studies, 170 7.7.1 Experimental Studies, 170 7.7.2 Cohort Studies, 172 7.7.3 Case–Control Studies, 174 Future Trends in Epidemiology and Drinking Water, 178

Application of Risk Assessments in Crafting Drinking Water Regulations

183

Bruce A. Macler

8.1 8.2 8.3 8.4

8.5 9

Introduction, 183 Risk Assessment Approaches for Drinking Water Regulations, 184 Risk Mandates from the Safe Drinking Water Act, 188 Developing MCLs and Treatment Techniques, 189 8.4.1 Maximum Contaminant Level Goals, 189 8.4.2 Identifying Candidate MCLs, 191 8.4.3 Health Risk Reduction and Cost Analysis, 193 8.4.4 Risk Assessments as Regulations, 193 8.4.5 Regulatory Reviews of NPDWRs, 194 Future Outlook, 195

‘‘Sound’’ Science and Drinking Water Regulation Frederick W. Pontius

9.1 9.2

Introduction, 197 Elements of ‘‘Sound’’ Science, 198 9.2.1 Objectivity, 199 9.2.2 Reason and Truth Claims, 199

197

CONTENTS

9.3 9.4 9.5 9.6 9.7 9.8

9.9 10

9.2.3 Clarity, 204 9.2.4 Critical Thinking, 205 Peer Involvement, 206 Scientific Disagreement, 209 ‘‘Junk’’ Science, 210 Causation and Causal Inference, 211 Science and SDWA Regulations, 214 Science and the Courts, 215 9.8.1 Judicial Review, 215 9.8.2 The Judicial Review Process, 216 9.8.3 Deference, 218 9.8.4 Example: Chloroform MCLG, 219 Future Developments and Trends, 221

Benefit–Cost Analysis and Drinking Water Regulation Robert S. Raucher

10.1 10.2 10.3 10.4 10.5

ix

Introduction, 225 Benefit–Cost Analysis (BCA) Under the SDWA, 226 Historical Application of BCA, 227 USEPA Policies and Practices, 228 Comparing Benefits to Costs, 229 10.5.1 Maximizing Net Benefits, 229 10.5.2 Incremental Benefits and Costs, 230 10.5.3 Accounting for System Size, 231 10.6 Measures of Risk Reduction Benefits, 233 10.6.1 Quantifying Risk Reduction Benefits, 233 10.6.2 Quality-Adjusted Life Years, 234 10.6.3 Valuing Risk Reduction Benefits, 235 10.6.4 Willingness to Pay: The Value of a Statistical Life, 236 10.6.5 Cost of Illness, 237 10.7 Benefits Transfer to Drinking Water, 238 10.7.1 Adjusting VSL, 239 10.7.2 Accounting for Latencies, 239 10.7.3 Discounting Costs and Benefits, 240 10.7.4 Adjusted VSLs to Reflect Latency, Discounting, and Income Growth, 241 10.8 Uncertainty and Variability, 242 10.8.1 What are Uncertainty and Variability?, 242 10.8.2 Addressing Uncertainties and Variabilities, 243 10.9 Precautionary Assumptions versus Central Tendencies, 244 10.10 Omitted or Unquantified Benefits and Costs, 246 10.11 Uncertain Costs, 247 10.12 Future Outlook, 247

225

x

11

CONTENTS

Public Involvement in Regulation Development

251

Frederick W. Pontius

11.1 11.2 11.3 11.4

11.5

11.6 11.7 11.8 11.9 PART III 12

Introduction, 251 Who is the Public?, 251 Objectives Determine Involvement Level, 252 Involvement during the Rulemaking Process, 253 11.4.1 Involvement Prior to Rule Proposal, 258 11.4.2 Involvement during Rule Proposal, 259 11.4.3 Involvement after Rule Proposal, 259 11.4.4 Ex Parte Communications, 260 Federal Agency Advisory Committees, 261 11.5.1 National Drinking Water Advisory Council (NDWAC), 264 11.5.2 USEPA Science Advisory Board, 265 Regulatory Negotiation, 266 Judicial Review, 268 USEPA’s Public Involvement Policy, 269 The Future of Public Participation, 271 CONTAMINANT REGULATION AND TREATMENT

Control of Drinking Water Pathogens and Disinfection Byproducts Stig E. Regli, Paul S. Berger and Thomas R. Grubbs

12.1 12.2 12.3

Introduction, 277 Control of Waterborne Pathogens Before the 1970s, 277 Control of Waterborne Pathogens and DBPs in the 1970s, 280 12.3.1 Total Coliform Rule (TCR), 281 12.3.2 Turbidity and Heterotrophic Bacteria, 282 12.3.3 Trihalomethanes (THMs), 283 12.4 Control of Waterborne Pathogens and DBPs in the 1980s, 284 12.4.1 Revised Total Coliform Rule, 285 12.4.2 Surface Water Treatment Rule (SWTR), 286 12.5 Control of Waterborne Pathogens and DBPs in the 1990s and Beyond, 289 12.5.1 1996 SDWA Amendments for Pathogen and DBP Control, 292 12.5.2 Information Collection Rule (ICR), 293 12.5.3 Stage 1 Disinfection Byproducts Rule (DBPR), 294 12.5.4 Strengthening the SWTR: The IESWTR, LT1ESWTR, and Filter Backwash Recycling Rule, 296 12.5.5 Ground Water Rule, 299 12.5.6 LT2ESWTR and Stage 2 DBPR, 300 12.6 A View Toward the Future, 301

275

277

xi

CONTENTS

13

Regulating Radionuclides in Drinking Water

307

David R. Huber

13.1 13.2 13.3

13.4 13.5 13.6 13.7

13.8

14

Introduction, 307 Radiation Basics, 310 SDWA Requirements for Radionuclide Standards, 312 13.3.1 Linear No-Threshold Assumption, 313 13.3.2 NonCancer Effects, 313 1976 Radionuclide Regulations, 314 1991 Proposed Radionuclides Rule, 317 1996 SDWA Amendments and Rule Revisions, 318 2000 Final Radionuclides Rule, 322 13.7.1 Alpha Emitters, 323 13.7.2 Radium 226=228, 323 13.7.3 Radium 224, 328 13.7.4 Uranium, 329 13.7.5 Beta and Photon Emitters, 332 Future Outlook, 336

Risk-Based Framework for Future Regulatory Decision-Making

339

Mark Gibson and Mike Osinsiki

14.1 14.2 14.3

14.4

14.5 14.6

14.7

Introduction, 339 SDWA Amendments of 1996, 340 Role of Third-Party Consultations in Regulatory Development, 342 14.3.1 The National Research Council (NRC), 342 14.3.2 The National Drinking Water Advisory Council (NDWAC), 343 Role of USEPA Programs, 344 14.4.1 National Drinking Water Contaminant Occurrence Database (NCOD), 344 14.4.2 Unregulated Contaminant Monitoring Program, 345 14.4.3 Drinking Water Research Plan, 346 Development of the First CCL, 347 Public Health Decisions from the 1998 CCL, 349 14.6.1 Applicability of Prioritization Schemes for CCL Contaminants, 351 14.6.2 Generalized Decisionmaking Framework, 351 14.6.3 NDWAC Regulatory Decisionmaking Protocols, 354 14.6.4 Regulatory Decisions from the 1998 CCL, 355 Development of Future CCLs, 356 14.7.1 Identifying Future Drinking Water Contaminants, 356 14.7.2 Classifying Future Contaminants for Regulation Consideration, 358

xii

CONTENTS

14.7.3 Overview of Classification Strategies, 363 14.7.4 PCCL to CCL: Attributes of Contaminants, 367 14.8 Illustration of a Prototype Classification Scheme, 368 14.8.1 The Training Data Set, 368 14.8.2 Attribute Scoring, 369 14.8.3 Prototype Classification Functions, 369 14.8.4 Classification Results Using a Linear Classifier, 372 14.8.5 Classification Results Using a Neural Network Classifier, 373 14.8.6 Examination of Misclassified Contaminants, 373 14.8.7 Validation Test Cases, 374 14.8.8 Prediction for Interesting Test Cases, 374 14.9 Virulence Factor–Activity Relationships (VFARs), 375 14.10 NRC Recommendations and Future Directions, 376 15

Selection of Treatment Technology for SDWA Compliance

381

Frederick W. Pontius

15.1 15.2

15.3 15.4

15.5 15.6 15.7 16

Introduction, 381 SDWA Requirements Affecting Technology Selection, 381 15.2.1 Best Available Technology (BAT), 382 15.2.2 Compliance and Variance Technologies, 383 15.2.3 Compliance Technology Lists, 384 15.2.4 Variance Technology Determinations, 384 Acceptance of New Technology, 385 Advanced Treatment Technology Overview, 386 15.4.1 Membranes, 387 15.4.2 Ultraviolet (UV) Disinfection, 390 15.4.3 Advanced Oxidation, 391 15.4.4 Ion-Exchange and Inorganic Adsorptive Media, 393 15.4.5 Biological Filtration, 394 Simultaneous Compliance, 395 Process Optimization, 396 Technology Selection, 396

SDWA Compliance Using Point-of-Use (POU) and Point-of-Entry (POE) Treatment Frederick W. Pontius, Regu P. Regunathan and Joseph F. Harrison

16.1 16.2 16.3 16.4 16.5 16.6

Introduction, 403 POU and POE Technology Benefits, 404 POU and POE Technology Limitations, 405 SDWA Requirements for POU and POE Technology, 407 Certification Programs, 408 POU and POE Technology Overview, 411

403

CONTENTS

xiii

16.6.1 16.6.2 16.6.3 16.6.4 16.6.5

POU Carbon Units, 411 POU Reverse-Osmosis Devices, 412 POU UV Devices, 414 POU Distillers, 414 POU Activated Alumina (AA) and Adsorptive Media Units, 415 16.6.6 Other POU Products, 415 16.6.7 POE Products, 415 16.7 Selecting POU and POE Technologies, 417 16.7.1 Pilot Testing, 418 16.7.2 Certification, 419 16.7.3 State and Local Regulations, 419 16.7.4 Negotiating Initial Costs, 420 16.7.5 Operation and Maintenance, 420 16.7.6 Residuals and Waste Disposal, 420 16.8 Installation and Maintenance, 420 16.9 Monitoring, 422 16.10 Implementation Issues and Strategies, 422 16.10.1 Public Relations, 423 16.10.2 Administration, 423 16.10.3 Operator Training and Certification, 424 16.10.4 Liability, 424 16.10.5 Equipment Reliability, 425 16.10.6 Waste Disposal, 425 16.10.7 Economics and Cost Estimating, 426 16.11 Future Outlook and Trends, 427

PART IV 17

COMPLIANCE CHALLENGES

Death of the Silent Service: Meeting Consumer Expectations Elisa M. Speranza

17.1 17.2 17.3 17.4 17.5

Introduction, 433 Who Are Water Utility Customers?, 433 Public Water Suppliers as a Monopoly, 436 Where Customers Obtain Information, 436 What Customers Think and Want, 437 17.5.1 Trust and Consumer Confidence, 438 17.5.2 Customer Satisfaction Surveys, 439 17.6 Gaining Customer Support, 441 17.7 Communicating with Customers, 441 17.7.1 Communicating Risk, 442 17.7.2 Consumer Confidence Reports, 443

431 433

xiv

CONTENTS

17.7.3 Strategic Communications Planning, 444 17.7.4 Stakeholder Involvement, 445 17.8 Benefits of Customer Communication, 446 18

Achieving the Capacity to Comply

449

Peter E. Shanaghan and Jennifer Bielanski

18.1 18.2

Introduction, 449 Water System Capacity, 450 18.2.1 Technical Capacity, 451 18.2.2 Managerial Capacity, 451 18.2.3 Financial Capacity, 452 18.2.4 Interrelationships among Capacity Dimensions, 452 18.3 Assessing Water System Capacity, 452 18.4 Enhancing System Capacity, 455 18.5 Future Outlook, 461 19

Achieving Sustainable Water Systems Janice A. Beecher

19.1 19.2

19.3

19.4

19.5

19.6

Introduction, 463 Sustainable Systems, 464 19.2.1 Systems Perspectives, 465 19.2.2 Water Systems as Systems, 466 19.2.3 Sustainability and System Size, 467 Sustainability and the SDWA, 468 19.3.1 The SDWA and Capacity, 469 19.3.2 The SDWA and Affordability, 469 19.3.3 The SDWA and Conservation Planning, 472 19.3.4 Implications, 472 Affordability and Sustainability, 473 19.4.1 Ability versus Willingness to Pay, 473 19.4.2 Affordability Thresholds, 474 19.4.3 Utility Assistance Programs, 474 19.4.4 Rate Design and Affordability, 475 19.4.5 The Role of Subsidies, 476 Pricing Theory, 477 19.5.1 Efficiency, 477 19.5.2 Prices, Income, and Demand, 478 19.5.3 Equity, 479 19.5.4 Sustainable Price Characteristics, 481 Rate Design, 481 19.6.1 Principles of Rate Design, 482 19.6.2 Cost Allocation, 483

463

xv

CONTENTS

19.6.3 Rate Design Options, 484 19.6.4 Implementation Strategies, 486 19.7 Future Trends in Achieving Sustainability, 487 20

Protecting Sensitive Subpopulations

491

Jeffrey K. Griffiths

20.1 20.2 20.3 20.4 20.5

Introduction, 491 Defining Sensitive Subpopulations, 491 Sensitive Subpopulations and the SDWA, 492 Identifying Sensitive Subpopulations, 493 What Makes a Person or Population Sensitive?, 495 20.5.1 Cancer or Adverse Reproductive Consequences, 495 20.5.2 Infections, 497 20.5.3 People with AIDS, 498 20.5.4 Transplantation, 500 20.5.5 Chemotherapy, 501 20.5.6 Immunosuppressive Therapy, 501 20.5.7 Diabetes, 502 20.5.8 Sensitivity to Exposure, 503 20.5.9 Genetic Predisposition, 504 20.6 Which Sensitive Subpopulations Are of Concern to Water Providers?, 505 20.7 Can or Should a Water Supplier Identify Who Belongs to a Sensitive Subpopulation?, 506 20.8 Nontransient and Transient Noncommunity Systems, 506 20.9 Public Health Concepts Relevant to Sensitive Subpopulations, 507 20.9.1 Reducing or Eliminating Exposure, 507 20.9.2 Acting on Suspicion, 508 20.9.3 Defining Increased Risk, 508 20.9.4 How Significant Is Increased Risk?, 508 20.9.5 Defining an Adverse Event or Outcome, 508 20.10 Future Outlook, 509 21

Environmental Justice and Drinking Water Regulation Frederick W. Pontius

21.1 21.2

Introduction, 513 Environmental Justice as a Movement, 513 21.2.1 National Environmental Justice Advisory Council, 516 21.2.2 Executive Order 12898, 516 21.3 Identifying Environmental Justice Situations, 517 21.3.1 Environmental Justice Communities, 517 21.3.2 Key Factors, 519

513

xvi

CONTENTS

21.3.3 Economic Tradeoffs, 519 21.3.4 Intergenerational Equity, 521 21.3.5 Quantitative Methods, 525 21.3.6 Scientific and Policy Limitations, 525 21.4 Environmental Justice and Contaminant Regulation, 526 21.5 Implications for Water Utilities, 528 21.6 Future Outlook, 529 22

What Water Suppliers Need to Know about Toxic Tort Litigation

533

Kenneth A. Rubin

22.1 22.2 22.3

Introduction, 533 Basics of Toxic Torts, 534 What Plaintiffs Must Prove, 538 22.3.1 Does the Water Contain a Contaminant, and Has the Plaintiff Been Exposed to It?, 538 22.3.2 Is the Level of the Contaminant Sufficient to Cause Harm?, 538 22.3.3 Has that Contaminant Caused the Injury?, 538 22.4 Key Steps in Litigation, 543 22.4.1 In the Beginning, 543 22.4.2 Threshold Requirements for a Class Action, 544 22.4.3 Discovery, 547 22.4.4 Trial, 548 22.5 Case Histories Involving Water Suppliers, 549 22.5.1 Sovereign Immunity, 549 22.5.2 Failure to Give Municipality Mandatory Advance Notice, 550 22.5.3 Federal, State, and PUC Preemption, 550 22.5.4 Consumer Confidence Reports and Litigation, 551 22.5.5 Court Action Regarding Treatment of Water, 552 22.6 Future Outlook for Tort Litigation, 552 23

Intellectual Property Laws and Water Technology Linda E. B. Hansen

23.1 23.2 23.3

Introduction, 555 Property, Copyrights, Trademarks, and Patents, 555 Patent Laws, 556 23.3.1 Historical Overview of Patent Protection, 557 23.3.2 The United States Patent System, 558 23.3.3 Basic Requirements for Patentability, 558 23.4 Obtaining a Patent, 563 23.5 Patent Infringement, 564 23.6 Future Outlook in Intellectual Property Law, 566

555

CONTENTS

24

Water System Security

xvii

567

Frederick W. Pontius

24.1 24.2 24.3

Introduction, 567 Threats to Public Water Systems, 568 SDWA Security Provisions, 570 24.3.1 Emergency Powers, 571 24.3.2 Tampering, 571 24.3.3 Vulnerability Assessments, 572 24.3.4 Emergency Response Plan, 573 24.3.5 Reviews and Information Sharing, 575 24.4 Department of Homeland Security, 576 24.4.1 DHS Organization, 576 24.4.2 Critical Infrastructure, 578 24.4.3 Homeland Security Advisory System, 579 24.5 Future Outlook, 580

APPENDIXES A

Summary Tables of Drinking Water Standards and Health Advisories

583

USEPA Office of Ground Water and Drinking Water and USEPA Office of Science and Technology

B C

1962 U.S. Public Health Service Standards

621

Section-by-Section Summary of the SDWA

635

Frederick W. Pontius

D

Text of the SDWA as Amended and Related Statutes

721

Compiled by Frederick W. Pontius

E

How Our Laws are Made

871

Charles W. Johnson

F

Enactment of a Law

923

Robert B. Dove

G

Listing of Drinking Water Federal Register Notices Compiled by Frederick W. Pontius, P.E.

953

xviii

H

CONTENTS

Outline of 40 CFR 141, 142, and 143

971

Compiled by Frederick W. Pontius

I

Example Capacity Development Tool

979

South Dakota Department of Environment and Natural Resources

J

U.S. Water Industry Statistics

995

USEPA Office of Ground Water and Drinking Water

INDEX

1009

PREFACE

Drinking water engineers and scientists generally receive extensive academic training in math, science, engineering, and technical subjects needed to pursue their chosen profession. In most cases, little formal training (and even then only a lecture or two) is provided on legislative and regulatory procedure and current issues prior to entering the workforce. By and large young (and old) professionals are essentially spoon fed, expected to accept what they are told by whatever a particular industry or lobbying group, professor or instructor, or agency they prefer or are compelled to believe. Few are able to take the time to understand and consider the issues discussed in this volume and draw thoughtful conclusions on their own, and if they do, many do not know where to begin. Such was the case for me as a young engineer. In the early 1980s, I attended my first meeting of the American Water Works Association (AWWA) Water Quality Division. There, Trustee Alan A. Stevens, then a USEPA research scientist (now retired), announced that a new drinking water regulation had just been issued, and then proceeded to hold up a copy of the Federal Register notice. What’s that? I asked. A few years later, at a national conference I sensed the excitement of many (and the disappointment of others) over the enactment of amendments to the Safe Drinking Water Act (SDWA) the day before. But lobbyists only knew the content of the new law, except for those who knew exactly where to look. [The Internet would not come into common use until well over a decade later.] The water industry’s need for a common sense understanding of regulatory and legislative procedure and issues was demonstrated to me in a series of events in the early 1990s. A paper I had written, ‘‘Complying with the New Drinking Water Quality Regulations,’’ was published in the February 1990 issue of Journal of the American Water Works Association, and won the 1990 AWWA Publications Award. Soon thereafter, Nancy Zeilig, then editor of the Journal AWWA, agreed to publish on a trial basis a monthly column on legislative and regulatory issues, known as Leg=Reg. The first article appeared in July 1990, ‘‘Surface Water Treatment Regulations.’’ In addition, I began preparing an annual review article on drinking water regulations, also published in the Journal AWWA. These articles became very popular after only a short period of time (several won Best Paper Awards), and they soon took on a life of their own. The Leg=Reg column was published monthly for xix

xx

PREFACE

over 10 years. With the support of Marcia Lacey, current editor of the Journal AWWA, the annual reviews are still published (as of 2003), although the nature of such reviews has changed each year given the flood of regulatory and legislative information now available on the Internet. Since the mid-1990s, use of the Internet has grown tremendously, especially within the water industry. Legislative and regulatory information is now more widely available than ever before. Unfortunately, this has resulted in a different problem— information overload. It is easy now to find information on regulatory and legislative matters. But it can be more difficult now to follow and understand the thinking, developmental work, and politics behind them. Indeed, federal and state regulators, as well as water utilities and consultants, spend most of their time and effort just keeping up with what requirements they must meet, let alone having time to fully understand the technical and policy basis behind them. This particular volume was developed to fill the current need for a professional reference text for water utilities, consultants, and regulators, regarding the regulation of drinking water in the United States. Basic principles are presented concerning the SDWA and drinking water regulation. It is not intended to be a detailed compliance guide to every regulation—nor does it cover the blow-by-blow of current political lobbying activities. Chapter authors for this first edition were intentionally selected from a cross-section of different agencies and organizations. In preparing their chapter, the author(s) worked independently to present the state of the knowledge in their subject area. Each chapter was peer reviewed prior to publication. By focusing only on certain foundational issues, this volume will hopefully provide for many the understanding they need to more effectively participate in the legislative and regulatory process, better determine what regulatory actions and activities are relevant to their water utility or agency, and thereby make better legislative and regulatory compliance decisions. Supplemented with additional reading and problem sets, this volume is also appropriate as a text for classroom use, either in undergraduate or graduate environmental engineering programs. By understanding the history and basic principles associated with drinking water legislation and regulation, and confronting current issues early in their career, students will be better prepared as they enter the workforce. In particular, professionals in the field who will spend at least some portion (and in some cases all) of their career working for a regulatory agency will benefit the most from early exposure to legislative and regulatory procedures and issues. Since enactment of the SDWA in 1974, great progress has been made in drinking water quality and regulation in the United States. It seems now that only the most difficult issues remain—protecting sensitive populations, achieving sustainable water systems, providing affordable drinking water for small systems, avoiding risk–risk tradeoffs, and controlling emerging waterborne pathogens, to name only a few. The need for creative thinking and innovation in drinking water regulation and legislation has never been greater. To that end, this volume is dedicated. Frederick W. Pontius Lakewood, Colorado

ACKNOWLEDGMENTS

This book would not have come about except for the dedication and persistence of the authors whose work is included herein. I have had the privilege over the prior 20þ years to know and, in many cases, work with them in differing circumstances that has created for me rich learning opportunities. My deep appreciation is especially expressed to Daniel Okun, Gunther Craun, Diane VanDe Hei, Tom Schaefer, Joyce Donohue, Jennifer Orme Zavaletta, Bruce Macler, Bob Raucher, Stig Regli, Tom Grubbs, Paul Berger, David Huber, Mark Gibson, Mike Osinski, Elisa Speranza, Peter Shanaghan, Jeff Griffiths, and Ken Rubin. There are many others who have inspired and contributed to my career that should be mentioned, but space would not allow it, so I can name only a few. David Preston, Executive Director of the American Water Works Association (AWWA) from 1979 to 1985, was instrumental early in my career, starting me down the path at AWWA in 1982 despite being stricken by illness (I remained at AWWA until 1999). I will always remember the early encouragement Abel Wolman provided to me at my first national conference. Before his well-deserved retirement (first from USEPA, then from the University of Houston), Dr. Jim Symons was a constant source of encouragement and instruction to me through the peaks and valleys I have experienced thus far in my career. Jack Sullivan, AWWA Deputy Executive Director (now retired), hired me as part of the AWWA Government Affairs program in 1989, and provided me many early tutorials. Though my principal objective was to care for my cancer-stricken father, and be with him when he died (both of which I was able to accomplish over a 5-year period), I had many valuable experiences under Jack during my term of service in Washington, D.C. Appreciation is expressed to Al Warburton, AWWA Director of Legislative Affairs, and other members of the AWWA Washington Office staff, for their support and advice during those years. During my term of service to them, the members of the Water Utility Council, Technical Advisory Group, and various Technical Workgroups, helped to shape my understanding of regulatory and legislative issues, water industry positions, and the political ways of Washington, D.C. More recently, John Regnier and the National Rural Water Association (NRWA) have provided support for my continuing work on a variety of national regulatory policy issues. Several of the chapters in this volume were adapted by the authors xxi

xxii

ACKNOWLEDGMENTS

from white papers they had prepared for NRWA. The support of NRWA is greatly acknowledged. Appreciation is expressed to the Office of Ground Water and Drinking Water of the U.S. Environmental Protection Agency (USEPA) for allowing this work to go forward and especially to the federal employees who took the time to contribute to this volume. Appreciation is also due to Bob Esposito, Executive Editor at John Wiley & Sons, Inc., for his patience and encouragement as the group labored to prepare their manuscripts, and to Christine Punzo, Associate Managing Editor at Wiley, for shepherding the manuscripts through review and production.

CONTRIBUTORS

Janice A. Beecher, Ph.D., Director, Institute of Public Utilities, Michigan State University, East Lansing, Michigan Jennifer Bielanski, Drinking Water Utilities Team, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Paul S. Berger, Ph.D., Microbiologist, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Rebecca L. Calderon, Ph.D., Office of Research and Development, National Health and Environmental, Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina Gunther F. Craun, P.E., M.P.H., D.E.E., Gunther F. Craun and Associates, Staunton, Virginia Michael F. Craun, P.E., M.S., Gunther F. Craun and Associates, Staunton, Virginia Joyce Morrissey Donohue, Ph.D., Toxicologist, Office of Water, Office of Science and Technology, U.S. Environmental Protection Agency, Washington, DC Floyd J. Frost, Ph.D., Director, Epidemiology Program, Lovelace Respiratory Research Institute, The Lovelace Institutes, Albuquerque, New Mexico xxiii

xxiv

CONTRIBUTORS

Mark Gibson, Program Officer, National Research Council, Water Science and Technology Board, Washington, DC Jeffrey K. Griffiths, M.D., M.P.H., T.M., Director, Graduate Programs in Public Health, Tufts University School of Medicine, Boston, Massachusetts Thomas R. Grubbs, P.E., Environmental Engineer, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Linda E. B. Hansen, Esq., Attorney, Patterson, Thuente, Skaar, and Christensen, L.L.C., Milwaukee, Wisconsin Joseph F. Harrison, P.E., CWS-VI, Technical Director, Water Quality Association, Lisle, Illinois David R. Huber, Regulation Manager, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Bruce A. Macler, Ph.D., Toxicologist, U.S. Environmental Protection Agency, Region 9, San Francisco, California Daniel A. Okun, Ph.D., Kenan Professor of Environmental Engineering, Emeritus, Department of Environmental Science and Engineering, University of North Carolina, Chapel Hill, North Carolina Jennifer Orme-Zavaleta, Associate Director for Science, Office of Research and Development, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oregon Michael Osinski, Drinking Water Utilities Team Leader, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Frederick W. Pontius, P.E., Pontius Water Consultants, Inc., Lakewood, Colorado Robert Raucher, Ph.D., Executive Vice President, Stratus Consulting Inc., Boulder, Colorado Stig E. Regli, Environmental Engineer, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Regu P. Regunathan, Ph.D., ReguNathan & Associates, Inc., Wheaton, Illinois

CONTRIBUTORS

xxv

Kenneth A. Rubin, Esq., Partner, Tort, Environmental, and Construction Practice, Morgan Lewis and Bockius, L.L.P., Washington, DC Thomas Schaefer, Regulatory Specialist, Association of Metropolitan Water Agencies, Washington, DC Peter E. Shanaghan, Chief of Staff, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Elisa M. Speranza, Vice President, CH2M Hill, New Orleans, Louisiana Diane VanDe Hei, Executive Director, Association of Metropolitan Water Agencies, Washington, DC

ACRONYMS

AA AF AIDS AMCL ANSI ANPRM AOP APA AR ARPA ASDWA ATSDR AWWA AWWARF AX

Activated alumina Attributable fraction Acquired immune deficiency syndrome Alternative maximum contaminant level American National Standards Institute Advance Notice of Proposed Rulemaking Advanced oxidation process Administrative Procedure Act Attributable risk Advanced Research Projects Agency Association of State Drinking Water Administrators Agency for Toxic Substances Diseases Registry American Water Works Association AWWA Research Foundation Anion exchange

B–C BAT BCA BMD BOM BT BTWC

Benefit–cost ratio Best available technology Benefit–cost analysis Benchmark dose Biodegradable organic matter Benefits transfer Biological and Toxins Weapons Convention

CA CCC

Cellulose acetate Chlorine Chemistry Council xxvii

xxviii

ACRONYMS

CCE CCL CCR CDC CERCLA

DEP DFO DHS DNA DWC DWCCL DWEL DWPL DWSRF

Carbon chloroform extract Contaminant Candidate List Consumer Confidence Report Centers for Disease Control and Prevention Comprehensive Environmental Response, Compensation and Liability Act Confidence interval Cost of illness Comprehensive performance evaluation Disinfectant residual concentration (C ) in milligrams per liter (mg=L) multiplied by the disinfectant contact time (T ) in minutes Clean Water Act Chemical Warfare Convention Community water system Clean Water State Revolving Loan Fund Community Water Supply Study Design, build, and operate Disinfection byproduct Disinfection Byproducts Rule Department of Environment and Natural Resources (South Dakota) Department of the Environment Designated Federal Official Department of Health Services Deoxyribonucleic acid Drinking Water Committee Drinking Water Contaminant Candidate List Drinking water equivalent level Drinking Water Priority List Drinking Water State Revolving Loan Fund

EBCT EDE EDF EDR EEAC EJ ELI EO ETV

Empty-bed contact time Effective dose equivalent Environmental Defense Fund Electrodialysis reversal Environmental Economics Advisory Committee Environmental justice Environmental Law Institute (Washington, DC) Executive Order Environmental Technology Verification

FACA FAS FBRR FGR

Federal Advisory Committee Act Federation of American Scientists Filter Backwash Recycling Rule Federal Guidance Report

CI COI CPE CT

CWA CWC CWS CWSRF CWSS DBO DBP DBPR DENR

ACRONYMS

GAC GAO GFH gpd GPRA GRAS GSA GWR

Granular activated carbon General Accounting Office Granular ferric hydroxide Gallons per day Government Performance and Results Act Generally recognized as safe General Services Administration Ground Water Rule

HA HAA5 HACCP HIV HPC HRL HRRCA HUD

Health Advisory Sum of five haloacetic acids Hazard assessment critical control point Human immunodeficiency virus Heterotrophic plate count Health reference level Health risk reduction and cost analysis Housing and Urban Development

IBMTR IBWA ICR IESWTR IOC IOM IQ IRIS ISAC IX

International Bone Marrow Transplant Registry International Bottled Water Association Information Collection Rule Interim Enhanced Surface Water Treatment Rule Inorganic contaminant Institute of Medicine Intelligence quotient Integrated Risk Information System Information Sharing and Analysis Center Ion exchange

LCR LED10 LNT LOAEL LP-LI LP-MI LSLR LTESWTR LT1ESWTR LT2ESWTR LYS

Lead and Copper Rule Lower limit on effective dose (producing an adverse effect in 10% of subjects exposed to a chemical) Linear nonthreshold Lowest-observed-adverse-effect level Low-pressure, low-intensity Low-pressure, medium-intensity Lead service line replacement Long Term Enhanced Surface Water Treatment Rule Long Term 1 Enhanced Surface Water Treatment Rule Long Term 2 Enhanced Surface Water Treatment Rule Life years saved

MCL MCLG MEGO MF mgd

Maximum contaminant level Maximum contaminant level goal My eyes glaze over Microfiltration Million gallons per day

xxix

xxx

ACRONYMS

MIB MMM MOE MP-HI MRDL MWB MWRA

Methylisoborneol Multimedia mitigation Margin of exposure Medium-pressure, high-intensity Maximum residual disinfectant level Metropolitan Water Board Massachusetts Water Resources Authority

NAE NAPA NAS NCOD NCSL NCWS NDWAC NEETF NEJAC NF NGA NIMBY NIPC NIPDWR NIRS NOAEL NODA NOM NOMS NORS NPDES NPDWR NRA NRC NRDC NRWA NSF NTNCWS NTU NWIS NYC DEP

National Academy of Engineering National Academy of Public Administration National Academy of Sciences National Contaminant Occurrence Database National Conference of State Legislatures Noncommunity Water System National Drinking Water Advisory Council National Environmental Education and Training Foundation National Environmental Justice Advisory Council Nanofiltration National Governors Association Not in my backyard National Infrastructure Protection Center (of FBI) National interim primary drinking water regulation National Inorganics and Radionuclides Survey No-observed-adverse-effect level Notice of data availability Natural organic matter National Organics Monitoring Survey National Organics Reconnaissance Survey National Pollutant Discharge Elimination System National primary drinking water regulation Negotiated Rulemaking Act National Research Council Natural Resources Defense Council National Rural Water Association National Science Foundation Nontransient noncommunity water system Nephelometric Turbidity Units National Water Information System New York City Department of Environmental Protection

OGWDW O&M OMB OPCW OR

Office of Ground Water and Drinking Water Operation and maintenance Office of Management and Budget Organization for the Prohibition of Chemical Weapons Odds ratio

ACRONYMS

PAC PCCL PCU PDWR POD POE POTWs POU PTA PUV PWS PWSS

Powdered activated carbon; Political Action Committee Preliminary Contaminant Candidate List Pinellas County Utilities Primary Drinking Water Regulation Point of departure Point of entry Publicly owned treatment works Point of use Packed-tower aeration Pulsed ultraviolet Public water system Public water system supervision

QALY QSAR

Quality-adjusted life years Quantitative structure–activity relationship

RD R&D RfC RfD RIA RMCL RO RR RSC

Rate difference Research and development Inhalation reference concentration Reference dose Regulatory impact analysis Recommended maximum contaminant level Reverse osmosis Relative risk Relative source contribution

SAB SAC SAR SARA SBA SDWA SDWR SDWIS SEB SEER SOCs SPAM SRLF SWT SWTR

Science Advisory Board Strong-acid cationic (resin) Structure–activity relationships Superfund Amendments and Reauthorization Act Strong-base anionic (resin) Safe Drinking Water Act Secondary Drinking Water Regulation Safe Drinking Water Information System Staphylococcal enterotoxin B Surveillance, Epidemiology, and End Result Synthetic organic chemicals (also compounds) Safety, participation, affordability, and management State Revolving Loan Fund Source water treatment Surface Water Treatment Rule

TCLP TCR TDS TFC THMs

Toxicity characteristic leaching potential Total Coliform Rule Total dissolved solids Thin-film composite Trihalomethanes

xxxi

xxxii

ACRONYMS

TNCWS TOC TON TRIP TRO TT TTHMs

Transient noncommunity water system Total organic carbon Threshold order number Trade-Related Aspects of Intellectual Property Temporary restraining order (for patentholders) Treatment technique Total trihalomethanes

UCC UCM UCMR UF UIC UL USC USCM USEPA USGS USPHS UV

Uniform Commercial Code Unregulated contaminant monitoring Unregulated Contaminant Monitoring Rule Ultrafiltration Underground injection control Underwriters Laboratory United States Code U.S. Conference of Mayors U.S. Environmental Protection Agency U.S. Geological Survey U.S. Public Health Service Ultraviolet

VA VFAR VOC VSL

Veterans Administration Virulence factor–activity relationship Volatile organic contaminant Value of a statistical life

WAC WBA WMD WQA WQR WTP

Weak-acid cationic (resin) Weak-base anionic (resin) Weapons of mass destruction Water Quality Association Water Quality Report Willingness to pay

PART I THE SAFE DRINKING WATER ACT AND PUBLIC HEALTH

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

1 DRINKING WATER AND PUBLIC HEALTH PROTECTION DANIEL A. OKUN, Sc.D., P.E. Kenan Professor of Environmental Engineering, Emeritus, University of North Carolina Chapel Hill, North Carolina

1.1

INTRODUCTION

The provision of drinking water for communities is an urban utility, but a utility with a difference. As with other urban utilities, such as electricity and gas, water for household use is a necessity that cannot readily be obtained by urban householders for themselves. The difference is that, while water may satisfy many household needs, including drinking, it has the potential of spreading disease, often without the knowledge of the consumer. As a result, water supplies have become subject to regulations for assuring adequate quality, regulations that are not faced by other municipal public utilities. Beginning with the water supply for Rome some 2000 years ago, the responsibility for water supply and its quality rested with the community. During the nineteenth century, with the beginning of the industrial era and the rapid growth of cities, public water supplies began to be provided by private entrepreneurs who sought profit in providing an essential service, frequently in competition with others. In the interest of getting a larger share of the market, they might provide a water of better quality than a competitor. The experiences with the provision of water for London from the Thames in the 1850s illustrate that the selection of the source of a water supply is Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

3

4

DRINKING WATER AND PUBLIC HEALTH PROTECTION

important. Then Dr. John Snow took advantage of the competition between two water suppliers to prove that water was responsible for the transmission of cholera. As cities grew, the need for large capital investments to provide adequate water supplies of high quality resulted in most cities abandoning private utilities when it became clear that they did not have the financial resources for the construction of the reservoirs, the long transmission lines, and the treatment plants. Decisions for selection of sources and treatment, which were introduced in the late nineteenth–early twentieth century, became the responsibility of the community, and not a regulatory body. Treatment in the form of filtration and then chlorination was widely introduced, although not primarily through regulation. City officials recognized that they had an obligation to their constituencies to provide water that would not spread typhoid and cholera. Some cities were slow to assume this responsibility, and, in the United States, some newly organized state health agencies began to institute regulations. The choice of sources between a costly high-quality upland supply and a polluted source and the treatment to be provided was local. The first nationwide water quality regulations in the United States were introduced by the federal government in 1909 to assure the safety of water to which the public was exposed in interstate and international traffic. Many states adopted these regulations even for smaller cities that did not have train or bus service. These federal regulations were upgraded over the years, and the bulk of this chapter is devoted to the nature of these regulations at the federal level until the passage of the Safe Drinking Water Act (SDWA) in 1974, after the U.S. Environmental Protection Agency (USEPA) became responsible for ensuring the safety of all public water supplies. This chapter recounts the high points in the history of the role that urban water supplies play in the health of those who are obliged to drink from public supplies, beginning with concerns with the water supply for Rome, followed by the story of the cholera outbreaks in London that led to the recognition that water was responsible for the spread of infectious disease, the introduction of successful public health measures to control infectious disease, and the explosion of the chemical revolution that became responsible for the spread of chronic disease through ingestion of public water supplies (Okun 1996).

1.2

WATER SUPPLY FOR THE CITY OF ROME

Among the major ancient cities of the world, none was better provided with water for its citizens than Rome. Initially, the city obtained its water from the Tiber River, which ran through the city. When it was apparent that the water had become heavily polluted, Appius Claudius built an aqueduct, the Aqua Appia, in 312 B.C. from the Tiber, about 11 miles upstream. Some 40 years later, the need was so great that a second aqueduct, 40 miles long, the Anio Novis, was built. Sextus Julius Frontinus, the water commissioner of Rome, wrote two books describing the water works of the city and their management (Frontinus A.D. 97). By A.D. 305, 14 aqueducts were serving the city.

1.3 THE MIDDLE AGES AND THE INDUSTRIAL REVOLUTION

5

The aqueducts fed the city by gravity with relatively short sections passing over valleys on stone arches, some three tiers high. Many of them carried water into the twentieth century. Such aqueducts remain throughout Europe and the Middle East as monuments to the early Romans. The water from the aqueducts passed through large cisterns and from these was distributed through lead pipes to other cisterns, to public buildings, baths, and fountains, and to a relatively small number of private residences. Incidentally, they also built stone sewers to carry off wastewater from bathtubs and toilets in the larger buildings. Frontinus questioned the wisdom of Augustus, whom he considered a most cautious ruler, in building one of the aqueducts, the Alsietinian, because the quality of its water was very poor and not suitable for the people. He speculated that Augustus built the aqueduct to serve nonpotable purposes and thereby ‘‘to avoid drawing on better sources of supply.’’ The most important nonpotable use was for a naumachia, an artificial lake that was used for exhibitions of sham naval battles (Fig. 1.1). This is also current practice in American cities that erect stadia for baseball, football, and basketball on behalf of the team owners. The surplus nonpotable water was used for landscape irrigation and fountains. Words from an inscription state: ‘‘I gave the people the spectacle of a naval combat . . . . Besides the rowers, three thousand men fought in these fleets.’’ Thus, Rome can claim to be the first city to employ a dual distribution system and to base the use of its water supply on its quality. The water quality from the aqueducts was variable, and the Marcia aqueduct carried the best water. Frontinus points out that it was ‘‘determined to separate (the aqueducts) and then to arrange that the Marcia should serve wholly for drinking purposes, and that the others should be used for purposes adapted to their special qualities.’’ It is interesting to note that, in 1958, some 2000 years later, the United Nations Economic and Social Council enunciated a policy (United Nations 1958): ‘‘No higher quality water, unless there is a surplus of it, should be used for a purpose that can tolerate a lower grade.’’

1.3

THE MIDDLE AGES AND THE INDUSTRIAL REVOLUTION

Beginning in the sixth century, the Roman Empire began to disintegrate and, up to the fourteenth century, infectious diseases rode rampant throughout Europe. Leprosy, bubonic plague, smallpox, diphtheria, measles, influenza, and countless other afflictions were epidemic, particularly in the cities. Water was only one of the many vectors for the spread of disease. Knowledge of the specific vectors was limited, and food received the most attention. Quarantine was the principal approach to control of the spread of disease. The lack of proper sanitation and the dense urban populations were largely responsible for the epidemics and there was little focus on water quality and its availability. The major accomplishment toward the end of the Middle Ages was the establishment of hospitals, often for specific diseases, by local governments and workers’

6

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Figure 1.1

Naumachia (from a coin of Domitian) (Frontinus

A.D.

97).

guilds. Little of importance with regard to the water environment and the public health emerged during that period.

1.4

THE GREAT SANITARY AWAKENING

In the middle of the nineteenth Century, the causes of the many common diseases of the day that afflicted the growing urban populations that accompanied the Industrial Revolution were still unknown. Water was beginning to be piped to houses of the well-to-do while the poor either carried their water from wells or bought water from purveyors who obtained the water at the most convenient sources. When the water was contaminated, which was its general condition in urban areas, the spread of disease was inevitable.

1.4 THE GREAT SANITARY AWAKENING

7

A more significant and serious situation resulted from the growing installation rate of piping and then water closets in homes and commercial establishments. In addition to the impact of a poor quality water for drinking was the necessity for disposing of the discharges of these new flush toilets. London had found it necessary to construct storm sewers to drain the streets to permit the conduct of commerce. The obvious solution was to discharge the household wastes from the toilets to the storm sewers, which, in turn, discharged directly into the Thames River, which ran through London and served as a source of water for several private companies that distributed the water to households. London was exemplary of the unsavory and squalid conditions in all cities in the early years of the century. The medical fraternity believed that the diseases were spread by poisons in the miasmatic air emanating from the ‘‘bowels of the earth.’’ The Thames at London at that time was a tidal river and the heavily polluted waters would flow very slowly to sea. In warm periods, Londoners avoided crossing London Bridge because the air was so foul. A headline of the period read ‘‘India is in Revolt and the Thames Stinks.’’ The drapery in the Houses of Parliament, located on bank of the Thames, needed to be soaked in chloride of lime to make the meeting room tolerable, and stirred the Parliament to establish the first of many committees to see to alleviating the situation. Two cholera outbreaks in the summer of 1854 were the greatest in London’s history. The first developed in Soho, a densely populated section in the heart of the city. Dr. John Snow, then physician to Queen Victoria Hospital, and reasonably the first epidemiologist, undertook to mark the deaths in the summer of 1953. In 2 days, 197 people died, and after 10 days more than 500 people died in an area only 250 yards across (Longmate 1966). Plotting the deaths on a map of the area (Fig. 1.2), the result resembled a target, with the greatest concentration of hits at the center. A church-owned well on Broad Street was identified at the site as being the source of the water ingested by the victims. The water had appeared to be of excellent quality. A woman living about a mile away regularly sent a cart to carry water to her home; she and a guest from outside London died of cholera in that epidemic. Dr. Snow examined the well site and concluded that a tannery on property owned by the church had a cesspool for discharge of its wastewaters. He ordered the church to remove the handle on the pump, ending the epidemic, but, by that time, the epidemic might well have been spent. At any rate, this demonstration was the first to suggest that drinking water was the source of the cholera. This was generations before the germ theory of disease had been elucidated, and Snow’s other studies in London were even more convincing. The John Snow Pub is on the site of the Broad Street pump, and these data decorate its walls. Annual death rates from cholera among households using Thames River water ranged from 10 to 110 per 10,000 households in 1832, increasing to 200 per 10,000 among those taking water from the downstream reaches of the river. While this justified the inference that water was responsible, Dr. Snow found a more definitive proof during the 1854 epidemic. Two private water companies, the Southwark and Vauxhall Company and the Lambeth Company were in direct competition, serving

8

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Figure 1.2 Map of Soho showing the location of those who died from cholera within the vicinity of the Broad Street pump in London 1854 (Cosgrove 1909).

piped water to the same area near the center of London but on the south side of the river. These water companies were characterized as ‘‘by far the worst of all those that take their water from the Thames, with 120 to 180 deaths per 10,000 households in 1849 for each of the two companies.’’ (Snow 1936) In 1852, however, the Lambeth Company, to attract more customers, improved the aesthetic quality of the Thames River water by moving its intake upstream above the heaviest pollution from London. Snow’s data showed that, in the 1854 epidemic, the death rate among those using Lambeth water was 37 deaths per 10,000 households as compared with 315 per 10,000 households for those using the intake downstream. During that period, the death rate in all of London was 59 per 10,000 households (256,423 deaths) among those taking water from all sources in London. In addition to establishing that the cholera outbreaks were caused by drinking water, Snow demonstrated the importance of source selection. As is pointed out below, almost a century later, some cities still chose to take water from run-of-river sources when better sources were available primarily because it was less costly. Professor Fair, in presenting his philosophy about water supply, characterized the

1.5 THE EMERGENCE OF WATER AS A PUBLIC HEALTH ISSUE

9

issue by declaring that he ‘‘preferred the virginal to the repentant,’’ a paraphrase of the philosophy of Allen Hazen, possibly the most important engineer in the early history of water supply in the United States, who put it: ‘‘Innocence is better than repentance.’’ (Okun 1991a).

1.5

THE EMERGENCE OF WATER AS A PUBLIC HEALTH ISSUE

The Industrial Revolution, beginning in the late eighteenth century in Britain and extending to Europe and the United States, was responsible for an explosive increase in urbanization with the development of the slums so ‘‘celebrated’’ by Dickens. It did eventually result in the English government and the northeastern states in the United States establishing agencies for addressing the terrible health conditions that emerged. Massachusetts, Pennsylvania, and New York established health boards to improve housing conditions; this resulted in the establishment of regulations for water supply and disposal of household wastes (Fig. 1.3). These efforts at regulating activities that might damage the environment led to the establishment of the public health movement. Two figures of lasting fame: Sir Edwin Chadwick, a lawyer, in England (Ives 1990), and Lemuel Shattuck, a physician (Fair 1945) in Massachusetts, who was inspired by Chadwick, were responsible for the creation of regulatory agencies and laws protecting the public from the wide range of microbial and chemical contaminants that inevitably found their way to the nearby streams and rivers that were drawn upon for water supply.

Figure 1.3 Simultaneous decline in typhoid fever death rate and rise in number of community water supplies in the United States (—— deaths per 100,000 population; water supplies: 1000s) (source: F. W. Pontius).

–

10

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Shattuck’s plan for the board of health for Massachusetts called for its membership to be composed of two physicians, one counselor-at-law, one chemist or natural philosopher, one civil engineer, and two persons of other professions or occupations. This comprehensive view of the needs for an agency for the protection of the public health was the springboard for the establishment of a sanitary engineering specialty within civil engineering. Shattuck had pointed out that competence in ‘‘planning and constructing public works’’ was essential to the provision of water supply and the disposal of household wastes. In 1886, the Massachusetts legislature passed ‘‘An Act to Protect the Purity of Inland Waters’’ and, to implement the Act, it called for the establishment of an engineering department in the State Board of Health. Among its activities was the establishment of the Lawrence Experiment Station, the first of its kind, which was instrumental in attracting engineers, chemists, and biologists from the Massachusetts Institute of Technology, many of whom were responsible not only for spreading the study of water-related diseases and their control but also in the introduction of community water supply systems. From a total of only 17 water supply systems in the state in 1869, the number grew to 138 in 1890 while the annual death rate from typhoid fever in the state dropped from 89 per 100,000 in 1873 to 37 in 1890, and by 1940 to 0.2 (Fair 1945). Despite the appearance of regulatory agencies, many years passed before they played a significant role in the monitoring of municipal water supply and wastewater collection, treatment, and disposal systems. Actually, there was, is, and should be far less need for regulation of drinking water quality than for regulation of wastewater discharges. In the early days of public water supplies, most were privately owned and needed to meet the requirements of the communities they served. When they were inadequate to the task, sometimes because they failed to satisfy the communities they served, but more generally because the rapid growth of the cities called for capital investments beyond the capacity of the private purveyors to meet, the community government became responsible for the water supply. When the community government itself was providing the water, there seemed to be little need for regulating the performance of their own utility as its objectives would naturally be to protect its citizenry from public health risks. A good example of this was the early history of the water supply for London, as mentioned above, where the private companies were generally loathe to invest in improvements. At the end of the nineteenth century, a Metropolitan Water Board (MWB) was created to take over responsibility for the water supply of London from eight private companies. In some other cities in England, private water companies continued to serve satisfactorily and continue to this day. The MWB established new technology and were seen to be at the leading edge of water supply technology and they set their own standards which were emulated by other communities. In the case of New York City, its early private purveyors also were inadequate to their responsibilities. The driving forces were the need to have water to prevent epidemics of yellow fever (which were not related to water) and to fight fires. One of the last private efforts was that inspired by Aaron Burr, who promised a

1.6 THE BEGINNING OF WATER TREATMENT

11

water supply as a condition of establishing the Bank of the Manhattan Company, the predecessor of the Chase Manhattan Bank. He had little interest in providing water and ‘‘this brilliant and unprincipled man suffered a series of political disasters that plunged him . . . to ruin and exile.’’ (Blake 1956). Burr’s plans were doomed. The city finally decided to develop its own supply and, after extensive study had to choose between two possible sources: the Bronx River very near the city and the Croton River some 40 miles distant. The former was considerably lower in cost but the latter promised a much better quality of water and a greater quantity for the future. The City Board of Water Commissioners Committee on Fire and Water, addressing this question in 1835 opted for the Croton and for public ownership in this language (Blake 1956): The question remains, ought the Corporation of the City of New York Committee to embark on this great work? The Committee are firmly of the opinion, that it ought to be done by no other body, corporate or personal . . . . The control of the water of the City should be in the hands of this Corporation, or in other words, in the hands of the people.

The City celebrated the delivery of high-quality water from the Croton Aqueduct to New York City by gravity at high pressure in ample quantity in 1842, then one of the largest water supplies in the world. It still provides about 15% of the water that the City uses. This costly choice was made by the city officials not to meet a regulation but to serve their constituency well. Another example is the city of Cincinnati, Ohio, which installed granular activated carbon (GAC) filtration in the 1980s though it is not required by regulations. Many cities do more than the existing regulations require because the regulations tend to be years behind our knowledge. Water officials desiring to serve their community best may find it wise to anticipate quality problems that will not be addressed by regulations for years. Unfortunately, the reception given new regulations is not always one of appreciation by many water officials but of concern for the costs that may be involved. Industry groups such as the American Water Works Association (AWWA) often challenge regulations that are in the process of being promulgated to reduce public health risks because it would increase costs and water rates. On the other hand, regulations for the quality and quantity of discharges of wastewaters to receiving waters are necessary, because the cost burden falls on the community while those who benefit are generally residents of other communities and not liable for the costs. This is also one of the reasons why the Clean Water Act (CWA) and similar earlier programs have been obliged to meet a significant share of the costs.

1.6

THE BEGINNING OF WATER TREATMENT

The relationship between source, water quality, and disease was demonstrated in the United States but much later than cholera in England and with much lower typhoid fever rates. Kober (1908) made a study of typhoid rates in American cities from

12

DRINKING WATER AND PUBLIC HEALTH PROTECTION

TABLE 1.1 Typhoid Rates in American Cities, 1902 Through 1908 Source Groundwater Impoundments and protected watersheds Small lakes Great lakes Mixed surface and groundwater Run-of-river supplies

Number of Cities

Death Rate per 100,000

4 18

18.1 18.5

8 7 5

19.3 33.1 45.7

19

61.6

Source: (Elms 1928).

1902 through 1908, summarized in Table 1.1. New York City, with its upland supply, had the lowest rate of the 61 cities with 15 typhoid deaths per 100,000 while Pittsburgh, with its run-of-river supply, suffered the highest death rate, 120 per 100,000. Filtration of water was introduced well before the turn of the nineteenth century in Europe, where run-of-river supplies were more common. An eightfold increase of filtration in the United States reduced the typhoid death rate from water supply from 1900 to 1913 by 55% (Ellms 1928). The availability of filtration mistakenly seemed to make the need for selecting better sources unnecessary. Philadelphia, which had been taking water from run-ofriver sources and had been one of the last of the large U.S. cities to adopt filtration, was suffering a typhoid death rate of 75 per 100,000 into the twentieth century. The city officials had contended that filtration was not as effective as boiling the water. In 1900, a reform mayor was determined to address the issue. A panel of distinguished engineers prepared ‘‘a report that was characterized as not having any surprises.’’ (McCarthy 1987). It recommended filtration and continued use of water from the lower Delaware and Schuykill Rivers. The report stated that ‘‘Water from upcountry sources might be preferable but the great cost of building aqueducts and reservoirs made that option very expensive and really unnecessary since filtration would provide safe water.’’ In 1911, before the filters were operational, a typhoid outbreak in Philadelphia resulted in 1063 deaths. After filtration, the death rate dropped to 13 per 100,000, still a relatively high figure. Philadelphia still takes most of its water from the ‘‘mouth’’ (more properly, the ‘‘anus’’) of the Delaware River and has had to adopt Herculean methods to deliver water of good quality ever since. A multimedia study of environmental health risks in Philadelphia in the 1980s determined that the water supply posed the highest risk of all sources of pollution in the city. Since then improved treatment processes along with stricter USEPA standards have been introduced. Many cities have no alternatives and are obliged to draw from run-of-river sources. Slow sand filtration became the treatment of choice in Massachusetts in

1.7 THE CHEMICAL REVOLUTION

13

the 1870s. In the 1890s, the Louisville Water Company, which took water from the Ohio River, introduced sedimentation of the water prior to filtration. For better removal of turbidity, they introduced chemical coagulation and rapid sand filtration. The introduction of chlorination for disinfection of water for municipal water supply took place in Boonton, New Jersey, in 1908 following decades of study of the use of chlorine in Europe and the United States (Baker 1948). It clearly was the greatest step in the reduction of the transmission of infectious diseases via water supply. An example of the role of chlorine was the effect it had in a city drawing water from a clear lake. Chlorine reduced the annual typhoid death rate from about 20 to 2 per 100,000 population, which was then reduced to virtually zero with the addition of filtration (Fair et al. 1968). Together with pasteurization of milk and better handling of human wastes, typhoid virtually disappeared in the United States by the middle of the twentieth century.

1.7

THE CHEMICAL REVOLUTION

While infectious disease was brought under control, although other diseases emerged later, two other problems arose. The first was that water treatment tools were believed to be so effective, engineers became sanguine about the need to seek waters of high quality; treatment would make it safe. The conventional treatment of the midtwentieth century, which remains the conventional treatment now, at the beginning of the twenty-first century—chemical coagulation, rapid sand filtration, and chlorination—does little to remove the trace synthetic organic chemicals in ambient water resulting from the post World War II surge in industrial development what has been labeled the ‘‘chemical revolution’’ (Okun 1996). The second problem is truly ironic—the life-saving treatment, chlorination, increases the risk from synthetic organic chemicals created by the chlorine itself. Other disinfection byproducts have surfaced and added to the problem of the trace synthetic organic chemicals discharged from industry and households using a wide range of such chemicals for house and garden. The first published material about disinfection byproducts (DBPs) emanated from Rook’s work at the Rotterdam water treatment plant, which drew water from the mouth of the Rhine River (Bellar et al. 1974). While it had been picked up quickly by USEPA, the potential had been recognized 5 years earlier. Dr. Joshua Lederberg, a Nobel Prize geneticist, who had been somewhat active in drinking water issues, wrote a 1969 syndicated column in the Washington Post. One column was headlined ‘‘We’re so accustomed to using chlorine that we tend to overlook its toxicity’’ (Lederberg 1969): What little we do know of the chemistry of chlorine reactions is portentous. It should sometimes react . . . to form substances that may eventually reach and react with genetic material, DNA, of body cells . . . That chlorine is also intended to inactivate viruses should provoke questions about the production of mutagens in view of the close similarity between viruses and genes.

14

DRINKING WATER AND PUBLIC HEALTH PROTECTION

USEPA was created in 1970, but Lederberg failed in attempts to attract funds to follow this up with his research team. The discovery of trihalomethanes and other disinfection byproducts and concerns regarding the potential cancer risks associated with chloroform would be a major driving force behind passage of the 1974 SDWA.

1.8

THE INTRODUCTION OF REGULATIONS

In the absence of regulations, many cities adopted practices that were believed to be the most appropriate for their own conditions on the recommendations of professional engineers and water scientists. As noted above, the spread of disease had been controlled in large measure by the quarantine of the sick. It was not unreasonable, therefore, for federal authority over the control of the spread of disease via water to be initially addressed by the U.S. Congress in the Interstate Quarantine Act of 1893 (United States Statutes 1893). Under the Act, the surgeon general of the U.S. Public Health Service (USPHS) was empowered ‘‘to make and enforce such regulations as are necessary to prevent the introduction, transmission, or spread of communicable disease from foreign countries into the states or possessions, or from one state or possession into any other state or possession.’’ Interstate regulations were first promulgated in 1894 with the first water-related regulation adopted in 1912, which prohibited the use of the common cup on carriers in interstate commerce (McDermott 1973). The first federal drinking water standards were adopted in 1914. The USPHS was then part of the U.S. Treasury Department and was charged with responsibility for the health care of the sailors of the Merchant Marine. The surgeon general of the USPHS recommended and the Treasury Department adopted standards for drinking water to be supplied to the public on interstate carriers, then called ‘‘Treasury Standards.’’ Because the group that was charged with developing the standards could not agree on physical and chemical parameters, only a bacterial standard of 100 microorganisms per milliliter was adopted. The organism adopted was Bacteria coli, now known as Escherichia coli. It was further stipulated that not more than one of five 10-mL portions (2 Bacteria coli per 100 mL) would be permitted. (Borchardt and Walton 1971) These coliform organisms were not themselves pathogenic but, originating in large numbers in the human colon and found in feces, they served as a surrogate for enteric pathogens because they were more resistant to removal and were present in large numbers and, if they were not present, it could be inferred that the enteric pathogens likely would not be present. Many local and state officials adopted the standard and monitored the water systems that served interstate carriers for themselves and on behalf of the Treasury Department. A federal commitment was made in 1915 to review the regulations on a regular basis. By 1925, most large cities drawing water from run-of-river sources were already using filtration and chlorination and having little difficulty in meeting the 1914 coliform standard of 2 100 mL 1 (2 coliforms per milliliter). Following a principle of attainability, the standard was tightened to 1 100 mL 1. In addition, standards

1.8 THE INTRODUCTION OF REGULATIONS

15

were established for physical and some chemical constituents, including lead, copper, zinc, and dissolved solids (USPHS 1925). The 1925 standards introduced the concept of relative risk. The preamble stated in part: The first step toward the establishment of standards which will ensure safety of water supplies conforming to them is to agree upon some criterion of safety. This is necessary because ‘‘safety’’ in water supplies, as they are actually produced, is relative and quantitative, not absolute. Thus, to state that a water supply is ‘‘safe’’ does not necessarily signify that no risk is ever incurred in drinking it. What is usually meant, and all that can be asserted from any evidence at hand is that the danger, if any, is so small that it cannot be discovered by available means of observation.

In 1941, an advisory committee for revision of the 1925 standards was assembled by the USPHS, composed of representatives of federal and state agencies, scientific associations, and members at large, which produced the 1942 standards (USPHS 1943). One new initiative was the introduction of requirements for monitoring microbial water quality in the distribution system, with specifications for the minimum number of samples to be collected each month according to the size of the community. Specifications for the laboratories and procedures involved were provided. Maximum permissible concentrations were established for lead, fluoride, arsenic, and selenium as well as for salts of barium, hexavalent chromium, heavy metals, and other substances having deleterious physiological effects. Maximum concentrations where other alternative sources were not available were set for copper, the total of iron and manganese, zinc, chlorides, sulfates, phenolic compounds, total solids, and alkalinity. Only minor changes were introduced in 1946 (USPHS 1946). Publication in the Federal Register was introduced, assuring more rapid dissemination of changes that might be made, one of which was the authorization in March 1957 of the use of the membrane filter procedure for the bacteriological examination of water samples. World War II (for the United States, 1942–1946) was the first war where deaths of American troops by infectious disease did not exceed deaths in combat. Steps had been introduced to reduce exposure to mosquitoes that were responsible for malaria and other related diseases in the tropics and facilities were provided to assure chlorination of the drinking water. In the postwar period, and driven by the need to make up for years during which the construction of state-side water-related civilian infrastructure had been dormant, attention was turned to making heavy investments for urban water supply. The need for standards was apparent. Dr. Abel Wolman (1960) addressed this issue thus: From its beginning, society by one means or another, has surrounded itself with restraints. These have had, for the most part, empiric origins—moral, ethical, economic, or spiritual. All the restraints have had the common basis of an assumed benefit to the

16

DRINKING WATER AND PUBLIC HEALTH PROTECTION

particular society establishing them. As societies became more complex and more sophisticated, efforts towards both standardization and restraint became more frequent, more necessary, and presumably more empiric, although examples of the last are not as numerous as one might expect.

He then went on to characterize the types of standards that are necessary:  Regularization of techniques of measurement;  Establishment of limits of concentration or density of biologic life and physical and chemical constituents;  Regularization of administrative practice;  Regularization of legislative fiat; and  Specification of materials.

The increasing complexity of the issues is exemplified in all that follows, including not only in the specific regulations required but also in the methodologies of reaching consensus among the many stakeholders involved. The beginning of the ‘‘chemical revolution’’ and regulating the thousands of synthetic organic compounds (SOCs) that are being invented annually and that find their way into the environment and into waters drawn on for drinking began with the 1962 update of the federal Drinking Water Standards. The establishment of the 1962 USPHS standards involved examining many new issues, including two important problems not previously addressed: radioactivity and SOCs. A new 18-member Advisory Committee was established representing 13 professional and scientific organizations that included consulting engineers, state officials, industry, academics, and water utility executives as well as personnel from the Food and Drug Administration and the U.S. Geological Survey. In addition, 10 officers of the USPHS formed a Technical Subcommittee that, with a six-member Task Force on Toxicology, were advisory to the main Committee (USPHS 1962). The 1962 USPHS standards were by far the most comprehensive to that date. They included three physical characteristics, odor, color, and turbidity; the last was the most controversial. The turbidity was established at 5 units over the objections of many on the committee from communities that were filtering their waters and who recommended 1 unit, which they could easily meet. Representatives from the northeast, where impounded surface sources were used without filtration, would have had to provide filtration, a measure they believed unnecessary. The bacteriological quality requirement was modified, essentially allowing no more than a monthly average of one coliform per milliliter when the membrane filter technique is used. The chemical standards were the most difficult to address. Fourteen parameters were listed, but the SOC problem was resolved with the introduction of a Carbon Chloroform Extract (CCE) standard of 0.2 mg L 1. A manual was prepared describing the procedure to be used; adsorption of organics by passing a sample of the water through a granular activated carbon (GAC) filter and then desorbing the filter with chloroform (Middleton et al. 1962). The standard was meaningless as a measure of

1.9 PRELUDE TO THE 1974 SAFE DRINKING WATER ACT

17

public health risk, because SOCs could not be distinguished from natural organics that are generally of little health consequence, except when they are precursors for chlorination and the creation of trihalomethanes (THMs). But the CCE standard was an attempt to address the SOC problem. The treatment to be provided to remove SOCs was the installation of GAC filters in the treatment train. Forty years later, only a handful of GAC filter plants are being used for treating the most vulnerable public water supplies, those drawing from run-of-river sources. It can be assumed that, at this writing, few supplies that draw from large rivers are removing SOCs that may be present. The 1962 Standards did introduce two principles that had not been incorporated in previous standards. The first was that ‘‘The water supply should be taken from the most desirable source which is feasible, and effort should be made to prevent or control pollution of the source.’’ The second issue was the absence of regulations related to availability of service. A community might be found to be violating the standards if one of the standards is not met but no violation is involved if water service is curtailed because of drought or mechanical failure. The 1962 Standards state ‘‘Approval of water supplies shall be dependent in part on: . . . adequate capacity to meet peak demands without development of low pressures or other health hazards.’’ The 1962 Standards were accepted by all the states, with minor modifications either as regulations or guidelines, but were binding only on about two percent of the communities, those that served interstate carriers (Train 1974).

1.9

PRELUDE TO THE 1974 SAFE DRINKING WATER ACT

On June 3, 1968, the keynote speaker at the Annual Conference of the AWWA quoted from a report of the Secretary of the U.S. Department of Health Education and Welfare (USDHEW 1967): ‘‘Fifty million Americans drink water that does not meet Public Health Service drinking water standards. Another 45 million Americans drink water that has not been tested by the Public Health Service.’’ The AWWA officials were reluctant to publish the paper because it appeared to be too critical of the water supply industry. They acceded only when the author happily agreed to allow rebuttals (Okun 1969). The task force that prepared the report was not satisfied that the USPHS drinking water standards adequately reflect the health need of the people of the United States. Several issues troubled them. Little information is available on the health implications of trace substances that may produce disease after exposure over long periods of time. Health experts have repeatedly pointed out that grave, delayed physical manifestations can result from repeated exposure to concentrations of environmental pollutants so small that victims do not report symptoms to a physician. Furthermore, an individually acceptable amount of water pollution, added to a bearable amount of air pollution, plus nuisances from noise and congestion, can produce a totally unacceptable health environment. It is entirely possible that the biological effects of these environmental hazards, some of which reach individuals

18

DRINKING WATER AND PUBLIC HEALTH PROTECTION

slowly and silently over decades or generations, will first begin to reveal themselves after their impact has become irreversible. In a prescient paper on cancer hazards, Hueper (1960) stated: It is obvious that with the rapidly increasing urbanization and industrialization of the country and the greatly increased demand on the present resources of water from lakes, rivers, and underground reservoirs, the danger of cancer hazards will grow considerably within the foreseeable future.

Hueper (1960) went on to report that studies in Holland revealed that cities drawing water from polluted rivers had higher cancer death rates than those taking water from higher-quality underground sources. At about the same time, the Genetic Study Section of the National Institutes of Health (NIH undated) reported that a number of widely used chemicals are known to induce genetic damage in some organisms and that chemicals mutagenic to one species are likely to be mutagenic to others. They believed that when large populations are exposed to highly mutagenic compounds, and they are not demonstrably mutagenic to individuals, the total number of deleterious mutations in the whole population over an extended period of time could be significant. In 1969, at the beginning of a review of the 1962 standards, the USPHS Bureau of Water Hygiene undertook a comprehensive survey of water supplies in the United States, known as the Community Water Supply Study (CWSS) (USPHS 1970a). A total of 969 public water systems, representing about five percent of the total number of systems in the United States serving 18 million people, about 12% of the population being served, were tested (USPHS 1970b). About 41% of the systems served did not meet the guidelines in the 1962 Standards. Deficiencies were found in source protection, disinfection, clarification, pressure in the distribution systems, and combinations of these. The small systems, mainly those serving fewer than 500 people, had the greatest difficulty in maintaining water quality. The study revealed that several million people were being supplied with water of inadequate quality and about 360,000 people were being supplied with potentially dangerous drinking water. The results of the CWSS generated interest in federal legislation that would bring all community water systems under the purview of federal regulations. In 1972, a report of an investigation of the quality of Mississippi River water, as withdrawn from the Carrolton filtration plant in New Orleans, extracted by GAC filtration and a solvent, revealed 36 organic chemicals in the finished water (USEPA 1972). Later, the U.S. General Accounting Office, an agency of the Congress, released a report of the results of an investigation of 446 community water supply systems in six states around the country and found that only 60 of them fully complied with the bacterial and sampling requirements of the 1962 Standards (Symons 1974). Bacteriological and chemical monitoring were inadequate in five of the states. In addition to government concern, public organizations and the press had begun to give attention to water supply issues. A three-part series in Consumer Reports drew attention to the organic contaminants in New Orleans drinking water (Harris

1.10 DRINKING WATER IN DEVELOPING COUNTRIES

19

and Brecher 1974) Several points were made at the outset of the series that are appropriate today: New Orleans, like many other American cities gets its drinking water from a heavily polluted source . . . . Many industries discharge their wastes into the river and many upriver cities discharge their sewage into it . . . runoff from farmland carries a wide variety of pesticides, herbicides, fertilizers, and other agricultural chemicals that swell the Mississippi’s pollution burden. Few New Orleans residents are alarmed. They have been repeatedly assured by city officials that their water, processed according to established water-treatment principles, meets the drinking water standards of the US Public Health Sevice and is ‘‘safe.’’ And so it probably is, if one takes ‘‘safe’’ to mean that the water won’t cause typhoid, cholera, or other bacterial diseases—the diseases that the standard water treatment is designed to prevent. In 1969, the Federal Water Pollution Control Administration sampled New Orleans drinking water . . . . Thirty six (organic compounds) were identified; others were found but could not be identified. Three of the organic chemicals (chloroform, benzene, and bis-chloroethyl ether) were carcinogens, shown to cause cancer in animal experiments. Three others were toxic, producing liver damage in animals when consumed even in small quantities for long periods. The long-term effects . . . are unknown.

The Environmental Defense Fund (EDF) conducted an epidemiologic study in the New Orleans area that compared cancer death rates from communities using lower Mississippi River water as a source with those from nearby communities that were using groundwater sources. The report indicating higher cancer rates among those using the Mississippi River Water was released to the press on November 7, 1974 (The States-Item 1974; Page et al. 1974, 1976). Further publicity followed on December 5, when Dan Rather on CBS aired nationally a program titled ‘‘Caution, drinking water may be dangerous to your health.’’ It is interesting to note that upon learning of this situation and the passage of the SDWA, the City of Vicksburg, which had been drawing its water from the Mississippi River, shifted its source to groundwater. These events, together with the revelation at the time that the chlorine used to make water microbiologically safe would create a family of compounds, trihalomethanes, that were themselves believed to be carcinogenic, led to the passage of the 1974 SDWA.

1.10

DRINKING WATER IN DEVELOPING COUNTRIES

The safety of drinking water cannot be examined without considering the problems of drinking water supply and safety in the countries of Asia, Africa, and Latin America. In the industrialized world, attempts are being made to eliminate the use

20

DRINKING WATER AND PUBLIC HEALTH PROTECTION

of chlorine for disinfection. Several cities in the Netherlands have abandoned chlorine and other disinfectants entirely because of their concern for DBPs. On the other hand, the situation in the developing world is so serious that the availability of chlorine for every water supply would reduce infant mortality by about 90%. In 1991, cholera broke out in the Pacific coast of Peru, most probably introduced by maritime traffic from Asia by the discharge of ballast water into the coastal zone from which fish are taken for food, often eaten uncooked. Within two weeks, most of the Peruvian coast, where half of the 22 million Peruvians reside, was host to the disease. Of the some 322,000 cases reported for the year, 55% occurred in the first 12 weeks of the epidemic. The case fatality rate was 0.9% signifying about 30,000 deaths in 1991. By the end of the year, 15 other countries in the Americas, including the United States and Canada, had reported outbreaks caused by the same strain of cholera (Salazar-Lindo and Alegre 1993). Because of its explosive and urban character, contaminated water was identified as the medium for the rapid spread and the intensity of the disease in the cities. Most of the cities had conventional water treatment plants with filtration for water drawn from surface sources. Investigation revealed that chlorination was curtailed and often entirely absent when well water was used. Some Peruvian officials blamed the USEPA for the failure to use chlorine because it had been trumpeting the cancer risks associated with chlorine in water supplies (Anderson 1991). A serious cholera outbreak occurred in early 2001 in KwaZulu-Natal in the Republic of South Africa, with more than 30,000 cases and some 100 deaths (Yahoo! 2001). At the height of the outbreak, more than 1000 cases were being reported daily. The reason stated was that the people do not have access to tapwater and are obliged to rely on water from very polluted streams. Even where ‘‘bleach’’ is available, it is not used because it is believed to interfere with fertility. Boiling is not feasible, as firewood is scarce. In 1980, only 44% of the total population of the developing countries was being served with water by any means, including carrying water of questionable quality long distances from standposts. In urban areas, 69% of the population was being served and very little of that can be considered safe because few cities maintained 24-hour service. When water pressure in distribution pipes is absent, which is most of the day, treated drinking water inevitably becomes contaminated from infiltration of groundwater that is highly contaminated because sewerage systems are absent or in poor condition. International agencies such as the World Health Organization, the World Bank, the regional development banks, and the developed countries along with the developing countries designated the 1980s the ‘‘international drinking water supply and sanitation decade,’’during which special efforts were to be made to bring water to the people of the developing world. Ten years later, the population in the developing countries with water supplies had increased to 69%, but the number of people unserved in urban areas had increased by 31 million (Okun 1991b). The rate of urbanization in Asia Africa and Latin America is so great that, even with intensified financial support in grants and loans, the number of urban residents without water service is growing. More important is that those who are counted as having water

1.11 THE FUTURE OF PUBLIC WATER SUPPLY

21

service do not have safe water by any standard. All that is required to reduce the infant death rate is the type of treatment facilities and their operation and maintenance that was conventional in the industrial world almost a hundred years ago. Given the nature of world travel today, it is clearly in the self-interest of the industrialized countries to help the developing countries provide water that at least meets 1925 U.S. standards. This would reduce infectious disease that is the major health risk to people and visitors in these countries.

1.11

THE FUTURE OF PUBLIC WATER SUPPLY

The history of the monitoring and control of drinking water quality from its earliest days through the present has lessons for those charged with protecting the public health, particularly for those responsible for providing the drinking water to their constituents. This volume demonstrates, if nothing else, that setting standards is a difficult and lengthy procedure. It may be many years, even decades between the time a new risk surfaces and regulations for its control are established and many years more before they are published. Also, years must be allowed for constructing the necessary facilities for eliminating the risks. It behooves the professionals in water utility leadership to educate themselves concerning new risks and prepare to address them before the standards appear in the Federal Register. The object is to minimize health risks to the public. Failure, or the perception of failure, drives the public to bottled water with its own risks and costs that are a hardship for a sizable fraction of the population. A not unrelated issue that is growing in importance as our population ages is the significant percent of the population that is more vulnerable to contaminants by virtue of compromised immune systems. Standards for this population may need to be promulgated. A similar solution is now being proposed in addressing the quality of water suitable for the potable reuse of wastewaters. Wastewaters contain a large number and a great variety of SOCs. The California Department of Health Services is proposing for the regulation of water quality for groundwater recharge with reclaimed wastewater to potable water aquifers drawn on for drinking water that total organic carbon (TOC) limits be set (California Code of Regulations 2001). Again, the carbon compounds may be innocuous or toxic, but in any case a maximum contaminant level (MCL) for TOC of wastewaters is hardly appropriate to assure drinking water safety. This principle carried over to the 1976 USEPA National Interim Primary Drinking Water Regulations, referred to by this language in Appendix A as ‘‘background used in developing the national interim primary drinking water regulations’’: Protection of water that poses no threat to the consumer’s health depends on continuous protection. Because of human frailties associated with protection, priority should be given to selection of the purest source. Polluted sources should not be used unless other sources are economically unavailable, and then only when personnel, equipment, and

22

DRINKING WATER AND PUBLIC HEALTH PROTECTION

operating procedures can be depended upon to purify and otherwise continually protect the drinking water supply.

This principle is being ignored today, in part because of our faith in treatment technology. Reclaimed wastewater is being proposed as a source for drinking water supplies. Wastewater is hardly likely to be the purest source, and its use for potable reuse is resisted by consumers. Use of reclaimed wastewater for nonpotable purposes is currently being practiced in many hundreds of communities in the United States (Okun 1997, 2000), and will be increasingly considered for relieving the pressure on limited high-quality resources.

REFERENCES Anderson, C. 1991. Cholera epidemic traced to risk miscalculation. Nature 354:255. Baker, M. N. 1948. The Quest for Pure Water. Denver: American Water Works Association. Bellar, T. A., J. J. Lichtenberg, and R. C. Kroner. 1974. The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66:703. Blake, N. M. 1956. Water for the Cities. Syracuse, NY: Syracuse University Press. Borchardt, J. A. and G. Walton. 1971. Water Quality and Treatment, 3rd ed. New York: McGraw-Hill. California Code of Regulations. 2001. Title 22. Draft Recycling Criteria. Sacramento: State Department of Health Services. Cosgrove, J. J. 1909. History of Sanitation. Pittsburgh, PA: Standard Sanitary Manufacturing Co. Ellms, J. W. 1928. Water Purification. New York: McGraw-Hill. Fair, G. M. 1945. Engineers and engineering in the Massachusetts State Board of Health. New Engl. J. Med. 232:443–446. Fair, G. M., J. C. Geyer, and D. A. Okun. 1968. Water Purification and Wastewater Treatment and Disposal, Vol. 2. New York: Wiley. Frontinus, S. J. A.D. 97. The Two Books on the Water Supply of the City of Rome., transl. Clemens Herschel, 1899. Boston: Dana Estes and Company. Harris, R. H. and E. M. Brecher. 1974. Is the water safe to drink? Part I. The problem. Part II. How to make it safe. Part III. What you can do. Consumer Reports 436 (June), 538 (July), 623 (August). Hueper, W. C. 1960. Cancer hazards from natural and artificial water pollutants. Proc. Conf. Physiol. Aspects Water Quality. Washington, DC: USPHS (U.S. Public Health Service). Ives, K. J. 1990. The Chadwick Centenary. The Life and Times of Sir Edwin Chadwick: 1800– 1890. London: University College. Kober, G. M. 1908. Conservation of life and health by improved water supply. Engineering Record 57. Lederberg, J. 1969. We’re so accustomed to using chlorine that we tend to overlook its toxicity. The Washington Post May 3, p. A15. Longmate, N. 1966. King Cholera: The Biography of a Disease. London: Hamish Hamilton.

REFERENCES

23

McCarthy, M. P. 1987. Typhoid and the Politics of Public Health in 19th-Century Philadelphia. Philadelphia: American Philosophical Society. McDermott, J. H. 1973. Federal drinking water standards—past, present, and future. J. Environ. Eng. Div. Am. Soc. Civil Eng. EE4(99):469. Middleton, F. M., A. A. Rosen, and Burttschell. 1962. Manual for recovery and identification of organic chemicals in water. J. Am. Water Works Assoc. 54:223–227. National Institutes of Health. Undated. Report of Chemical Mutagens as a Possible Health Hazard. Bethesda, MD: NIH Genetics Study Section. Okun, D. A. 1969. Alternatives in water supply. J. Am. Water Works Assoc. 61:5:215–224. Okun, D. A. 1991a. Clean water and how to get it. J. New Engl. Water Works Assoc. 105(1):110. Okun, D. A. 1991b. A water and sanitation strategy for the developing world. Environment 33(8):16. Washington, DC: Heldref Publications. Okun, D. A. 1996. From cholera to cancer to cryptosporidiosis. J. Environ. Eng. 122:453–458. Okun, D. A. 1997. Distributing reclaimed water through dual systems. J. Am. Water Works Assoc. 89(11):52–64. Okun, D. A. 2000. Water reclamation and unrestricted nonpotable reuse: A new tool in urban water management. Annual Rev. Public Health 21:223–245. Page, T., E. Talbot, and R. H. Harris. 1974. The Implication of Cancer-Causing Substances in Mississippi River Water. Washington, DC: Environmental Defense Fund. Page, T., R. H. Harris, and S. S. Epstein. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55. Salazar-Lindo, E. and M. Alegre. 1993. The Peruvian cholera epidemic and the role of chlorination in its control and prevention. In Safety of Water Disinfection; Balancing Chemical and Microbial Risks, G. Craun, ed. Washington, DC: International Life Sciences Institute. Snow, J. 1936. Snow on Cholera. England: Oxford Univ. Press. Symons, G. E. 1974. That GAO Report. J. Am. Water Works Assoc. 66:275. The States-Item. 1974. Cancer victims could be reduced—deaths tied to New Orleans water 98(129)1 (Nov. 7). Train, R. S. 1974. Facing the real cost of clean water. J. Am. Water Works Assoc. 66:562. United Nations. 1958. Water for Industrial Use. UN Report E=3058ST=ECA=50. New York: Economic and Social Council. United States Statutes. 1893. Interstate Quarantine Act of 1893. U.S. Statutes at Large. Chap. 114, Vol. 27, p. 449, Feb. 15. USDHEW (U.S. Department of Health, Education and Welfare). 1967. A Strategy for a Livable Environment. Washington, DC: HEW. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas: USEPA Region VI. USPHS. 1925. Report of the Advisory Committee on Official Water Standards. Public Health Reports 40:693. USPHS. 1943. Public Health Service Drinking Water Standards and Manual of Recommended Water Sanitation Practice. Public Health Reports 56:69. USPHS. 1946. Public Health Service Drinking Water Standards. Public Health Reports 61:371.

24

DRINKING WATER AND PUBLIC HEALTH PROTECTION

USPHS. 1962. Public Health Service Drinking Water Standards. Washington, DC: HEW. USPHS. 1970a. Community Water Supply Study: Analysis of National Survey Findings. PB214982. Springfield, VA: National Technical Information Service. USPHS. 1970b. Community Water Supply Study: Significance of National Findings. PB215198=BE. Springfield, VA: National Technical Information Service. Wolman, A. 1960. Concepts of policy in the formulation of so-called standards of health and safety. J. Am. Water Works Assoc. 52:11. Yahoo! Asia—News. 2001. Cholera infections sky-rocket in South Africa (Feb. 1).

2 IMPROVING WATERBORNE DISEASE SURVEILLANCE FLOYD J. FROST, Ph.D. The Lovelace Institutes, Albuquerque, New Mexico

REBECCA L. CALDERON, Ph.D. National Health and Environmental Effects Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina

GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun and Associates, Staunton, Virginia

2.1

INTRODUCTION

Public health surveillance has played a key role in controlling the spread of communicable disease and identifying the need for specific public health practices, such as the filtration and chlorination of drinking water supplies. However, the characteristics of waterborne outbreaks since the early 1990s have raised questions about whether current water treatment practices can prevent transmission of some enteric pathogens (D’Antonio et al. 1985, Hayes et al. 1989, Leland et al. 1993, MacKenzie et al. 1994). In addition, one analysis suggested that a significant fraction of all enteric disease in the United States may be due to drinking water (Bennett et al. 1987). Another study Disclaimer: The views expressed in this chapter are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The chapter has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

25

26

IMPROVING WATERBORNE DISEASE SURVEILLANCE

found evidence that consuming surface-derived drinking water which meets current U.S. Environmental Protection Agency (USEPA) drinking water standards may significantly increase the risk of enteric illness (Payment et al. 1991). These concerns have motivated the U.S. Congress to require USEPA to prepare a report on the magnitude of epidemic and endemic waterborne disease in the United States. Even as the needs increase for better information about waterborne disease occurrence and causes, some have suggested that our disease surveillance system is in a state of crisis and may possibly collapse (Berkelman et al. 1994). Another study revealed that state health departments often cannot dedicate any staff to enteric disease surveillance (Frost et al. 1995). Current concerns over the preparedness for detecting and controlling bioterrorism attacks have also motivated interest in the adequacy of waterborne disease surveillance. In this chapter, issues relating to disease surveillance and outbreak investigations are presented to assist readers in understanding the strengths and weaknesses of current waterborne disease surveillance and outbreak detection programs and to suggest additional steps to strengthen the system. With limited public health resources available, it is important to carefully consider the goals and approaches to waterborne disease surveillance. In addition to addressing the information needs of governmental disease control programs, it is essential to ensure that the information needs of the drinking water industry, the regulatory agencies, and the public are best served. It may also be essential for drinking water utilities to participate in and, perhaps, help fund these surveillance systems.

2.2

BACKGROUND

It is increasingly accepted that additional information is needed about the occurrence and causes of waterborne disease, both epidemic and endemic. The Centers for Disease Control (CDC) funded ‘‘emerging pathogen’’ surveillance projects in selected state health departments, in part to improve surveillance for several important waterborne agents. In New York City (NYC), the Department of the Environment (DEP), responsible for drinking water treatment and delivery, convened a panel of public health experts in 1994 to evaluate current health department disease surveillance programs. The panel recommended specific waterborne disease surveillance activities and epidemiologic studies to determine endemic waterborne disease risks associated with use of unfiltered surface water sources (Table 2.1) (Craun et al. 1994). Efforts to improve NYC waterborne disease surveillance are funded by the NYC DEP, the first time this has occurred for a drinking water utility in the United States. An option for improving waterborne disease surveillance is to build on the current surveillance programs in place in most state and local health departments. This system is based on voluntary disease reporting by healthcare providers and clinical laboratories. However, a number of limitations of the system have been identified, and other factors may have already significantly reduced the effectiveness of traditional disease surveillance programs. Some pathogens, such as Cryptospori-

2.2 BACKGROUND

27

TABLE 2.1 New York City Panel Recommendations on Waterborne Disease Surveillance Designate an individual who is specifically responsible for coordinating waterborne disease surveillance Conduct special surveillance studies of nursing and retirement home populations Conduct surveillance in managed care populations Monitor visits to emergency rooms Conduct surveillance of high-risk populations Monitor sales of prescription and nonprescription medications

dium, are often difficult to diagnose, and other pathogens may exist for which there are no known diagnostic tests or no tests available for routine use. Changes in healthcare access and delivery practices may reduce the number of patients seeking healthcare and, also, the chances that medically attended diseases are confirmed by laboratory tests. An outbreak resulting in many medically attended illnesses in a large city could be unrecognized, as almost happened in the Milwaukee outbreak. In that outbreak, a large increase in the occurrence of diarrheal illness occurred around March 30–31, 1993. On Thursday, April 1, 1993 a pharmacist noted a dramatic increase in sales of over-the-counter antidiarrheal and anticramping medications. Normally his pharmacy sold $30 a day of such medications. Starting that Thursday, drug sales increased to approximately $500–$600 a day, or 17–20 times the normal sales. The increased sales continued on Friday, as a result of which the pharmacy sold most of its supply of antidiarrheal medications. The pharmacist called the health department to inquire about excessive reports of diarrhea or intestinal illness. The health department was unaware of any outbreak. On Saturday the increased sales continued so the pharmacist contacted the three local television stations to report what he believed to be a major occurrence of diarrheal disease in the city. On Sunday night his report was carried on the evening news for one station and by Wednesday, April, 7, the outbreak was confirmed by the Milwaukee Health Department. In the case of the Milwaukee outbreak, few of the people sought medical care for their diarrhea. However, even in situations where care was sought, it is possible that no one physician would notice an outbreak. For example, if many different healthcare providers treated the patients, it is possible that no one provider would recognize excess occurrences of illness. In addition, the existence of health effects in a small but extremely susceptible subpopulation might be difficult to detect because of the small number of people at risk. As some changes have made it more difficult to detect outbreaks, other changes present new disease surveillance opportunities. Computerization of patient records, healthcare and laboratory workloads, prescription and nonprescription pharmaceutical sales, and calls to nurse hotlines are potential new tools for more effective and less costly disease surveillance. Technological advancements, such as detection of antigen or antibodies specific to a pathogen in sera, stools, and other secretions, may

28

IMPROVING WATERBORNE DISEASE SURVEILLANCE

improve detection of etiological agents. These may also allow detection of infections in the absence of disease. To better evaluate the current and alternative surveillance opportunities, five questions have been selected for discussion in this chapter: 1. What are the limitations of our current disease surveillance systems? 2. Should the early detection of outbreaks be the primary goal of a surveillance system and, if so, how can it be best achieved? 3. What is meant by endemic or background rates of disease, can some of this endemic disease be attributable to drinking water, and what should water utilities do to better understand these risks? 4. Can findings from outbreak investigations be used to estimate the unreported burden of enteric disease attributable to drinking water? 5. Since only a fraction of infected persons become ill from most enteric infections, should expanded surveillance programs monitor infection rather than illness?

2.3 LIMITATIONS OF THE CURRENT DISEASE SURVEILLANCE SYSTEMS What are the limitations of our current disease surveillance systems? Detection of waterborne disease outbreaks depends, in part, on a state–federal system of notifiable or reportable diseases. Disease reporting is primarily the responsibility of healthcare providers and diagnostic laboratories. State or local laws require the reporting of certain diseases. Primary responsibility for disease surveillance rests with the state or local public health authorities. Most state surveillance systems are ‘‘passive,’’ in that reports are sent to the state or local health department by cooperative health care providers or laboratories. Providers and laboratories usually receive little encouragement from the health department to report illnesses. Government enforcement of reporting requirements is minimal. An ‘‘active’’ system will routinely contact some or all healthcare providers and laboratories, asking for illness reports (Table 2.2) (Foster 1990). It has long been recognized that both passive and active disease reporting incompletely ascertain the level of disease in the community. The level of completeness varies by disease, by state, and by areas or populations within a state (Corba et al. 1989). For example, reporting is likely to be more complete for severe diseases such as hemorrhagic E. coli than for milder infections, such as Norwalk virus gastroenteritis. Laboratories tend to be much better at reporting their findings than are physicians (Foster 1990). Even within an area, there can be great variations in reporting, depending on the interest of clinical laboratories and the dedication of diagnosing physicians (Corba et al. 1989). For example, for pathogens that are new or where there are questions about the mode of transmission, reporting may be more

2.3 LIMITATIONS OF THE CURRENT DISEASE SURVEILLANCE SYSTEMS

TABLE 2.2 Mandatory reporting Passive Active

Enhanced

29

Surveillance System Definitions A diagnosed case of disease is required, by law, to be reported; for example, in the case of cryptosporidiosis, all diagnosed cases are to be reported Disease reports are submitted by providers and=or laboratories without specific follow-up by the health department Providers and=or laboratories are contacted to encourage diseases reporting; because of resource requirements, this is usually done as a special project for a limited duration of time Special additional efforts are made to encourage disease reporting; this might include news releases, posters at strategic locations, presentation to special populations, or health surveys in communities with water quality problems

complete than for agents that are common, where the mode of transmission is well known and where public health intervention is less necessary. In addition to incomplete reporting of diagnosed illnesses, only a portion of all infections will ever be medically attended. As illustrated in Figure 2.1, only a fraction of infections will lead to illness. These infected persons may be unaware of their infection. In other cases, such as sometimes occurs as a result of childhood Giardia infection, the child fails to thrive but experiences none of the classic symptoms of giardiasis. When symptoms occur, they may be mild and=or may resolve in a short period of time. In this case, the person may not seek medical care or may simply visit a pharmacy to obtain medication to alleviate their symptoms. In the case of Milwaukee, despite the large number of reported cases of cryptosporidiosis, very few people visited their physician and few stool specimens were positive for Cryptosporidium oocysts.

Figure 2.1

Disease pyramid.

30

IMPROVING WATERBORNE DISEASE SURVEILLANCE

If the person seeks healthcare, the physician may fail to correctly diagnose the infection, since in many cases symptoms are not sufficiently specific to accurately identify the pathogen. If misdiagnosed and the infection resolves itself, the patient may not seek additional healthcare and no report of an infection will be generated. Even when the physician correctly diagnoses the illness and prescribes the appropriate medication, a confirmatory laboratory test may not be ordered. If ordered, the patient may not submit the sample to the laboratory, since many patients are unwilling to submit stool specimens for laboratory analysis. Since laboratories are the primary source of disease reports for surveillance systems, without a laboratoryconfirmed diagnosis, a report may never be filed. When a stool or blood sample is submitted for laboratory analysis, it can also test negative because of analytical or specimen collection error, untimely collection or because the material submitted was, by chance, free of the pathogen (Chappell et al. 1996). Laboratory proficiency can vary considerably. This may be more of a problem for laboratories that run only a small number of the ordered test. For persons infected with enteric parasites, single stools may often be free of the parasite or have insufficient numbers of parasites to assure laboratory detection. In some cases, even multiple stools may be pathogen-negative. If a sufficient number of cases of illness from the same pathogen are reported to the health department at about the same time and if the epidemiologist is alert to an increase in case reports, an outbreak may be identified. Because of the time required to perform the diagnostic tests and to report the results, outbreak recognitions may occur weeks after the onset of the actual outbreak. Many outbreaks are first detected by an alert clinician. For example, in 1976, a Camas, Washington physician’s son had returned from Russia with giardiasis. The physician later recognized that several of his patients had similar symptoms. This lead to the identification of a waterborne giardiasis outbreak (Kirner et al. 1978). As mentioned earlier, in Milwaukee, Wisconsin a pharmacist noted a dramatic increase in sales of antidiarrheal medication. In California and Arizona, diarrheal illnesses reported to health agencies by 65 campers who had visited an Arizona park initiated an investigation that implicated contaminated water as the source of an outbreak that affected 1850 people (Starko et al. 1986). The fortuitous circumstances surrounding the detection of many outbreaks raises concerns about how many medium to large outbreaks are never detected. Small outbreaks may seldom be detected, especially among travelers who consume water from noncommunity systems or who swim in multiple locations. Limitations of the current disease surveillance systems prompted a series of studies in the early 1980s to evaluate potential improvements in disease reporting and to evaluate the efficacy of active surveillance programs. A three state study of various approaches to active disease surveillance, funded by USEPA, detected no additional waterborne disease outbreaks in two states (Washington and Vermont) (Harter et al. 1985). However, in one state (Colorado) a greater than threefold increase in the number of detected waterborne outbreaks occurred (Hopkins et al. 1985). The reasons why Colorado was able to identify so many more outbreaks than either Washington State or Vermont are unclear. An intense effort was made to

2.4 EARLY DETECTION OF OUTBREAKS

31

increase disease reporting in all states and dramatic increases in reports of enteric diseases were observed in all three states. It is possible that a combination of poor quality water supplies plus an exposed tourist population, without protective immunity, may have allowed Colorado to identify more outbreaks than the other two states. In summary, active disease reporting can increase reporting of diagnosed illnesses only from providers and laboratories. All the other barriers to disease identification and reporting will still remain (Fig. 2.1). If healthcare access declines over time or, to reduce healthcare costs, physicians use fewer laboratory diagnostic services, then the number of diagnosed reportable illnesses will decline. This will occur despite the efforts of health departments to insure that most diagnosed illnesses are reported.

2.4

EARLY DETECTION OF OUTBREAKS

Should the early detection of outbreaks be the primary goal of a surveillance system, and, if so, how can it be best achieved? The occurrence of a waterborne disease outbreak is an exciting, newsworthy, and politically important event. Affected populations may experience severe illness and a large number of people may become ill. As a result of the investigation, much is often learned about the cause of major failures in water treatment or distribution. However, when the excitement has subsided, water system deficiencies have been corrected and the outbreak is officially said to be over, has the problem been solved or is disease continuing to occur but at a reduced level, below what is detectable by traditional surveillance activities? For example, a waterborne disease outbreak investigation detected major problems with the filtration system of an anonymous small community water supply. The system was, at the time of its installation, considered adequate. However, high turbidity levels were observed in treated water at the time of the outbreak, suggesting poor operation of the filtration facility. Optimization of treatment by consulting engineers allowed the plant to dramatically improve pathogen removal. This improvement reduced the number of new cases of disease, and the outbreak officially ended. However, 2 years later a serological survey of the town’s residents revealed the continued occurrence of infection by the same etiologic agent responsible for the earlier outbreak. These new data presented both philosophical and technical problems. Should all outbreaks be followed by such a survey? Is evidence of continuing infection sufficient reason for further intervention? If the serological survey were not conducted, there would be no evidence of increase risk of infection. If the plant was already optimized, what are the remaining intervention options without new filtration or disinfection technology? This scenario assumes that the continued high serological levels resulted from waterborne transmission. In fact, without a follow-up epidemiologic investigation, it is not possible to distinguish waterborne from other routes of transmission. In addition, without improved surveillance activities, we know little about the absence of symptomatic disease. Low levels of disease from exposure to waterborne microbes over a period of many years can result in a much larger health burden

32

IMPROVING WATERBORNE DISEASE SURVEILLANCE

for a community than the number of disease cases that might occur during a detected outbreak. However, exposure to some waterborne pathogens at levels that boost the immune response may prevent symptomatic illness. These concerns must all be considered when developing a surveillance system. Without clear goals and a commitment to conduct epidemiologic investigations and take appropriate actions, a better surveillance system will not improve public health. Failure to detect low levels of disease transmission may provide a false sense of security. For example, why should an outbreak such as occurred in Milwaukee not have been preceded by many smaller outbreaks? Is it possible that in each of the cities experiencing a large waterborne cryptosporidiosis outbreak, prior undetected smaller outbreaks occurred? In fact, is it possible that lower levels of waterborne Cryptosporidium infection had occurred years prior to the outbreak? At the time of the detected outbreak, a higher number of oocysts may have passed through the treatment system or a more virulent strain of the pathogen emerged. If so, relying on disease surveillance systems that can only detect large outbreaks will seldom provide public health officials and the industry early warnings of emerging new diseases. This may be equivalent to basing the science of meteorology only on the study of hurricanes. The detection of an outbreak can also affect future disease reports in an area. For example, it is possible that overreporting of symptoms consistent with the disease of interest could occur. If so, similar outbreaks may be detected in neighboring areas. Given the increased popularity of bottled water use, it is possible that the at-risk population could change following an outbreak if a significant fraction of the population discontinued drinking tapwater. Therefore, decreases in the occurrence of reported waterborne disease may not reflect better control of the contamination but a reduction in the number of exposed individuals.

2.5

ENDEMIC DISEASE

What is meant by endemic or background rates of disease and can some of this endemic disease be attributed to drinking water? Endemic level of disease is defined by the CDC as a persistent low to moderate level of disease occurrence. A persistently high level of occurrence is called hyperendemic while an irregular pattern of occurrence is called sporadic (Fig. 2.2). For most enteric infections, endemic disease results from a statistical averaging of small to moderate-sized undetected outbreaks or clusters of infection. There is little information to suggest that endemic levels of disease remain constant over time or across geographic areas, nor is there reason to believe the endemic level of disease is unimportant. Over the past century, the importance of endemic disease has become increasingly recognized. Following World War I, an attempt was made to estimate the prevalence of parasite infections in both the returning British soldiers and the British population who remained at home (Smith and Mathews 1917). To the surprise of the researchers, a high prevalence of asymptomatic infection was found among persons who had never left Britain. Later, a survey of Wise County, Virginia in 1930 revealed that half of the population carried Entamoeba histolytica and that 38% carried

2.5 ENDEMIC DISEASE

Figure 2.2

33

Epidemic versus endemic disease.

Giardia lamblia (Faust 1930). A study to determine the incidence of Cryptosporidium infection among Peace Corp workers to be sent overseas revealed that almost 30% had possibly experienced infection prior to leaving the United States. (Ungar et al. 1989). More recent work we conducted suggests that endemic rates of Cryptosporidium infection may be very high, but that rates of cryptosporidiosis may be low (Frost 1998, Frost et al. 2001). Data derived from disease surveillance systems cannot be used to compare endemic disease levels between areas or populations with different water systems. Whether observed differences in disease reports are due to the differences in the completeness of reporting or to differences in the occurrence of the disease or the infection cannot be answered, even with improved surveillance systems. In addition, it has become increasingly recognized that populations can develop protective immunity to infectious agents. If so, rates of infection may remain high while rates of illness remain low (Frost et al. 2001). The absence of disease in a population may, therefore, not mean that there is an absence of infections. Epidemiologic studies must be specifically designed and conducted to address the association of endemic disease with water system type or quality. Several epidemiologic studies have reported waterborne disease associated with public water systems in the absence of a reported waterborne outbreak. In New Zealand, the incidence of laboratory-confirmed giardiasis was found to be higher in a part of the city receiving chlorinated, unfiltered surface water compared to the part where surface water was treated by coagulation, flocculation, granular filtration, and chlorination (Frasher and Cooke 1989). In Vermont, a higher incidence of endemic giardiasis was found in municipalities using unfiltered surface water or wells than in municipalities with filtered surface water (Birkhead and Vogt 1989). A Canadian study attempted to estimate how much endemic enteric illness was due to drinking water (Payment et al. 1991). The fraction of illness attributable to drinking water was estimated by comparing rates of reports of ‘‘highly credible gastrointestinal illnesses’’ among persons drinking tapwater with rates among

34

IMPROVING WATERBORNE DISEASE SURVEILLANCE

people drinking water from reverse osmosis filtration units. Although different rates of illness could have resulted from reporting biases, if the findings are confirmed by future studies, then drinking water may significantly contribute endemic disease in at least one community. Unfortunately, a study using a similar design conducted in Melbourne, Australia, did not provide evidence of endemic waterborne disease (Hellard et al. 2001). A variety of approaches have been proposed for estimating the burden of endemic diarrheal disease from drinking water sources. In addition to the Australian replication of the Payment design, a small pilot household intervention study in California has recently been completed (Colford et al. 2001). That study concluded that it was possible to blind families as to the type of treatment device they had, and although the study was not powered to examine illness rates, the families with true home treatment devices reported a lower rate of illness. A larger randomized household intervention study is under way in the United States. The advantage of the randomized household interventions is that the design precludes reporting biases and assignment biases, assuming that people do not know whether they are in the intervention or the control group. A major disadvantage of this approach is that only household drinking water quality is altered. Drinking water from other sources, such as work or at restaurants, is not altered. Another limitation is that long-term healthy residents are usually recruited and these people may have the lowest risk of suffering illness from waterborne infections. Therefore, negative results are difficult to interpret. Household intervention studies are limited in generalizability because they are conducted in single communities, although the study design would be amenable to national randomized trial. Another proposed approach is to relate variations in the occurrence of health events, such as emergency room visits and hospitalization, with variation in drinking water turbidity levels (Schwarz et al. 1997, Morris et al. 1998). This approach has some merit; however, the results are difficult to interpret since no causal agents are identified. There are also concerns that the optimized statistical modeling cannot be statistically evaluated. Therefore, many of the claimed associations may be spurious. Another approach uses planned changes in drinking water treatment and then evaluates the occurrence of potentially waterborne disease before and after intervention. The advantage of this approach is that most or all drinking water from an area is changed. This avoids one of the problems with household interventions. One disadvantage of this approach is that the sites receiving new water treatment technologies are not randomly assigned. For example, most unfiltered drinking water systems in the United States use high-quality source water. Adding filtration may not dramatically change the health risks from the drinking water. Another is that the community intervention looks at only one city or one pair of cities, so the sample size is restricted.

2.6

APPLICABILITY OF OUTBREAK INVESTIGATIONS

Can findings from outbreak investigations be used to estimate the burden of enteric disease attributable to drinking water? Epidemic disease is defined as an unusual

2.6 APPLICABILITY OF OUTBREAK INVESTIGATIONS

35

occurrence or clustering of a specific illness. Between 1971 and 1994 there were 737 documented waterborne disease outbreaks (Craun 1992, Moore et al. 1994, Kramer et al. 1996). Almost half of these were due to unknown etiological agents that caused acute gastrointestinal illness. Among these outbreaks, the relative importance of different etiologic agents (viruses, bacteria, protozoa, and chemicals) can be estimated. For example, the etiologic agents most commonly associated with waterborne disease in the United States include, in descending order, undefined gastroenteritis, giardiasis, shigellosis, viral gastroenteritis, and hepatitis A. This ranking is based on outbreaks and may or may not reflect the relative importance of these etiologic agents for all waterborne disease. For diseases where outbreaks account for the majority of illnesses, the outbreak is of primary interest. However, for many waterborne pathogens, outbreaks account for only a small fraction of all illnesses. For example, in a 1.5-year period during the late 1970s in Washington State, 1347 laboratory confirmed cases of giardiasis were reported to the state health department (Frost et al. 1983). Extensive follow-up of these cases (Table 2.3) revealed that clusters or possible small outbreaks accounted for only 16% of all cases of giardiasis reported during this time period. These data suggest that ‘‘endemic giardiasis’’ was overwhelmingly more abundant than ‘‘epidemic giardiasis’’ in Washington State during this time period. There are a number of problems with extrapolating the characteristics of cases involved in outbreaks to revise all cases of illness, including the following: 1. If there is variation in the virulence of a pathogen, then detected outbreaks may predominantly be caused by the more virulent strains of the pathogen. This may overestimate the severe morbidity or mortality associated with the pathogen. 2. By examining only detected outbreaks, one may overestimate the importance of drinking water as a route of transmission. Because of the large number of cases often involved, waterborne outbreaks may be more detectable than TABLE 2.3 1977–1978

Case Clusters of Giardiasis in Washington State

Number of Cases 10 14 11 12 17 8 24 73 51 220

Etiology Untreated streamwater consumption Untreated water consumption at a work camp One small community water system Tourists returning from a resort in Mexico One-daycare center outbreak One-daycare center outbreak Among 10 different daycare centers Multiple cases among 21 families Nonfamily association with another case Total in all clusters

36

IMPROVING WATERBORNE DISEASE SURVEILLANCE

outbreaks from other routes of transmission. Even a severe day care outbreak would involve only a few cases. Within family clusters usually involve too few cases to be a detectable outbreak. 3. Outbreak detection is often more difficult for common or endemic diseases than for uncommon diseases. For example, two cases of cholera anywhere in the United States might be considered an outbreak whereas 50 cases of cryptosporidiosis widely dispersed in a large U.S. city during a week might easily be absorbed as expected background cases of diarrhea and not recognized as an outbreak (Craun et al. 1994). Outbreaks of short duration of illnesses (e.g., some viruses) are more difficult to detect and study than are outbreaks of long duration illnesses (e.g., giardiasis, shigellosis, hepatitis A). Therefore the importance of acute, self-limited gastrointestinal illness of undetermined etiology and short duration may be underestimated relative to outbreaks of parasitic infections and some bacterial or viral pathogens with a longer duration of symptoms. Pathogens with long incubation periods are difficult to investigate since the conditions that allowed transmission of the pathogen may have changed between the time of infection and the time when the outbreak was detected. Underascertaining waterborne sources for disease outbreaks caused by these agents is likely.

2.7

MONITORING INFECTION VERSUS DISEASE

Only a fraction of infected persons become ill from the most commonly occurring enteric infections. Of the people that become ill, only a fraction of cases will be reposted (Fig. 2.3). Should expanded surveillance programs attempt to monitor infection rather than disease? The existence of asymptomatic carriers of infections has been known for some time (e.g., Typhoid Mary). However, the number of asymptomatic carriers for many infections has only relatively recently been appreciated. The parasite prevalence surveys in Britain (Smith and Matthews 1917) and in Virginia (Faust 1930) found more asymptomatic infected persons than expected. Even as late as 1952, in New Hope, Tennessee, 10.6% of the general population was infected with Giardia lamblia (Eyles et al. 1953). Following a 1966 giardiasis outbreak in Aspen, Colorado, a stool survey found that 5% of the population was infected with Giardia (Gleason et al. 1970). A survey of Boulder, Colorado, also conducted following an outbreak, found a prevalence of 5% (Wright et al. 1977). Most of the individuals participating in these surveys were asymptomatic. A stool survey of one to 3-year-old Washington State children was conducted in 1980 (Harter et al. 1982). This survey found that 7% of the children were infected with Giardia lamblia. All participating children were reported as healthy at the time of the survey. The Seattle Virus Watch program, conducted during the 1960s and early 1970s monitored virus infections among a sample of people in selected U.S. cities. This study found that illness was reported in less than half of all enterovirus infections (Elveback et al. 1966).

2.7 MONITORING INFECTION VERSUS DISEASE

Figure 2.3

37

Events in reporting an individual infection.

New serological tools have been developed since the early 1980s to better monitor the prevalence of prior infections among the population. Even though infection may not result in moderate or severe illness, there are several reasons for considering infection rather than disease, including the following: 1. Information on infections can provide a much expanded understanding of the relative importance of various routes of transmission and provide an early warning for risks of outbreaks. 2. Serological epidemiologic studies of infection can better estimate the extent of endemic waterborne disease. These studies are statistically more powerful to detect low risks in moderate-size populations. 3. Just as the occurrence of a coliform test indicates the potential of disease risk for a drinking water source, the waterborne transmission of pathogens, even when infection is predominantly asymptomatic, can provide critical information for evaluating water treatment systems and may help identify correctable problems in water source protection and=or treatment. 4. Widespread, unrecognized transmission of infection in the general population may indicate a devastating outbreak for a susceptible subpopulation.

38

IMPROVING WATERBORNE DISEASE SURVEILLANCE

Another advantage of serological surveillance occurs during an outbreak. An outbreak of cryptosporidiosis was detected in Las Vegas, Nevada in Spring 1994. Although this was clearly a cryptosporidiosis outbreak, the inability to detect problems with the water treatment system and publicity prior to the investigation that suggested the outbreak was waterborne raised questions over whether the outbreak could be classified as waterborne (Craun and Frost 2002, Craun et al. 2001, Rodman et al. 1998). Since the majority of the diagnosed cases also suffered from HIV or AIDS, the extent of the outbreak was unclear. Had asymptomatic infected persons been identified serologically, the effects of reporting bias would be reduced since asymptomatic cases would have no motivation to explain an asymptomatic infection.

2.8

IMPROVING DISEASE SURVEILLANCE

Several options are available for enhanced waterborne disease surveillance. The option or combination of options selected will depend on the specific goals for disease surveillance. The currently used national system of surveillance, based on diagnosed illness, has a long-established record of both performance and nonperformance for detecting outbreaks (Table 2.4). Because the current system is both inexpensive to maintain and currently operational, it has considerable appeal among public health practitioners. However, monitoring pharmaceutical sales, nurse hotline calls, or physician visits is a potential enhancement to the traditional disease surveillance programs (Table 2.5) (Rodman et al. 1997, 1998). TABLE 2.4 Advantages and Disadvantages of the Current Waterborne Disease Surveillance System Advantages In-place and operational across the nation Extensive health department experience using the system Inexpensive to maintain An operational nationwide network, operated by the Centers for Disease Control (CDC), for summarizing and reporting findings Methodological development of algorithms for detecting excess occurrences of disease Disadvantages Inability to detect outbreaks when diagnosed cases are not reported to the health department Delays in detecting outbreaks due to the time required for laboratory testing and for reporting of findings Undetected outbreaks where health problems are not medically treated or where infection results in only mild or no illness Limited opportunities for system improvement Possible long-term trend in healthcare delivery that may reduce its efficacy

2.8 IMPROVING DISEASE SURVEILLANCE

TABLE 2.5 Systems

39

Advantages and Disadvantages of New Waterborne Surveillance

Advantages They may detect outbreaks where few patients seek healthcare or where the illness is of sufficiently short duration that healthcare is unimportant They are relatively fast in reporting outbreaks since the time delay between the onset of symptoms and the purchase of drugs or calls to nurses is likely to be short They are relatively inexpensive to maintain, especially if nationwide retail pharmacies are involved or common nurse hotline software is programmed for reporting Disadvantages Since only symptoms are ascertained, they will not usually identify an etiologic agent Although inexpensive to maintain, initial computer programming and establishing data sharing agreements would require some investment The specificity of the system for outbreak detection (e.g., number of false leads) is untested

The goal of our current disease surveillance system is outbreak detection. Unfortunately, there is little rigorous evaluation of its capability to detect outbreaks. Furthermore, the common occurrence of fortuitous situations that lead to the outbreak detection raise questions about the sensitivity of the system. To improve the sensitivity to detect small to medium-size outbreaks or to provide early information on the occurrence of an outbreak, these alternative approaches mentioned have promise. Over-the-counter pharmaceutical sales may be useful, but it has some significant limitations (Rodman et al. 1997). The use of nurse hotline calls to continuously monitor the occurrence of infectious disease has tremendous promise, but no efforts have been made to use this surveillance tool (Rodman et al. 1998). Better linkages with infectious disease specialists in healthcare organizations may also improve disease surveillance. None of the traditional or enhanced surveillance tools will provide much useful information on low-level or endemic risk of enteric pathogen infection. However, new serological tests have increased the feasibility of studies to estimate the incidence of new infections or the prevalence of antibody response to pathogens and to relate this information with modes of transmission. In the early 1970s, the Seattle Virus Watch program examined occurrences of viral infections among volunteers in selected communities (Gleason et al. 1970). Similar approaches to monitoring the occurrence of Giardia (Nulsen et al. 1994) and Cryptosporidium (Moss and Lammie 1993) infections have been developed since then. More work is needed to evaluate these new tools as well as to develop other tests. We also need to design cost-effective approaches to their widespread implementation. These tools may give us an opportunity to greatly improve our understanding of the importance of various

40

IMPROVING WATERBORNE DISEASE SURVEILLANCE

modes of transmission and identify reasons why one population group has a higher endemic level of disease than another. It is likely that as more is known about the modes of transmission, a better understanding will emerge of both drinking water and nondrinking water routes of pathogen transmission. Healthcare reforms may reduce the use of diagnostic laboratory services, reducing the value of laboratory-based disease surveillance. However, new opportunities for improved disease surveillance, including both individual and community disease reporting and surveillance of endemic infections, may also result. To fully exploit these opportunities, a new public health partnership with distributed responsibilities may be needed between healthcare providers, health maintenance organizations (HMOs), pharmacies, and the traditional public health agencies. The increasing age of our population has resulted in increases in the number of immunosuppressed persons. Some of this immunosuppression may result from chronic diseases, while some may result from medically induced immunosuppression following treatment for other conditions. For example, many cancer patients have temporary periods of immunosuppression following treatment. These populations may be at especially high risk of adverse consequences of infection. Since diarrheal disease in this population is also relatively common, many infections may not be detected. Infectious disease surveillance systems are operated by state and local public health agencies with little or no direct contact with healthcare providers. To improve disease surveillance system, it will likely be necessary to better integrate healthcare delivery systems with those disease surveillance programs. This integration can only occur if both the state public health agencies and the healthcare providers recognize benefits from this cooperation and barriers to data sharing are reduced.

REFERENCES Bennett J. V., S. D. Holmberg, and M. F. Rogers. 1987. Infectious and parasitic diseases. Closing the Gap: The Burden of Unneccessary Illness. Edited by R. W. Amler and H. B. Dull. New York: Oxford University Press. Berkelman, R. L., R. T. Bryan, M. T. Osterholm, J. W. LeDuc, and J. M. Hughes. 1994. Infectious disease surveillance: A crumbling foundation. Science 264:368–370. Birkhead, G. and R. L. Vogt. 1989. Epidemiologic surveillance for endemic Giardia lamblia infection in Vermont. Am. J. Epidemiol. 129:762–768. Chappell, C. L., P. C. Okhuysen, C. R. Sterling, and H. L. DuPont. 1996. Cryptosporidium parvum: Intensity of infection and oocyst excretion patterns in healthy volunteers. J. Infect. Diseases 173:232–236. Colford, J. M., J. R. Rees, T. J. Wade, A. Khalakdina, J. F. Hilton, I. J. Ergas, S. Burns, A. Benker, C. Ma, C. Bowen, D. C. Mills, D. J. Vugia, D. D. Juranek, and D. A. Levy. 2001. Participant blinding and gastroinntestinal illness in a randomized, controlled trial of an in-home drinking water intervention. Emerging Infect. Diseases 8(1):29–36. Chorba, T. L., R. L. Berkelman, S. K. Safford, N. P. Gibbs and P. E. Hull. 1989. Mandatory reporting of infectious diseases by clinicians. J. Am. Med. Assoc. 262:3018–3026.

REFERENCES

41

Craun, G. F. 1992. Waterborne disease outbreaks in the United States of America: Causes and prevention. World Health Statistics Quart. 45:192–196. Craun, C. L., G. Birkhead, S. Erlandsen, et al. 1994. Report of New York City’s Advisory Panel on Waterborne Disease Assessment. New York: The New York City Department of Environmental Protection. Craun, G. F., F. J. Frost, R. L. Calderon, H. Hilborne, K. R. Fox, D. J. Reasoner, C. Poole, D. J. Rexing, S. A. Hubbs, and A. P. Dufour. 2001. Improving waterborne disease outbreak investigations. Int. J. Environ. Health Research, 11:229–243. Craun, G. F. and F. J. Frost. 2002. Possible information bias in a waterborne outbreak investigation. Int. J. Environ. Health Research, 12:5–15. D’Antonio, R. G., R. E. Winn, J. P. Taylor, T. L. Gustafson, G. W. Gray, W. L. Current, R. A. Zajac and M. M. Rhodes. 1985. A waterborne outbreak of cryptosporidiosis in normal hosts. Annals Int. Med. 103:886–888. Elveback, L. R., J. P. Fox, A. Ketler, C. D. Brandt, F. E. Wassermann, and C. E. Hall. 1966. The Virus Watch program; a continuing surveillance of viral infections in metropolitan New York families. 3. Preliminary report on association of infections with disease. Am. J. Epidemiol. 83:436–454. Eyles, D. E., F. E. Jones, and S. C. Smith. 1953. A study of Entamoeba histolytica and other intestinal parasites in a rural west Tennessee community. Am. J. Trop. Med. 2:173–190. Faust, E. C. 1930. A study of the intestinal protozoa of a representative sampling of the population of Wise County, southwestern Virginia. Am. J. Hygiene 11:371–384. Foster, L. R. 1990. Surveillance for waterborne illness and disease reporting: State and local responsibilities. In Methods for Investigation and Prevention of Waterborne Disease Outbreaks, G. F. Craun, ed. EPA=600=1-90=005a. Cincinnati: USEPA Office of Research and Development. Frasher, G. G. and K. R. Cooke. 1989. Endemic giardiasis and municipal water supply. Am. J. Public Health 79:39–41. Frost, F. 1998. Two-city Cryptosporidium study. Am. Water Works Assoc. Research Found.— Drink. Water Research 8(6):2–5. Frost, F. J., R. L. Calderon, R. L., and G. L. Craun. 1995. Waterborne disease surveillance: Findings of a survey of state and territorial epidemiology programs. J. Environ. Health. 58(5):6–11. Frost, F., L. Harter, B. Plan, K. Fukutaki, and B. Holman. 1983. Giardiasis in Washington State. USEPA Report 83-134-882. Springfield, VA: National Technical Information Service. Frost, F. J., T. Muller, G. F. Craun, R. L. Calderon, and P. A. Roeffer. 2001. Paired city Cryptosporidium serosurvey in the southwest USA. Epidemiol. Infect. 126:301–307. Frost F. J., T. Muller, G. F. Craun, D. Fraser, D. Thompson, R. Notenboom, and R. L. Calderon. 2000. Serological analysis of a cryptosporidiosis epidemic. Internatl. J. Epidemiol. 29:376–379. Gleason N. N., M. S. Horwitz, L. H. Newton, and G. T. Moore. 1970. A stool survey for enteric organisms in Aspen, Colorado. Am. J. Trop. Med. Hygiene 19:480–484. Harter, L., F. Frost, and W. Jakubowski. 1982. Giardia Prevalence among 1-to-3 year-old children in two Washington State counties. Am. J. Public Health 72:386–388. Harter, L., F. Frost, R. Vogt, A. Little, R. Hopkins, B. Gaspard, and E. Lippy. 1985. A threestate study of waterborne disease surveillance techniques. Am. J. Public Health 75: 1327–1328.

42

IMPROVING WATERBORNE DISEASE SURVEILLANCE

Hayes, E. P., T. D. Matte, T. R. O’Brien, T. W. McKinley, G. S. Logsdon, J. B. Rose, B. L. P. Ungar, D. M. Word, P. F. Pinksky, M. L. Cummings, M. A. Wilson, E. G. Long, E. S. Hurwitz, and D. D. Jaranek. 1989. Large community outbreak of cryptosporidiosis due to contamination of a filtered public water supply. New Engl. J. Med. 320:372–376. Hellard, M. E., M. I. Sinclair, A. B. Forbes, and C. K. Fairley. 2001. A randomized blinded controlled trial investigating the gastrointestinal health effects of drinking water quality. Environ. Health Perspect. 109:773–778. Hopkins, R. S., P. Shillam, B. Gaspard, L. Eisnach and R. S. Karlin. 1985. Waterborne disease in Colorado: three years surveillance and 18 waterborne outbreaks. Am. J. Public Health 75:254–257. Kirner, J. C., J. D. Littler, and L. A. Angelo. 1978. A waterborne outbreak of giardiasis in Camas. J. Am. Water Works Assoc. 70:35–40. Kramer, M. H., B. L. Herwaldt, G. F. Craun, R. L. Calderon, and D. D. Juranek. 1996. Surveillance for waterborne disease outbreaks—United States, 1993–1994. J. Am. Water Works Assoc. 88:66–80. Leland, D., J. McAnulty, W. Keene, and G. Terens. 1993. A cryptosporidiosis outbreak in a filtered water supply. J. Am. Water Works Assoc. 85:34–42. MacKenzie, W. R., N. J. Hoxie, M. E. Proctor, M. S. Gradus, K. A. Blair, D. E. Peterson, S. S. Kazmierczak, D. G. Addiss, K. R. Fox, J. B. Rose, and J. P. David. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public drinking water supply. New Engl. J. Med. 331(3):161–167. Moore, A. C., B. L. Herwaldt, G. F. Craun, R. L. Calderon, A. K. Highsmith, and D. D. Juranek. 1994. Waterborne disease in the United States, 1991 and 1992. J. Am. Water Works Assoc. 84(2):87–99. Morris, R. D. F., E. N. Naumova, and J. K. Griffiths. 1998. Did Milwaukee experience waterborne cryptosporidiosis before the large documented outbreak in 1993? Epidemiology 9:264–270. Moss, D. M. and P. J. Lammie. 1993. J. Am. Soc. Trop. Med. Hygiene 49:393. Nulsen, M. F., P. G. Tilley, L. Lewis, H. Z. Zhang and J. L. Isaac-Renton. 1994. The humeral and cellular host immune responses in an outbreak of giardiasis. Immunol. Infect. Diseases 4:100–105. Payment, P., L. Richardson, J. Siemiatycki, et al. 1991. A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiologic standards. Am. J. Public Health 81(6):703–708. Rodman, J. S., F. Frost, I. D. Burchat, D. Fraser, J. Langer, and W. Jakubowski. 1997. Pharmacy sales—a method of disease surveillance. J. Environ. Health 60(4):8–14. Rodman, J. S., F. J. Frost, and W. Jakubowski. 1998. Using nurse hot line calls for disease surveillance. Emerg. Infect. Diseases 4:1–4. Schwarz, J., R. Levin, and K. Hodge. 1997. Drinking water turbidity and pediatric hospital use for gastrointestinal illness in Philadelphia. Epidemiology 8:615–620. Smith, A. M. and J. R. Matthews. 1917. The intestinal protozoa of non-dysenteric cases. Annals Trop. Med. Parasitol. 10:361–390. Starko, K. M., E. C. Lippy, L. B. Dominguez, C. E. Haley and H. J. Fisher. 1986. Campers’ diarrhea traced to water-sewage link. Public Health Reports 101:527–531.

REFERENCES

43

Ungar, B. L., M. Milligan, and T. B. Nutman. 1989. Serologic evidence of Cryptosporidium infection in U.S. volunteers before and during Peace Corps service in Africa. Arch. Int. Med. 149:894–897. Wright, R. A., H. C. Spender, R. E. Brodsky, and T. M. Vernon. 1977. Giardiasis in Colorado: An epidemiologic study. Am. J. Epidemiol. 105:330–336.

3 WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000 GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun & Associates, Staunton, Virginia

REBECCA L. CALDERON, Ph.D. U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Research Triangle Park, North Carolina

MICHAEL F. CRAUN, P.E., M.S. Gunther F. Craun & Associates, Staunton, Virginia

3.1

INTRODUCTION

In this chapter, the causes of 1010 waterborne outbreaks reported in the United States during the period 1971 to 2000 are reviewed. Most (74%) of the outbreaks were associated with contaminated drinking water; 648 outbreaks were reported in public drinking water systems, and 103 were reported in individual water systems. An additional 259 (26%) outbreaks were associated with water recreation, primarily swimming. Disclaimer: The views expressed in this article are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The article has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

45

46

3.2

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

WATERBORNE DISEASE OUTBREAK SURVEILLANCE SYSTEM

National statistics on waterborne outbreaks have been reported in the United States since 1920. In 1971, the Centers for Disease Control and Prevention (CDC), the U.S. Environmental Protection Agency (USEPA), and the Council of State and Territorial Epidemiologists began a collaborative surveillance program for the collection and reporting of data on the occurrence and causes of waterborne outbreaks (Lee et al. 2002, Barwick et al. 2000, Levy et al. 1998, Kramer et al. 1996, Moore et al. 1993, Herwaldt et al. 1993, Craun 1990). Each year, state and territorial epidemiologists or persons designated as waterborne outbreak surveillance coordinators voluntarily report information about waterborne outbreaks to the CDC and USEPA. The surveillance system records information about the epidemiology of the outbreak, etiologic agents, types of water system, system deficiencies, water sources, and water quality. These surveillance data are useful for evaluating the relative degrees of risk associated with different types of source water and systems and the adequacy of current technologies and regulations. To be defined as a waterborne outbreak, at least two persons must have experienced a similar illness after the ingestion of drinking water or after exposure to water used for recreational purposes, and epidemiologic evidence must implicate water as the probable source of the illness. The exceptions are single case outbreaks of chemical poisoning (e.g., methemoglobinemia), if water-quality data indicate contamination by the chemical, and single case outbreaks of laboratory-confirmed, primary amebic meningoencephalitis. Waterborne outbreaks are classified according to the strength of the epidemiological evidence implicating water and the available information about water quality, sources of contamination, and system deficiencies. Epidemiological information is weighted more heavily than information about water quality or the water system. The circumstances of each outbreak investigation differ. Not all outbreaks are rigorously investigated, and information is often incomplete. Even when adequate information is available to classify the outbreak, the investigation may not have been optimal in terms of an epidemiological, engineering, or water quality evaluation. Water samples are usually collected for bacteriological or chemical contaminants, but in only a few outbreak investigations are attempts made to isolate a pathogen from water samples. Most reported outbreaks were associated with water used or intended for drinking or domestic purposes, but outbreaks were also associated with the ingestion of water not intended for consumption (e.g. the use of springs and creeks by backpackers and campers, and accidental ingestion of water while swimming, diving, or other waterassociated recreation). Recreational waters encompass swimming pools, water parks, interactive play fountains, wading pools, and naturally occurring fresh and marine surface waters. Although the surveillance system records whirlpool- and hot tubassociated outbreaks of dermatitis, these outbreaks are not included in the analysis presented here. The surveillance system collects information about but does not record waterborne outbreaks caused by contamination of water or ice at its point of use (e.g., a contaminated water faucet or serving container). Outbreaks caused by contaminated ice, faucets, and containers are included in the analysis. Point-of-use

3.2 WATERBORNE DISEASE OUTBREAK SURVEILLANCE SYSTEM

47

outbreaks, along with outbreaks caused by consumption of water not intended for drinking, are classified under miscellaneous causes. Although waterborne outbreak surveillance data are useful for evaluating the adequacy of approaches for providing safe drinking and recreational water, the data underestimate the true incidence of waterborne outbreaks. Not all outbreaks are recognized, investigated, and reported to the CDC or USEPA, and the extent to which these outbreaks are not recognized and not reported is largely unknown. The likelihood that individual cases of illness will be detected and epidemiologically associated with water is dependent on many factors including a) public awareness, b) the likelihood that persons who are ill will consult the same rather than different health-care providers, c) availability and extent of laboratory testing, d) local requirements for reporting cases of particular diseases, and e) the surveillance and investigative activities of state and local public health and environmental agencies. The states that report the most outbreaks during any single year might not be those where the most outbreaks occur. Outbreaks of acute illness characterized by a short incubation period are more readily identified. Outbreaks involving serious illness are most likely to receive the attention of health authorities and to be investigated. Outbreaks associated with community water systems are more likely to be recognized than those associated with non-community systems because the latter serve mostly nonresidential areas and transient populations. Outbreaks associated with individual systems are the least likely to be recognized and reported because they generally involve few persons. Recreational water and non-community outbreaks that result from persons congregating in one venue then dispersing are often difficult to recognize and investigate. Each water system associated with a waterborne outbreak was classified as a public or individual system and as having one of the following deficiencies: untreated surface water; untreated groundwater; treatment deficiency (e.g., temporary interruption of disinfection, inadequate disinfection, and inadequate or no filtration); distribution system deficiency (e.g., cross-connection, contamination of water mains during construction or repair, and contamination of a storage facility); and unknown or miscellaneous deficiency (e.g., contaminated ice, faucets, containers, or bottled water). Water sources were identified as either surface water, groundwater, or mixed (both surface water and groundwater sources). Water not intended for drinking was classified as an individual water system, since the person consuming the water chose to drink from these sources. Although non-community systems were not classified as transient or non-transient, most of these systems primarily served visitors and would be considered non-transient. In outbreaks associated with water used for recreation, information was collected about the venue, suspected source of contamination, and factors that may have contributed to the outbreak. State and local health departments and other agencies have jurisdiction over recreational waters, including public swimming and wading pools. The USEPA has established a guideline (monthly geometric mean must be
33=100 mL for enterococci or
126=100 mL for Escherichia coli.) for microbial water quality of fresh waters (e.g., lakes, ponds) used for recreational activities (Lee et al. 2002, Dufour 1984, Cabelli 1983). However, states and localities can have

48

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

either more or less stringent guidelines or regulations and can post warning signs to alert potential bathers until water quality improves. The CDC recently cautioned consumers about health risks associated with swimming and wading pools and how to reduce these risks (CDC 2001).

3.3 3.3.1

WATERBORNE OUTBREAK STATISTICS Type of Water System

During the 30-year period 1971–2000, 308 outbreaks and 517,944 illnesses were reported in community water systems; 340 outbreaks and 54,893 illnesses were reported in non-community systems (Table 3.1). An estimated 21,740 persons became ill during 259 outbreaks associated with water recreation; 103 outbreaks and 1600 cases of illness were reported in individual water systems. In the United States, public water systems are classified as either community or non-community and are regulated by the USEPA. A community water system serves year-round residents of a community, subdivision, or mobile-home park that has 15 or more service connections or an average of 25 or more residents. A non-community water system is used by the general public for greater than 60 or more days per year and at least 15 service connections or serves an average of 25 or more persons. Non-community water systems are also classified as non-transient and transient. Non-transient systems (e.g., in factories and schools with their own water systems) serve 25 or more persons for at least six months of the year. Transient non-community water systems do not serve at least 25 of the same persons over six months per year (e.g., many restaurants, rest stops, parks). Individual water systems, which are not owned or operated by a water utility and serve less than 15 connections or less than 25 persons, are not regulated by the USEPA; each state or county develops regulations for these systems. The reporting of waterborne outbreaks varied considerably during the 30-year period (Figure 3.1). More outbreaks (n ¼ 220) were reported during the five-year period of 1976–80 than during any other 5-year period. The largest number (n ¼ 90) TABLE 3.1 Waterborne Outbreaks Reported in the United States by Type of System or Activity, 1971–2000 Water System Type Non-Community Community Recreational Individual All Water Systems

Outbreaks

Cases of Illness

Emergency Room Visits

Hospitalizations

Deaths

340 308 259 103 1010

54,893 517,944 21,740 1600 596,177

116 904 36 5 1061

868 1056 206 93 2223

4 65 28 3 100

3.3 WATERBORNE OUTBREAK STATISTICS

49

Figure 3.1 Waterborne outbreaks reported in the United States by type of water system or water recreation, 1971–2000. I ¼ individual water system; NC ¼ non-community water system; C ¼ community water system; R ¼ recreation water.

of community system outbreaks was reported during 1981–85, and the smallest number (n ¼ 24) was reported during 1996–2000. Slightly more than half (56%) of the outbreaks in community systems and almost half (46%) of the outbreaks in non-community systems were reported during the ten-year period 1976–85. Most (62%) of the outbreaks associated with recreational water have been reported since 1991. Apparent trends in the occurrence of outbreaks are most likely due to differences in the recognition, investigation, and reporting by public health agencies during the 30-year period. Because of recent publicity about waterborne pathogens such as Cryptosporidium and E. coli 0157:H7, both the public and health officials are more aware of the potential for an outbreak when persons suffer gastrointestinal symptoms, especially during or after swimming. The magnitude of waterborne illness reported in outbreaks varied considerably during the 30-year period (Figure 3.2). In community water systems, outbreaks resulted in as few as 59 illnesses in 1997 to as many as 404,013 illnesses in 1993. In 1997, three community outbreaks caused an average of 20 illnesses, but a single outbreak of cryptosporidiosis in Milwaukee (MacKenzie et al. 1994) was responsible for 403,000 illnesses in 1993. This is the largest number of illnesses reported in a single outbreak since the collection of waterborne outbreak statistics began in 1920. During each of the five-year periods through 1995, more than 10,000 illnesses were reported in community systems. In 1980, 18,958 illnesses were reported in 26 community system outbreaks. In 1991, and 1987, two and eight outbreaks in community systems caused 10,049 and 16,922 illnesses. During each of the five-year periods through 1995, more than 5000 illnesses were reported in non-community systems; in each of two five-year periods (1981–85, 1986–90), more than 10,000 illnesses were reported. In 1983, 12,007 illnesses were reported

50

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.2 Cases of illness reported in drinking water and water recreational outbreaks, 1971–2000.

3.3 WATERBORNE OUTBREAK STATISTICS

51

in nine non-community system outbreaks, and in 1998 only one outbreak of four illnesses was reported. The mean and median numbers of illness reported in community system outbreaks during the 30-year period were 17,265 and 1979, respectively compared to a mean of 1830 and median of 1316 illnesses in non-community systems. The large mean for community system outbreaks is due primarily to the Milwaukee outbreak. On average, each community system outbreak caused 1682 illnesses; however, if the Milwaukee outbreak is excluded, this statistic is reduced to 373 illnesses per outbreak. Outbreaks in non-community and individual systems each caused, on average, 161 and 16 illnesses per outbreak. Outbreaks in individual water systems were generally small; however, during four of the five-year periods, more than 200 illnesses were reported. Outbreaks in individual water systems resulted in a mean of 53 and median of 33 illnesses. Almost one-fourth of the illnesses in individual systems were reported in 2000. In outbreaks associated with recreation water, most cases of illness were reported during 1991–2000; almost half (49%) of the illnesses occurred during the three years 1994, 1995, and 1996. Water recreation was responsible, on average, for 84 illnesses per outbreak. As anticipated, outbreaks associated with recreation water occurred primarily in the summer months; 90% of the outbreaks occurred during June, July, and August (Figure 3.3). Most (71%) outbreaks in non-community systems occurred from May through September. Outbreaks in individual water systems occurred primarily (61%) from June to September. It is not certain whether outbreaks in individual systems were associated with increased water contamination during the late spring and summer months or increased exposures during these months. A recent paper suggested precipitation events might contribute to the increased risk of outbreaks (Rose et al. 2000). Increased exposures for more susceptible persons are probably an important factor for the large number of outbreaks in non-community systems, whereas in individual systems, the increased potential for water contamination is probably more important. Although fewer outbreaks were reported during the months of February and December, no seasonal distribution was apparent for outbreaks in community system.

3.3.2

Type of Water Source

A disproportionate percentage of outbreaks were reported in public water systems that use surface water sources. Although only slightly more than 13,000 (8%) public water systems use surface water sources, 186 (29%) outbreaks were reported in these systems (Table 3.2). Over 156,000 (92%) systems rely primarily on ground water sources; 402 (62%) of the outbreaks in public systems were associated with ground water sources. A water source was not identified for 53 outbreaks in non-community and community systems, and in seven systems, both surface and ground water sources were used. In recreation water outbreaks, most outbreaks occurred in surface water sources. Of the 259 outbreaks reported, 126 (63%) were associated with recreational activities in lakes, ponds, rivers, and streams; 104 (40%) were associated with swimming and wading pools (Table 3.3).

52

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.3

Seasonal distribution of waterborne outbreaks, 1971–2000.

3.3 WATERBORNE OUTBREAK STATISTICS

53

TABLE 3.2 Waterborne Outbreaks in Drinking Water Systems by Type of System and Water Source, 1971–2000 Number of Outbreaks

Water Source Ground Watera Surface Waterb Mixed Unknown Totals a b

Community Systems

Non-Community Systems

Individual Systems

All Systems

130 142 6 30 308

272 44 1 23 340

65 23 1 14 103

467 209 8 67 751

Surface water ¼ lakes, reservoirs, rivers, streams. Ground water ¼ wells and springs.

3.3.3

Outbreak Etiologies

Ninety-one (9%) outbreaks and 4517 ( < 1%) cases of illness were classified as acute chemical poisoning while a bacterial, viral, or protozoan etiology was identified in 513 (51%) outbreaks and 505,189 (85%) cases of illness (Tables 3.4, 3.5). In 405 (40%) outbreaks and 86,457 (14%) cases of illness, an infectious etiologic agent was suspected but not identified. Illnesses associated with drinking water outbreaks included acute gastroenteritis due to a wide variety of pathogens, typhoid fever, hepatitis, and cholera. Illnesses associated with recreational water outbreaks included acute gastroenteritis, dermatitis, primary amebic meningoencephalitis, leptospirosis, otitis externa, pharyngitis, typhoid fever, and hepatitis. In community systems, most outbreaks were caused by protozoa (31%) and undetermined etiologic agents (32%). Chemical contaminants caused 18% and bacterial agents caused 13% of the outbreaks in community systems. Most (67%) outbreaks in non-community systems were classified as acute gastroenteritis of TABLE 3.3 Waterborne Outbreaks Associated with Water Recreation Activities, 1971–2000 Water Source Lakes, Ponds Swimming and Wading Pools Othera Rivers, Streams Springs Water Slides and Wave Pools Interactive Water Fountains Unknown Totals a

Number of Outbreaks 116 104 14 10 7 4 3 1 259

Canals, puddles, ocean, dunking booth at fair, waste water holding pond, and mixed sources.

54

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

TABLE 3.4

Waterborne Outbreaks by Type of Etiology, 1971–2000 Number of Outbreaks of Specified Etiology

Water System Type

Unidentified Agents

Protozoa

Viruses

Bacteria

Chemical

Non-Community Community Recreationala Individual All Water Systems

228 98 40 39 405

31 96 97 16 240

27 20 18 9 74

43 40 97 18 198

11 54 5 21 91

a

One outbreak attributed to algae and one outbreak attributed to bacteria and protozoa not included in table.

undetermined etiology. Bacterial agents caused most of the remaining non-community system outbreaks. Protozoan and bacterial agents caused most (75%) outbreaks associated with water recreation. Recognizing that any observed trends may be due to reporting differences and other reasons (e.g., availability of laboratory facilities), the etiologies of waterborne outbreaks were examined during the 30-year period (Figure 3.4). Fewer outbreaks of acute gastroenteritis of unknown etiology and viral etiology are now being reported. An etiologic agent was identified in 76% of all outbreaks reported during 1996–2000, but in the previous 25-year period, an etiologic agent was identified in only 57% of the outbreaks. Protozoan and bacterial agents caused 60% of the outbreaks reported during 1991–2000 but only 36% of the outbreaks reported during 1971–1990. In the late 1970s and early 1980s, Giardia was an important waterborne protozoan. Although Giardia continues to cause outbreaks, Cryptosporidium was the more important waterborne protozoan in the 1990s. A viral agent was identified in 9% of the outbreaks reported during 1971–1990 but only 5% of outbreaks reported during 1991–2000.

TABLE 3.5

Cases of Waterborne Illness by Type of Etiology, 1971–2000 Cases of Illness in Outbreaks of Specified Etiology

Water System Type

Unidentified Agents

Protozoa

Viruses

Bacteria

Chemical

Community Non-Community Recreationala Individual All Water Systems

48,320 34,371 2966 800 86,457

445,882 3907 12,701 136 462,626

5435 10,076 1433 247 17,191

14,633 5830 4548 323 25,334

3674 709 40 94 4517

a One outbreak attributed to algae (14 cases) and one outbreak attributed to bacteria and protozoa (38 cases) not included in table.

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

55

Figure 3.4 Trend in identifying etiologic agents in waterborne outbreaks, 1971–2000. AGI ¼ acute gastroenteritis of undetermined etiology. One recreational outbreak attributed to algae (1981) and one recreational outbreak attributed to both bacteria and protozoa (1999) were excluded from the analysis.

3.3.4

Severity of Illness

The severity of illness also varied during the 30-year period. Although information was not always available about hospitalizations or emergency room treatment, illness was severe enough in 239 outbreaks for 2223 persons to be admitted to the hospital. Various etiological agents were responsible for the hospitalizations; bacterial agents were identified for almost half (46%) of the hospitalizations (Figure 3.5). In 21 outbreaks, an additional 1061 persons were treated in a hospital emergency room. One hundred deaths were associated with the reported outbreaks; most (83%) were associated with outbreaks in community system or recreational waters (Table 3.1). In Milwaukee, it was estimated that 50 deaths were associated with the cryptosporidiosis outbreak. Twenty-seven persons died from primary amebic meningoencephalitis after becoming infected with Naegleria fowleri while swimming or diving. Sixteen deaths were associated with bacterial infections, and six deaths were attributed to acute chemical poisoning.

3.4 3.4.1

CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS Etiology of Drinking Water Outbreaks

Giardia was the most frequently identified etiologic agent for outbreaks reported in public water systems (Table 3.6). Giardia was responsible for 83 (27%) of the outbreaks in community systems and 29 (9%) of the outbreaks in non-community

56

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.5 Hospitalizations in waterborne outbreaks by etiologic agent, 1971–2000. AGI ¼ acute gastroenteritis of undermined etiology. One recreational outbreak attributed to both bacteria and protozoa (4 hospitalizations) was excluded from the analysis.

TABLE 3.6 1971–2000

Etiology of Waterborne Outbreaks by Type of Drinking Water System,

Community Water Systems Etiologic Agent Undetermined Giardia Chemical Shigella Cryptosporidium Salmonella, non-typhoid Hepatitis A virus Campylobacter Norwalk virus Escherichia coli 0157:H7 Rotavirus Salmonella, typhoid Cyclospora V. cholerae E. histolytica Yersinia Plesiomonas shigelloides Escherichia coli 06:H16 SRSV Total a

Non-Community Water Systemsa

Individual Water Systems

Outbreaks Illnesses Outbreaks Illnesses Outbreaks Illnesses 98 83 54 14 11 11 10 9 9 4 1 1 1 1 1

308

48,320 25,001 3674 5715 420,856 3044 241 5353 3433 451 1761 60 21 11 4

517,944

228 29 11 24 2 2 10 7 16 4

34,371 3329 709 3417 578 72 369 120 9637 66

39 14 21 6 2 2 8 3 1 3

800 97 94 64 39 87 217 132 30 12

1

210

3

12

1

17

1 1 1 1 339

8787 60 1000 70 54,112

1

16

103

1600

One outbreak of Escherichia coli and Campylobacter with 781 cases is not included in this table.

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

57

systems. Shigella, Hepatitis A, non-typhoid Salmonella, Norwalk-like viruses, Campylobacter, and Cryptosporidium were identified as etiologic agents in 64 (21%) community system outbreaks. Shigella, Hepatitis A, Norwalk-like viruses, and Campylobacter were identified in 57 (17%) outbreaks in non-community systems. In individual systems, Giardia was the second most frequently (14%) identified etiologic agent causing outbreaks; Shigella and Hepatitis A were also important etiologic agents. Chemical contaminants caused 20% of the outbreaks in individual systems and 18% of the outbreaks in community water systems. Acute chemical poisonings were caused by arsenic, benzene, chlordane, chlorine, chromium, copper, ethyl acrylate, ethylene glycol, fluoride, gasoline, hydroquinone, lead, morpholine, nitrate=nitrite, oil, polychlorinated biphenyls, phenol, selenium, liquid soap, sodium hydroxide, sodium metaborate, trichloroethylene, and unidentified herbicides. Nineteen chemical poisonings were reported in community water systems during 1991–2000. High levels of copper caused eight outbreaks, and high levels of fluoride caused three outbreaks. The remaining outbreaks were caused by nitrite, chlorine, sodium hydroxide, sodium metaborate, and liquid soap. Five outbreaks of gastroenteritis were reported in Wisconsin where high copper levels were found in new and remodeled homes with copper pipe (Levy et al. 1998, Kramer et al. 1996). In Pennsylvania, elevated copper levels in a hotel were associated with at least 43 illnesses (Kramer et al. 1996). In two other outbreaks, improper wiring and plumbing caused the leaching of copper from restaurant pipes, and a defective check value and power outage at the water treatment facility led to the release of high levels of sulfuric acid which caused corrosion and leaching of copper (Barwick et al. 2000). A large outbreak of acute fluoride poisoning in 262 persons and was attributed to improperly installed equipment and inadequate monitoring which resulted in excessive fluoride levels (Moore et al. 1993). In two other outbreaks, excessive levels of fluoride were siphoned into the system due to inadequate controls at the feed pump and due to a cross-connection; fluoride levels of 200–220 mg=L were measured in these outbreaks (Kramer et al. 1996). Thirty persons developed chemical burns in their mouths after they drank water contaminated with sodium hydroxide accidentally released from a surface water treatment plant (Levy et al. 1998). In a similar outbreak, lack of a check valve allowed approximately 200 gallons of sodium hydroxide to spill into a community well over a period of several hours (Lee et al. 2002). Although some 100–1000 persons may have been exposed, only two persons reported illness. In Florida, one person became ill after drinking water obtained from a drive-through window of a restaurant; median chlorine levels were 4.5 mg=L (Levy et al. 1998). The source of the high chlorine levels was not determined. In two outbreaks of nitrite poisoning, defective check valves for the prevention of backflow allowed chemicals used to treat water in a chilling system and a boiler to contaminate the drinking water system (Levy et al. 1998). Among employees and visitors to a hospital cafeteria, seven persons became ill 1–5 minutes after drinking a carbonated beverage (Lee et al. 2002). The investigation discovered a cross-connection in the plumbing system that might have allowed water from the cooling tower, which had been recently shock-treated with sodium metaborate, to enter the drinking water system. Sodium metaborate has been associated with nitrate

58

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

poisoning and methemoglobinemia in past incidents (Lee et al. 2002). Thirteen persons at a health-care facility developed burning in their mouths after drinking water contaminated with a concentrated liquid soap (Levy et al. 1998). Vacuum breakers to prevent backsiphonage had been incorrectly installed at soap dispensers. Since 1991, seven chemical poisonings have been reported in individual systems. Four outbreaks of methemoglobinemia were attributed to nitrate contamination of wells. In one outbreak, high levels of nitrate were found in treated well water where a reverse-osmosis membrane filter was used to reduce nitrate levels in the water source (Moore et al. 1993). Three single-case outbreaks involving infants with high blood levels were detected through a lead screening program (Herwaldt et al. 1991). Water stored at the homes of the infants was found to be corrosive, leaching lead from fittings and lead-soldered seams in the storage tanks.

3.4.2

Water System Deficiencies

Distribution system deficiencies and inadequate or interrupted disinfection of unfiltered surface water caused slightly more than half (52%) of the outbreaks in community systems (Table 3.7). Contaminants entered the distribution system through cross-connections, backsiphonage, corrosion and leaching of metals, broken or leaking mains, storage deficiencies, and construction or repair of mains. The most important chemical contaminant causing distribution-system outbreaks was copper. Outbreaks of acute illness occurred primarily as a result of corrosion in home and building plumbing systems and plumbing deficiencies in soft drink mixing machines. Microbial contaminants also entered distribution systems to cause outbreaks. Most infectious disease outbreaks were caused by unidentified pathogens; the most frequently identified pathogen in distribution system outbreaks was Giardia. In unfiltered surface water systems, disinfection was either inadequate to inactivate waterborne protozoa or interrupted so that undisinfected water was distributed. In many instances, the source water quality was such that filtration should have been provided to remove protozoan cysts and oocysts in addition to disinfection. When surface waters are filtered, however, care should be taken to ensure that facilities are adequately designed and operated. Inadequate or interrupted filtration of surface water sources was responsible for almost 10% of the outbreaks in community systems. Inadequately treated or untreated ground water caused about one-fourth of the outbreaks. The remaining outbreaks in community systems were associated with the use of untreated surface water, inadequate control of chemicals added during treatment, and miscellaneous=undetermined causes. In non-community water systems, almost three-quarters of the outbreaks were reported in ground water systems that were either not disinfected or inadequately disinfected (Table 3.7). Although fewer non-community systems use surface water sources and most non-community distribution systems are less extensive and complicated than those in community systems, outbreaks were still associated with inadequate treatment of surface water (1%) and distribution system deficiencies (7%).

59

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

TABLE 3.7 2000

Waterborne Outbreaks and Deficiencies in Public Water Systems, 1971–

Community Systems Type of Contamination Distribution System Contamination Inadequate or Interrupted Disinfection; Disinfection Only Treatment, Surface Watera Inadequate or Interrupted Disinfection; Disinfection Only Treatment, Ground Water Untreated Ground Water Inadequate or Interrupted Filtration, Surface Water Miscellaneous=Unknown Inadequate Control of Chemical Feed Untreated Surface Water Inadequate or Interrupted Filtration, Ground Water Inadequate Control of Disinfection Total a

Non-Community Systems

Outbreaks

Percent

Outbreaks

96

31.2

24

7.1

64

20.8

22

6.5

42

13.6

101

29.7

34 30

11.0 9.7

140 4

41.2 1.2

21 11

6.8 3.6

32 4

9.4 1.2

6 3

1.9 1.0

13 —

3.8 —

1

0.3

—

—

340

100

308

100

Percent

Three outbreaks mixed source.

More than 71% of the outbreaks occurred in individual systems that did not provide treatment of ground or surface water sources (Table 3.8). In only two of the reported outbreaks, were the wells disinfected. It is not known how many individual systems in the United States are disinfected, but it is suspected that the few outbreaks reported in disinfected systems reflect the small number of individual systems that are disinfected. Distribution system contamination also caused outbreaks in individual systems (8%). 3.4.3

Water Quality During Outbreaks

Water quality information was reviewed for 665 known or suspected infectious disease outbreaks in public and individual water systems to determine whether coliform bacteria were detected during the outbreak investigation. In many of the investigations in public water systems, more coliform samples were collected than would be required by the Total Coliform Rule (TCR) (USEPA 1989a) for routine monitoring purposes. Water samples were collected during a relatively short period of time, often during a one to two week period. Water samples were usually collected during

60

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

TABLE 3.8 Waterborne Outbreaks and Deficiencies in Individual Water System, 1971–2000 Type of Contamination

Outbreaks

Percent

55 19 17 8 2

53.4 18.4 16.5 7.8 1.9

1 1

1.0 1.0

Untreated Groundwater Untreated Surface Water Miscellaneous=Unknown Distribution System Contamination Inadequate or Interrupted Disinfection; Disinfection Only Treatment of Ground Water Inadequate Chemical Removal Inadequate or Interruption of Filtration, Ground Water Total

103

100

the early stages of an outbreak, but in some investigations, samples were collected several weeks to a month after the beginning of the outbreak. Water samples were sometimes collected from well water sources rather than the distribution system. When such samples were obtained from undisinfected or inadequately disinfected well water systems, information from source water samples was used as a substitute for tap water samples. It was presumed that tap water samples, if they had been collected, would likely provide similar results. Information was available about the presence of coliform bacteria in the water system during the investigation of 459 (69%) outbreaks in public and individual water systems (Table 3.9). Total and=or fecal coliforms were detected during 359 (78%) of these outbreaks. Coliforms were detected during the investigation of 84% of non-community system outbreaks and 94% of individual system outbreaks but during only 65% of community system outbreaks.

TABLE 3.9 Total Coliform Data Collected During Drinking Waterborne Outbreak Investigations, 1971–2000a Number of Outbreaks

Water System Non-community Community Individual Totals a

Outbreaks of Known or Suspected Infectious Etiology 329 254 82 665

Outbreaks with Total Coliform Data 241 160 49 459

(73%) (67%) (60%) (69%)

Outbreaks where Total or Fecal Coliforms were Detected 203 110 46 359

(84) (65) (94%) (78%)

Excluded from analysis are chemical outbreaks and outbreaks associated with water recreation.

3.5 OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

61

Water samples were collected for pathogen analysis during the investigation of 81 outbreaks. Pathogens were isolated from water samples in 72 outbreaks. Cryptosporidium and Giardia were the pathogens most frequently isolated from these water samples, however, E. coli, Shigella, Salmonella, Campylobacter, Yersinia, and enteric viruses were also isolated.

3.5

OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

An etiologic agent was identified in most (85%) water recreation outbreaks (Table 3.10). The four most frequently identified etiologic agents in water recreation outbreaks were Cryptosporidium (15%), Pseudomonas (14%), Shigella (13%), and N. fowleri (11%). 3.5.1

Lakes

Most outbreaks were associated with swimming and bathing activities in lakes, ponds, and reservoirs, and Shigella was the most frequently (24%) identified etiology of these outbreaks. Other important illnesses associated with bathing in lakes included primary amebic meningoencephalitis caused by N. fowleri, gastroenteritis caused by E. coli 0157:H7 and Norwalk-like viruses, and Schistosoma dermatitis (swimmer’s itch). Fourteen (50%) of the twenty-eight single-case outbreaks of amebic meningoencephalitis were associated with swimming in lakes or ponds. The remaining cases were associated with swimming in rivers and canals, facial immersion in a puddle during a fight (Moore et al. 1993), and bathing in a hot spring that had been associated with two previous cases (Moore et al. 1993). N. fowleri infections are generally acquired during the summer months, when the temperature of fresh water is favorable for the multiplication of the organism. The ameba can enter a person’s body through the nasal passages when water is forced up the nose, especially during underwater swimming and diving (Barwick et al. 2000). Four outbreaks of schistosomal dermatitis were associated with swimming in Oregon lakes; in two outbreaks, geese were the suspected source of these parasites. Other outbreaks of schistosomal dermatitis occurred after swimming in lakes in California, New Jersey, Utah, and Wyoming. One swimming-associated dermatitis outbreak was associated with ocean water in Delaware where local snails were found to contain cercariae of Austrobilharzia variglandis, an avian schistosome implicated as a cause of cercarial dermatitis (Moore et al. 1993). E. coli 0157:H7 caused fifteen outbreaks, eleven (75%) of which occurred while swimming in lakes. 3.5.2

Pools

A significant number of outbreaks were associated with swimming or wading pools at various locations including community centers, parks, water theme parks, motels, country clubs, day-care centers, schools, hospitals. The three most frequently identified etiologies of outbreaks in pools were Cryptosporidium (32%), Pseudomonas

62 2243 14 77 334 252 572 404 595 40

11

7559

29 14 4 11 12 7 4

1 1

1

116

9 40 81 30 20

1 5 2 2 1

6 3 13,795

589 51

10 3

1 1 110

11,058 518 1325 65

Cases

36 9 35 4

Outbreaks

Swim Poola,b

b

Includes wading pools and pools and other activities at water parks and interactive water fountains. One outbreak of shigella and cryptosporidium (38 cases) at an interactive water fountain not included. c Includes dunking booth, natural springs, canal, ocean, unknown, and mixed sources.

a

649 2368

Cases

4 28

Outbreaks

Lakes or Pond

32

1 2 1

3 1 1 14 2 1 1 2 3

Outbreaks

348

6 11 14

80 50 21 14 18 2 30 80 22

Cases

River and Other c

40 39 36 34 28 16 15 13 10 7 5 3 3 2 2 1 1 1 1 258

Outbreaks

11,707 2966 1375 2329 28 684 387 282 661 426 40 676 70 26 11 14 11 6 3 21,702

Cases

All Recreational

Etiology of Recreational Waterborne Outbreaks, Outbreaks and Cases of Illness by Type of Water Source, 1971–2000

Cryptosporidium Undetermined Pseudomonas Shigella Naegleria Giardia Escherichia coli 0157:H7 Schistosoma Norwalk-like virus Leptospira Chemical Adenovirus Enterovirus Hepatitis A Salmonella, typhoid Microcoleus Escherichia coli 0121:H19 Campylobacter Jejuni Salmonella, non-typhoid Total

Etiologic Agent

TABLE 3.10

3.5 OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

63

(32%), and Giardia (9%). Three pool-associated E. coli outbreaks were reported. One outbreak was associated with swimming in a poorly maintained and poorly chlorinated indoor pool, and another occurred among children using an unchlorinated wading pool where a fecal accident had occurred. In the remaining outbreak seven of 23 ill persons developed hemolytic uremic syndrome; a fecal accident in a children’s pool in the water park was suspected to be the cause. Outbreaks of chemical dermatitis or keratitis were associated with bromine, chlorine, incorrect dosing of chemicals to adjust the pH of swimming pool water, and the addition of chemicals to remove excess chloramines. An outbreak of Salmonella enterica was reported among persons using a scuba dive pool that had been filled with fish. 3.5.3

Recreational Outbreaks Reported Since 1991

Since most (62%) of these outbreaks were reported during 1991–2000, they were examined more closely. Thirty-five (90%) of the 39 outbreaks caused by Cryptosporidium since 1991 were associated with treated recreational water including swim pools, wading pools, water slides, wave pools, and interactive fountains; only four outbreaks were associated with swimming in lake water. Four of the cryptosporidiosis outbreaks and 8485 cases of illness occurred in water theme parks. One cryptosporidiosis outbreak occurred at a local zoo where 369 persons became ill after playing in a sprinkler fountain (Barwick et al. 2000). The fountain was originally designed as a decorative fountain and had become a popular interactive play area for children. Water was sprayed through the air, drained through grates, collected, passed through a sand filter, and chlorinated and recirculated. Another outbreak where illness was linked to playing in an interactive fountain was attributed to both Cryptosporidium and Shigella. The fountain’s recirculation, filtration, and disinfection systems were inadequate or not completely operational. Samples of the fountain water were positive for coliform bacteria (Lee et al. 2002). Since 1991, Shigella has caused 15 outbreaks, thirteen of which were associated with swimming in lake water. One S. sonnei outbreak was associated with a wading pool that included a sprinkler fountain. The system recirculated chlorine-treated water, and many diaper-aged children were observed sitting in the wading pool. Six of the ten outbreaks of gastroenteritis caused by Norwalk-like virus were reported during 1991–2000; three were associated with swimming in lakes and two with bathing in hot springs. Fecal coliforms were detected in the water during the investigation of the three lake outbreaks, but the source of contamination was not identified. In 1991, an outbreak of leptospirosis was associated with swimming in a rural pond in Illinois; Leptospira interrogans was found in urine specimens from cases and in pond water. The largest outbreak of leptospirosis ever reported in the United States occurred in Illinois during 1998 (Barwick et al. 2000). Among competitors in a triathlon, 375 persons became ill after swimming in a lake; 28 were hospitalized. The most recent outbreak of leptospirosis was reported among 21 persons who participated in an adventure race in Guam in July 2000. These persons reported multiple outdoor exposures, including running through jungles and savannahs,

64

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

swimming in a river and a reservoir, and bicycling and kayaking in the ocean. Leptospira was confirmed by serology, and an epidemiologic investigation demonstrated that swimming in the reservoir, submerging one’s head in the water, and swallowing water while swimming were risk factors for illness. Adenovirus serotype 3 was implicated from clinical and water samples collected in a 1991 outbreak of 595 persons with conjunctivitis, pharyngitis, and fever after swimming in a pond in North Carolina. Information was available about the source of contamination and other factors that contributed to 85 of the recreational outbreaks reported since 1991 (Table 3.11). Forty-six outbreaks were associated with pools, and 39 outbreaks were associated with untreated surface water. Multiple possible causes were noted in many of the outbreaks, and each potential source of contamination or deficiency was tabulated in Table 3.11. Poor maintenance and operation (e.g., inadequate chlorination or filtration, excessive application of pool chemicals) were identified in 23 (50%) of 46 outbreaks associated with swimming or wading pools. Fecal accidents or use of pools by diaper-age children were suspected or identified in 24 (52%) poolassociated outbreaks. Outbreaks associated with lakes, ponds, rivers, canals, and other waters used for recreational purposes were caused by a variety of problems. Fecal accidents or ill bathers were responsible for 14 (36%) of 39 untreated surface-water-associated outbreaks. Other suspected causes of surface-water outbreaks included contamination from diapers (28%), over crowding (28%), animal or bird contamination (23%), and sewage contamination of the bathing area (13%). Floodwaters contaminated a canal that was used by children for swimming in one outbreak, and heavy rains contaminated a spring in another outbreak. Fecal accidents were identified or suspected in 32 outbreaks. Fourteen (44%) of these outbreaks were caused by Cryptosporidium, six (19%) outbreaks were caused by E. coli 0157:H7, and five (16%) outbreaks were caused by Shigella. The remaining outbreaks where fecal accidents were identified were caused by an undetermined

TABLE 3.11 Causes of Waterborne Disease Outbreaks Associated with Recreational Water During 1991–2000 Source of Contamination or Deficiency Fecal Accident, Ill Bathers Poor Maintenance, Inadequate Treatment, or Operation of Swimming or Wading Pool Children in Diapers Bather Overload or Crowding Animals Seepage or Overflow of Sewage Floods a

Some outbreaks have multiple deficiencies.

Number of Outbreaks Containing Deficiencya 32 23 20 15 10 6 2

3.6 OUTBREAK TRENDS

65

infectious agent (9%), Norwalk-like virus (6%), E. coli 0121:H19 (3%), and Giardia (3%).

3.6

OUTBREAK TRENDS

Distribution system contamination was the most frequent (31%) identified cause of outbreaks reported in community systems. Unfiltered surface water and inadequate or interrupted filtration of surface water sources were responsible for 21% and 12% of the outbreaks in community systems. Inadequately treated ground water and untreated ground water caused 14% and 11% of the outbreak deficiencies. Since 1995, distribution system contamination has increased in importance as a cause of outbreaks. During 1995–2000, distribution system deficiencies caused almost half (47%) of all outbreaks reported in community water systems while unfiltered surface water caused only 3% of the outbreaks. During this same period, inadequate or interrupted filtration of surface water sources was responsible for 9% of the outbreaks in community systems, and untreated or inadequately disinfected ground water caused 9% and 18% of the outbreaks reported during 1995–2000. These statistics suggest that USEPA regulations and other actions have helped decrease the importance of unfiltered surface water systems as a cause of outbreaks. However, distribution system deficiencies as a cause of outbreaks increased in importance, and the contamination of ground water sources and inadequate operation of surface water filtration facilities continued to be important causes of outbreaks. Additional regulations and increased attention may be needed to reduce these outbreaks risks. Most distribution-system-related outbreaks were associated with cross-connections or problems with backflow prevention devices (i.e., they had not been installed, had been inappropriately installed, or inadequately maintained). Other outbreaks were caused by contaminants entering through mains and storage facilities or leaching of metals from plumbing and pipes because of corrosive water. To reduce the risk of outbreaks, increased attention should be paid to maintaining the integrity of the distribution system, reducing corrosion byproducts, and preventing contamination from cross-connections and backsiphonage, inadequately protected storage reservoirs and tanks, and the repair and construction of water mains. The maintenance of a chlorine residual throughout the system can also help protect against many sources of distribution system contamination. Frequent monitoring of chlorine residuals in the system is important as a way to detect contamination sources. Water utilities should consider a program to investigate potential distribution system contamination when chlorine residuals decline or suddenly disappear. The relative importance of inadequately filtered surface water during recent periods is similar to previous periods. The Interim Enhanced Surface Water Treatment Rule (USEPA 1998) should help reduce outbreak risks in filtered surface water systems, however, operators should also strive to maintain filtration efficacy. In non-community water systems, most of the outbreaks during the most recent period were reported in ground water systems that were inadequately protected from

66

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

sources of contamination and inadequately disinfected. Contaminated ground water also continues to be an importance cause of outbreaks in community systems. Wells and springs should be protected from surface water run-off, septic-tank effluents, and other sources of contamination, and the location of wells should consider potential sources of contamination. Periodic sanitary surveys, along with appropriate corrective measures, and a hydrogeologic assessment can help identify systems with a high possibility of fecal contamination. The location of wells should consider potential sources of contamination. The disinfection of ground water may also be required to reduce the occurrence of outbreaks, particularly for small systems where intermittent contamination of wells and springs is difficult to detect or prevent. When disinfection is provided, it must be adequate in terms of concentration and contact time based on the anticipated contamination. Disinfection must also not be interrupted. USEPA has proposed ground water requirements for sanitary surveys, disinfection of groundwater for vulnerable systems, and additional source water monitoring (USEPA 2000). Despite improved drinking water treatment for surface waters, Giardia and Cryptosporidium continue to pose waterborne risks in the United States. Cryptosporidiosis outbreaks are increasingly being reported in ground water systems. Outbreaks of giardiasis continue to be associated with both surface and ground water systems. Among the outbreaks reported in drinking water systems during 1991–1994, four (50%) of the cryptosporidiosis outbreaks and five (56%) giardiasis outbreaks were associated with ground water contamination. In 1997 and 1998, two of the three reported giardiasis outbreaks occurred in well water systems, and in 1998, both of the reported outbreaks of cryptosporidiosis were caused by sewage contamination of well water. During 1999–2000, two cryptosporidiosis and six giardiasis outbreaks were reported. One of the cryptosporidiosis outbreaks was associated with distribution system contamination; the other was caused by contamination at the point of water use. Two of the giardiasis outbreaks occurred in untreated ground water systems and two in filtered systems where filtration was by-passed. The remaining outbreaks were associated with distribution system contamination and use of water not intended for drinking. Recent outbreaks of Giardia and Cryptosporidium emphasize the importance of assessing sources of ground water contamination. Ground water sources found to be contaminated by protozoa or subject to the direct influence of surface water must meet the filtration requirements of the Surface Water Treatment Rule (SWTR) (USEPA 1989b). The continued occurrence of protozoa outbreaks in surface water systems emphasizes the importance of requiring water systems to meet USEPA’s new turbidity standards and other provisions regarding filtration efficacy (USEPA 1998). More stringent USEPA regulations for acceptable turbidity values for surface water systems have become effective since the Milwaukee outbreak in 1993 and many water utilities have conducted composite performance evaluations and corrections programs to optimize treatment plant performance to consistently achieve good removal of microorganisms and turbidity. Water contamination was documented in the majority of public water systems during the outbreak investigation. Although these statistics show that the coliform

3.6 OUTBREAK TRENDS

67

test can detect water contamination during an outbreak investigation, there are acknowledged problems with using coliform bacteria as an indicator for waterborne protozoa. Studies have suggested that the USEPA’s TCR monitoring and maximum contaminant level may not be effective in assessing the outbreak vulnerability of public water systems (Craun et al. 1997, Nwachuku et al. 2002). The number of samples required for non-community and small community systems may not be sufficient, and coliform monitoring of the distribution system alone may not be adequate to assess outbreak vulnerability. Dose-response studies have shown that relatively few organisms of Cryptosporidium, Giardia, Shigella and E. coli 0157:H7 are required to cause infection. Thus, the unintentional ingestion of a single mouthful of contaminated water while swimming and bathing could cause illness, even in non-outbreak settings (Calderon et al. 1991, Seyfried et al. 1985). Most recreational outbreaks occurred while swimming in lakes and swimming pools that were contaminated by either bather crowding, fecal accidents, or children in diapers. Swimming pool outbreaks were associated with inadequate treatment and poor maintenance and operation. Outbreaks attributed to bacteria, such as Shigella and E. coli 0157:H7, were associated primarily with swimming in fresh water (i.e., lakes, ponds, reservoirs). In contrast, most of the outbreaks caused by Cryptosporidium and Giardia were reported in chlorinated, filtered pool water. USEPA has published criteria for evaluating the quality of both marine and fresh water used for recreation (Lee et al. 2002, Dufour 1984, Cabelli 1983). Fresh and marine waters are subject to contamination not only from bathers but also from sewage discharges, watershed runoff from agricultural and residential areas, and floods. Microbial monitoring has been recommended for recreational areas potentially contaminated by sewage. Overt fecal accidents and soiled bodies can also cause fecal contamination of the water, however, the utility of routinely monitoring water for fecal contamination caused by bathers has not been established. Efforts have focused on providing adequate toilet and diaper-changing facilities at recreational areas, requiring showers before bathing, and limiting the number of bathers. Although difficult to enforce, an important measure is to prevent persons, especially young children from entering recreational waters if they are experiencing or convalescing from a diarrheal illness. Limiting the amount of water forced into the nasal passages during jumping or diving (e.g., holding the nose or wearing nose plugs) could reduce the risk for primary amebic meningoencephalitis. Cryptosporidium and Giardia are resistant to disinfection at levels generally used in swimming pools, and some pool filtration systems might not be effective in removing oocysts. Even pools with filters and disinfection practices capable of removing or killing these parasites may require hours or even a day to completely recirculate and disinfect the pool water once it becomes contaminated. Swimmers remain at risk until all of the water is recirculated through an effective water treatment process. Although the reporting of outbreaks is incomplete and the accuracy of case counts vary, waterborne outbreak surveillance data has helped identify the types of water systems, their deficiencies, and the respective etiologic agents associated

68

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

with the outbreaks. These data are important for evaluating the adequacy of current source water protection strategies, water treatment technologies and drinking and recreational water regulations and for influencing research priorities. Because the surveillance system is voluntary and does not include data for sporadic cases of disease that may be waterborne, the statistics do not reflect the true incidence of waterborne outbreaks or disease. Observed trends in the occurrence of waterborne outbreaks are likely to be a reflection of surveillance activities of local and state health agencies. Waterborne outbreaks continue to occur in the United States despite additional regulatory requirements, increased monitoring efforts, and improved water treatment facilities. These statistics provide a reminder that in this new century, waterborne problems of the previous centuries are likely to continue. New challenges remain from emerging and re-emerging pathogens that can quickly be transported with relatively ease from one part of the globe to another. Thus, vigilance and advanced preparation are needed to solve not only old problems but also anticipate new ones.

REFERENCES Barwick, R.S., D.A. Levy, G.F. Craun, M.S. Beach, and R.L. Calderon. 2000. Surveillance for waterborne-disease outbreaks—United States, 1997–1998, Morbid. Mortal. Weekly Report 49(SS-4):1. Cabelli, V.J. 1983. Health Effects Criteria for Marine Recreational Waters. EPA publication 600=1-80-031. Research Triangle Park, N.C.: USEPA. Calderon, R.L., E.W. Mood, and A.P. Dufour. 1991. Health effects of swimmers and nonpoint sources of contaminated water. Internatl. J. Environ. Health Research 1:21. CDC. 2001. Notice to readers: responding to fecal accidents in disinfected swimming venues. Morbid. Mortal. Weekly Report 50(20):416–417. Craun, G.F., ed. 1990. Methods for the Investigation and Prevention of Waterborne Disease Outbreaks. EPA 600=1-90=005a. Cincinnati, OH: USEPA. Craun, G.F., P.S. Berger, and R.L. Calderon. 1997. Coliform bacteria and waterborne disease outbreaks. J. Am. Water Works Assoc. 89(3):96. Dufour, A.P. 1984. Health Effects Criteria for Fresh Recreational Waters. EPA 600=1-84-004. Research Triangle Park, N.C.: USEPA. Herwaldt, B.L., G.F. Craun, S.L. Stokes, and D.D. Juranek. 1991. Waterborne-disease outbreaks, 1989–1990. Morbid. Mortal. Weekly Report 40(SS-3):1. Kramer, M.H., B.L. Herwaldt, G.F. Craun, R.L. Calderon, and D.D. Juranek. 1996. Surveillance for waterborne-disease outbreaks—United States, 1993–1994. Morbid. Mortal. Weekly Report 45(SS-1):1. Lee, S.H., D.A. Levy, G.F. Craun, M.J. Beach, and R.L. Calderon. 2002. Surveillance for waterborne-disease outbreaks—United States, 1999–2000. Morbid. Mortal. Weekly Report 51(SS-8):1. Levy, D.A., M.S. Bens, G.F. Craun, R.L. Calderon, and B.L. Herwaldt. 1998. Surveillance for waterborne-disease outbreaks—United States, 1995–1996. Surveillance for waterbornedisease outbreaks—United States, 1993–1994. Morbid. Mortal. Weekly Report 47(SS-5):1.

REFERENCES

69

MacKenzie, W.R., N.J. Hoxie, M.E. Proctor, M.S. Gradus, K.A. Blair, D.E. Peterson, J.J. Karmierczak, D.G. Addiss, K.R. Fox, J.B. Rose, and J.P. David. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. New Eng. J. Med. 331:161. Moore, A.C., B.L. Herwaldt, G.F. Craun, R.L. Calderon, A.K. Highsmith, and D.D. Juranek. 1993. Surveillance for waterborne disease outbreaks—United States, 1991–1992. Morbid. Mortal. Weekly Report 42(SS-5):1. Nwachuku, N., G.F. Craun, and R.L. Calderon. 2002. How effective is the TCR in assessing outbreak vulnerability. J. Am. Water Works Assoc. 94(9):88–96. Rose, J.B., S. Dauschner, D.R. Easterline, F.C. Curriero, S. Lele, and J.A. Patz. 2000. Climate and waterborne disease outbreaks. J. Am. Water Works Assoc. 92(9):77–87. Seyfried, P.L., R.S. Tobin, N.E. Brown, and P.F. Ness. 1985. A prospective study of swimmingrelated illness. I. Swimming-associated health risk. Am. J. Public Health 75:1068. USEPA. 1989a. Drinking Water; National Primary Drinking Water Regulations; Total Coliforms (Including Fecal Coliforms and E. coli); Final Rule. Fed. Reg. 54:27544–27568. USEPA. 1989b. Drinking Water; National Primary Drinking Water Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionella, and Heterotrophic Bacteria; Final Rule. Fed. Reg. 54:27486–27541. USEPA. 1998. National Primary Drinking Water Regulations; Interim Enhanced Surface Water Treatment Rule; Final Rule. Fed. Reg. 63:69478–69521. USEPA. 2000. National Primary Drinking Water Regulations; Ground Water Rule; Proposed Rules. Fed. Reg. 65:30194–30274.

4 HISTORY OF THE SAFE DRINKING WATER ACT (SDWA) FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

4.1

INTRODUCTION

Drinking water quality standards and regulations define in quantitative terms water that is ‘‘safe’’ for human consumption. Drinking water must be free from organisms capable of causing disease. It must not contain minerals and organic substances at concentrations that could produce adverse physiological effects. Drinking water should be aesthetically acceptable; it should be free from apparent turbidity, color, and odor and from any objectionable taste. It should also have a reasonable temperature. Water meeting these conditions is called ‘‘potable.’’ The term ‘‘drinking water standards’’ typically refers to numerical limits that define the maximum concentration of contaminants that water may contain to be considered potable (i.e., safe to drink). Drinking water standards may or may not be mandatory or enforceable, depending the agency issuing the standards and the legislative authority under which they are issued. Drinking water regulations are set by a regulatory agency under the authority of federal, state or local law. The Safe Drinking Water Act (SDWA) is the principal law governing drinking water safety in the United States. Enacted initially in 1974 (SDWA 1974), the SDWA as amended (Table 4.1) authorizes the U.S. Environmental Protection Agency (USEPA) to establish comprehensive national drinking water regulations to ensure drinking water safety. This chapter reviews the history of the SDWA from a legal Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

71

72

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.1

SDWA and Amendmentsa

Year 1974 1977 1979 1980 1986 1988 1996 2002

a

Law P.L. P.L. P.L. P.L. P.L. P.L. P.L. P.L.

93-523 95-190 96-63 96-502 99-339 100-572 104-182 107-188

Date

Act

Dec. 16, 1974 Nov. 16, 1977 Sept. 6, 1979 Dec. 5, 1980 Jun. 16, 1986 Oct. 31, 1988 Aug. 6, 1996 June 12, 2002

SDWA SDWA Amendments of 1977 SDWA Amendments of 1979 SDWA Amendments of 1980 SDWA Amendments of 1986 Lead Contamination Control Act SDWA Amendments of 1996 Public Health Security and Bioterrorism Preparedness and Response Act of 2002

Codified generally as 42 USC 300f-300j-11.

and regulatory perspective. VanDe Hei and Schaefer (this volume, Chapter 5) provide an insightful review of the development of the SDWA from a sociopolitical perspective. USEPA drinking water regulations require public water systems in the United States to meet specified drinking water quality standards. Regulations may also require that compliance monitoring be conducted, specified treatment be applied, and reports be submitted documenting that regulations are being met. To ensure compliance with water quality regulations, a water utility usually must produce water of a better quality than a standard or regulation would demand. Hence, each water utility needs its own water quality goals to ensure compliance while producing the highest quality tapwater possible within its financial, technical, and managerial capacity.

4.2

EARLY DEVELOPMENT OF DRINKING WATER STANDARDS

Okun (this volume, Chapter 1) has reviewed the early development of drinking water standards. Historically, civilizations began and located within regions of abundant water supplies. Water quality was not very well documented, and little was known about disease as it related to water quality. Early treatment was performed only to improve the appearance or taste of drinking water. No defined standards of quality other than general clarity or palatability were recorded by ancient civilizations (Borchardt and Walton 1971). The growth of community water supply systems in the United States began in Philadelphia. In 1799, a small section was first served by wooden pipes and water was drawn from the Schuylkill River by steam pumps. By 1860, over 400 major water systems had been developed to serve the nation’s major cities and towns. Although municipal water supplies were growing in number during this early period of the nation’s development, healthy and sanitary conditions did not begin

4.3 THE SAFE DRINKING WATER ACT OF 1974

73

to improve significantly until the turn of the century. By 1900, an increase in the number of water supply systems to over 3000 contributed to major outbreaks of disease because pumped and piped supplies, when contaminated, provide an efficient means for spreading pathogenic bacteria throughout a community. In the mid- to late 1800s, acute waterborne disease of biological origin was still prevalent in the United States. Following the lead of European investigators, slow sand filters were introduced in Massachusetts by the mid-1870s. Empirical observations showed that this improved the aesthetics of water quality. In the mid-1890s, the Louisville Water Company, in Kentucky combined coagulation with rapid sand filtration, significantly reducing turbidity and bacteria in the water. The Louisville studies refined the knowledge of the process and showed the essential need for pretreatment, including sedimentation for Ohio River water. The next major milestone in drinking water technology was the use of chlorine as a disinfectant. Chlorination was first used in 1908 and was introduced in a large number of water systems shortly thereafter. U.S. drinking water standards have developed and expanded since the early 1900s as knowledge of the health effects of contaminants has increased and the treatment technology to control contaminants has improved. Protection of public health has provided the principal driving force behind development of drinking water standards and regulations. In the United States, federal authority to establish drinking water regulations originated with the enactment by Congress in 1893 of the Interstate Quarantine Act (U.S. Statutes 1893). The first water-related regulation, adopted in 1912, prohibited the use of the common cup on carriers of interstate commerce, such as trains (McDermott 1973). In 1914, the Treasury Standards were established by the U.S. Public Health Service (USPHS), then a part of the Treasury Department. The USPHS revised these standards in 1925, 1942, 1946, and 1962 (USPHS 1925, 1943, 1946, 1962). For a discussion of these early standards prior to the SDWA, see Chapter 1.

4.3

THE SAFE DRINKING WATER ACT OF 1974

Results of a Community Water Supply Study (CWSS) in 1969 by the USPHS generated congressional interest in federal safe drinking water legislation. The first series of bills to give the federal government power to set enforceable standards for drinking water were introduced in 1970. Congressional hearings on legislative proposals concerning drinking water were held in 1971 and 1972 (Kyros 1974). In September 1972 the U.S. Senate passed S. 3994, an original bill reported by the Committee on Commerce, requiring establishment of minimum Federal drinking water standards with enforcement by the states and a program of grants to support state drinking water programs. The House took no action on this bill in the 92nd Congress. Enactment of the initial SDWA is inextricably intertwined with the discovery of trihalomethanes and organic contaminants in drinking water. Symons (2001a,

74

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

2001b) and Okun (Chapter 1) provide an insightful chronicle of this period. Researchers in the Netherlands and at USEPA discovered that a class of compounds, trihalomethanes (THMs), were formed as a byproduct when free chlorine was added for disinfection (Rook 1974; Bellar et al. 1974). Johannes Rook (Netherlands) and Thomas Bellar (USEPA), working independently, in a similar timeframe, on projects not related to chloroform or disinfection byproducts, accidentally discovered that chloroform was created during the disinfection of drinking water (Symons 2001a). Although unrelated, publicity surrounding reports of the discovery of the formation of THMs coincided with the finding of synthetic organic chemicals (SOCs) in the New Orleans water supply. On Nov. 8, 1974, the date of the USEPA Region VI press conference concerning the New Orleans study, USEPA simultaneously announced that a nationwide survey would be conducted to determine the extent of the THM problem in the United States (Symons et al. 1975). This survey was known as the National Organics Reconnaissance Survey (NORS), and was completed in 1975 (discussed below). The true health significance of THMs and SOCs in drinking water was not known, and questions still remain today regarding the health significance of low concentrations of organic chemicals and disinfection byproducts. After more than 4 years of effort by Congress, federal legislation was enacted to develop a national program to protect the quality of the nation’s public drinking water systems. On Nov. 19, 1974, the House debated and passed by voice vote H.R. 13002, a clean bill reported by the House Committee on Interstate and Foreign Commerce. Language of the House bill was then inserted into S. 433. The Senate considered and amended S. 433 on Nov. 26. The House agreed to the Senate amendment Dec. 3, 1974. President Ford signed the SDWA on Dec. 16, 1974 as Public Law 93-523 (Congressional Research Service 1982). Symons (2001a) notes that enactment of the 1974 SDWA generated much confusion and resentment with regard to the national publicity criticizing drinking water quality, and many people were unhappy about passage of the SDWA. USEPA was accused of deliberately planting the THM story just to get the SDWA passed, which was unfounded. Many people within the drinking water industry had difficulty believing that trace concentrations of chemicals with difficult names could be harmful (J. M. Symons 1974). The 1974 SDWA established a cooperative program among local, state, and federal agencies. The act required the establishment of primary drinking water regulations designed to ensure safe drinking water for the consumer. These regulations were the first to apply to all public water systems in the United States, covering both chemical and microbial contaminants. Except for the coliform standard under the Interstate Quarantine Act mentioned previously, drinking water standards were not legally binding until passage of the SDWA. The SDWA mandated a major change in the surveillance of drinking water systems by establishing specific roles for the federal and state governments and for public water suppliers. The federal government, specifically the USEPA, was authorized to set national drinking water regulations, conduct special studies and research, and oversee implementation of the act. The state governments, through their health departments and environmental agencies, are expected to accept the

4.3 THE SAFE DRINKING WATER ACT OF 1974

75

major responsibility, called primary enforcement responsibility or primacy, for the administration and enforcement of the regulations set by USEPA under the Act. Public water suppliers have the day-to-day responsibility of meeting the regulations. To meet this goal, routine monitoring must be performed, with results reported to the regulatory agency. Violations must be reported to the public and corrected. Failure to perform any of these functions can result in enforcement actions and penalties. The 1974 Act specified the process by which USEPAwas to adopt national drinking water regulations. Interim regulations [national interim primary drinking water regulations (NIPDWRs)] were to be adopted within 6 months of its enactment. Within about 2 years (by March 1977), USEPA was to propose revised regulations (revised national drinking water regulations) on the basis of a study of health effects of contaminants in drinking water conducted by the National Academy of Sciences (NAS). Establishment of the revised regulations was to be a two-step process. First, the agency was to publish recommended maximum contaminant levels (RMCLs) for contaminants believed to have an adverse health effect based on the NAS study. RMCLs were to be set at a level such that no known or anticipated health effect would occur. An adequate margin of safety was to be provided. These levels were to act only as health goals and were not intended to be federally enforceable. USEPA then established maximum contaminant levels (MCLs) as close to the RMCLs as the agency thought feasible. The agency was also authorized to establish a required treatment technique instead of an MCL if it was not economically or technologically feasible to determine the level of a contaminant. The MCLs and treatment techniques comprise the National Primary Drinking Water Regulations (NPDWRs) and are federally enforceable. The regulations were to be reviewed at least every 3 years. 4.3.1

The National Interim Primary Drinking Water Regulations

Interim regulations were adopted Dec. 24, 1975 (USEPA 1975b) on the basis of the 1962 USPHS standards with little additional health effects support. The interim rules were amended several times before the first primary drinking water regulation was issued (see Table 4.2). The findings of the NORS (mentioned previously) were published in November 1975 (Symons et al. 1975). The four trihalomethanes (THMs)—chloroform, bromodichloromethane, dibromochloromethane, and bromoform—were found to be widespread in the chlorinated drinking waters of 80 cities studied. USEPA subsequently conducted the National Organics Monitoring Survey (NOMS) between 1976 and 1977 to determine the frequency of specific organic compounds in drinking water supplies (USEPA 1978a). Included in the NOMS were 113 community water supplies representing different sources and treatment processes, each monitored 3 times during a 12-month period. NOMS data showed that THMs were the most widespread organic contaminants in drinking water, occurring at the highest concentrations. From the NORS, NOMS, and other surveys, more than 700 specific organic chemicals had been identified in various drinking waters (Cotruvo and Wu 1978).

76

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.2

History of the NIPDWRs

Regulation

Promulgation Date

NIPDWRs (USEPA 1975b)

Dec. 24, 1975

June 24, 1977

1st NIPDWR Amendment (USEPA 1976a) 2nd NIPDWR Amendment (USEPA 1979) 3rd NIPDWR Amendment (USEPA 1980)

July 9, 1976

June 24, 1977

Inorganic, organic, and microbiological contaminants and turbidity Radionuclides

Nov. 29, 1979

Varied depending on system size

Total trihalomethanesa

Aug. 27, 1980

Feb. 27, 1982

Feb. 28, 1983

March 30, 1983

Special monitoring requirements for corrosion and sodium Identifies best generally available means to comply with THM regulations

4th NIPDWR Amendment (USEPA 1983)

a

Effective Date

Primary Coverage

The sum of chloroform, bromoform, bromodichloromethane, plus dibromochloromethane.

On June 21, 1976, the EDF petitioned the USEPA, alleging that the initial interim regulations set in 1975 did not sufficiently control organic compounds in drinking water. In response, USEPA issued an Advance Notice of Proposed Rulemaking (ANPRM) on July 14, 1976, requesting public input on how THMs and SOCs should be regulated (USEPA 1976b). On Feb. 9, 1978, USEPA proposed a two-part regulation for the control of organic contaminants in drinking water (USEPA 1978b). The first part concerned the control of THMs. The second part concerned control of source water SOCs and proposed the use of GAC adsorption by water utilities vulnerable to possible SOC contamination. The next day, Feb. 10, 1978, the U.S. Court of Appeals, District of Columbia Circuit, issued a ruling in the EDF case filed June 21, 1976 (U.S. Court of Appeals 1978). The court upheld USEPA’s discretion to not include comprehensive regulations for SOCs in the NIPDWRs, but as a result of new data being collected by USEPA, the court told the agency to report a plan for amending the interim regulations to control organic contaminants. The court stated (U.S. Court of Appeals 1978): ‘‘In light of the clear language of the legislative history, the incomplete state of our knowledge regarding the health effects of certain contaminants and the imperfect nature of the available measurement and treatment techniques cannot serve as justification for delay in controlling contaminants that may be harmful.’’

4.3 THE SAFE DRINKING WATER ACT OF 1974

77

The agency contended that the proposed rule published the day before satisfied the court’s judgment. Reaction to the proposed regulation on GAC adsorption treatment varied. Federal health agencies, environmental groups, and a few water utilities supported the proposed rule. Many state health agencies, consulting engineers, and most water utilities opposed it (Symons 1984). USEPA responded to early opposition to the GAC proposal by publishing an additional statement in the July 6, 1978, Federal Register (USEPA 1978c). Nevertheless, significant opposition continued based on several technical considerations (Pendygraft et al. 1979a, 1979b, 1979c). USEPA promulgated regulations for the control of THMs in drinking water on Nov. 29, 1979 (USEPA 1979), but subsequently, on March 19, 1981, withdrew its proposal to control organic contaminants by GAC (USEPA 1981). 4.3.2

National Academy of Sciences (NAS) Study

As required by the 1974 SDWA, USEPA contracted with the NAS to have the National Research Council (NRC) assess human exposure via drinking water and the toxicology of contaminants in drinking water. The NRC Committee on Safe Drinking Water published their report, Drinking Water and Health, in 1977 (NAS 1977). Five classes of contaminants were examined: microorganisms, particulate matter, inorganic solutes, organic solutes, and radionuclides. This report, the first in a series of nine, served as the basis for revised drinking water regulations. USEPA published the recommendations of the NAS study on July 11, 1977 (USEPA 1977). The 1977 amendments to the SDWA called for revisions of the NAS study ‘‘reflecting new information which has become available since the most recent previous report [and which] shall be reported to the Congress each two years thereafter’’ (SDWA 1977). NAS reports issued on drinking water related issues are listed in Table 4.3. USEPA often funds the NAS to conduct independent assessments of drinking water contaminants, typically directed by Congress to do so in legislation or by conference committee report. 4.3.3

1977–1980 SDWA Amendments

The SDWA was amended and=or reauthorized in 1977, 1979, and 1980 (Congressional Research Service 1982). At the beginning of the 95th Congress (1977), jurisdiction for the SDWA was transferred from the Senate Committee on Commerce to the Senate Committee on Environment and Public Works. In November 1977 Congress enacted amendments to the 1974 SDWA that reauthorized and revised certain provisions. S. 1528, containing amendments to the SDWA, was signed into law by President Carter Nov. 16, 1977 as Public Law 95-190 (SDWA 1977). Congress again reauthorized the SDWA in 1979. S. 1146, a 3-year extension of authorizations for appropriations for the SDWA, was signed into law by President Carter on Sept. 6, 1979 as Public Law 96-63 (SDWA 1979). During the 96th Congress the House Commerce subcommittee on Health and Environment held oversight hearings on the SDWA. On Sept. 19, 1980, the Committee on Interstate and Foreign Commerce reported a clean bill, H.R. 8117, which the

78

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.3

Drinking Water Studies Completed by the National Academy of Sciences

Study Drinking Water and Health (NAS 1977) Drinking Water and Health, Vol. 2 (NAS 1980a) Drinking Water and Health, Vol. 3 (NAS 1980b)

Drinking Water and Health, Vol. 4 (NAS 1982) Drinking Water and Health, Vol. 5 (NAS 1983) Drinking Water and Health, Vol. 6 (NAS 1986)

Drinking Water and Health, Vol. 7 (NAS 1987a) Drinking Water and Health, Vol. 8 (NAS 1987b) Drinking Water and Health, Vol. 9 (NAS 1989) Health Effects of Ingested Fluoride (NAS 1993) Nitrate and Nitrite in Drinking Water (NAS 1995) Safe Water from Every Tap: Improving Water Service to Small Communities (NAS 1997) Issues in Potable Reuse: The Viability of Augmenting Drinking Water Supplies with Reclaimed Water (NAS 1998) Setting Priorities for Drinking Water Contaminants (NAS 1999a) Identifying Future Drinking Water Contaminants (NAS 1999b) Risk Assessment of Exposure to Radon in Drinking Water (NAS 1999c)

Scope Examines microorganisms, particulate matter, inorganic solutes, and radionuclides Evaluated disinfectants, disinfection byproducts, and granular activated carbon Evaluates several epidemiologic studies, assesses the toxicology of selected drinking water contaminants, and examines the contribution of drinking water to the mineral nutrition in humans Examines distribution system water quality and toxicity of selected inorganic and organic contaminants Reviews the toxicology of selected synthetic organic chemicals, uranium, arsenic, and asbestos Examines developmental effects, reproductive toxicology, neurotoxic effects, mechanisms of carcinogenesis, dose–response extrapolations, risk assessment issues, and the toxicology of selected contaminants Again addresses disinfectants and disinfection byproducts Focuses exclusively on the application of pharmacokinetics in risk assessment Complex mixtures Evaluates the MCLG and MCL for fluoride Evaluates the MCLG and MCL for nitrate and nitrite Presents institutional and technological options for improving the management efficiency and financial stability of small water systems Reviews current issues associated with the potable use of reclaimed water

Evaluates decision processes for selecting contaminants for regulation Evaluates options for selecting contaminants for future regulation Evaluates health risks of radon in drinking water (continued )

4.4 1986 SDWA AMENDMENTS

TABLE 4.3

79

(Continued)

Study Arsenic in Drinking Water (NAS 1999d) Copper in Drinking Water (NAS 2000a) Re-Evaluation of Drinking Water Guidelines for Diisopropyl Methylphosphonate (USEPA 2000b) Classifying Drinking Water Contaminants for Regulatory Consideration (NRC 2001)

Scope Health risk assessment of arsenic Evaluates the MCLG for copper Reevaluates drinking water guidelines

Recommends a contaminant selection process

House passed by voice vote on Sept. 23. The House-passed bill was referred to the Senate Committee on Environment and Public Works, which took no action. On Nov. 19, 1980, the Senate discharged the Committee from consideration of H.R. 8117 and passed the bill by voice vote. The bill was signed into law by President Carter Dec. 5, 1980 as Public Law 96-502 (SDWA 1980). 4.4

1986 SDWA AMENDMENTS

Congress severely underestimated the time required for USEPA to develop credible regulations. USEPA’s slowness in regulating contaminants and its failure to require GAC treatment for organic contaminants served as a focal point for discussion of possible revisions to the law. Reports in the early 1980s of drinking water contamination by organic contaminants and other chemicals (Westrick et al. 1984) and pathogens such as Giardia lamblia (Craun 1986) aroused congressional concern over the adequacy of the SDWA. The rate of progress made by USEPA to regulate contaminants was of particular concern. Both the House and Senate considered various legislative proposals beginning in 1982 that informed the SDWA debate and helped to shape the SDWA amendments enacted in 1986. Four oversight hearings were held in 1982 by the Senate Environment and Public Works Subcommittee on Toxic Substances and Environmental Oversight. Congress began considering broad amendments to the SDWA in 1983. The SDWA Amendments of 1983 (H.R. 3200) was introduced in the 98th Congress. The House Energy and Commerce Subcommittee on Health and the Environment held hearings on the SDWA. Issue-specific legislation was introduced to provide for the protection of sole source underground drinking water supplies. The House passed an SDWA reauthorization bill (H.R. 5959) on Sept. 18, 1994 (House Report 98-1034). An SDWA reauthorization bill (S. 2649) was passed by the Senate on Sept. 28, 1984 (Senate Report 98-641). However, the 98th Congress ended before a conference agreement could be reached (Congressional Research Service 1993).

80

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

The 99th Congress built on the previous Congress’ efforts to reauthorize the SDWA. S. 124 was introduced Jan. 2, 1985, reported by the Senate Environment and Public Works Committee on May 15, 1985 (Senate Report 99-56), and passed by the Senate on May 16, 1985. The companion bill, H.R. 1650, was introduced March 21, 1985, and reported by the House Energy and Commerce Committee June 11, 1985 (House Report 99-168). On June 17, 1985, the House considered and passed H.R. 1650, passed S. 124 with amendments, and tabled H.R. 1650. A conference committee was formed and the conference report on S. 124 (House Report 99-575) was debated and passed in the House on May 13, 1986 and in the Senate on May 21, 1986. The President signed S. 124 into law on June 19, 1986 as Public Law 99-339 (SDWA 1986). To strengthen the SDWA, especially the regulation-setting process and groundwater protection, most of the original 1974 SDWA was amended in 1986. Major provisions of the 1986 amendments included (Cook and Schnare 1986; Dyksen et al. 1988; Gray and Koorse 1988):  Mandatory standards for 83 contaminants by June 1989.  Mandatory regulation of 25 contaminants every 3 years.  National interim drinking water regulations were renamed national primary drinking water regulations.  Recommended maximum contaminant level goals (RMCLs) were replaced by maximum contaminant level goals (MCLGs).  Required designation of best available technology for each contaminant regulated.  Specification of criteria for deciding when filtration of surface water supplies is required.  Disinfection of all public water supplies.  Monitoring for contaminants that are not regulated.  A ban on lead solders, flux, and pipe in public water systems.  New programs for wellhead protection and protection of sole source aquifers.  Streamlined and more powerful enforcement provisions. The 1986 amendments significantly increased the rate at which USEPA was to set drinking water standards. Resource limitations and competing priorities within the agency prevented USEPA from fully meeting the mandates of the 1986 amendments. Figure 4.1 summarizes the growth of regulated contaminants under the SDWA from initial enactment to the present. The mandate to regulate 25 contaminants every 3 years simply could not be met, and after 1992 regulations ceased to be issued until the law was amended in 1996. 4.5

1988 LEAD CONTAMINATION CONTROL ACT

On Dec. 10, 1987, the House Subcommittee on Health and Environment held a hearing on lead contamination of drinking water. At that hearing the U.S. Public

4.6 1996 SDWA AMENDMENTS

Figure 4.1

81

Growth of regulated contaminants.

Health Service warned that some drinking water coolers may contain lead solder or lead-lined water tanks that release lead into the water they distribute. Data submitted to the subcommittee by manufacturers indicated that close to 1 million water coolers were in use at that time that contained lead. A subcommittee hearing was subsequently held on July 13, 1988 to consider H.R. 4939, the Lead Contamination Control Act. The bill had widespread support and moved swiftly through the House and Senate (Congressional Research Service 1993). The Lead Contamination Control Act was enacted Oct. 31, 1988 as Public Law 100-572 (LCCA 1988). This law amended the SDWA to, among other things, institute a program to eliminate lead-containing drinking water coolers in schools. Part F—Additional Requirements to Regulate the Safety of Drinking Water—was added to the SDWA. USEPA was required to provide guidance to states and localities to test for and remedy lead contamination in schools and daycare centers. It also contains specific requirements for the testing, recall, repair, and=or replacement of water coolers with lead lined storage tanks or with parts containing lead. Civil and criminal penalties for the manufacture and sale of water coolers containing lead are set.

4.6

1996 SDWA AMENDMENTS

The 1986 SDWA amendments authorized congressional appropriations for implementation of the law through fiscal year 1991. Reauthorization was not completed until 1996. 4.6.1

Reauthorization Issues Emerge

Several studies following the 1986 SDWA amendments set the stage for potential changes to the SDWA. A 1988 study sponsored and supported by consumer advo-

82

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

cate Ralph Nader drew attention to trace organic chemicals found in drinking water (Conacher 1988). A study by the National Wildlife Federation released in 1988 (Dean 1988) and updated in 1989 (Dean 1989) captured media attention by highlighting violations of the SDWA. Both reports characterized USEPA’s enforcement of the SDWA as virtually nonexistent. Studies such as these raise the issue of whether public health is threatened by the failure of water utilities to comply with SDWA regulations. Although noncompliance occurs mostly in small systems, the issue of noncompliance raises concerns about the adequacy of the SDWA and USEPA’s drinking water program and the ability of water suppliers to provide safe drinking water to their customers. The GAO assessed the implementation of the SDWA program by USEPA and the states at the request of the Subcommittee on Environment, Energy, and Natural Resources (Committee on Government Operations, House of Representatives). The GAO report (USGAO 1990) was released June 8, 1990, and was the subject of Subcommittee oversight hearings held Aug. 2, 1990 (Hembra 1990). The GAO found that published USEPA data indicate that most water systems are complying with monitoring and MCL requirements and that the relatively few violating systems have generally committed minor infractions but that considerable noncompliance existed. USEPA sponsored a workshop in September 1990 that served as a starting point for USEPA to identify issues related to the SDWA (Schnare 1990). Representatives from the drinking water community, state agencies, USEPA, environmental organizations, and others presented their views on policy and technical issues that could be addressed during reauthorization. The National Drinking Water Advisory Council (NDWAC) compiled comments from a survey it conducted on changes to the SDWA (Kessler and Schnare 1991) and developed recommendations (NDWAC 1993). 4.6.2

GAO Studies Note Deficiencies

The GAO released three studies in 1992 regarding implementation of the SDWA. An audit of 28 water systems in six states revealed high rates of noncompliance with SDWA public notification requirements (USGAO 1992a). The public notification requirements themselves were cited as a major cause of noncompliance, particularly for small systems, because the requirements have been difficult to understand and implement. A July 6, 1992, GAO report examined the gap between available resources and drinking water program needs (USGAO 1992b). Funding shortages at the federal, state, and water system levels were found to contribute to implementation and compliance problems. It is estimated that by 1995, the total state and federal program requirements will exceed state and federal resources by $150 million. The solesource aquifer program was examined in another GAO report issued Oct. 13, 1992 (USGAO 1992c). The principal finding was that mechanisms used to identify projects for possible USEPA review were weak. The GAO released three additional studies in 1993. The wellhead protection program was the focus of a GAO report issued April 14, 1993 (USGAO 1993a).

4.6 1996 SDWA AMENDMENTS

83

Several barriers were found to hinder states’ efforts to develop and implement wellhead protection programs, including (1) opposition at the local level against states’ enactment of land-use controls and (2) a general lack of public awareness about the vulnerability of drinking water to contamination and the need to protect wellhead areas. A severe shortage of funds was identified as the underlying cause of these barriers and the primary problem affecting state wellhead protection programs. GAO conducted a nationwide questionnaire and reviewed 200 sanitary surveys conducted in four states (Illinois, Montana, New Hampshire, and Tennessee). Their report, issued April 9, 1993, disclosed that sanitary surveys are often deficient in how they are conducted, documented, and=or interpreted (USGAO 1993b). Many of the 200 sanitary surveys revealed recurring problems with water systems’ equipment and management, particularly among small systems. Regardless of system size, deficiencies previously disclosed frequently went uncorrected. The gap between the needs and available resources of state drinking water programs was a major barrier severely affecting states’ capabilities to conduct sanitary surveys. Severe resource constraints have made it increasingly difficult for many states to effectively carry out the monitoring, enforcement, and other mandatory activities to retain primacy. A June 25, 1993, GAO report concluded that the funding difficulties faced by states are likely to worsen and that resolving the primacy issue involves bringing the program’s costs in line with resources (USGAO 1993c). State funding needs represent only a fraction of the expenditures that public water systems must make to comply with SDWA requirements. Results of a survey released in 1993 by the Association of State Drinking Water Administrators (ASDWA) identified an immediate need of $2.738 billion for SDWA-related infrastructure projects in 35 states (ASDWA 1993). Insufficient funding, political interference, and mismanagement were cited in a 1993 study by the Center for Resource Economics as the three main obstacles preventing USEPA from fully meeting its environmental statutory mandates (Center for Resource Economics 1993). On March 9, 1994, GAO released the results of an audit of the ability of small systems to comply with SDWA regulations (USGAO 1994). The study found that states are experimenting with technology- and management-based approaches to help small community drinking water systems comply with SDWA regulations, but that barriers exist. The report recommended that USEPA revise its priorities to place greater emphasis on developing and maintaining viability programs.

4.6.3

102nd Congress

Minimal formal activity on the SDWA took place during the 102nd Congress. Rep. Henry Waxman (D–CA) introduced H.R. 2840, Lead Contamination Control Act Amendments, which was intended to rewrite USEPA’s lead rule. A companion bill, S. 1445, Lead in Drinking Water Reduction Act, was introduced in the Senate by Frank Lautenberg (D–NJ). The House Subcommittee on Health and Environment held a hearing May 10, 1991 on progress in carrying out the SDWA provisions for control of drinking water contamination.

84

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

During the closing months, Senator Pete Domenici (R–NM) introduced bill S. 2900, that would have established a moratorium on implementation of drinking water regulations by prohibiting USEPA from spending money to implement and enforce regulations in place retroactive to December 1989. Senator Domenici offered S. 2900 as a floor amendment when the Senate considered the Veterans’ Administration (VA), Housing and Urban Development (HUD), and Independent Agencies appropriation bill for fiscal year (FY) 1993. This bill provides funding for USEPA and several other government agencies. To counter Senator Domenici, Senators John Chafee (R–RI) and Frank Lautenberg (D–NJ) offered an alternative amendment. The Chafee–Lautenberg amendment required USEPA to conduct a study on the implementation of the SDWA. It also required the agency to conduct a separate study on radon. The Domenici amendment was narrowly defeated in the Senate by only six votes. The Chafee–Lautenberg amendment was adopted (Congressional Record 1992) and signed into law Oct. 6, 1992 (Public Law 102-389). 4.6.4

103rd Congress

Activity increased in the 103rd Congress with the introduction of several proposed bills. The first comprehensive reform bill was S. 767, introduced by Senator Don Nickles (R–OK). An identical companion bill, H.R. 2344, was introduced in the House by Rep. James Walsh (R–NY). These proposals did not move forward and simply served to stimulate discussion on various SDWA issues. State Revolving Loan Fund Proposed Debate on the SDWA began in earnest when proposals were introduced in the House to authorize a drinking water state revolving loan fund (DWSRF) for drinking water. The proposal for a DWSRF was included in President Clinton’s economic stimulus package offered early in 1993. A jurisdictional dispute arose between two House committees vying for control over the DWSRF, and two competing bills were introduced, one for each committee. Although the president’s package was eventually defeated by Congress, the DWSRF bills moved forward. Separate DWSRF bills were not introduced in the Senate because of the desire to deal with DWSRF funding at the same time that other SDWA issues were considered. H.R. 1701, introduced by Rep. Henry Waxman (D–CA), authorized an DWSRF as part of the SDWA. A proposal to amend the Clean Water Act (CWA) to expand the scope of the existing clean water state revolving loan fund (CWSRF) to include drinking water was introduced by Rep. Norman Mineta (D–CA), who chaired the House Committee on Public Works and Transportation that has jurisdiction over the CWA. Both bills were reported out of committee. A DWSRF was included in President Clinton’s fiscal year 1994 budget, initially funded at $600 million, with additional funding planned at $1 billion per year thereafter. Because authorizing legislation for this money was not in place, it could not be appropriated and spent. Therefore, Congress decided to include the money in the budget with a condition that it could not be spent until authorizing

4.6 1996 SDWA AMENDMENTS

85

legislation was passed. This meant that such legislation must have been in place before Oct. 1, 1994, or the money could not be appropriated for fiscal year 1994. USEPA Recommendations and Reports to Congress USEPA’s recommendations for reauthorization were released by Administrator Carol Browner on Sept. 8, 1993, during a speech before the National Association of Towns and Townships (Browner 1993). A list of 10 recommendations was issued based on USEPA’s report to Congress on SDWA implementation released a few days later. USEPA’s report on SDWA implementation was submitted to Congress in September 1993 (USEPA 1993). In this report, USEPA estimated that compliance with the standards for 84 contaminants regulated to date is expected to cost public water systems approximately $1.4 billion (in 1991 dollars) per year by 1995 (Auerbach 1994). Individual household costs to comply with federal drinking water rules were estimated to range from a few dollars per year in metropolitan areas to several hundred dollars per year in small communities that have contamination problems. The report estimated that the 1993 state funding shortfall for implementing federal drinking water requirements was about $162 million; needs totaled $304 million, yet only $142 million was available from state and federal sources. USEPA’s report to Congress on radon was published in March 1994 (USEPA 1994a). The report revised the agency’s risk and cost assessments for radon in drinking water. USEPA estimated that approximately 19 million people are exposed to a radon level above the then-proposed MCL of 300 pCi=L. The total cost to treat radon in drinking water to below the then proposed MCL was estimated at $272 million. Natural Resources Defense Council (NRDC) Report The Natural Resources Defense Council (NRDC) released a report, Think before You Drink, the Failure of the Nation’s Drinking Water System to Protect Public Health, on Sept. 17, 1993, which highlighted violations of the SDWA (Olson 1993). The report reviewed a number of problems associated with SDWA implementation and presented a set of proposals for SDWA reforms. The NRDC report made many serious claims regarding the quality of U.S. drinking water supplies and served as the first shot fired in an intense battle over the SDWA. In response, the National Rural Water Association (NRWA) issued a statement claiming NRDC sensationalized the report findings (Carroll 1993). The NRDC report attracted media attention, including a page-one story in the Sept. 27, 1993, USA Today (USA Today 1993a). Subsequent letters to the editor challenged the report’s findings and conclusions (USA Today 1993b, Wade 1993, Ronnebaum 1993). House and Senate Consider Reauthorization Bills On Oct. 14, 1993, Senator Max Baucus (D–MT) introduced S. 1547 and a hearing on this bill was held Oct. 27, 1993, by the Senate Committee on Environment and Public Works, which Senator Baucus chaired. Senator John Chafee (R–RI), a key player in the SDWA debate, decided not to cosponsor S. 1547 because of disagreement over some of the

86

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

bill’s provisions. S. 1547 received a mixed response from environmental groups, and some of its provisions were generally opposed by various interest groups from both sides. On Oct. 27, 1993, the problem of unfunded federal mandates received national attention at a press conference held by national public interest groups representing state and local governments. These groups included the National Governors Association (NGA), the U.S. Conference of Mayors (USCM), the National Association of Counties, and the National Conference of State Legislatures (NCSL). Unfunded federal mandates are laws passed by the U.S. Congress imposing requirements on state and local governments without providing adequate federal funds to implement those requirements. The cost of complying with environmental laws in general and the SDWA in particular, in the absence of federal, state, and local financial resources, caused many groups to pressure Congress for relief. Concurrently with the October 27 Capitol Hill press conference on unfunded mandates, H.R. 3392 was introduced by Rep. Jim Slattery (D–KS) and Rep. Thomas Bliley (R–VA). This bill received the most support of all of the proposed SDWA bills. H.R. 3392 was supported by the NGA, the National League of Cities, USCM, NCSL, ASDWA, NRWA, AWWA, the National Water Resources Association, the Association of Metropolitan Water Agencies, and the National Association of Water Companies. The bill was opposed by the National Wildlife Federation, the NRDC, Friends of the Earth, Alliance to End Childhood Lead Poisoning, National Education Association, and the National Parent Teacher Association. A key issue proposed by H.R. 3392 was a change in how drinking water standards are set. Proponents of H.R. 3392 argued that changes to the process are needed so that rational standards can be developed to maximize health protection with the limited funds available. Opponents of H.R. 3392 argued that it merely served to roll back existing standards (Waxman 1994). On Nov. 22, 1993, H.R. 3686 was introduced by Rep. Pat Roberts (R–KS). This bill would suspend the requirements of the SDWA until the cost to state and local governments of implementing its requirements was fully funded by the federal government. Although this bill did not receive serious consideration, it expressed the strong attitude many elected officials had regarding SDWA funding. On March 10, 1994, Senator Pete Domenici (R–NM) introduced S. 1920. The bill was similar to H.R. 3392, but included provisions for a drinking water State Revolving Loan Fund (SRLF). Although the bill was referred to the Senate Environment and Public Works Committee, it was not considered during markup of S. 1547. On March 24, 1994, Senator Domenici announced his intention to negotiate for inclusion of provisions from S. 1920 to give water utilities more relief from unfunded mandates (Domenici 1994). Senator Domenici offered several amendments during floor deliberations on S. 2019. On April 18, 1994, Reps. Lambert (D–AR), Synar (D–OK), and Studds (D–MA) introduced H.R. 4314. The provisions of this bill generally followed the Clinton administration recommendations. H.R. 4314 served as an alternative bill for those representatives who desired to support an SDWA bill, but did not want to support H.R. 3392 because of opposition by Rep. Waxman (D–CA).

4.6 1996 SDWA AMENDMENTS

87

The Senate Environment and Public Works Committee developed changes to S. 1547 in response to testimony at the hearing and other comments. The committee met to consider amendments to S. 1547 in March 1994. On March 24, the committee completed markup and ordered reported, by a unanimous vote, an original bill (S. 2019) that incorporated the amendments to S. 1547. After substantial floor amendment, S. 2019 was passed by the Senate May 19, 1994. With action completed in the Senate, the focus of attention shifted to the House of Representatives. Rep. Henry Waxman (D–CA), who strongly opposed the standardsetting provisions of H.R. 3392, threatened a legislative stalemate (Waxman 1994). Rep. Waxman chaired the House Subcommittee on Health and Environment, in which H.R. 3392 and other SDWA legislation was first considered in the House. Environmental Groups Oppose House Bill At the same time H.R. 3392 received strong support in the House, environmental interest groups mounted a strong campaign to defeat it. On Feb. 23, 1994, a coalition that included the NRDC, Friends of the Earth, the Environmental Defense Fund, National Wildlife Federation, National Audubon Society, Sierra Club, Citizen Action, U.S. Public Interest Group, and Physicians for Social Responsibility wrote to members of the House of Representatives urging them to oppose H.R. 3392. A March 4, 1994, memo from Erik Olson, a lobbyist for NRDC, to the heads of the NRDC, National Audubon Society, National Wildlife Federation, the Environmental Defense Fund, Friends of the Earth, and the Sierra Club cited specific actions to defeat reauthorization of the SDWA through ‘‘immediate CEO meetings to ask for a delay’’ and ‘‘to allow time to organize a stronger media, grass roots and lobbying campaign’’ (Olson 1994a). Environmental lobbyists plan ‘‘to pour major resources’’ into their efforts on the SDWA, and ‘‘may move to a kill strategy.’’ This strategy was successful in achieving delay of the markup for S. 1547; environmentalists convinced Senator Baucus to delay the markup from March 15 to March 24, which allowed time for the coalition of environmental groups to place a full-page ad in the New York Times. The ad appeared the day markup began and denounced actions by water utilities that the environmental groups believed to be aimed at weakening the SDWA health standards. On March 14, 1994, NRDC released a report titled Victorian Water Treatment Enters the 21st Century (Cohen and Olson 1994). This study was designed specifically to influence SDWA reauthorization. It presents a critique of current water treatment practice and proceeds to make the judgment that water utilities have been irresponsible in their choices for treatment and maintenance. This charge was rejected by water suppliers in general (Parmelee 1994). The report encouraged opposition to H.R. 3392 and S. 1920. However, the report contained inconsistencies and was characterized as being ‘‘laced with the language of propaganda’’ (Parmelee 1994, Waterweek 1994). The NRDC hosted a press conference on July 17, 1994 to release a 1992=93 update of their report, Think before You Drink (Olson 1994b). The report stated that between 1992 and 1993, one out of five Americans drank water contaminated by unlawfully high levels of toxic chemicals, microbes, and other pollutants, or water

88

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

that was inadequately treated for those pollutants. The report was released simultaneously in more than 50 locations throughout the United States. Prior to the press conference, Erik Olson, NRDC senior attorney, and USEPA Administrator Carol Browner appeared on Good Morning America to discuss the NRDC report. Reauthorization Dies in Closing Days With the end of the 103rd Congress close at hand, pressure to take action on an SDWA measure increased. After many months of negotiation and delay, the House Health and Environment Subcommittee finally took action to mark up H.R. 3392 on Sept. 20, 1994, more than one year after its introduction. The full House Committee on Energy and Commerce also took action that day to pass H.R. 3392, after making substantial amendments. The full House of Representatives passed H.R. 3392 under suspension of the rules on Sept. 27, 1994, less than 2 weeks before adjournment. Disagreement over procedural strategy and legislative language killed the 103rd Congress’ chances to reauthorize the SDWA. Because of limited time, convening a formal conference committee was not possible. This meant that the committee staffs were faced with developing a compromise between S. 2019 and H.R. 3392 that would be acceptable to both chambers. Such a task was an impossible dream. For example, S. 2019 included a risk assessment amendment offered by Senator Johnston (D–La) that was strongly opposed by environmental groups. The Senate overwhelmingly passed this amendment and (judging by earlier votes in the House of Representatives on USEPA cabinet legislation) the majority of the House would also have supported it. Congressman Henry Waxman (D–CA) personally visited the Senate floor and lobbied senators to block the SDWA bill in the closing days of Congress to try to avoid having to consider a Senate-passed bill with risk assessment provisions. Disputes over amendments regarding takings, private property rights, and Davis–Bacon labor provisions also contributed to doom passage of an SDWA bill in the 103rd Congress. In the end, pure election-year politics regarding non-SDWA issues killed the SDWA. 4.6.5

USEPA Redirection of Regulatory Priorities

Limited resources forced USEPA to determine how many regulations can be funded and in what order. The agency initiated a process in late 1994 to redirect its regulatory priorities, which had a significant effect on discussions regarding SDWA reauthorization. A draft strategic plan was prepared in December 1994 (USEPA 1994b). Based on discussions of this plan, USEPA asked the U.S. District Court for Oregon to extend certain regulatory deadlines so that new priorities may be set for the highest-risk substances (BNA 1995). The request for an extension was submitted Jan. 9, 1995, in an amended consent decree signed by Robert Perciasepe, USEPA’s assistant administrator for water (U.S. Court of Appeals 1995). An extension was granted until Aug. 1, 1995, for USEPA to develop new rulemaking schedules. This deadline was extended several times because of Congressional delays in finalizing the agency’s FY 1996 budget. USEPA initiated discussions on possible realignment of its priorities with the public at a meeting held Jan. 19, 1995. Select groups were asked to help the agency

4.6 1996 SDWA AMENDMENTS

89

select and shape a range of potential actions to refocus and redesign the nation’s drinking water program.

Strengthening the Safety of Our Drinking Water To initiate the comprehensive review and redirection of the federal safe drinking water program, USEPA released a report, Strengthening the Safety of Our Drinking Water, in March 1995 (USEPA 1995a). This report included the following agenda for action intended to influence reauthorization of the SDWA: 1. 2. 3. 4. 5.

Give Americans more information about our drinking water Focus safety standards on the most serious health risks Provide technical assistance to protect source water and help small systems Reinvent federal–state partnerships to improve drinking water safety Invest in community drinking water facilities to protect human health

Also included in the report was a discussion of the following eight topics that were to be the focus of stakeholder meetings: regulatory reassessment, scientific data needs, treatment technology, health assessment, analytical methods, source water protection, small systems capacity building, and focusing and improving implementation.

Stakeholder Meetings USEPA held a series of public meetings in 1995 to gain input on how USEPA should redirect and improve its drinking water programs (USEPA 1995b, 1995c, 1995d). These meetings resulted in the development of a priority ranking of contaminants to be regulated that was released June 21, 1995 (Auerbach 1995a). Stakeholders indicated at the initial regulatory reassessment meeting on March 13, 1995, that they did not want to address existing regulations. USEPA recognized that a statutory mandate to review existing rules exists. However, the agency did not have the resources to conduct these reviews, and had no schedule to do so. The agency had planned to consider contaminants regulated in the past as candidates for future priority lists when new information indicates that they should be rereviewed (Auerbach 1995b).

USEPA Drinking Water Redirection Plan On Nov. 29, 1995 (USEPA 1995e), USEPA released for public comment a draft comprehensive drinking water program redirection plan (USEPA 1995f ). This document reported the results of the stakeholder meetings mentioned above and included a priority listing of activities. The priorities and principles proposed in this document served as a basis for discussion of needed revisions to the SDWA. The final National Drinking Water Program Redirection Strategy report was issued in June 1996 (USEPA 1996).

90

4.6.6

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

104th Congress Activity

The 104th session of Congress began on Jan. 4, 1995, but the signs of change were evident following the November 1994 elections. New members, new committee structures, and a new political order set the stage for Congress to shape and consider legislative proposals inconceivable in prior years. House and Senate Pass Reauthorization Bills In the House of Representatives, the SDWA was charged to the Commerce Committee (previously the Energy and Commerce Committee), chaired by Rep. Thomas Bliley (R–VA). Bliley cosponsored H.R. 3392 in the 103rd Congress. The Health and Environment Subcommittee, chaired by Michael Bilirakis (R–FL), first considered SDWA legislative proposals in the House. To stimulate discussion, Rep. John Dingell (D–MI) introduced H.R. 226, a bill identical to H.R. 3392 as passed by the House in the 103rd Congress. On Dec. 7, 1995, Rep. E.G. (Bud) Shuster (R–PA) introduced H.R. 2747. This bill amended the Federal Water Pollution Control Act (Clean Water Act or CWA) to create water supply infrastructure accounts within existing CWSRF for the state’s use in making loans for the construction of and improvements to drinking water supply infrastructure. This bill rekindled a longstanding jurisdictional dispute between the House Committee on Transportation and Infrastructure and House Commerce Committee, which is responsible for the SDWA. In the Senate the SDWA was under the jurisdiction of the Environment and Public Works Committee chaired by John Chafee (R–RI). Senator Chafee chaired the committee when the SDWA was amended in 1986 and has a more liberal view of environmental protection than his conservative Republican colleagues. The Subcommittee on Drinking Water, Fisheries, and Wildlife, chaired by Dirk Kempthorne (R–ID), was charged with drafting an SDWA reform bill. S. 1316 was introduced Oct. 11, 1995, and hearings were held Oct. 19, 1995. Markup of the bill took place on Nov. 7, 1995 (Senate Report No. 104-169), and the bill was passed by the Senate Nov. 29, 1995. Following passage of S. 1316, the focus of SDWA reauthorization activity shifted to the House of Representatives. On Dec. 12, 1995, Rep. Timothy P. Johnson (D–SD), introduced H.R. 2762 to address concerns regarding regulation of sulfate. The provisions of this bill mirrored S. 1316. Hearings on the SDWA held Jan. 31, 1996 by the House Health and Environment Subcommittee provided the basis for discussions to develop a bipartisan bill in the House. Discussion among Commerce Committee staff to develop a bipartisan SDWA reauthorization proposal progressed for several months. On March 6, 1996, Rep. Pomeroy (D–ND) introduced H.R. 3038, a bill similar to S. 1316. On March 26, 1996, the majority staff floated a comprehensive proposal, which stimulated several additional proposals and counterproposals in an attempt to negotiate a bipartisan agreement. On April 18, 1996, Rep. Waxman introduced H.R. 3280, the Water Quality Public Right-to-Know Act of 1996. This bill required each community water system to issue a report at least once annually to its consumers on the level of

4.6 1996 SDWA AMENDMENTS

91

contaminants in the drinking water purveyed by that system. On April 23, 1996, Rep. Lowey introduced H.R. 3293, the Safe Drinking Water Estrogenic Substances Screening Program Act, to establish a screening program for estrogenic substances. The House Subcommittee on Health and Environment met on June 6, 1996, in an open markup session to consider draft SDWA legislation. The Subcommittee unanimously approved the introduction of a clean bill for full consideration. The bipartisan bill, H.R. 3604, the Safe Drinking Water Act Amendments of 1996, was introduced by Rep. Bliley (R–VA) on June 10, 1996. The House Commerce Committee met in an open markup session on June 11, 1996, and ordered H.R. 3604 to be reported to the House, as amended (House Report 104-632). H.R. 3604 was passed by the House of Representatives on June 25, 1996, under suspension of the rules. Public Law 104-182 Enacted A conference committee was formed to resolve the differences between the Senate and House SDWA bills. The conference committee report was filed on Aug. 1, 1996 (Conference Report 104-741). The House and Senate both approved the conference report on Aug. 1, 1996. The SDWA Amendments of 1996 were signed into law as Public Law 104-182 by President Clinton on Aug. 6, 1996. The SDWA amendments of 1996 made substantial revisions to the SDWA and 11 new sections were added (SDWA 1996, Pontius 1996). Statutory requirements in the 1996 SDWA amendments and related deadlines are summarized in Table 4.4. USEPA has aggressively pursued implementation of the 1996 SDWA amendments. A section-by-section summary of the SDWA is provided in Appendix C, and the full text appears in Appendix D. However, one issue is worthy of discussion to conclude this section. During deliberations on the SDWA amendments in the 104th Congress, both the U.S. Senate and the House of Representatives approved legislative changes and report language that would have changed the SDWA legislative history regarding maximum contaminant level goals for carcinogens. At the conference, the Senate receded from its legislative provision and report language (found in Senate Report 104-169, pp. 30–33) with respect to maximum contaminant level goals for carcinogens. The House receded from all its report language on the same subject (House Report 104-632, the first paragraph on p. 28). The conferees agreed that the SDWA Amendments of 1996 make no changes to the provision or legislative history for MCLGs (Congressional Record 1996).

4.7 PUBLIC HEALTH SECURITY AND BIOTERRORISM PREPAREDNESS AND RESPONSE ACT In the aftermath of the Sept. 11, 2001, terrorist attacks on the World Trade Center in New York City and the Pentagon in Alexandria, Virginia, Congressional staff and committees conducted investigations and hearings to identify needed measures to ensure the security of public water systems in the United States. Several bills were introduced in each chamber that addressed in some way drinking water system

92

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.4 Statutory Requirements in the 1996 SDWA Amendments and Related Deadlines 1997 February 2, 1997 Report to Congress—Drinking Water Infrastructure Needs Survey (including Indian tribes) Develop plan for additional research on cancer risks from exposure to low levels of arsenic (consult with NAS, other stakeholders) Develop study plan to support development of the DBPs=microbial pathogen rules (in consultation with the Secretaries of HHS and Agriculture) Complete review of existing state capacity development efforts and publish information to assist states and PWSs with capacity development efforts Initiate partnership with states, PWSs, and the public to develop information for states on recommended operator certification requirements (released 2=28=97) Drinking Water State Revolving Fund (DWSRF) Guidelines (no statutory deadline) (agreement 2=28=97) Contract with NAS to conduct peer-reviewed assessment of the health risk reduction benefits associated with various radon mitigation alternatives (no statutory deadline) (released 3=12=97) Develop allotment formula for states on the basis of 1997 Drinking Water Needs Survey (no statutory deadline) August 6, 1997 Issue guidelines for alternative monitoring requirements Guidance establishing procedures for state application for groundwater protection grants Publish list of technologies that meet the Surface Water Treatment Rule for systems serving 10,000–3300 persons, 3300–500 persons, and 500–25 persons Guidance to states for developing source water assessment programs Guidance to states to assist in developing source water petition programs State Primacy Agencies submit to USEPA a list of community water systems and NTNC water systems that have a history of significant noncompliance and reasons for noncompliance 1998 January 1, 1998 States submit to USEPA first (annual) compliance report February 6, 1998 Publish a list of contaminants not subject to any proposed or final national primary drinking water regulation (must include sulfate) Publish information to assist states in developing affordability criteria Publish information on recommended operator certification requirements, resulting from partnership with states, public water systems, and the public July 1, 1998 Issue first (annual) report summarizing and evaluating state compliance reports August 6, 1998 Publish guidelines for small system water conservation programs Promulgate regulation on consumer confidence reports Review and revise as necessary existing monitoring requirements for not fewer than 12 contaminants

4.6 1996 SDWA AMENDMENTS

TABLE 4.4

93

(Continued)

Publish guidance on variance technologies for existing regulations for systems serving 10,000–3300 persons, 3300–500 persons, and 500–25 persons Promulgate regulations for variances and exemptions Publish list of technologies that achieve compliance for existing rules (except SWTR) for systems serving 10,000–3300, 3300–500, 500–25 Publish guidance on capacity development describing legal authorities and other means to ensure that new community water systems and nontransient, noncommunity water systems demonstrate capacity Conduct waterborne disease occurrence studies (with the Centers for Disease Control and Prevention) End of transition period for water suppliers determined to be public water system as a result of modifications to Sec. 1401(4) (constructed conveyances) November 1998 Promulgate Stage I Disinfectants and Disinfection Byproducts Rule Promulgate Interim Enhanced Surface Water Treatment Rule 1999 February 1999 Complete sulfate study with the Centers for Disease Control and Prevention to establish a reliable dose–response relationship Publish guidelines specifying minimum standards for certification and recertification of water system operators Publish health risk reduction benefits=cost analysis for potential radon standards Deadline for State Primacy Agency submission of programs for source water assessments August 6, 1999 Report to Congress on state groundwater protection programs Propose radon standard Establish National Contaminant Occurrence Data Base Promulgate final regulation establishing criteria for a monitoring program for unregulated contaminants September 1999 UIC Class V study (judicial deadline) October 1999 Final determination on whether states have legal authorities or other means in place and are implementing to ensure new system capacity (for purposes of DWSRF withholding determination) UIC Class V rule (judicial deadline) December 1999 Promulgate rule on public notification 2000 January 1, 2000 Propose standard for arsenic August 2000 Promulgate a regulation for filter backwash recycling within the treatment process of a PWSS, unless addressed in SWTR Report to Congress on DWSRF transfer of funds Promulgate final radon standard (continued )

94

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.4

(Continued)

Conduct studies to identify subpopulations at greater risk and report to Congress October 2000 Determine whether states have met Sec. 1419 requirements related to capacity development strategy (for purpose of DWSRF withholding determinations) November 2000 Promulgate Final LT1 Enhanced Surface Water Treatment Rule Promulgate final rule on radionuclides (judicial deadline) Promulgate final rule on groundwater determining when disinfection is necessary (USEPA schedule) 2001 January 1, 2001 Promulgate final standard for arsenic February 2001 2nd Needs Survey Report to Congress 2nd Needs Survey for Indian Tribes August 2001 Determine state compliance with operator certification guidelines for purposes of DWSRF withholding Determine whether to regulate at least five contaminants from contaminant candidate list State Primacy Agencies Report to USEPA on success of enforcement mechanisms and assistance efforts in capacity development November 2001 States complete local source water assessments With Fiscal year (FY) 2003 Budget Report to Congress—evaluation of effectiveness of state DWSRF loan funds 2002 May 2002 Promulgate Stage II Disinfection Byproducts Rule (delayed) Promulgate LT2 Enhanced Surface Water Treatment Rule (delayed) Promulgate Phase II rule on UIC Class V wells September 2002 States submit publically available report to governors on efficacy of state capacity development strategy and progress in implementation 2003 May 2003 Extension deadline for states to complete local source water assessments August 2003 Propose MCLG and national primary drinking water regulation for any contaminant selected for regulation from contaminant candidate list 2005 February 2005 Final MCLG and rule for any contaminant selected for regulation from contaminant candidate list 3rd Drinking Water Needs Survey for States and Tribes

4.8 FUTURE OUTLOOK

95

security. On Dec. 11, 2001, Rep. Billy Tauzin (R–LA) introduced H.R. 3448 to improve the ability of the United States to prevent, prepare for, and respond to bioterrorism and other public health emergencies. On Dec. 20, 2001, the provisions of S. 1765, introduced by Sen. Bill Frist (R–TN) were incorporated into H.R. 3448. After consideration by both chambers, a conference committee was formed, that generate conference report H. Report 107-481. The conference report was passed by the House and Senate on May 22, 2002 and May 23, 2002, respectively. The Public Health Security and Bioterrorism Preparedness and Response Act of 2002 was signed into law June 12, 2002, as Public Law 107-188. Title IV, Drinking Water Security and Safety, requires water systems to conduct vulnerability assessments, develop emergency response plans, and take other actions. Water system security provisions are reviewed in Chapter 24.

4.8

FUTURE OUTLOOK

The SDWA has been revised and rewritten since it was first enacted in 1974 to establish the federal program administered by USEPA to ensure safe drinking water in the Unites States. The text has grown and expanded to address unforeseen issues (Fig. 4.2) and provide authorization of federal funding to administer drinking water programs and provide for the DWSRF (Fig. 4.3). When funding authorizations end, reauthorization is needed, affording legislators an opportunity to make other revisions to the law. Infrastructure funding needs will continue to draw congressional attention. On April 11, 2002, the House Subcommittee on Environment and Hazardous Materials held a hearing on drinking water needs and infrastructure. A unique combination of social, scientific, and political forces shape the content of the SDWA. In this regard, it is no different than any other major piece of legislation passed by Congress. The SDWA will always be a work in progress, needing periodic amendments to meet current needs. In Chapter 5, the general U.S. govern-

Figure 4.2

Growth of the SDWA text.

96

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Figure 4.3

Growth of SDWA authorized funding.

mental structure and legislative process are explained in more detail and the authorization and appropriations process discussed. The interplay of social, scientific, and political forces that have shaped the SDWA are explored. The forces of change and how those changes may shape future versions of the law are reviewed.

ACKNOWLEDGMENTS This chapter was adapted and updated from ‘‘The History of the Safe Drinking Water Act’’ prepared by the same author and published in the public domain by USEPA on the Agency’s Website for the SDWA 25th Anniversary, displayed during calendar year 1999.

REFERENCES ASDWA. 1993. Over $2.7 billion needed for SDWA infrastructure this year. ASDWA Update VIII:1. Auerbach, J. 1994. Costs and benefits of current SDWA regulations. J. Am. Water Works Assoc. 86:69. Auerbach, J. 1995a. Letter from J. Auerbach, USEPA OGWDW, to regulatory reassessment stakeholders. Regulatory Reassessment: Final Summary and Rankings, June 21, 1995. Auerbach, J. 1995b. Letter from Janet L. Auerbach, USEPA Drinking Water Standards Division, Washington, DC, to Fred Pontius, American Water Works Association. Denver, Aug. 22, 1995. Baker, M. N. 1981. The Quest for Pure Water, 2nd ed., Vol. I. New York: McGraw-Hill and American Water Works Association.

REFERENCES

97

Bellar, T. A., J. J. Lichtenberg, and R. C. Kroner. 1974. The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66:12:703. BNA. 1995. BNA National Environment Daily. Drinking Water: More Time to Regulate Contaminants Requested by EPA, Agency Officials Say. Jan. 24, 1995. Washington, DC: The Bureau of National Affairs, Inc. Borchardt, J. A. and G. Walton. 1971. Water Quality. In Water Quality and Treatment, 3rd ed. New York: McGraw-Hill and American Water Works Association. Browner, C. M. 1993. Annual Conf., National Association of Towns and Townships, Washington, DC, Sept. 8, 1993. Carroll, B. 1993. Letter to state members and NRWA directors regarding SDWA reauthorization battle, Sept. 28, 1993. Center for Resource Economics 1993. Annual Review of the U.S. Environmental Protection Agency. Washington, DC: Center for Resource Economics. Cohen, B. A. and E. D. Olson. 1994. Victorian Water Treatment Enters the 21st Century. Washington, DC: Natural Resources Defense Council. Conacher, D. 1988. Troubled Waters on Tap: Organic Chemicals in Public Drinking Water Systems and the Failure of Regulations. Washington, DC: Center for Study of Responsive Law. Conference Report 104-741. Conf. Report on S. 1316, Safe Drinking Water Act Amendments of 1996. Congr. Record (House), pp. H9678–H9703, Aug. 1, 1996. Congressional Record. 1992. Congr. Record (Senate), pp. S15103, Sept. 25, 1992. Congressional Record. 1996. Conf. Report on S. 1316, SDWA Amendments of 1996. Cong. Record (H. Rept. 104-741), pp. H9678–H9703, Aug. 1, 1996. Congressional Research Service. 1982. A Legislative History of the Safe Drinking Water Act. Washington, DC: U.S. Government Printing Office. Congressional Research Service. 1993. A Legislative History of the Safe Drinking Water Act Amendments 1983–1992. Washington, DC: U.S. Government Printing Office. Cook, M. B. and D. W. Schnare. 1986. Amended SDWA marks new era in the water industry. J. Am. Water Works Assoc. 78:66–69. Cotruvo, J. A. and C. Wu. 1978. Controlling organics: Why now? J. Am. Water Works Assoc. 70:590. Craun, G. F. 1986. In Waterborne Diseases in the United States, G. F. Craun, ed. Boca Raton, FL: CRC Press. Dean, N. L. 1988. Danger on Tap, the Government’s Failure to Enforce the Federal Safe Drinking Water Act. Washington, DC: National Wildlife Federation. Dean, N. L. 1989. Update. Danger on Tap, the Government’s Failure to Enforce the Federal Safe Drinking Water Act. Washington, DC: National Wildlife Federation. Domenici, P. V. 1994. Domenici wants changes to Safe Drinking Water Act. Press release, March 24, 1994. Dyksen, J. E., D. J. Hiltebrand, and R. F. Raczko. 1988. SDWA Amendments: Effects on the water industry. J. Am. Water Works Assoc. 80:30–35. Gilbertson, W. E. 1989. Letter to the Editor. J. Am. Water Works Assoc. 81:4. Gray, K. F. and S. J. Koorse. 1988. Enforcement: USEPA turns up the heat. J. Am. Water Works Assoc. 80:47–49.

98

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Harris, R. H. and E. M. Brecher. 1974. Is the water safe to drink? Part I: The problem. Part II: How to make it safe. Part III: What you can do. Consumer Reports 436 (June), 538 (July), 623 (Aug.). Hembra, R. L. 1990. Compliance Problems Undermine EPA’s Drinking Water Program. Testimony before the Subcommittee on Environment, Energy, and Natural Resources. Committee on Government Operations. U.S. House of Representatives GAO=T-RCED90-97. Washington, DC: General Accounting Office. House Report 98-1034. Safe Drinking Water Act Amendments of 1984 (Sept. 18, 1984). Washington, DC: U.S. Government Printing Office. House Report 99-168. Safe Drinking Water Act Amendments of 1995 (June 11, 1985). Washington, DC: U.S. Government Printing Office. House Report 99-575. Safe Drinking Water Act Amendments of 1996 (May 5, 1986). Washington, DC: U.S. Government Printing Office. House Report 104-632. Safe Drinking Water Act Amendments of 1996 (June 24, 1996). Washington, DC: U.S. Government Printing Office. Kessler, C. and D. Schnare. 1991. A Section by Section Analysis of Comments on Safe Drinking Water Act Reauthorization (March 8, 1991). Prepared for the National Drinking Water Advisory Council. Washington, DC: USEPA Office of Ground Water and Drinking Water. Kyros, P. N. 1974. Legislative history of the Safe Drinking Water Act. J. Am. Water Works Assoc. 66:566. LCCA. 1988. Lead Contamination Control Act of 1988. Public Law 100-572, Oct. 31, 1988. Washington, DC: U.S. Government Printing Office. McDermott, J. H. 1973. Federal drinking water standards—past, present, and future. J. Envirn. Eng. Div.—ASCE EE4(99):469. NAS. 1977. Committee on Safe Drinking Water. Drinking Water and Health. Washington, DC: National Academy Press. NAS. 1980a. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 2. Washington, DC: National Academy Press. NAS. 1980b. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 3. Washington, DC: National Academy Press. NAS. 1982. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 4. Washington, DC: National Academy Press. NAS. 1983. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 5. Washington, DC: National Academy Press. NAS. 1986. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 6. Washington, DC: National Academy Press. NAS. 1987a. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 7. Washington, DC: National Academy Press. NAS. 1987b. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 8. Washington, DC: National Academy Press. NAS. 1989. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 9. Washington, DC: National Academy Press. NAS. 1993. Subcommittee on Health Effects of Ingested Fluoride. Health Effects of Ingested Fluoride. Washington, DC: National Academy Press.

REFERENCES

99

NAS. 1995. Subcommittee on Nitrate and Nitrite in Drinking Water. Nitrate and Nitrite in Drinking Water. Washington, DC: National Academy Press. NAS. 1997. Committee on Small Water Supply Systems. Safe Water from Every Tap: Improving Water Service to Small Communities. Washington, DC: National Academy Press. NAS. 1998. Committee to Evaluate the Viability of Augmenting Potable Water Supplies with Reclaimed Water. Issues in Potable Reuse: The Viability of Augmenting Drinking Water Supplies with Reclaimed Water. Washington, DC: National Academy Press. NAS. 1999a. Committee on Drinking Water Contaminants. Setting Priorities for Drinking Water Contaminants. Washington, DC: National Academy Press. NAS. 1999b. Identifying Future Drinking Water Contaminants. Washington, DC: National Academy Press. NAS. 1999c. Committee on Risk Assessment of Exposure to Radon in Drinking Water. Risk Assessment of Exposure to Radon in Drinking Water. Washington, DC: National Academy Press. NAS. 1999d. Subcommittee on Arsenic in Drinking Water. Arsenic in Drinking Water. Washington, DC: National Academy Press. NAS. 2000a. Copper in Drinking Water, Washington, DC: National Academy Press. NAS. 2000b. Re-Evaluation of Drinking Water Guidelines for Diisopropyl Methylphosphonate. Washington, DC: National Academy Press. NDWAC. 1993. Safe Drinking Water Act Reauthorization Issues. A Draft White Paper, April 13, 1993. NRC. 2001. Classifying Drinking Water Contaminants for Regulatory Consideration. Washington, DC: National Academy Press. Oleckno, W. A. 1982. The National Interim Primary Drinking Water Regulations, Part I— Historical Development. J. Environ. Health 44:5. Olson, E. D. 1993. Think before You Drink. Washington, DC: Natural Resources Defense Council. Olson, E. D. 1994a. Memo to the Heads of the NRDC, National Audubon Society, National Wildlife Federation, the Environmental Defense Fund, Friends of the Earth, and the Sierra Club Regarding SDWA Reauthorization, March 4, 1994. Olson, E. D. 1994b. Think before You Drink: 1992–1993 Update. Natural Resources Defense Council, Washington, DC, July 27, 1994. Page, T., E. Talbot, and R. H. Harris. 1974. The Implication of Cancer-Causing Substances in Mississippi River Water: A Report by the Environmental Defense Fund. Washington, DC. Page, T., R. H. Harris, and S. S. Epstein. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55. Parmelee, M. A. 1994. NRDC report skews utility operations, goals. Mainstream 38(4):1. Pendygraft, G. W., F. E. Schegel, and M. J. Huston. 1979a. The EPA-proposed granular activated carbon treatment requirement: Panacea or Pandora’s box? J. Am. Water Works Assoc. 71(2):52. Pendygraft, G. W.; F. E. Schegel, and M. J. Huston. 1979b. Organics in drinking water: A health perspective. J. Am. Water Works Assoc. 71(3):118. Pendygraft, G. W., F. E. Schegel, and M. J. Huston. 1979c. Maximum contaminant levels as an alternative to the GAC treatment requirements. J. Am. Water Works Assoc. 71(4):174.

100

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Pontius, F. W. 1996. Overview of the Safe Drinking Water Act Amendments of 1996. J. Am. Water Works Assoc. 88:22–27, 30–33. Pontius, F. W. 1999. Complying with future water regulations. J. Am. Water Works Assoc. 91(3): 46–58. Ronnebaum, E. 1993. Letter to the Editor. USA Today (Sept. 29, 1993). Rook, J. J. 1974. Formation of haloforms during chlorination of natural water. Water Treat. Exam. 23:234–243 (Part 2). Schnare, D. W. 1990. Summary of Comments, Safe Drinking Water Act Implementation and Reauthorization Meeting. USEPA, OGWDW, Washington, DC, Sept. 26–27, 1990. SDWA. 1974. The Safe Drinking Water Act of 1974. Public Law 93-523, Dec. 16, 1974. Washington, DC: U.S. Government Printing Office. SDWA. 1977. SDWA Amendments of 1977. Public Law 95-190, Nov. 16, 1977. Washington, DC: U.S. Government Printing Office. SDWA. 1979. SDWA Amendments of 1979. Public Law 96-63, Sept. 6, 1979. Washington, DC: U.S. Government Printing Office. SDWA. 1980. SDWA Amendments of 1980. Public Law 96-502, Dec. 5, 1980. Washington, DC: U.S. Government Printing Office. SDWA. 1986. Safe Drinking Water Act Amendments of 1986. Public Law 99-339, June 19, 1986. Washington, DC: U.S. Government Printing Office. SDWA. 1996. Safe Drinking Water Act Amendments of 1996. Public Law 104-182, Aug. 6, 1996. Washington, DC: U.S. Government Printing Office. Senate Report 98-641. Safe Drinking Water Act Amendments of 1984. 98th Congress, 2nd Session. Sept. 28 (Legislative Day Sept. 24), 1984. Washington, DC: U.S. Government Printing Office. Senate Report 99-56. Safe Drinking Water Act Amendments of 1985. 99th Congress, 1st Session. May 15 (Legislative Day April 15), 1985. Washington, DC: U.S. Government Printing Office. Senate Report 104-169. Safe Drinking Water Act Amendments of 1995, Nov. 7, 1995. Washington, DC: U.S. Government Printing Office. Subcommittee on Health and Environment. 1985. Hearings before the Subcommittee on Health and the Environment of the Committee on Energy and Commerce, House of Representatives. Ninety-Ninth Congress, First Session. Safe Drinking Water Act Amendments of 1985—H.R. 1650. May 1, 1985. Serial No. 99-28. Washington, DC: U.S. Government Printing Office. Symons, G. E. 1974. That GAO Report. J. Am. Water Works Assoc. 66(5):275. Symons, J. M. 1974. Chlorinated Organics Workshop. In Proc. 2nd AWWA Water Quality Technology Conf. Denver: American Water Works Association (AWWA), Dec. 2–3, 1974. Symons, J. M. 1984. A history of the attempted federal regulation requiring GAC adsorption for water treatment. J. Am. Water Works Assoc. 76(8):34. Symons, J. M. 2001a. The early history of disinfection by-products: A personal chronicle (Part I). Environ. Eng. (Jan.). Symons, J. M. 2001b. The early history of disinfection by-products: A personal chronicle (Part I). Environ. Eng. (April). Symons, J. M., T. A. Bellar, J. K. Carswell, J. DeMarco, K. L. Kropp, G. G. Robeck, D. R. Seeger, C. L. Slocum, B. L. Smith, and A. A. Stevens. 1975. National Organics Reconnaissance Survey for Halogenated Organics. J. Am. Water Works Assoc. 67:634.

REFERENCES

101

The States-Item. 1974. Cancer victims could be reduced—deaths tied to New Orleans water. The States-Item 98(129):1, New Orleans, LA, Nov. 7, 1974. Train, R. S. 1974. Facing the real cost of clean water. J. Am. Water Works Assoc. 66:562. USA Today. 1993a. USA Today (Sept. 27, 1993). USA Today. 1993b. Water crisis? Well, sort of. USA Today (Sept. 30, 1993). U.S. Court of Appeals. 1978. Environmental Defense Fund v. Costle, No. 752224, 11 ERC 1214, U.S. Court of Appeals, D.C. Circuit, Feb. 10, 1978. U.S. Court of Appeals. 1995. Donison v. Browner, DC Ore, CV 92-6280-HO; Miller v. Browner, DC Ore, CV 89-6328-HO; Frohwerk v. Browner, DC Ore, CV 90-6363-HO, Citizens Interested in Bull Run v. EPA, DC Ore, CV 92-1587-MA; Frohwerk v. Browner, DC Ore, CV 91-6549-TC, Jan. 9, 1995. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas, TX: USEPA Region VI, Surveillance and Analysis Division. USEPA. 1975a. New Orleans Area Water Supply Study. EPA Report EPA-906=9-75-003, Dec. 9, 1975. Dallas: USEPA. USEPA. 1975b. National Interim Primary Drinking Water Regulations. Fed. Reg. 40:59566– 59588. USEPA. 1976a. Promulgation of Regulations on Radionuclides. Fed. Reg. 41:28402–28409. USEPA. 1976b. Organic Chemical Contaminants; Control Options in Drinking Water. Fed Reg. 41:28991. USEPA. 1977. Recommendations of the National Academy of Sciences. Fed. Reg. 42:35764– 35779. USEPA. 1978a. National Organics Monitoring Survey. Cincinnati: USEPA Technical Support Division, Office of Drinking Water. USEPA. 1978b. Control of Organic Chemicals in Drinking Water. Proposed Rule. Fed. Reg. 43(28): 5756. USEPA. 1978c. Control of Organic Chemicals in Drinking Water. Notice of Availability. Fed. Reg. 43(130): 29135. USEPA. 1979. Control of Trihalomethanes in Drinking Water. Final Rule. Fed. Reg. 44: 68624. USEPA. 1980. Interim Primary Drinking Water Regulations; Amendments. Fed. Reg. 45:57332–57357. USEPA. 1981. Control of Organic Chemicals in Drinking Water. Notice of Withdrawal. Fed. Reg. 46:17567. USEPA. 1983. National Interim Primary Drinking Water Regulations; Trihalomethanes. Final Rule. Fed. Reg. 48:8406–8414. USEPA. 1993. Technical and Economic Capacity of States and Public Water Systems to Implement Drinking Water Regulations. EPA 810-R-93-001. Washington, DC: Office of Water. USEPA. 1994a. Report to the United States Congress on Radon in Drinking Water; Multimedia Risk and Cost Assessment of Radon. EPA 811-R-94-001. Washington, DC: Office of Water. USEPA. 1994b. Redirecting the Drinking Water Program. Draft. Office of Ground Water and Drinking Water. Washington, DC: Office of Water. USEPA. 1995a. Strengthening the Safety of Our Drinking Water. EPA 810-R-95-001. Washington, DC: Office of Water.

102

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

USEPA. 1995b. Notice of Public Meetings on Drinking Water Issues. Fed. Reg. 60:10391– 10393. USEPA. 1995c. Public Meeting on Drinking Water, Consumer Awareness Project. Fed. Reg. 60:30538. USEPA. 1995d. Public Meeting on Drinking Water Paperwork Burden Reduction. Fed. Reg. 60:37894–37895. USEPA. 1995e. Comprehensive Drinking Water Program Redirection Plan Availability of Draft Document and Request for Comment. Fed. Reg. 60:61254. USEPA. 1995f. Drinking Water Program Redirection Proposal. A Public Comment Draft. EPA 810-D-95-001. Washington, DC: Office of Water. USEPA. 1996. National Drinking Water Program Redirection Strategy. EPA 810-R-96-003. Washington, DC: Office of Water. USGAO. 1990. Drinking Water: Compliance Problems Undermine EPA Program as New Challenges Emerge. GAO=RCED-90-127. Washington, DC: U.S. General Accounting Office. USGAO. 1992a. Drinking Water: Consumers Often Not Well-Informed of Potentially Serious Violations. GAO=RCED-92-135. Washington, DC: U.S. General Accounting Office. USGAO. 1992b. Drinking Water: Widening Gap between Needs and Available Resources Threatens Vital EPA Program. GAO=RCED-92-184. Washington, DC: U.S. General Accounting Office. USGAO. 1992c. Drinking Water: Projects that May Damage Sole-Source Aquifers Are Not Always Identified. GAO=RCED-93-4. Washington, DC: U.S. General Accounting Office. USGAO. 1993a. Drinking Water: Stronger Efforts Needed to Protect Areas around Public Wells from Contamination. GAO=RCED-93-96. Washington, DC: U.S. General Accounting Office. USGAO. 1993b. Drinking Water: Key Quality Assurance Program Is Flawed and Underfunded. GAO=RCED-93-97. Washington, DC: U.S. General Accounting Office. USGAO. 1993c. Drinking Water: States Face Increased Difficulties in Meeting Basic Requirements. GAO=RCED-93-144. Washington, DC: U.S. General Accounting Office. USGAO. 1994. Drinking Water: Small Systems. GAO=RCED-94-40. Washington, DC: U.S. General Accounting Office. USPHS. 1925. Report of the Advisory Committee on Official Water Standards. Public Health Rept. 40:693 (April 10, 1925). USPHS. 1943. Public Health Service Drinking Water Standards and Manual of Recommended Water Sanitation Practice. Public Health Reports 58:69 (Jan. 15, 1943). USPHS. 1946. Public Health Service Drinking Water Standards. Public Health Rept. 61:371 (March 15, 1946). USPHS. 1962. Drinking Water Standards. Fed. Reg. 2152–2155 (March 6, 1962). USPHS. 1970a. Community Water Supply Study: Analysis of National Survey Findings. Pb214982. Springfield, VA: National Technical Information Service. USPHS. 1970b. Community Water Supply Study: Significance of National Findings. PB215198=BE, Springfield, VA: National Technical Information Service. U.S. Statues. 1893. Interstate Quarantine Act of 1893. U.S. Statutes at Large, Chap. 114, Vol. 27, p. 449, Feb. 15, 1893.

REFERENCES

103

VanDe Hei, D. and T. Schaefer. 2002. SDWA: Looking to the future. In Drinking Water Regulation and Health, F. W. Pontius, ed. New York: Wiley. Wade, S. 1993. Letter to the Editor. USA Today (Sept. 30, 1993). Waterweek. 1994. Enviros charge utilities with treatment technology failures. Waterweek 3(7):7 (March 28, 1994). Waxman, H. A. 1994. The next water crisis. The Washington Post (Jan. 19, 1994). Westrick, J. J., J. W. Mello, and R. F. Thomas. 1984. The groundwater supply survey. J. Am. Water Works Assoc. 76:52.

5 SDWA: LOOKING TO THE FUTURE DIANE VANDE HEI Executive Director, Association of Metropolitan Water Agencies, Washington, DC

THOMAS SCHAEFFER Regulatory Specialist, Association of Metropolitan Water Agencies, Washington, DC

5.1

INTRODUCTION

A unique combination of social, scientific and political forces shape the content of the Safe Drinking Water Act (SDWA). In this regard, it is no different than any other major piece of legislation passed by Congress. The impact these forces have on any given piece of legislation can be convoluted and at times mystifying, but the legislative process itself is rather straightforward. It is prescribed by procedures established in the both the U.S. Senate and the U.S. House of Representatives. In this chapter, the general U.S. governmental structure and legislative process are explained and the authorization and appropriations process discussed. This will provide a framework for exploring how the interplay of social, scientific, and political forces have shaped the SDWA since it was first passed in 1974, building on the SDWA history provided in Chapter 4. Using that foundation, the forces of change and how those changes may shape future versions of the law are reviewed. 5.2

U.S. GOVERNMENTAL STRUCTURE

The structure of the U.S. government is truly unique. It is set forth by the Constitution of the United States, initially adopted on Sept. 17, 1787, and subsequently Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

105

106

SDWA: LOOKING TO THE FUTURE

amended (Congress 1993). The United States is a federal, democratic republic, an indivisible union of 50 sovereign States. The government is at the local, state, and national levels ‘‘democratic’’ because the people govern themselves; ‘‘representative’’ because people choose elected delegates by free and secret ballot; and ‘‘republican’’ because the government derives its power from the will of the people. There are three primary branches of the U.S. government: the executive branch, the legislative branch, and the judicial branch, described briefly below, and fully detailed in the US Government Manual (GPO 2001).

5.2.1

The Executive Branch

The President of the United States is the administrative head of the executive branch of the government. The executive branch includes numerous agencies, both temporary and permanent, as well as 15 executive departments. The Cabinet is composed of the heads of the 15 executive departments (the Secretaries of Agriculture, Commerce, Defense, Education, Energy, Health and Human Services, Homeland Security, Housing and Urban Development, Interior, Labor, State, Transportation, Treasury, and Veterans Affairs, and the Attorney General). The U.S. Environmental Protection Agency (USEPA) is an independent agency under the executive branch, whose administrator is appointed by the President subject to confirmation by the Senate.

5.2.2

The Legislative Branch

All legislative powers are vested by the Constitution in a Congress of the United States that consists of a Senate and a House of Representatives. The Senate is composed of 100 members, 2 from each state, who are elected to serve for a term of 6 years. There are three ‘‘classes’’ of Senators, and a new class is elected every 2 years. The House of Representatives comprises 435 Representatives. The number representing each state is determined by population. Every state is entitled to at least one Representative. Members are elected by the people for 2-year terms, all terms running for the same period. The work of preparing and considering legislation is done largely by committees of both Houses of Congress. There are standing committees, select committees, joint commissions, special investigating committees, and joint committees. Table 5.1 lists the current (2002) committees of the U.S. Congress.

5.2.3

The Judicial Branch

Judicial power is vested by the Constitution in one Supreme Court, and in such inferior courts as the Congress may establish. The Supreme Court is composed of the Chief Justice and the number of Associate Justices as fixed by Congress,

5.3 HOW LAWS ARE MADE

107

TABLE 5.1 Committees of the U.S. Congressa U.S. House of Representatives Agriculture Appropriations Armed Services Budget Education and the Workforce Energy and Commerce Financial Services Government Reform House Administration International Relations Judiciary Resources Rules Science Small Business Standards of Official Conduct Transportation and Infrastructure Veterans Affairs Ways and Means

U.S. Senate Agriculture, Nutrition, and Forestry Appropriations Armed Services Banking, Housing, and Urban Affairs Budget Commerce, Science, and Transportation Energy and Natural Resources Environment and Public Works Finance Foreign Relations Governmental Affairs Health, Education, Labor, and Pensions Indian Affairs Judiciary Rules and Administration Small Business Veterans Affairs

a Committee homepages and schedules can be accessed through THOMAS (http:==thomas.loc.gov=), a legislative information service on the internet provided by the Library of Congress and named after Thomas Jefferson.

currently eight. The President nominates the Justices with the advice and consent of the Senate. The United States is divided geographically into 12 judicial circuits, including the District of Columbia. Each circuit has a court of appeals, created to relieve the Supreme Court of having to consider all appeals in cases originally decided by the federal trial courts. 5.3

HOW LAWS ARE MADE

The making of a law in the United States requires both the House of Representatives and the Senate pass an identical act (bill), that the act receive final approval, and that the law be made known to the people who are to be bound by it. By far, the most demanding part of the process is the approval of identical legislation by both the House and the Senate. Figure 5.1 summarizes the federal legislative process. 5.3.1

How Legislation Originates

This legislative process starts with the thought that federal legislation is necessary in some areas. This thought may be very basic and unexplored, or it may take the form

108

SDWA: LOOKING TO THE FUTURE

Figure 5.1

The federal legislative process.

of proposed legislation. The ideas for legislation may come from all corners, from individual constituents, citizens groups, business, state legislatures, industry, associations and similar groups, the administration, individual representatives, or combinations of these groups and others.

5.3.2

The Committee–Subcommittee Process

Once the idea is translated to legislative form and introduced by a sponsor in the House or Senate, the real action starts. Although the House and Senate procedures for considering and passing legislation differ greatly, they contain the same basic elements. The submitted bill is referred to a committee with jurisdiction over the subject matter for consideration. The committee, in turn, will normally refer the bill to one of its subcommittees. In the House, the bill or portions of it may be referred to a primary committee and one or more additional, secondary committees when the subject matter crosses committee jurisdiction. The committee process is perhaps the most important phase of a bill’s life. Many bills that lack support or general interest will languish in committee. For bills considered important, committees or subcommittees will meet to consider the bill and normally hold hearings to seek input from a variety of interested parties on the bill’s content. The committee or subcommittee may seek input from the U.S. General Accounting Office (USGAO) on the necessity for or desirability of the proposed

5.3 HOW LAWS ARE MADE

109

legislation. Government departments and agencies affected by the proposed legislation will develop positions on the bill for committee consideration. These positions will be reviewed by the U.S. President’s Office of Management and Budget (OMB) for consistency with the President’s program prior to submission. After hearings, the committee or subcommittee will meet in what is called a ‘‘markup’’ session to put the bill into final form prior to voting on it. If handled in subcommittee, it will then go back to the full committee as written or as amended in the markup with a favorable, unfavorable, or no recommendation. The subcommittee can also recommend that the bill be ‘‘tabled’’ or held indefinitely. The full committee, when either acting on a bill itself or receiving a report from a subcommittee, will also consider the bill and hold a meeting to consider amendments. The full committee may report the bill back to the House or Senate with a favorable or unfavorable recommendation or hold the bill in committee. Both the House and the Senate have procedures for discharging a bill held by a committee if it is a bill of sufficient interest. 5.3.3

Floor Action on Bills

Bills sent back to the parent body are accompanied by a committee report that describes the purpose and scope of the bill, and a section-by-section analysis explaining the intent of the section. This ‘‘report language’’ is a significant part of the legislative history and is often used by regulatory agencies and courts in determining ‘‘the intent of Congress’’ once a bill becomes law. Reports are printed and available prior to further consideration of a bill. The next step for a bill reported from committee is normally consideration and debate before the House or Senate, although there are several ways to expedite the process on noncontroversial bills. After general debate, amendments are accepted, debated, and voted on. Finally, the measure as amended is voted on and approved or disapproved by the full body. The bill is then sent to the other house for consideration. The other house may have already been considering a companion bill or a similar measure or may take up the bill of the other house for consideration. In any event, after a similar process leading to a final or engrossed bill it is extremely unlikely that the House and Senate versions are identical. One body may choose to adopt the differences or changes in the bill of the other house. But working out differences between the two versions is normally handled through the Conference Committee process. 5.3.4

The Conference Committee Process

The fact that there is a bill from each house dealing with the same subject is not sufficient to start the conference process. One house must first amend and pass the bill of the other house and request a conference to work out differences. Often this amendment may be in the nature of substituting the entire bill from the other body with the bill as originally passed by the first. Once both houses agree to a conference, conferees are appointed and meet. The Conference Committee is limited to

110

SDWA: LOOKING TO THE FUTURE

consideration of issues where differences exist and may not modify areas of agreement or add new subject matter. In most instances a compromise is reached between the two versions of the bill, and it is reported back to each house for action. In some cases where agreement cannot be reached, items in disagreement are reported back to each house, which may elect to agree with all or part of the other house’s actions. This may lead to several iterations of the Conference Committee process before each house has an identical bill to vote on. Once agreement is reached, a conference report detailing the actions of the committee in resolving differences is prepared. This report is also a key part of the legislative record. 5.3.5

Final Passage, Approval, and Publication

The bill becomes an enrolled bill after final approval by both houses and is ready for forwarding to the President. The debate each time a bill comes to the floor is captured in the Congressional Record, and becomes another key element in the legislative history. Each bill passed by Congress must be presented to the President for approval. The work of the Congress becomes law if the President signs it. The President has 10 days to sign passed legislation (excluding Sundays) or it becomes law without his signature. Within the 10 days, the President may veto the legislation by returning it to Congress with his objections. The act can still become law should the House and the Senate both override the veto by a two-thirds or higher vote. If the Congress adjourns during the 10-day period, the President is precluded from returning the bill with his objections, and it does not become law. This is known as a ‘‘pocket veto.’’ The first official publication of the statute is in the form known as the ‘‘slip law.’’ In this form, each law is published separately as an unbound pamphlet immediately after the law is approved. Each law is also published in the United States Statutes AtLarge, a collection of all laws passed by each session of Congress, and in the United States Code (USC), containing a consolidation and codification of general and permanent laws. As noted previously, this is a very simplified outline of how laws are passed. The procedures of both the House and Senate contain many nuances and differ greatly. House and Senate procedures are presented in Appendixes E and F, respectively. While some consider these procedures archaic, they have been developed over the years since the first Congress to ensure that laws receive full debate, that the views of the minority are aired, and that a standard of civility is maintained. 5.3.6

Authorization and Appropriation Measures

The authorization and appropriations process is derived from Senate and House rules that seek to bring discipline to the overall budget process. Laws such as the SDWA are authorizing measures; that is, they constitute the legislation that creates or continues an agency or program and authorizes the subsequent funding of those programs through the appropriations process. Authorizing legislation in the discretionary budget typically lays out funding for various programs for a 5-year period. Limiting the time on authorized funding is intended to ensure that programs are

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

111

reviewed and updated on a continuing schedule or eliminated when no longer needed. Although the rules of both houses prohibit the appropriation of funds for programs that are not authorized, the prohibition is enforceable only through the raising of points of order in each body. Therefore, although a 5-year period may have been exceeded and no funds are authorized for subsequent years, Congress typically appropriates funds for important programs anyway. For instance, the 1986 Amendments to the SDWA authorized funding through fiscal year 1992, yet the bill was not reauthorized until the Amendments of 1996. Funds were appropriated annually regardless of the lack of authorization. Often, reauthorization of statutes will wait until there are sufficient driving forces to significantly modify the underlying law. Laws such as the annual Veterans Administration (VA), Housing and Urban Development (HUD), and Independent Agencies Appropriations bill, which provides funding for the USEPA, are appropriation measures; that is, they provide funding for programs previously authorized. Interestingly, appropriations for specific programs are frequently only a fraction of the amount called for in authorization legislation. For example, the 1996 Amendments to the SDWA authorized one billion dollars per year for State Revolving Loan Funds to assist water systems in complying with the Act. Funds appropriated for that purpose for the years following the Act’s passage have been in the neighborhood of three-quarters of that amount.

5.4

FORCES SHAPING THE SDWA AND AMENDMENTS

The SDWA was developed and driven by political, social, and scientific forces working in tandem. As discussed in prior chapters, federal involvement in drinking water standards stems from the Interstate Quarantine Act of 1893. The Act, among other things, authorized the director of the U.S. Public Health Service (USPHS) to make and enforce regulations to prevent the introduction, transmission, or spread of communicable disease between states. This authority was first used regarding drinking water in 1912, when regulations were issued banning the use of common drinking cups on interstate carriers. Through the early 1960s, the USPHS issued a variety of standards covering both microbial and chemical contaminants, including coliform bacteria, lead, arsenic, fluoride and sulfate. The standards were applicable only to water supplied to interstate carriers and placed no legal obligations on individual community water systems. Nevertheless, all states adopted, with few changes, the USPHS standards as state regulations or guidelines. 5.4.1

The Setting for the 1974 SDWA

Quoting Charles Dickens (A Tale of Two Cities): It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope, it was the winter of despair.

112

SDWA: LOOKING TO THE FUTURE

Although Dickens could hardly have imagined the 1960s and 1970s in the United States, he was correct in his conclusion that his characterization of the age of the French Revolution, the setting for his novel, was no different from his own time or, indeed, any time. Nevertheless the two decades marking the last years of the industrial era and the beginnings of the information era were quite remarkable and shaped the formation of the original SDWA. The period witnessed the Apollo program resulting in the first person on the moon, and the nationally divisive war in Vietnam. It held the civil rights movement as well as the shift of the conservation ethic to environmental activism. The period marked the height of the Cold War, and the Watergate scandal. It witnessed the beginning and growth of the Women’s Movement and the Great Society programs intended to ensure that the less fortunate in American society were provided a safety net as well as help for the future. The period contained near its beginning the Cuban Missile Crisis, and at its end the Arab Oil Crisis. The year 1974 marked a year that would significantly change the Congress. That year 75 freshmen Democrats were elected to the U.S. House of Representatives. The ‘‘Class of 1974’’ was far more activist and liberal than the Democrats in the House, particularly the leadership. The freshmen led a revolt against that leadership displacing several powerful committee chairmen, transferring many of the leadership’s prerogatives to the Democratic Caucus, and setting up the now familiar subcommittee process to spread the power base. Where once the administration or lobbyists had to deal only with the House leadership and the heads of committees to affect legislation, now it was necessary to deal with the House almost on a member-by-member basis. This weakening of the old lobbying links opened the door for political advocacy organizations and political action committees (PACs) and their contributions. The Class of ’74 is still playing a major role in directing environmental legislation implementation through oversight functions and in subsequent environmental law development and reauthorization. The postwar boom provided individuals with both an overall prosperity and an increase in leisure time. Additionally, it was during this period that the baby boom generation was moving through college or into the workforce. The activism of the period, if not engendered solely by this group, had a lasting impact on their thoughts and values. In short, it was a time ripe for political activism, an activism that called into question the previous general faith in the government and its leaders. In this regard, it mirrored Dickens’ observation. At the same time there was declining faith in the ability of government to solve problems, there were increased demands for government to step in with federal solutions. It was also a time of remarkable technological achievement. Of these, perhaps the most important were those that paved the way for the information era to grow and accelerate. The transistor came into commercial use and was quickly outdated by the integrated circuit and microchip. Apple developed the first personal computer. The space program paved the way for communications satellites. And the military, through the Advanced Research Projects Agency (ARPA), developed ARPA net, the precursor to the Internet, email (electronic mail), and the World Wide Web. The space program paved the way for communications satellites, and the fax (Facsimile) machine was commercialized.

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

113

On the environmental front, the country suffered from the effects of the post– World War II industrial boom. One did not need to take air or water samples and look for contaminants in parts per billion concentrations. Individuals in large cities could see and smell the effects of air pollution. People could see and touch lakes, rivers, and streams with surfaces covered by petroleum products and other industrial wastes. Many waters, including Lake Erie, were reported to be dead zones or were quickly approaching that point with little aquatic life and numerous, massive fish kills. It was during this time that the Cuyahoga River was so polluted by industrial wastes that it burst into flames from spontaneous combustion; in 1961, the World Wildlife Fund was formed; in 1961, the Environmental Defense Fund; and in 1970, the Natural Resources Defense Council. Science added to the obvious signs of pollution, detailing invisible and heretofore immeasurable threats of heavy metals, radiation, hydrocarbons, pesticides, and chlorinated solvents. In 1962, Rachel Carson published Silent Spring, an indictment of pesticide use that caught the imagination of both the public and the press and became a national bestselling book. The 1960s witnessed the growth of local, grassroots organizations dedicated to solving local pollution problems. With the growth of these groups in number and size, media coverage of environmental issues increased on the local and state levels, bringing with it an increase in local and state action. Nevertheless, concern over the environment failed to rise to a national political level until the end of the decade. One national politician, Gaylord Nelson, then Senator from Wisconsin, had taken notice of environmental pollution and conceived of a national ‘‘teach-in’’ to raise environmental awareness and promote the issue to the national political level. The first Earth Day, which grew from his vision and work, succeeded far beyond what he could have imagined. On April 22, 1970, an estimated 20 million people gathered for Earth Day teach-ins, rallies, speeches, and protests. So anxious were members of Congress to be part of the action that Congress was adjourned for the day. This one event raised the environment to a national political issue almost overnight. The time marks the transition of the conservation movement and its organizations to more general environmental concerns as well as the birth of numerous new national and local environmental groups. In rapid succession, then-President Nixon proposed and formed the USEPA, the Council on Environmental Quality and the National Oceanic and Atmospheric Administration. Congress was exceptionally prolific, passing the National Environmental Policy Act; the Clean Air Act; the Resource Conservation and Recovery Act; the Federal Water Pollution Control Act; the Noise Control Act; the Marine Protection, Research and Sanctuary Act; the Federal Insecticide, Fungicide and Rodenticide Act; the Endangered Species Act; the Coastal Zone Management Act; the Port and Waterways Safety Act; and the Marine Mammal Protection Act. USEPA Administrator Russell E. Train characterized this progress in a 1975 address, ‘‘Never before in history has a society moved so rapidly and so comprehensively to come to grips with such a complex set of problems.’’ This legislation resulted from a bipartisan effort in Congress driven by broad public consensus that environmental laws were needed. This consensus included many leaders in business and industry who would later rethink their support as the laws were turned to regulations that affected their bottom line.

114

SDWA: LOOKING TO THE FUTURE

It was in this setting that the SDWA took form. As awareness grew through the 1960s of the impacts of industrial pollution and agricultural pesticide runoff, concern mounted over what impacts this contamination would have on drinking water drawn from polluted sources. A series of federal studies was undertaken to explore the issue. Perhaps the most important from the point of view of driving later legislation was the 1969 Community Water Supply Study by the USPHS, discussed in Chapter 4. The study found that bacterial and chemical contamination of community drinking water was widespread, particularly in small communities. Only 60% of systems checked met the few standards then in effect. The study also found that monitoring was rarely practiced and that the quality of state supervision varied greatly from state to state. The SDWA (Public Law 93-523) was passed in 1974 in response to the public and congressional concern, and preceding chapters have expanded on this early history. As USEPA proceeded to implement the law, the agency made significant progress in advancing the state of the science on all fronts. But not surprisingly, few major regulations were issued through the mid-1980s other than the interim standards based on USPHS work. 5.4.2

The Setting for the 1986 Amendments

Quoting Douglas M. Costle, USEPA Administrator, in the Carter administration: What does a reasonably prudent person do in the face of scientific uncertainty? The answer is: You take reasonable precautions. That is the essence of many EPA decisions. But it is not, unfortunately, how most of these laws are written.

Although the SDWA was amended in 1977, 1979 and 1980, few significant changes were made. The first major revision came in 1986. The 1986 amendments are often said to have been a reflection of dissatisfaction in Congress over the small number of contaminants USEPA had managed to regulate since 1974. But much more was going on. The Oil Embargo at the end of the 1970s was a major national crisis. Fuel shortages and price controls were the order of the day. It was a time of doubledigit inflation and interest rates over 20%. The energy industry was called on to develop alternative sources of energy and reduce dependence on foreign oil. The industry complained that environmental regulations severely impeded their ability to do so. This theme was adopted by other businesses and industries and expanded to the sound bite that overregulation was strangling the American economy. Although the goals of environmental and consumer legislation had been accepted by many leaders in these sectors during the 1970s, they were less enchanted when the goals were translated into required actions and expenditures through specific regulations. Even though the regulatory impact was small compared to that caused by the overall economic climate, calls for regulatory relief came from all sides. The breakdown of the old lobbying links that started with the changes in the House of Representa-

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

115

tives’ committee structure led to a boom in businesses and industry establishing offices in Washington, DC, hiring lobbyists and forming associations to make sure that their views were heard by Congress and the administration. In 1980, Ronald Reagan was elected President, promising both an economic turnaround and regulatory reform in his platform. Shortly after taking office, he signed an Executive Order requiring a cost=benefit analysis be submitted to OMB for review with each new proposed regulation. This Executive Order had no direct effect on environmental regulations because few of them allowed considerations of costs in establishing standards. But it had the indirect effect of making USEPA look for cost-effective solutions. The administration pursued regulatory reform so strongly that its efforts paradoxically ensured that little real reform would take place. Environmental groups were energized by the administration’s efforts, which they characterized as attempts to undo environmental protection. The ranks of these organizations grew dramatically, as did the amount of donations they received. The groups used public relations campaigns to focus public opinion against any changes in environmental laws. Although such groups were initially on the defensive against the administration, their efforts within a few years put the administration on the defensive. Two administration appointments aided this effort. Interior Secretary James Watt and USEPA Administrator Anne Burford, became lightening rods for environmental forces because of their approach to deregulation and past statements and actions on environmental issues. The two were subjected to intense scrutiny by House environmental subcommittees, and were, rightly or wrongly, demonized by the media. Both eventually resigned from their positions. The combination of all these forces turned the administration’s regulatory reform into regulatory retreat. Additionally, throughout this period, public support for protecting the environment remained strong. It was in this setting that the 1986 Amendments to the SDWA were considered. Congress, particularly the environmental subcommittees, had expected USEPA to accomplish more in the way of regulating contaminants. By 1986, only 23 contaminants were regulated. The majority of those were interim standards based on the USPHS standards in existence prior to 1974. Although blamed for not developing more standards, USEPA had not been idle during the period. The agency completed the National Organics Reconnaissance Survey in 1975, the National Organics Monitoring Survey (1976=77), the National Screening Program for Organics (1977–1981), and the Ground Water Supply Survey (1980=81). These surveys focused on chemical contamination, continuing the focus of the 1974 law and the perceived public health threat of these contaminants. The House Energy and Commerce and Senate Environment and Public Works Committees led the 1986 reauthorization of the Act. In the House, members of the environmentally active Class of ’74 now held senior positions in these committees and subcommittees. After their battles with the Reagan administration over regulatory reform, the Class of ’74 wanted to ensure that the administration had little flexibility to delay regulations. In addition, moderate House Republicans were not eager to jump on a failing regulatory reform effort, nor did they want to add to the Democrats’ reelection advantage. Democrats and Republicans joined forces, vowing

116

SDWA: LOOKING TO THE FUTURE

that neither party would accept an amendment to the SDWA that the other did not agree with. The Senate Committee united in much the same way as the House, working to rein in the administration’s regulatory flexibility. The Senate, generally proenvironmental, with proenvironmental Republicans on the environmental committees, more than compensated for other Republican members following the administration’s lead on regulatory reform. It was no surprise then that the Congress passed the 1986 SDWA Amendments by overwhelming majorities and that the most significant changes were to provide little flexibility for USEPA in pursuing contaminant standards. Nor was it a surprise that the President approved the legislation. That approval was driven, at least in part, by the need for the administration to dispel the negative effects of appearing antienvironmental, but mostly by the overwhelming (veto-proof) majorities by which the law passed both Houses. Ironically, it was also during this time period that USEPA solidified its policy establishing a health goal of zero for carcinogens in drinking water. This policy stood unchallenged until 1999, when USEPA proposed a goal other than zero for chloroform—a known human carcinogen. The 1986 law was very prescriptive to the point of limiting even the use of common sense on contaminants that did not need regulation. This overkill would later drive changes in the 1986 Amendments. Among other requirements, the 1986 law specified 83 contaminants to be regulated in stages by 1989 (with limited ability to substitute up to seven other contaminants with greater health risks). The USEPA Administrator at the time had advised Congress against pursuing this path, noting that doing so would preempt decisions based on good scientific evidence and could lead to unsound and unwarranted regulations. The amendments also required USEPA to establish a list of potential drinking water contaminants by 1988 and to regulate at least 25 of them by 1991. The list was to be regularly updated, and USEPA was required to regulate an additional 25 contaminants every 3 years starting in 1994. The 1986 Amendments to the Act also required USEPA to promulgate regulations requiring disinfection as a treatment technique for all public water systems, and to establish criteria under which surface water systems would be required to filter. 5.4.3

The Setting for the 1996 Amendments

The USEPA Science Advisory Board (USEPA 1990) stated that There are heavy costs involved if society fails to set environmental priorities based on risk. If finite resources are expended on lower-priority problems at the expense of higher-priority risks, then society will face needlessly high risks. If priorities are established based on the greatest opportunities to reduce risk, total risk will be reduced in a more efficient way, lessening threats to both public health and local and global ecosystems.

Environmental groups continued to grow in strength after passage of the 1986 Amendments until the William Clinton presidency, at which point their membership

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

117

declined, perhaps driven by the perception that a Democratic president would take care of environmental issues. Contrary to expectations, the entire issue of regulatory reform, which saw so little movement during the Reagan–Bush years, would resurface with a vengeance. The public was still concerned about the environment, but with the end of the Cold War, began looking inward at other social problems. Although the environment remained a national issue, it ranked far behind the economy, Social Security, healthcare, and education in the minds of the public and consequently Congress. USEPA’s policy in implementing the 1986 Amendments was that it was prudent to err on the side of safety in the face of insufficient scientific data for a clear-cut decision. The policy was tempered by the need for agency actions to be scientifically and legally defensible as well as technically and economically feasible. On one hand, not ensuring a minimum scientific basis for standards would lead to lawsuits; on the other, requiring communities to expend funds far in excess of any reasonable benefit or for no known benefit at all would lead to a backlash against the entire drinking water program. Working within these constraints led to delays in meeting the Act’s deadlines, resulting in lawsuits by environmental groups over missed deadlines. When USEPA did issue regulations in the early 1990s for the bulk of the 83 contaminants specified by Congress, the USEPA Administrator noted in press releases that, for most of the contaminants, the common factor was that they rarely occurred in drinking water and seldom at levels of public health concern. Even though this was true, water systems were required to monitor for the contaminants. To the surprise of many, these monitoring requirements, which were particularly expensive for small water systems, would prove to be one of the driving forces in the 1996 reauthorization effort. In the early 1990s, the issue of unfunded federal mandates came into focus. The genesis of the issue was both a reduction in federal grants to states and localities since the early 1990s coupled with an ever-growing list of federal requirements for state and local governments to spend money on programs mandated by Congress. Effectively, those officials saw increasing proportions of their budgets dedicated to federal requirements reducing their discretion to deal with local problems they considered more pressing. Organizations representing state and local governments prepared a number of studies detailing costs and lack of flexibility of various regulations. A typical study would describe programs or services that cities had to forego because of mandated programs, as well as listing requirements they deemed had minimal, if any, benefit. Some reports went on to contend that resources taken from local governments to meet mandates actually increased risks to public health and safety because of higher-priority programs that could not be funded. Drinking water programs were typically included in the studies because of the regulations governing chemicals that were rarely found. The organizations and the cities and states themselves took their issue to Congress, calling for mandate relief, through either additional financial assistance, fewer requirements, or greater flexibility in implementing laws and regulations. The unfunded mandates issue also generated political awareness of benefit-cost considerations. One way to reduce regulatory burdens is to ensure that the benefits of

118

SDWA: LOOKING TO THE FUTURE

regulation are justified by its costs. Additionally, proponents of benefit-cost considerations believed that its use would help ensure that the most important priorities were addressed first—those where states and localities could get the most ‘‘bang for the buck.’’ Another related issue was that of risk assessment. In order to develop the benefits side of benefit-cost equations, risk assessment was a necessary part. However, risk assessment can employ many assumptions or conservative default positions. This, in turn, led to calls for the use of sound science in developing regulations. Environmental regulations, which had first targeted business and industry, now were impacting not only states and localities but also individuals. This impact on individuals, particularly in their use of property, led to calls for government compensation to address ‘‘takings.’’ In the early 1990s, regulatory reform was defined as addressing the three issues just discussed: unfunded mandates, risk and benefit-cost analysis, and takings. Environmental groups began calling the three the ‘‘unholy trinity,’’ reflecting their concerns that these issues were merely being used as an excuse to roll back or ‘‘gut’’ environmental protections. Although such groups had outlined their issues for the Clinton administration including elevating USEPA to Cabinet status, addressing nonpoint sources of water pollution, dealing with environmental justice issues, and fostering ecosystem management, the ‘‘unholy trinity’’ argument put them on the defensive. The calls for regulatory reform were bipartisan in nature. They came from Republican and Democrat governors and mayors alike and the administration and Congress took notice. There were numerous bills addressing regulatory reform during the Bush administration at the beginning of the 1990s and throughout the following Clinton administration. The Clinton administration, which took office in 1993, expressed sympathy for the issue and made it a part of their program of ‘‘reinventing government.’’ One of the administration’s key efforts on the environmental front was elevating USEPA to Cabinet status. This effort was derailed in Congress by efforts to add risk and benefit-cost considerations to USEPA’s governing legislation. Environmental groups considered the threat of incorporating these issues so great that they dropped elevating USEPA as an issue, and it did not come up again during the two terms of the administration. Also at the beginning of the Clinton administration, there were a number of environmental laws, including the SDWA, that were due for reauthorization. The debate over inclusion of risk and benefit-cost considerations as well as the other regulatory reform issues in these statutes resulted in no action being taken despite Democratic control of the House and White House. Missteps by the Clinton administration, particularly over healthcare reform, and frustration that nothing was getting done in Congress, left an opening for Republicans in the 1994 Congressional elections. After 40 years of being in the minority, Republicans took power in the House. Republicans responded with their ‘‘Contract with America,’’ a list of 10 major issues they proposed for passage within the first 100 days of taking power. Phase II of the regulatory reform movement was about to begin.

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

119

In a virtual replay of the first years of the Reagan administration, the Republicans pressed their agenda very hard and very fast solidifying opposition on the Democratic side. Even regulatory reform, which had been a bipartisan issue, became partisan. Once again the approach to regulatory reform virtually assured that little real reform would take place. For example, the Republicans were able to pass the Unfunded Mandates Reform Act of 1995. Although the Act purports to deal with limiting Congress’ ability to place unfunded mandates on States and localities, in reality even if the requirements of the statute are invoked a simple majority vote by Congress is all that is needed to impose such mandates. During the early 1990s, problems with the 1986 Amendments began to surface. Small systems compliance problems increased as the number of regulated contaminants grew. Additionally, the fact—acknowledged by USEPA—that contaminants were being regulated that seldom occurred in drinking water and rarely at levels of public health concern added fuel to the fire. Paradoxically, USEPA was having a great deal of trouble regulating those contaminants that did occur in drinking water at levels of public health concern such as radon, arsenic, and the radionuclides. This was both because these regulations would involve significant costs to a large number of communities and because the science underlying any regulatory level was unclear. In the case of radon, concerns with regulating the generally very low levels that occurred in drinking water—low even when compared to ambient outdoor air levels—led Congress to write language into the fiscal year 1994, 1995, and 1996 USEPA appropriations bills prohibiting the agency from issuing a regulation. Another problem arose from the focus of the 1974 Act and the 1986 Amendments on chemical contaminants. As the agency began to focus on regulation of disinfectants and disinfection byproducts in the early 1990s, it realized that steps taken by water systems to reduce byproducts, such as reducing levels of disinfectant used, could effectively reduce microbial protection. The agency characterized this as a risk–risk tradeoff situation, which was not covered by the SDWA. In fact, the Act would require the reduction of the chemical pollutants regardless of the introduction of countervailing risks. Additionally, there was a great deal of scientific uncertainty surrounding the health effects of disinfectants and disinfection byproducts. Accordingly, USEPA looked for an innovative way to address the problem, settling on negotiating a regulation through the Federal Advisory Committee Act (FACA) process. The Federal Advisory Committee that was formed consisted of representatives from the drinking water community as well as from environmental and consumer groups and state and local governments. This group provided a forum that underscored shortcomings in the Act, including the risk–risk tradeoff issue, the lack of overall flexibility in the Act and the need to consider risk and cost-effectiveness of regulations. Additionally, another, but related problem, arose from the way standards were required to be set. The 1986 Amendments required USEPA to set maximum contaminant levels (MCLs) as close to maximum contaminant level goals (MCLGs) as feasible. Very little flexibility was allowed in considering relative costs and benefits at different MCL levels. MCLGs for suspected carcinogens were routinely set at zero

120

SDWA: LOOKING TO THE FUTURE

with the corresponding MCL set at the limit dictated by analytical measurement method capability. Many in the drinking water community feared that as analytical measurement capability improved, MCLs would constantly be lowered with no brake on the process. They knew that chasing zero would eventually become very expensive as lower and lower standards caused shifts to more expensive technologies, and that reasonably a point would be reached where any theoretical marginal benefit of the reduction would be minimal. They believed that consideration of incremental benefits and costs, as different regulatory levels were considered, would help inform USEPA’s risk management process and lead to more appropriate regulatory decisions. During this same time period, environmental groups sued USEPA for failure to meet statutory deadlines, and states became increasingly vocal about the expense of retaining primacy over the drinking water program. USEPA worked with the courts to extend deadlines and with the states to set priorities, but pressure to reform the 1986 Act was at an all-time high. It was in this context that the 1996 Amendments were considered. The Act was modified in 1988, but like the 1977, 1979 and 1980 changes, the modifications were incremental in nature. In this case, a requirement to recall lead-lined water coolers found in schools was added. After that, a number of SDWA reauthorization bills making significant changes were submitted but none were seriously considered until 1994. During 1994, both the House and Senate worked on legislation to reform the SDWA only to fail when Congress adjourned sine die on October 7 to campaign for the upcoming elections. The final days of the 103rd Congress provide a glimpse into how social and political forces can collide, bringing Congress to a standstill. On September 27, 1994, the House leadership blessed a plan to send the Safe Drinking Water Act Amendments of 1994 (H.R. 3392) to the floor on the Suspension Calendar. Since the Democratically controlled House did not want to go to a formal conference with the Senate because its bill contained risk assessment provisions, the USEPA Cabinet status bill and unwanted ‘‘property takings’’ language, the House chose to pursue a ‘‘take it or leave it’’ strategy. Rather than stripping S. 2019 of its content and sending it back to the Senate with the House bill inserted, the House simply passed its own bill, H.R. 3392, exactly as reported out of committee. By following this course of action, the House ensured that if the Senate wanted to reauthorize the SDWA in the few days remaining in the session, their only choice was to pass H.R. 3392 as its own. However, the Senate responded by attempting to ‘‘hotline’’ its own bill. The Senate stripped S. 2019 of the Cabinet bill, risk assessment, and property takings language, hoping to get it to the Senate floor on the consent calendar on Friday, September 30. The effort was unsuccessful. Not willing to call it off, Senate staff worked over the weekend to redraft S. 2019 incorporating many of the House provisions. Their only hope was to craft a bill that had complete agreement by all senators. In the end, the bill never made it to the Senate floor. For very different reasons, members of the Class of ’74 and regulatory reform proponents worked to defeat the measure not willing to concede the SDWA as a vehicle for pursuing their political agenda in the next Congress. In this way, the SDWA reauthorization debate moved to 1995 and a new Congress.

5.5 FUTURE AMENDMENTS TO THE SDWA

121

Pressure to reform the SDWA continued in 1995. In that year, both the House and Senate passed separate SDWA reauthorization bills, but the bills did not make it to conference before the end of the session. The House and Senate finally passed and reconciled bills in 1996. The final bill was supported by the Administration and signed by the President. It was also supported by the water supply community, state and local governments, and environmental groups. The 1996 amendments contained many compromises arising from the need for a bipartisan bill. Nevertheless, it marked a significant departure from past versions and from most other environmental statutes by allowing consideration of costs and benefits in setting standards, provisions to improve the scientific basis of decisions, and a State Revolving Loan Fund (SRLF) to provide loans and grants to systems needing to make improvements under the Act. 5.4.4

The Setting for the 2002 Amendments

Following the terrorist attacks of September 11, 2001, Congress and the Administration devoted themselves to security issues. Hearings were held, legislative proposals were introduced. At issue was how well prepared were U.S. agencies to prevent and respond to future terrorist activity. The water industry, USEPA, and the Centers for Disease Control had been working on security issues in general, and specific to drinking water supplies, for several years prior to 2001. This activity greatly accelerated, with additional funding and attention given by USEPA. Attention to potential threats to drinking water systems began to increase following the posting of a series of articles on the Internet by MSNBC in January 2002, which included a complete vulnerability analysis of threats to water systems. The public, the administration, and bipartisan Congressional concern for security in the broadest sense, including water supplies, provided the context for the Public Health and Bioterrorism Prevention and Response Act of 2002, discussed in Chapter 24. As mentioned below, ensuring water system security will continue to be a significant force shaping the SDWA. 5.5

FUTURE AMENDMENTS TO THE SDWA

As in the past, future Amendments to the SDWA will be driven by political, social, and scientific forces working in tandem. Although each force is complex on its own, any combination of factors could lead to innumerable future outcomes. Nevertheless, some general conclusions can be drawn from the past and some potential future directions projected. 5.5.1

Political Dimension

Political forces that shaped the early SDWA and its reauthorization discussed above provide insight about why those laws took the form they did. Because of the way the U.S. government is structured and the way laws are made in the United States, defeating proposed legislation is far easier than passing it. Measures that do survive to become laws are typically the products of bargaining and compromise involving not only legislators and the administration but also interest groups.

122

SDWA: LOOKING TO THE FUTURE

Environmental legislation, in particular, is seldom considered unless there is significant public support and demand for action. Laws normally evolve in incremental steps such as those seen in the 1977, 1979, 1980, and 1988 Amendments to the SDWA. In general, major changes to laws, such as those seen in the 1986 and 1996 Amendments, are relatively rare since they require an urgency for immediate action and strong public and bipartisan support. The Public Health Protection and Bioterrorism Prevention and Response Act of 2002 is a notable exception. Public concern over security of public water supplies and other public health measures in the aftermath of the Sept. 11, 2001, terrorist attacks on the United States drove bipartisan support and congressional action. The original 1974 Act was drafted in an era when the need for action on the environmental front was clear, and the Act had strong bipartisan support. The same Republican Nixon administration that formed the USEPA strongly supported the SDWA crafted by a House and Senate controlled by Democrats. The 1986 Amendments were as much a reaction to the Republican Reagan administration’s overreaching on regulatory reform in environmental areas as frustration with USEPA’s progress in issuing regulations. Its restrictiveness in specifying specific contaminants to be regulated and unrealistic timeframes for regulation, however, contained within it the seeds of its ultimate failure. The 1996 Amendments were driven by a series of implementation problems in the 1986 law and shaped by a bipartisan call from states and municipalities for a more rational, cost-effective way to approach environmental regulations. The 2002 Amendments embodied in the Public Health Protection and Bioterrorism Prevention and Response Act of 2002 were clearly driven by public concern over security of public water supplies and other public health measures in the aftermath of Sept. 11, 2001. Security concerns will likely shape Congressional initiatives and action for many years to come. Currently (2002) political campaigns for the 2002 midterm Congressional elections are beginning to gear up. Historically, midterm elections have turned against the party holding the White House, with seats held in Congress increasing for the party seeking to take the presidency in the next presidential election. Speculation is high regarding potential shifts in the balance of power within the U.S. Congress. Recent (at the time of writing) corporate scandals as well as difficult economic times will have an unmistakable impact on the 2002 midterm elections, as well as any future amendments to the SDWA.

5.5.2

Social Dimension

In order for an issue to demand political attention of the type leading to a new law or major amendments to existing law, it must rise to national prominence. To do this, it must become an issue in the eyes of the public and lead to public support for action. Public demand sometimes happens on its own, but more often is generated through special interests. Crisis situations compel Congress, the administration and even special interests to work together toward a solution. Issues, without crisis, grind

5.5 FUTURE AMENDMENTS TO THE SDWA

123

slowly through the process in search of support and often languish until a driving force (whether political or social) arises. For most environmental laws, the political power of environmental groups acts as a counterbalance to business and industry groups. In the case of the SDWA, the majority of the entities affected are states and municipalities, not business and industry. The political dynamics between environmental groups and public entities is often less harsh than with private industry, since they may share common goals although at times very different approaches. Major tensions between the groups arise more often over the details of protecting public health rather than the overall objective. Environmental groups strongly supported the 1986 Amendments with their emphasis on little flexibility and strict deadlines. Much of the power of environmental groups is gained through the enforcement of environmental laws through the courts. Through settlements, they have a say in the final schedule for regulations and, at times, portions of their content. Environmental groups have also sought to increase their power by promoting access to the regulatory process in environmental laws. The most obvious legal provisions are those that allow citizen suits under the various statutes. The public since the 1960s has increasingly identified themselves as supporting efforts to improve the environment, and today the vast majority of people do so. However, in specific cases, that general support may depend on who pays for improvements, where facilities to address pollution are located, and the degree of impact on an individual’s lifestyle. For example, the majority of those identifying themselves as environmentalists do not choose to drive the most fuel-efficient cars. Most environmental statutes approach the ‘‘who pays’’ issue from the ‘‘polluter pays’’ point of view. The public has generally supported these laws even when realizing that costs may eventually be passed to consumers through higher costs for goods and services. The public has also supported environmental improvement efforts funded by income tax or other general revenues. Support has not been as strong when costs must be borne directly by individuals. Rate increases to be borne by consumers for water and wastewater are often opposed. Opposition has also been seen when facilities need to be built to meet environmental requirements. Although the public generally supports such facilities they have resisted having such facilities located near them. The ‘‘not in my backyard’’ (NIMBY) phenomenon is a further indication that the environment is a priority as long as efforts to improve it do not significantly impact an individual’s lifestyle. Additionally, support for the environment as a national issue tends to vary as general economic conditions vary with support decreasing in worsening economic conditions. Overall, support for environmental efforts appears to be very broad, but not necessarily very deep. 5.5.3

Scientific Dimension

Although an important factor, science has taken a back seat to political and social forces in shaping the SDWA. The 1974 Act included few provisions related to

124

SDWA: LOOKING TO THE FUTURE

science. The 1986 Amendments tended to ignore the improvements in the science of risk assessment that had developed since passage of the 1974 Act. The Act’s requirement to regulate a specific list of contaminants regardless of their public health implications is testament to a lack of trust in science or those who would use it. Incorporation of science provisions in the 1996 Amendments was less a recognition by Congress of the importance of scientific and technical issues than a desire to deal with the problem of costs, particularly to small systems. In the face of bipartisan criticism that regulation of congressionally specified contaminants without regard to costs or benefits did not make sense, inclusion of provisions for the use of ‘‘good science’’ became a political necessity. Consideration of the political, social, and scientific forces involved in shaping the SDWA over more than three decades indicates that future changes to the SDWA are likely to be incremental in nature. This is the way laws normally evolve. Environmental groups are likely to oppose any tightening of the now mostly discretionary cost=benefit provisions of the law or the still very general scientific provisions. States and communities are likely to oppose any weakening of these provisions as well as any major changes to the overall regulatory process. Major crisis situations that may develop are likely to involve only one contaminant. If these are addressed at all in changes to the law, the changes will probably be specific to that contaminant rather than a major change in the structure of the statute. 5.5.4

Unresolved Issues

Two of the leading drivers behind the 1996 Amendments to the SDWA were not resolved: small system compliance and state primacy costs. Without some impetus to restructure small systems to improve economies of scale, their inability to finance compliance costs will persist. Monitoring relief, the State Revolving Loan Fund, grants for capacity development, small system variances, and other provisions of the 1996 Amendments simply provided a short-term bandage for what is a more pervasive structural and political problem. A further unresolved issue, and perhaps the most difficult to ultimately resolve, is the capacity of both USEPA and states to deal with the legislative and regulatory requirements of the Act. The demands put on USEPA and the states by all versions of the Act have been significant. USEPA’s levels of funding for the drinking water program have not been in line with the congressional demands placed on it. This is particularly true in the research and development (R&D) area but extends to all areas from regulatory development to enforcement. This is not peculiar to the drinking water program. Congress typically places more responsibility on federal departments and agencies than they are willing to fund since approved budgets reflect political and economic realities that may or may not be in line with mandated requirements. In the R&D area, this problem was foreseen in the 1996 Amendments. Expenditures on R&D are essential to the regulatory process outlined in the Act in order to develop the scientific and technical basis for decisions on standards. The Act

5.5 FUTURE AMENDMENTS TO THE SDWA

125

requires USEPA to set aside $10 million per year from the SRLF to fund health effects studies, but because of political considerations, the funds have never been reserved or utilized. The overall underfunding of the science necessary to carry out the Act may drive future changes to the Act. Since 1974, states have been given primacy in implementing and enforcing regulations under the Act. The Act and its amendments have provided grants to states to carry out these requirements. Public water system supervision (PWSS) grants have proven far too small to effectively implement the requirements placed on states even when supplemented by state funds. States have been reluctant to significantly increase their investment in state drinking water programs given their other funding needs and the federal nature of drinking water mandates. The lack of federal commitment to unfunded federal mandates was another of the drivers of the 1996 Amendments, and although PWSS grants were increased, the total is far below what is required for an effective program. Recent regulations, such as the Lead and Copper Rule, those dealing with microbial and disinfection byproducts, and the Filter Backwash Rule require extensive interaction between water systems and states in the form of studies. This trend is expected to continue with the growing realization that a one-size-fits-all regulatory approach can lead to very cost ineffective regulations. However, such regulations will serve to increase demands on already stressed state programs.

5.5.5

Emerging Issues

Future challenges to the drinking water community that may lead to changes in the SDWA are many and varied, and to a large extent interrelated. On September 11, 2001 the terrorist attacks on the World Trade Center in New York and the Pentagon in Washington, DC, had a profound impact on the nation, its economy, sense of well-being and influence around the world. Federal, state, and local governments and the private sector reordered priorities to enhance security against potential future attacks. Banking, finance, and telecommunications moved rapidly to harden security against cyber attack. The water, energy and transportation sectors deployed additional personnel and resources to protect critical infrastructures from both physical and cyber attacks. Nationally, Congress provided $40 billion for defense and domestic security, including approximately $80 million for drinking water vulnerability assessments. Congress has enacted amendments to the SDWA requiring water systems to conduct vulnerability assessments and emergency response plans. Additional enforcement authority and increased funding for security-related research is also being considered. As the nation learns more about terrorism and the potential for physical and biological, chemical, and radiological contamination of water supplies, the SDWA may be amended even further. On the contaminant front, endocrine disrupting chemicals and pharmaceuticals in drinking water will receive increased attention. Both the SDWA Amendments of

126

SDWA: LOOKING TO THE FUTURE

1996 and the Food Quality Protection Act of 1996 require a screening and testing program to determine if contaminants present in food and water may have endocrine-disrupting effects. The endocrine system regulates metabolism and a wide range of biological processes, such as control of blood sugar, growth and function of reproductive systems, so the occurrence of such contaminants in drinking water at levels of public health concern would cause major changes in drinking water treatment. The science on the issue is presently unclear, particularly at low contaminant dose levels. Pharmaceuticals have also been found at low levels in drinking water sources. A great deal of research is necessary before determining if these contaminants constitute a threat to treated drinking water. Population growth is leading to increased demands for water for both drinking and agricultural purposes. In some areas, these uses directly compete with efforts to restore water flow for ecological purposes or for protection of endangered species. Some sources of water will not be available for development or further development for drinking water purposes because of endangered species problems. Many groundwater aquifers are being drawn down for agricultural and drinking water purposes at rates far in excess of their natural recharge rates. While these problems are most visible in the arid Western states, they occur in all regions of the country. Increasing and conflicting usage demands on water sources will rise to the level where congressional action will be called for. While any congressional fix is likely to focus on other statutes, changes to the SDWA that encourage or require conservation measures are possible. Changes that address the increasing practice of wastewater recycle and reuse are also possible. Water infrastructure replacement challenges are already being felt and will increase in the future. Water distribution and wastewater collection systems have been constantly built and expanded over the past century. Such systems have a very long lifetime, but thousands on thousands of miles of the systems, particularly in the nation’s largest cities, are now reaching or have passed their useful life. The required investment is of such magnitude, in addition to the investments required by new regulations, that Congress has begun exploring the problem and a possible federal role in solutions through hearings. The infrastructure replacement challenge may spur changes to the SDWA’s SRLF provisions and funding levels. Environmental laws such as the SDWA were developed on a media by media (air, water, land) or contaminant (pesticides) basis. This structure has long been acknowledged to be inefficient for dealing with environmental problems that tend to occur across media. This arbitrary segmenting under the laws complicates or prohibits coordinated approaches to solving environmental problems. Although a watershed management approach to meet multiple environmental objectives has been recognized as the best approach, existing laws do not facilitate such an approach. There are strong forces at the federal level against integration of environmental laws since it would reduce the authority of the congressional committees overseeing each law as well as the turf of the various departments, agencies, and individual offices within those agencies that oversee regulations under each law. The 1996 Amendments to the SDWA required source water assessment programs and included a source water

5.6 OUTLOOK FOR MAJOR CHANGE

127

petition program to attempt to initiate at the state and local levels the very same coordination that the federal laws impede. The programs as presented in the law are not very effective, and the petition program in particular is likely to see little use. This area is one, however, that may see further work in future amendments. The trend in communications and information availability most evidenced by the dramatic growth of the Internet will continue. Advocacy groups as well as water suppliers will have enhanced opportunities to provide information to consumers. The so-called Consumer Right-to-Know provisions of the 1996 Amendments will likely be enhanced in future revisions to the law as communications technology advances. Already, there have been pilot projects involving water systems providing real-time or near-real-time water quality data to consumers via the Internet.

5.6

OUTLOOK FOR MAJOR CHANGE

Forces that have shaped SDWA legislation since the early 1970s indicate that any changes to the SDWA will occur incrementally. However, looking toward the future with an eye on the past doesn’t necessarily tell us where we’re headed in the long run. Major forces on the drinking water industry could bring about an abrupt transformation in legislation and regulation. Today, the industry is being forced to change in ways not previously considered. Globabilization, competition, security, and new demands for service and efficiency are moving utilities in new directions. As pressure on the resource itself grows from agriculture, population growth, and the environment, utilities will evolve in new and different ways. Corporate corruption, bankruptcies, and the economic downturn currently (2003) being experienced may affect the financial capacity of water systems to afford needed improvement. Out of necessity, drinking water utilities will be working with a larger regional or even national community to ensure that basic human and environmental needs are met, as Kader Asmai, former Minister for Water Affairs and Forestry in South Africa, put it, ensuring that there is ‘‘some for all, instead of more for some.’’ These trends could restructure the water industry and drive it to even higher levels of service, rendering irrelevant the contaminant-by-contaminant standard-setting process painstakingly detailed in the SDWA. Arguably, the SDWA has become what it is today because of the fragmented structure of the water industry in the United States. The law establishes various requirements based on system size and ability; it sets forth a standard-setting process targeted at large system capabilities, with provisions that are intended to provide latitude for smaller systems. Some will argue that it has failed, but Congress has attempted to develop a statute that creates programs that are implementable by a very fragmented and diverse industry. Restructuring of the water industry, along with new local pressures forcing industry changes to new technology, could result in the statute, as currently crafted, becoming obsolete.

128

SDWA: LOOKING TO THE FUTURE

REFERENCES AMWA. 2000. Safe Drinking Water Act Primer. Washington, DC: Association of Metropolitan Water Agencies. Blodgett, J. 1994. Environmental Reauthorizations and Regulatory Reform: Recent Developments. Report for Congress (95-3 ENR). Washington, DC: Congressional Research Service. Congress. 1993. Our American Government. 102D Congress, 2d Session. H.Doc. 102-192. Washington, DC: U.S. Government Printing Office. Copeland, C. 1996. Reinventing the Environmental Protection Agency: EPA’s Water Programs. Report for Congress (96-283 ENR). Washington, DC: Congressional Research Service. Dove, R. 2001. Enactment of a Law. U.S. Senate. Washington, DC: Library of Congress. GPO. 2001. US Government Manual 2001=2002. Washington, DC: U.S. Government Printing Office. Heniff, B., Jr., 1999a. Overview of the Congressional Budget Process. Report for Congress (RS20368). Washington, DC: Congressional Research Service. Heniff, B., Jr., 1999b. Overview of the Authorization—Appropriation Process. Report for Congress (RS20371). Washington, DC: Congressional Research Service. Johnson, C. W. 2000. How Our Laws Are Made. U.S. House of Representatives. Washington, DC: Library of Congress. Lee, M. R. 1994. Environmental Protection and the Unfunded Mandates Debate. Report for Congress (94-739 ENR). Washington, DC: Congressional Research Service. Lewis, J. 1990. The spirit of the First Earth Day. EPA J. (Jan.=Feb.). Nelson, G. 1980. Earth Day ’70: What it meant. EPA J. (April). Rosenbaum, W. A. 1995. Environmental Politics and Policy. Washington, DC: Congressional Quarterly Press. Streeter, S. 1999. The Congressional Appropriations Process: An Introduction. Report for Congress (97-684 GOV). Washington, DC: Congressional Research Service. SDWA. 1974. Safe Drinking Water Act, 42 U.S.C., s=s 300f et seq. USDHEW. 1970. Community Water Supply Study: Analysis of National Survey Findings. Washington, DC: U.S. Department of Health, Education, and Welfare. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas, TX: USEPA Region VI, Surveillance and Analysis Division. USEPA. 1973. EPA voices support for Safe Drinking Water Act. EPA press release, March 8. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1975. Train stresses long-range planning as the environmental movement comes of age. EPA press release, April 22. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1977a. EPA safe drinking water standards go into effect today. EPA press release, June 25. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1977b. History Office, oral history interview, Douglas M. Costle, USEPA Administrator, Carter Administration. USEPA. 1986. President signs Safe Drinking Water Act Amendments. EPA press release, June 20. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1990. Reducing Risk: Setting Priorities and Strategies for Environmental Protection. EPA Science Advisory Board. Washington, DC: U.S. Environmental Protection Agency.

REFERENCES

129

USEPA. 1992. EPA issues 23 final drinking water standards. EPA press release, May 19. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1999. 25 Years of the Safe Drinking Water Act: History and Trends. EPA 816-99-007. Washington, DC: U.S. Environmental Protection Agency. USEPA. 2001. Annual Report FY 2000. EPA-190-R-01-001. Washington, DC: U.S. Environmental Protection Agency. Whitaker, J. C. 1988. Earth Day recollections: What it was like when the movement took off. U.S. Environ. Protect. Agency J. (July=Aug.).

PART II REGULATION DEVELOPMENT

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

6 TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT JOYCE MORRISSEY DONOHUE, Ph.D. Office of Water, Office of Science and Technology, U.S. Environmental Protection Agency, Washington, DC

JENNIFER ORME-ZAVALETA Office of Research and Development, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oregon

6.1

INTRODUCTION

The U.S. Environmental Protection Agency (USEPA) is charged with protecting human health and the environment. Environmental protection decisions are often guided by risk assessments that are used to develop regulatory policy and other related guidance. Historically, in environmental protection, risk assessments were developed to protect humans from carcinogenic effects that could result from inhalation or ingestion exposures to specific chemicals. Risk assessments have since evolved to address endpoints other than cancer as well as stressors other than chemicals. Toxicological concepts used to develop risk assessments for drinking water contaminants are discussed in this chapter. Disclaimer : This chapter has been reviewed within the USEPA and represents the views of the authors; it does not necessarily reflect Agency policy. Any mention of trade names or products does not constitute endorsement. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

133

134

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Figure 6.1

NAS risk assessment paradigm.

In 1983, the National Academy of Sciences (NRC 1983) formalized the process of human health risk assessment into four organizing steps: hazard identification, dose–response assessment, exposure assessment, and risk characterization (Fig. 6.1). Hazard identification, later renamed hazard characterization (NRC 1994), is a qualitative evaluation of whether exposure to a substance such as a drinking water contaminant would produce an adverse or otherwise undesirable effect. The data used to make such a determination usually come from animal studies. In a few instances, human epidemiologic, occupational, clinical, or case studies may be available for the contaminant of interest. Having causally linked exposure with an effect, the next step, dose–response assessment, involves a more quantitative evaluation of the empirical evidence relating a specific exposure dose to the effect of interest. In particular, the available data are examined to determine the relationship between the magnitude of the exposure and the probability of the observed effect. The exposure assessment step involves an evaluation of human exposure information. This includes a comparison of exposure before and after regulatory controls, as well as a characterization of the environmental fate and transport of the contaminant from the source to the exposed population by different media (e.g., air, water, food) and different exposure routes (e.g., ingestion, inhalation, dermal contact). All of this information is captured in a summary statement, or risk characterization. The risk characterization contains a description of the overall nature and magnitude of risk posed to human populations exposed to a particular contaminant. Included in the description is a discussion of what is known and not known about the hazards posed by the substance, what models were used to quantify the risk and why they were selected, assumptions and uncertainties associated with the qualitative and quantitative aspects of the assessment, and general level of confidence in the assessment.

6.2 TOXICOLOGICAL EVALUATION OF DRINKING WATER CONTAMINANTS Risk assessments for drinking water contaminants involve a toxicological evaluation of the contaminant. Toxicology, by definition, is the study of poisons and their effect on living organisms. It encompasses many scientific disciplines, including chemistry, biochemistry, epidemiology, physiology, pathology, statistics, modeling, and

6.2 TOXICOLOGICAL EVALUATION OF DRINKING WATER CONTAMINANTS

135

ecology. Environmental toxicology is that branch that studies the biological effects of environmental chemicals. Environmental exposure to chemicals can occur through the air, food or water, or dermal contact. Upon entering the body through one of these routes, a chemical can interact with various biological systems to produce an effect. The type of interaction and resulting effect depend on the chemical itself and the dose. Basically, everything has the potential to be toxic once a certain dose has been reached. Some chemicals may produce an effect once a certain dose or concentration of the chemical has reached a particular site in the body. Doses below this level will not result in the effect. This type of effect is commonly referred to as a threshold effect. In some cases, an effect could theoretically result after exposure to just one molecule of a chemical. These effects are commonly referred to as nonthreshold effects, and are often assumed as the mechanism by which some chemicals cause cancer. Effects can also be categorized as reversible or irreversible. Reversible effects disappear when an organism is no longer exposed to the chemical. Effects that involve a permanent change in the structure or function of a biological system or persists after the exposure ends are considered irreversible. Chemical contaminants entering an organism may cause an effect either directly or indirectly, after the chemical has been modified by the organism. Organisms are able to biologically transform chemicals to different forms usually with enzymes. As a result, a particular chemical contaminant can undergo a number of chemical changes, forming metabolites that are chemically distinct from the original chemical contaminant (Sipes and Gandolfi 1986). In some cases, it is the modified chemical rather than the parent compound that produces a toxic effect within the organism. The following are some of the biological effects that can result from chemical exposure:
Lethality—the ability of a chemical to cause death
Liver, kidney, or other organ effects—effects that alter the structure or function of the organ
Biochemical effects—changes in the activity or concentration of biomolecules that are indicative of tissue damage and=or impaired cellular function
Carcinogenicity—the ability of the chemical to cause cancer
Mutagenicity—the ability of the chemical to cause changes in genetic material
Reproductive effects—effects on the ability of an organism to reproduce
Developmental effects—effects on the developing organism
Neurological effects—effects on the structure and function of the nervous system

These effects are often the endpoints of concern observed in epidemiologic or other human studies and evaluated in laboratory or other types of experimental research.

136

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

In addition to dose, toxicologic effects may be dependent on the duration of chemical exposure. Therefore, studies evaluate the potential occurrence of these effects following certain types of exposure, including single, intermittent, or continuous exposure over a certain proportion of an organism’s lifetime. The following lists common experimental study durations: Acute toxicity—effects observed after one or a few exposures Subchronic toxicity—effects observed after repeated exposure for a portion of the animal’s or human’s lifetime (for rodents, this is approximately one-tenth of the lifetime or 90 days) Chronic toxicity—effects observed after repeated exposures for most of an animal or human’s lifetime As noted above, toxicology data can come from human or animal studies and are discussed separately. 6.2.1

Human Studies

In risk assessments to protect human health, human data are preferred over animal data. Human studies include epidemiologic, clinical, occupational, or case studies. Epidemiology is simply defined as the study of epidemics, and in particular, the causes that explain patterns of disease frequency in humans (Rothman and Greenland 1998). There are two broad types of epidemiologic studies: descriptive and analytic. Descriptive studies include correlational or ecologic studies (those that compare geographic regions), cross-sectional studies, and case reports. The descriptive studies collect information on groups of people and help to generate hypotheses associating chemical exposure with an effect. (Epidemiology is discussed further in Chapter 7.) Analytic studies are designed to test specific hypotheses related to chemical exposure causing a particular disease or effect. One type of analytic study is called an intervention study. Like clinical studies, intervention studies are generally used to test therapeutic agents. Another type of analytic study more commonly used to study environmental toxicants is the observational study. Observational studies are categorized into two types: case–control or cohort. Case–control studies involve a comparison of persons with the condition of interest (e.g., high blood pressure) compared with a reference or control group who do not have the condition. Cohort studies follow groups of individuals who have been exposed or not exposed to the factor(s) of interest over time. Incidence rates for the condition of interest are compared between groups with different exposure levels. Although epidemiology studies provide information on the potential effects of chemical exposure in human populations, uncertainties can affect the interpretation of study results. In these studies, several factors that could affect the study findings, such as age, gender, or other types of exposure (e.g., occupational exposure, cigarette smoke), occur in addition to the exposure of interest. All of these potential

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

137

confounders could make it difficult to attribute an effect or disease to any one chemical exposure. 6.2.2

Animal Studies

Experimental animal studies are the primary means for evaluating the toxicity of a chemical. Toxicity studies using laboratory animals provide a direct method for testing a cause-and-effect relationship between exposure and effect. The animals used most often are mice and rats, because a number of genetically homogeneous strains are available and it is possible to test a sufficiently large number of animals to observe statistically as well as biologically significant effects. In some circumstances, other experimental animals such as monkeys, guinea pigs, hamsters, rabbits, and dogs are used. Selection of a particular animal species depends on the type of toxicity test to be conducted and how well the selected species would predict a human response. It is generally assumed that if an animal metabolizes a chemical in much the same way as do humans, it is a good surrogate for predicting human toxicity. In general, monkeys and dogs are more similar to humans than rodents, but are more difficult to study because, for example, there is potentially a greater expense for maintenance and resulting limitations on the numbers of animals that can be accommodated in each dose group. In controlled experiments, animals are exposed to several different doses of a chemical to see which dose produces an effect and the type of effect that results. Toxicity studies often utilize high-dose exposures to increase the likelihood of detecting any adverse effect that could result from exposure to the chemical. Short-term range-finding studies are generally conducted in small groups of animals in order to identify likely effects and select the doses for longer-term test protocols. Ideally, the lowest dose tested will not produce an effect, while the intermediate dose would result in mild observable effects, and the highest dose producing more pronounced toxic effects. Use of animal studies in estimating human risk has been criticized for many reasons, including the introduction of additional uncertainties. Examples include the extrapolation of effects in genetically homogeneous laboratory strains to the genetically heterogeneous human population and extrapolation of high experimental doses to predict human risk at lower environmental concentrations. These uncertainties are taken into account in quantifying risk (see text below).

6.3

USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

When a risk assessment is conducted, all available toxicity data are gathered from the published literature or other sources, such as data submitted to regulators as confidential business information. Data that describe cancer effects or those that are relevant to the induction of cancer are assessed separately from noncancer effects. For cancer effects, data are assessed qualitatively (with respect to hazard identification) and quantitatively (in terms of dose–response assessment). The qualitative

138

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

evaluation involves an assessment of the weight of evidence for the chemical’s potential to cause cancer in humans and accounts for the mode of action by which the chemical produces cancer in the body (USEPA 1996, 1999). The data considered in the risk assessment include both human and animal studies (if available). The method for quantifying the potential carcinogenic risk depends on the mode of action, that is, whether the chemical is thought to produce cancer through a mutagenic mechanism, or is secondary to cellular toxicity. In cases where the mode of action is unclear, both approaches may be used to quantify risk. Risk may be assessed for different routes of exposure such as inhalation, oral, or dermal.

6.3.1

Cancer Risk Guidelines

The 1986 USEPA Guidelines for Carcinogen Risk Assessment (USEPA 1986) established five alphanumeric cancer categories (Table 6.1). Although these classifications will be phased out in the future by the Agency’s 1996=1999 guidelines, they are now still widely used. USEPA proposed revisions to the 1986 guidelines in the proposed 1996=1999 Guidelines for Carcinogen Risk Assessment (USEPA 1996, 1999). These revisions include replacing the alphanumeric system with descriptors and a narrative describing a chemical’s potential to produce cancer in humans. Under the new proposed guidelines, there are five descriptors as follows:
Known human carcinogen
Likely to be carcinogenic to humans
Suggestive evidence of carcinogenicity, but not sufficient to assess human carcinogenic potential
Data inadequate for an assessment of human carcinogenic potential
Not likely to be carcinogenic to humans

TABLE 6.1 Group A B

C D E

USEPA Cancer Classification Categories Category Human carcinogen based on epidemiologic or other human data Probable human carcinogen: B1 indicates limited human evidence with sufficient evidence in animals B2 indicates sufficient evidence in animals and inadequate or no evidence in humans Possible human carcinogen based on limited or suggestive evidence in animals Not classifiable as to human carcinogenicity Evidence of noncarcinogenicity for humans

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

6.3.2

139

Effects Other than Cancer

For effects other than cancer, USEPA develops either an oral reference dose (RfD) or an inhalation reference concentration (RfC). For drinking water contaminants, the oral RfD is determined. The RfD is defined as an estimate of a daily exposure that is not expected to produce adverse effects over a person’s lifetime. To develop an RfD, one must evaluate the available data from human or animal studies. For drinking water contaminants, oral drinking water studies are preferred, but studies using other routes such as food or inhalation may be considered. For each of these studies, the highest dose that causes no adverse effect, the no-observed-adverse-effect level (NOAEL), and the lowest dose that produces an adverse effect, the lowestobserved-adverse-effect level (LOAEL) are identified for each appropriate (i.e., relevant to humans) test species. The RfD has generally been estimated by dividing a NOAEL or LOAEL by an uncertainty factor (Fig. 6.2). The NOAEL (or LOAEL in the absence of a NOAEL) is selected on the basis of the relevance of the test species to humans, the sensitivity of the response relative to those identified from other studies, and whether the NOAEL is supported by other data. The NOAEL is divided by an uncertainty factor that accounts for the differences in response to toxicity within the human population as well as differences between humans and animals, if animal data are used. If the study selected as the basis for the RfD involves an exposure that is less than the animals’ lifetime, another factor may be applied. Similarly, if a LOAEL is used in estimating the RfD, a factor may be applied to account for the absence of a NOAEL. Professional judgment may suggest the use of an extra uncertainty factor

Figure 6.2 Example of RfD determination for noncarcinogenic effects. The uncertainty factor (UF) will differ depending on whether the point of departure (PoD) for the RfD calculation in a NOAEL, LOAEL, or BMDL (x is the selected response level: 10%, 5%, etc.).

140

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

because of an insufficient database for that chemical, or there may be special properties of the chemical that would require an adjustment of the uncertainty factor. In selecting the uncertainty factor, one must evaluate each area of uncertainty and assign a value of 1, 3, or 10 depending on the strength of the data. A threefold factor is used when data reduce the need to apply a 10-fold unit of uncertainty. For example, a LOAEL that is an early biomarker of toxicity or a nearly complete toxicity data set may require a threefold uncertainty factor rather than a 10. An uncertainty factor of 1 is employed when the data are clearly from the most sensitive members of the population, eliminating the need for an intraspecies adjustment, or when there are data to demonstrate that the responses of the animals are the same as those of humans, eliminating the need for an interspecies adjustment. The net uncertainty factor is the product of the individual factors applied. Uncertainty factors tend to range from 1 to 3000-fold (Table 6.2). Uncertainty factors greater than 3000 may indicate too much uncertainty to have any confidence in the risk assessment. The RfD may also be determined using a benchmark method instead of the traditional NOAEL=uncertainty factor approach. The benchmark dose (BMD) is defined as the lower statistical limit for the dose corresponding to a specified increase in the level of the critical health effect over the background level (Crump 1984). In other words, the BMD approach does not utilize a single dose such as a NOAEL for estimating risk, but considers the available data from which a dose level corresponding to an increase in the incidence of an adverse health effect is identified. Statistical modeling of the dose–response curve is used to determine the dose that corresponds to a specific population response level. Frequently, a 10% response above the population background is the response level identified, although other response levels from the low end of the dose–response curve can also be selected when supported by the data. The lower confidence bound on this dose, referred to as the effective dose or LED10, is then divided by an uncertainty factor to estimate the RfD. The uncertainty associated with this method for estimating risk is generally less than the NOAEL=uncertainty factor approach. However, the data requirements for using this approach are more rigorous. The general risk assessment conducted for both cancer- and non-cancer-related endpoints provides the foundation for developing drinking water-specific risk assessments. These risk assessments are used to establish the health-based guidelines and standards discussed below.

TABLE 6.2 Uncertainty Factors Factor 1, 3, or 1, 3, or 1, 3, or 3 or 10 1, 3, or

Area of Uncertainty Addressed 10 10 10 10

Differences within the human populations Differences between humans and animals Use of less than lifetime data for estimating lifetime risk Use of a LOAEL in the absence of a NOAEL Data gaps or other chemical-specific uncertainty

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

6.3.3

141

Maximum Contaminant Level Goal (MCLG)

The RfD, cancer classification, and method for estimating carcinogenic risk (e.g., slope factor for nonthreshold chemicals or point of departure for threshold-like chemicals) are the critical elements in establishing the MCLG for a drinking water contaminant. The MCLG is defined as a concentration of a contaminant in drinking water that is anticipated to be without adverse effects over a lifetime. The methodology used in establishing an MCLG will differ depending on the nature of the critical adverse effect. For the purpose of establishing the MCLG, contaminants fall into one of five categories:
Linear (genotoxic) carcinogens
Carcinogens with an accepted nonlinear mode of action
Carcinogens for which a mode of action has not been established and that are thus treated using a linear approach
Contaminants with a threshold, non-cancer-critical effect
Chemicals with a threshold, non-cancer-critical effect that may have some tumorigenic activity This arbitrary view of the world of chemical contaminants is presently in a state of flux, a factor that somewhat complicates understanding the basis of an MCLG. The unsettled status of the procedures used in development of an MCLG is a product of the fact that the 1996=1999 drafts of the USEPA new cancer guidelines have not been finalized. In addition, there is activity under way to harmonize the approaches used in the risk assessment for carcinogens and noncarcinogens. Completion of these two activities is likely to reduce the number of categories above from five to three by removing the distinction between carcinogens and noncarcinogens and focusing the risk assessment on whether a chemical acts through a linear, nonlinear, or unidentified mode of action. Under present USEPA policies, the MCLG for a linear carcinogen or a carcinogen with an unidentified mode of action is zero. For example, the MCLG for bromoform in the 1998 disinfectant and disinfectant byproduct rule was established at zero because all evidence pointed to the fact that it caused the development of tumors through mutations to DNA, a linear mode of action. The MCLG for arsenic is zero because, although there is some evidence to suggest that it may be tumorigenic through a nonlinear mode of action, there are not enough data to identify either the mode of action or the shape of the dose–response curve at doses below those associated with effects that can be detected with statistical confidence. Cancer risks associated with exposure to such chemicals are estimated by extrapolation using a multistage dose–response model (Fig. 6.3). Chloroform, on the other hand, has been a different story. In 1998, the Agency established an MCLG of zero for chloroform in order to be protective, although there was strong evidence to demonstrate that it was tumorigenic through a nonlinear mode of action. In other words, chroroform does not appear to cause cancer through a mutagenic mode of action (i.e., because it causes change in the structure of DNA). Instead, data indicate that the tumorigenic properties of chloroform are a conse-

142

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Figure 6.3

Example of cancer risk extrapolation using the linear dose–response model.

quence of its ability to cause cell death, thereby stimulating rapid cell division and tissue repair. It is theorized that rapid cell division leads to a series of DNA replication errors that eventually culminate in tumor development. In 2000, the Court of the District of Columbia decreed that USEPA remove the zero MCLG for chloroform and establish a value based on its nonlinear mode of action. A nonzero MCLG will be proposed by the Agency as part of the Stage II rule for disinfection byproducts. The risk assessment supporting the nonzero MCLG was accepted by the Agency and posted on the Agency Integrated Risk Information System (IRIS) in October 2001 (USEPA 2001). For noncarcinogens, the MCLG has traditionally been derived from the RfD under the assumption that there is a threshold below which there are no effects of exposure. As described above, the RfD is developed from the dose–response data for the critical effect(s) observed in a well-designed toxicologic or epidemiologic study, or from a collection of human data that clearly demonstrate a NOAEL following exposure. The following equation is applied in calculating the MCLG from the RfD:  MCLG ¼ where

 RfD  (body weight)  relative source contribution (RSC) drinking water intake RfD body weight drinking water intake RSC

¼ reference dose ¼ 70 kgðadultsÞ ¼ 2 L=day ¼ the portion of the total exposure contributed by water. The default value is 20%:

6.4 HEALTH ADVISORIES

143

For drinking water regulations, chemicals have been grouped in three categories (Category I, II, or III). Category I chemicals are those that were categorized as Groups A or B under the 1986 cancer guidelines, and thus, have a zero MCLG. Chemicals characterized as Group C, possible carcinogens under the 1986 cancer guidelines, were placed in Category II. Rather than treat these compounds as carcinogens in determining the MCLG, USEPA traditionally added a risk management factor of 10 to the denominator of the MCLG equation above, thereby lowering the projected no-effect concentration in drinking water to a tenth of that suggested by the RfD. Category III chemicals are those with no evidence of carcinogenicity via the oral route (Groups D and E) and have an MCLG based on the RfD following the equation above. In the future, MCLG values for chemicals found to have tumorigenic effects as a result of a nonlinear mode of action will be determined using the threshold approach outlined above. However, in place of the RfD, the point of departure (PoD) from the dose–response curve for either tumors or a precursor preneoplastic event will provide the basis for the MCLG. The PoD will be divided by a margin of exposure (MoE), which is similar to the uncertainty factor in the RfD equation, and will replace the RfD term in the equation above. At present, it is difficult to predict whether the risk management factor of 10 will continue to be applied for chemicals lacking clear-cut evidence on carcinogenicity. Exposures to most chemicals occur from other media in addition to water. Many of the contaminants in water are also found in foods. Others, particularly volatile compounds, are present in ambient air. The relative source contribution (RSC) adjusts the MCLG so that only a portion of the total allowable exposure is allocated to drinking water. The RSC can be applied using either a percentage approach or a subtraction approach. The percentage approach was used for all current MCLG values. The RSC is preferably based on data regarding the exposures that occur from food, air, and other important media such as personal care products or pharmaceutical agents. However, the required data are often limited leading to the use of default RSC values. In the Ambient Water Quality Criteria Methodology Document for Human Health, a decision-tree approach for evaluating data adequacy and determining the appropriate default when key data elements are lacking is described. The approach described in the Ambient Water Quality Criteria Methodology is being applied in the derivation of MCLG values for upcoming regulations. When the lack of key data elements prevents using a data-derived percentage or subtraction allocation, default adjustments of 20%, 50%, or 80% are possible when supported by the appropriate data. In the absence of appropriate data, the 20% default is applied.

6.4

HEALTH ADVISORIES

Many chemicals that can contaminate drinking water have not been regulated by USEPA. In some cases, contamination is the result of an accidental spill or a

144

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

temporary problem. The USEPA Health Advisory (HA) program was developed to assist local officials and utilities when dealing with episodic drinking water contamination problems and=or contamination with unregulated contaminants. HA values describe nonregulatory concentrations of drinking water contaminants at which adverse health effects would not be anticipated to occur over specific exposure durations. They serve as informal technical guidance to federal, state, and local officials responsible for protecting public health when emergency spills or contamination situations occur. They are not legally enforceable federal standards and are subject to change as new information becomes available (USEPA 1989). Currently available (2002) HA values are provided in Appendix A. HA values are developed for 1-day, 10-day, longer-term (approximately 7 years, or 10% of an individual’s lifetime), and lifetime exposures on the basis of data describing noncarcinogenic endpoints of toxicity. For those substances that are known or probable human carcinogens, according to the Agency’s 1986 classification scheme for carcinogens (Group A or B), lifetime HAs are not recommended (USEPA 1989). For Group A or B carcinogens, the carcinogenic risk estimates for drinking water are presented in the HA documents. The 1-day and 10-day values are established for a 10-kg (22-lb) child based on the premise that this group is the most sensitive to acute toxicants. Longer-term exposures, estimated to be 7 years or one-tenth of an average lifetime, are calculated for both the 10-kg child and adults. Each of these HA calculations assumes that drinking water is the only source of exposure to the chemical. The lifetime HA is established only for the adult and, as indicated by its name, assumes that exposure occurs over the entire lifetime. The lifetime HA is calculated using the same approaches applied in deriving the MCLG for chemicals with a threshold toxic effect. As is the case with the MCLG, the lifetime HA is adjusted for other sources of exposure to the contaminant by applying an RSC factor. The lifetime HA is the most conservative of the suite of HA values and is the equivalent of the MCLG for unregulated contaminants. A lifetime HA is generally not established for a contaminant that is a known or probable carcinogen (Categories A and B of the 1986 USEPA cancer guidelines). Lifetime HA values for a Group C carcinogen include a 10% reduction in the calculated lifetime value as a risk management adjustment to protect against possible human carcinogenicity. Treatment of acute exposures is one unique feature of the HA program. The acute HAs were developed specifically for dealing with episodic drinking water contamination incidents resulting from spills, accidental releases, or equipment malfunction that lead to contamination of drinking water. Because many of these contamination incidents persist for only a short period of time, it is important to provide guidelines that apply to acute exposures. A 1-day HA is an estimate of the concentration of a chemical in drinking water that is not expected to cause any adverse noncarcinogenic effects for up to 1 day of exposure. The 10-day HA value is the concentration of a chemical in drinking water that is not expected to cause any adverse noncarcinogenic effects for up to 10 days of exposure. Both are established for a 10-kg child consuming 1 L of water per day,

6.5 FUTURE OUTLOOK

145

because the child is expected to be the most sensitive to acute exposures (USEPA 1989). The approximate body weight of a 1-year-old child is 10 kg. Ideally, the less-than-lifetime HAs are developed from a study in humans or animals that provides dose–response data for the critical effect using the desired duration (i.e., 1 day, 10 days, 90 days) (USEPA 1989). Examination of all the data from the appropriate short-term studies provides information on the critical effect. It is important that a complete data set be available for identifying the spectrum of health effects resulting from short-term exposures and their dose–response characteristics. The dose–response data are used to identify a NOAEL, LOAEL, or benchmark dose (LED10 or lower as justified by the data) for the critical effect associated with the duration of interest. The LOAEL is used for calculation only if a NOAEL has not been identified and if the observed effect is an early marker of toxicity rather than a frank (severe) effect. The HA value is derived using the following equation: 

ðNOAEL or LOAELÞ  10 kg Less-than-lifetime HA ¼ UF  ð1 L=dayÞ



where UF is the uncertainty factor. In the derivation of the less-than-lifetime HAs, uncertainty factors are most often employed for the intraspecies adjustment, interspecies adjustment, and use of a LOAEL in place of a NOAEL. The uncertainty factors are usually in multiples of 10. An uncertainty factor to adjust for study duration deficiencies is not applied. Instead the HA value for the next-higher duration is used in lieu of the shorter duration (i.e., a 10-day value in place of the 1-day value). The longer-term HA is calculated for both a child and an adult. The less than lifetime HA values do not include an RSC adjustment.

6.5

FUTURE OUTLOOK

Risk assessment approaches and their application to regulatory toxicology are consistently changing as advances in science reduce the uncertainties in extrapolating data from studies in animals or limited observations in humans to the entire regulated population. Improvements in the design of both animal and epidemiologic studies have reduced the impact of confounding variables on results and improved the reliability of the data. Increased understanding of the biological changes responsible for cancer and noncancer adverse effects improves the ability of a risk assessor to postulate a mode of action for a toxic event and model dose– response curves below the ability of a study to measure change. As research allows hypothesis to become theory and improves the precision of NOAEL, LOAEL, and benchmark dose estimates, the risk assessment methodologies applied in establishing MGLG and Health Advisory values will change as well. The major impact of improvement is likely to be in the application of data-derived uncertainty factors and the narrowing of the uncertainty component reflected in the MCLG or HA value.

146

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Regulations and guidelines that apply to drinking water and air are important. Humans and animals require a daily intake of water to live. Unlike contaminants that affect specific foods or commercial products where avoidance is one measure that can be used to modulate the risk to sensitive populations, this option is limited when it comes to drinking water. The drinking water MCLG and HA values must err on the side of protection. Accordingly, changes in risk assessment approaches will be introduced only when supported by a strong body of scientific data.

ACKNOWLEDGMENTS The authors would like to thank Julie Du, Hend Galal-Gorchev, and Edward Ohanian for their thoughtful comments and insights in reviewing this document.

REFERENCES Crump, K. S. 1984. A new method for determining allowable daily intakes. Fund. Appl. Toxicol. 4:854–871. NRC. 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. NRC. 1994. Science and Judgment in Risk Assessment. Washington, DC: National Academy Press. Rothman, K. J. and S. Greenland. 1998. Modern Epidemiology, 2nd ed. Baltimore: Lippincott Williams & Wilkins. Sipes, I. G. and A. J. Gandolfi. 1986. Biotransformation of toxicants. In Casarett and Doull’s Toxicology. The Basic Science of Poisons, 3rd ed. C. D. Klaassen, M. O. Amdur, and J. Doull, eds. New York: Macmillan. USEPA. 1986. Guidelines for Carcinogenic Risk Assessment. Fed. Reg. 51:33992–34003. USEPA. 1989. Guidelines for Authors of EPA Office of Water Health Advisories for Drinking Water Contaminants. Washington, DC: Office of Drinking Water, Office of Water. USEPA. 1996. Proposed Guidelines for Carcinogen Risk Assessment. EPA=600=P-92=003C. Washington, DC: Office of Research and Development. USEPA. 1999. Guidelines for Carcinogen Risk Assessment. Review Draft. Washington, DC: USEPA, Risk Assessment Forum. USEPA. 2000. Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health. Technical Support Document. Vol. 1, Risk Assessment. EPA-822-B-00-005. Washington, DC: Office of Science and Technology, Office of Water. USEPA. 2001. Toxicological Review of Chloroform. Washington, DC: Office of Research and Development.

7 EPIDEMIOLOGIC CONCEPTS FOR INTERPRETING FINDINGS IN STUDIES OF DRINKING WATER EXPOSURES GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun and Associates Staunton, Virginia

REBECCA L. CALDERON, Ph.D. National Health and Environmental Effects Laboratory, U.S. Environmental Protection Agency Research Triangle Park, North Carolina

FLOYD J. FROST, Ph.D. The Lovelace Institutes, Albuquerque, New Mexico

7.1

INTRODUCTION

To the inexperienced, environmental epidemiology may appear to be an uncomplicated, straightforward approach to studying exposure–disease associations in human populations. Epidemiologic studies can provide useful information about the risks of environmental exposures that human populations may actually experience, but the study designs and their conduct are not as simple as supposed. Many of the issues are complex and subtle, and this needs to be realized so that the studies can be Disclaimer : The views expressed in this chapter are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The chapter has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

147

148

EPIDEMIOLOGIC CONCEPTS

properly designed and findings will be meaningful. Because the studies appear to be so straightforward, they are sometimes conducted by investigators with little training and experience, often leading to results that are difficult to interpret. Also, scientists with little knowledge of epidemiology feel comfortable explaining the importance of findings to the public, and this can lead to conflicting interpretations of the findings. A hypothetical example can help explain. A government agency releases statistics that show high cancer mortality in certain counties in the United States. A chemist wonders whether increased cancer mortality is related to environmental exposures. The chemist has compiled a computer file of chemical analyses reported by public water utilities. She decides to conduct a study because in some of the counties with high mortality rates water utilities have occasionally recorded high levels of some chemical constituents. The chemist uses a widely available software program for a statistical analysis of possible associations between reported levels of chemicals in water and cancer rates. She finds no correlation between any of the specific water quality parameters and cancer mortality; however, a statistically significant correlation is observed between cancer mortality and chlorinated surface water. The chemist concludes that chlorination byproducts, which were not included in the database, are responsible for increased cancer mortality and prepares a paper for publication. The article is peer-reviewed by an epidemiologist, who points out that a large elderly population has migrated to many of the counties in the past 10 years and in some counties the workforce is employed primarily in the chemical industry. Both of these factors could be responsible for much, if not all, of the increased cancer mortality in these counties. In addition, the reviewer notes that site-specific cancer rates should be evaluated rather than all cancers and that, among other factors, cancer incidence and survival should be considered in the analysis. Being an experienced environmental epidemiologist, the reviewer also notices that most of the population in several counties with high mortality rates uses individual wells, not public water systems. The chemical analyses in the database applied only to public water systems. The article is rejected for publication. The author is informed of these potential problems, and recommendations are made to help the author improve the analysis. Before reaching any conclusions about the observed association, additional efforts must be undertaken to evaluate potential sources of bias and confounding and improve the assessment of exposure. The chemist, on reading these comments, becomes confused and angry that her finding is not given a high priority for publication and decides the finding is so important that the public must be made aware of her research. The study is reported on page 1 of the local newspaper, and the chemist explains the findings in a 30-second news bite on television. Also interviewed is another scientist who describes the study’s flaws and says that the findings are uninformative. Now, the public is confused, and a local politician wants to know who hired the scientist to dispute such an important finding. A study that seemed so straightforward has now become the center of controversy. Whom is the public to believe? What risks are posed by chlorinating drinking water? There are many questions but few answers. Recently, results of environmental studies have increasingly provoked controversy about the need for regulatory actions. The public is often confronted by conflicting results and conflicting interpretations of these results, as water system

7.2 WHAT IS EPIDEMIOLOGY?

149

managers, engineers, and scientists debate the contradictory results. Unlike toxicologists, who can conduct replicate experiments with genetically identical strains of animals randomly assigned to various exposure groups, epidemiologists must rely on observations in human populations. Except in clinical trials, the perfect experimental situation is seldom found in human populations. People are never randomly distributed in a particular geographic location or neighborhood, nor is each person in a study likely to have similar lifestyles, diets, or other behaviors. Neither are exposures likely to be randomly distributed among persons who have similar behavioral or demographic characteristics. Since epidemiologists deal with the complexities of real-life, human experiences, they are usually conservative in their interpretation and cautious about their study conclusions. Comments made by Wade Hampton Frost in 1936 (Snow 1965) continue to apply: Epidemiology at any given time is something more than the total of its established facts . . . it is not easy, when divergent theories are presented, to distinguish immediately between those which are sound and those which are merely plausible.

Since epidemiologists are increasingly studying water contaminants, it is important for drinking water professionals to become more familiar with methods. The information in this chapter should help readers understand frequently used terminology, why certain types of studies are conducted, sources of possible bias, some of the reasons why results can be controversial, and the complexity of interpreting exposure–disease associations. Readers should also gain an appreciation of the complexity and importance of epidemiology in assessing health risks.

7.2

WHAT IS EPIDEMIOLOGY?

The term epidemiology, derived from the Greek roots epi (on), demos (people), and logos (study), is the study of the distribution and determinants of disease and injuries in human populations and the application of this knowledge to the prevention and control of health problems (Last 1995; Rockett 1994). Whereas clinicians consider the unique problems of diagnosing, treating, and preventing disease in individual patients, epidemiologists view disease primarily at the population level, describing its occurrence and statistically associating exposures, demographic characteristics, or behaviors with the disease in order to develop public health control measures. Besides describing temporal trends, geographic clustering, and other patterns of disease occurrence, epidemiologists seek explanations for disease etiologies, obtain information about characteristics and behaviors that may increase, or reduce, the risk of disease, and evaluate public health conditions or therapeutic interventions.

7.3

HISTORICAL ORIGINS

The origins of epidemiology date back over 2000 years to a manuscript (On Airs, Waters, and Places attributed to Hippocrates) describing the influence of environ-

150

EPIDEMIOLOGIC CONCEPTS

mental factors on the occurrence of disease (Rockett 1994). However, it was not until the seventeenth century with John Graunt’s observations about births and deaths in England that a firm foundation was laid for epidemiology. In the mid nineteenth century John Snow, William Farr, and Ignaz Semmelweis moved beyond describing disease trends and patterns and began to offer possible explanations for their observations (Snow 1965, Rockett 1994). Snow’s investigation of the occurrence and location of cholera deaths in the Golden Square district convinced the Board of Guardians to order the pump handle removed from the Broad Street well (Snow 1965). His hypothesis that sewage contaminated drinking water was responsible for cholera was strengthened by evidence that cholera mortality was higher in London households served by the Southwark and Vauxhall Water Company than in households served by the Lambeth Company. The Lambeth Company obtained water from the River Thames at a location free of London’s sewage; the Southwark and Vauxhall Company obtained water contaminated by sewage. Snow was able to compute mortality rates for his comparison because he could identify both the fatalities and their household water supplier. Information about mortality was available largely because of Farr’s efforts to systematically collect information about deaths. In Hungary, Semmelweis tested his hypothesis that medical students transmitted childbirth fever, by requiring medical personnel in his wards to wash their hands and soak them in chlorinated lime before conducting pelvic exams. Within 7 months, lower mortality rates were found in these birthing wards than in wards where the intervention was not implemented, thus offering a control measure to prevent disease long before its etiologic agent was identified. Although these two examples are almost 150 years old, they serve as dramatic reminders of the powerful and unique role epidemiology can play in the identification of health risks and prevention of disease, even though the specific etiologic agent may not be known. More recently, epidemiologists warned of lung cancer and other health risks in cigarette smokers many years before analytical chemists and toxicologists were able to identify the specific chemicals in tobacco smoke that may be carcinogens (Doll and Hill 1964).

7.4

DISEASE MODELS

Epidemiologists use disease models to help evaluate and explain disease etiologies. The simplest of these models is the triad (Fig. 7.1). The host, agent, and environment coexist independently, and disease occurs only when there is interaction between them (Rockett 1994). Many diseases have multiple agents, exposures, or risk factors that cause the disease or influence the course of the disease, and these must all be considered. Thus, a slightly more complex disease model must now be presented for cholera, which although it is often caused by contaminated water, can also be transmitted in other ways (Fig. 7.2). The etiologic agent and all relevant social, physical, or biological environments (e.g., personal behaviors, cultural practices, hygiene and sanitation practices, climate, and reservoirs of infection) combine to ‘‘cause’’ disease if the host is susceptible. Host susceptibility can be affected by

7.4 DISEASE MODELS

Figure 7.1

151

Host–agent–environment relationship.

personal characteristics such as occupation, income, education, immune status, behavior, and genetic traits. The presence (or absence) of the agent is necessary for disease to occur (or be prevented). The environment must support the agent, and the agent must be transmitted to a susceptible host in an appropriate time, manner, and dose sufficient to cause infection and disease. For example, infection and serious illness may occur among AIDS patients when a water source is contaminated with

Figure 7.2 Causes of cholera [adapted from: Beaglehole et al. (1993)].

152

EPIDEMIOLOGIC CONCEPTS

Figure 7.3 Web of causation applied to cardiovascular disease [adapted from: Rockett (1994)].

Cryptosporidium oocysts from human or animal feces, sufficient numbers of oocysts survive in the source water, the oocysts are inadequately removed or inactivated by the water system, and an infective dose of oocysts is ingested while drinking tapwater. The disease process is often complex, and this complexity can be illustrated in a more detailed model, sometimes referred to as the ‘‘web of causation.’’ The relationship between water exposures and other risk factors for cardiovascular disease is shown in Fig. 7.3. This model places less emphasis on the role of the agent or water contaminant in favor of other factors that may be important in the onset of disease. Epidemiologists have found lower cardiovascular disease mortality in areas where water hardness (e.g., levels of calcium and magnesium) is high, and some studies have associated water constituents with decreased blood pressure. However, evidence is not yet available to fully understand the role of water constituents or where to place this exposure within the web of causation. Thus, the use of a dotted line is shown for water exposures in Figure 7.3. Additional research is required to better understand how water constituents may affect cardiovascular disease or blood pressure.

7.5

BASIC MEASURES OF DISEASE FREQUENCY

To evaluate and compare disease and other health conditions in populations, epidemiologists use several measures of disease frequency. The most important are disease prevalence and incidence. Incidence measures the rate at which new cases of disease occur in a group of people who do not have the disease during a defined period of time; prevalence measures both new and existing cases in a population with and without disease (Last 1995, Beaglehole et al. 1993). A meaningful measure

7.5 BASIC MEASURES OF DISEASE FREQUENCY

153

of disease frequency requires the accurate compilation of cases of disease (the numerator) and an estimate of the susceptible population or population at risk (the denominator). The cases of disease must all arise from the population at risk. Although often referred to as a ‘‘rate,’’ prevalence is the proportion of people in the population who have a specific disease, condition, or infection at any specified time (e.g., on Jan. 1, 1996, or during Jan. 1–May 31, 1995). For example, a stool survey in Washington State estimated a 7% prevalence of giardiasis during 1980 in young children in diapers (Harter et al. 1982). The incidence rate requires an estimate of the amount of time people are at risk of contracting the disease, and person– time is specified in the denominator. For example, in a cohort of 118,539 women, 30–55 years of age, free from coronary heart disease and stroke in 1976, and followed for almost 8 years, 274 stroke cases were identified during the 908,447 person-years of follow-up (Beaglehole et al. 1993). In this study, the overall incidence rate of stroke in females was 30.2 per 100,000 person-years [computed as (274=908,447)
(100,000)]. Incidence can be compared among populations, and in this cohort the incidence of stroke was studied among smokers, ex-smokers, and nonsmokers. The incidence rate of stroke among current women smokers (49.6=100,000 person-years) was almost 3 times the nonsmokers’ rate (17.7= 100,000 person-years) and almost twice the ex-smokers’ rate (27.9=100,000 person-years). Incidence can be estimated from prevalence data when the average duration of disease is known, as prevalence is approximately equal to the incidence rate times the average duration of illness. The attack rate measures the cumulative incidence of disease in a particular group observed for a limited time and under special circumstances (i.e., during an epidemic or outbreak). The period of time for observation of cases can vary but should begin at the presumed time of exposure and continues over a time interval that allows for the occurrence of all possible cases that may be attributable to the exposure. In communicable disease outbreaks, secondary transmission can occur. The secondary attack rate refers to cases among familial, institutional, or other contacts following exposure to a primary case; the denominator includes only susceptible contacts. An example of the potential importance of secondary transmission is provided by an outbreak of E. coli 0157:H7 gastroenteritis caused by consumption of contaminated, undercooked hamburger (Grimm et al. 1995). In waterborne outbreaks caused by E. coli 0157:H7, Norwalk-like viruses, and Cryptosporidium, secondary transmission of infection and illness have been documented among personal contacts with primary cases. Waterborne etiologic agents such as these, which have a low infectious dose, are often transmitted by person-to-person contact, and a search should be conducted for secondary cases even though the primary cases may have been transmitted by contaminated water. The number of secondary cases may exceed the number of primary cases. Geography-specific (Kent et al. 1988) and food-specific attack rates are frequently used to help identify the vehicle or mode of transmission of disease during an outbreak investigation (Tables 7.1–7.3). For example, area-specific attack rates for populations were used to help identify the source of contamination in a waterborne outbreak of giardiasis in Pittsfield, MA (Kent et al. 1988). Attack rates were 4–5 times higher in populations served exclusively by reservoir C than in those served by

154

EPIDEMIOLOGIC CONCEPTS

TABLE 7.1 Geography-Specific Attack Rates from a Waterborne Outbreak of Giardiasis in Massachusetts Water Source

Cases

Population

Attack Rate per 1000

Reservoir A Reservoir B Reservoir C Mixed

68 14 126 427

9405 2309 4200 34,351

7.2 6.1 30.0 12.4

Total

635

50,265

12.6

Source: Kent et al. (1988).

reservoirs A and B (Table 7.1). In an outbreak in Milwaukee, MacKenzie et al. (1994) found that the attack rate for watery diarrhea was higher among persons living in the area served by the Milwaukee Water Works (39%) than in those living outside this area (15%). The Water Works service area receives water from two different treatment plants, and the attack rate was also found to be higher in the

TABLE 7.2 Dose–Response Attack Rates from a Waterborne Outbreak of Chronic Diarrhea Water Consumption during Outbreak (Glasses)

Attack Rate per 100

1–10 11–30 >30

12.5 37.5 57.1

Source: Parsonnet et al. (1989).

TABLE 7.3 Outbreak

Hypothetical Vehicle-Specific Attack Rates during a Waterborne Ate or Drank

Did Not Eat or Drink

Item

Ill

Not Ill

Attack Rate per 100 (a)

Water Coffee Fruit cup Beef

60 40 25 30

20 30 40 40

75 57 38 43

a

For instance, risk ¼ 75=13 ¼ 5.8.

Ill 4 19 17 20

Not Ill

Attack Rate per 100 (b)

Risk (a=b)

26 21 28 20

13 48 38 50

5.8a 1.2 1.0 0.9

7.5 BASIC MEASURES OF DISEASE FREQUENCY

155

southern area (52%) than in the northern area (26%). Contamination was suspected to have entered the water system at the southern treatment plant. Attack rates can also be computed based on estimated water consumption (Parsonnet et al. 1989). In an outbreak of chronic diarrhea, the attack rate was higher among ill person who consumed more water (Table 7.2). Hypothetical food-specific attack rates presented in Table 7.3 illustrate the typical methodology used to identify the vehicle responsible for an outbreak. In this example, water is implicated because the rate of illness among those who drank water was almost 6 times greater than the rate of illness among those who did not drink water, while attack rates were similar among those who did or did not consume other foods or beverages that may have been possible vehicles of infection. Mortality and case–fatality rates are also important measures of disease frequency. The mortality rate is a measure of deaths from a select disease or from all causes in a given period, usually a calendar year. The denominator is the average total population in which the deaths occurred, and the number of deaths is usually multiplied by 100,000 (or another multiple of 10) to produce a rate per 100,000 people. The case–fatality rate is the percent of individuals diagnosed with a specific disease who die as the result of that disease. For example, Bennett et al. (1987) report a case–fatality rate of 0.2% for shigellosis, a bacterial disease that may be waterborne. Although a lower case–fatality rate is reported (Bennett et al. 1987) for campylobacteriosis, another bacterial disease that may be waterborne, the mortality rate for campylobacteriosis is higher than that for shigellos because its incidence is higher (Table 7.4). Mortality and morbidity frequency may be determined for the total population (usually called an overall or crude rate) or for specific groups in the population. The crude rate is not used to evaluate long-term health trends or compare the health of different population groups because it does not take into account demographic characteristics, such as age, gender, and race, which may differ among the groups. An age-, gender-, or race-specific rate can be computed for comparison purposes or the crude rate can be standardized using a weighted averaging of specific rates. For example, an age-adjusted rate is a summary measure of the disease rate a population would have if it had a standardized age structure. Table 7.5 illustrates how use of crude rates can be misleading. In Finland, the crude mortality rate for diseases of the

TABLE 7.4 Estimated Incidence and Mortality for Campylobacteriosis and Shigellosis, 1985, USA Disease

Incidence (person-years)

Case–Fatality Rate (%)

Mortality

Mortality Rate

Campylobacteriosis

883.8=100,000

0.1

2100

Shigellosis

126.3=100,000

0.2

600

8.8 per 100,000 people 2.5 per 100,000 people

Source: Bennett et al. (1987).

156

EPIDEMIOLOGIC CONCEPTS

TABLE 7.5 Mortality for Diseases of the Circulatory System, 1980, Finland and Egypt Country

Crude Rate

Age-Adjusted Rate

Age-Specific Ratea

Finland Egypt

491=100,000 192=100,000

277=100,000 299=100,000

204=100,000 301=100,000

a

For subjects aged 45–54 years. Source: Beaglehole et al. (1993).

circulatory system is higher than in Egypt (Beaglehole et al. 1993), but the ageadjusted or age-standardized rate is higher in Egypt. This is because Finland has a larger proportion of older people. Age-specific rates show that mortality is higher among 45–54-year-old persons in Egypt than in Finland. Knowing the frequency or magnitude of disease in the population is just the beginning of the epidemiologist’s search. Time, place, and person must be considered to identify possible associations, risk groups, and risk factors. The collection of this information provides the general framework for an effective disease surveillance system. Knowing the time of disease occurrence assists epidemiologists in the detection of outbreaks, assessment of possible exposures that may have occurred, and evaluation of seasonal and long-term disease trends. For example, laboratory surveillance of stool specimens for Cryptosporidium parvum alerted health officials in Jackson County, Oregon, to an unusual number of cases; in the first 4 months of 1992, 46 cases of cryptosporidiosis were reported compared to 27 cases for all of 1991. The subsequent investigation identified waterborne transmission (Oregon Health Division 1992). Knowledge of the place of disease occurrence can help detect clusters of disease and allows the epidemiologist to compare disease rates among various geographic areas, such as countries, states, counties, census tracts, institutions, or water service districts. Information about people who become ill (e.g., through exposure) and their demographic characteristics is necessary to develop and test hypotheses about exposure–disease associations (Craun 1990, IAFMES 1996).

7.6

TYPES OF EPIDEMIOLOGIC STUDIES

Both observational and experimental studies have been conducted for drinking water exposures (Table 7.6). Experimental studies include population intervention studies and clinical trials of the efficacy of medications, medical therapies, and public health controls. Both studies consider the effect of varying some characteristic or exposure that is under the investigator’s control. Comparable individuals are assembled, randomly assigned to a treatment or intervention group, and observed for disease outcome, much as in a toxicologic study. The major difference is that experiments in human populations seek only cures or ways to help prevent disease. Ethical concerns must be fully addressed, and risks must be carefully weighed

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

TABLE 7.6

157

Types of Epidemiologic Studies

Experimental Clinical Population Observational Descriptive Disease surveillance and surveys Correlational or ecological Analytical Longitudinal Cohort or follow-up Case–control Cross-sectional Source: Adapted from Monson (1980).

against potential benefits. Examples of clinical and population experimental studies with results of interest to water officials are available (DuPont et al. 1995; Chappell et al. 1996; Dann et al. 2000; Okhuysen et al. 1998; Frost and Craun 1998; Muller et al. 2001; Ward et al. 1986; Moe et al. 1999; Payment et al. 1991, 1997). To determine the median infective dose of C. parvum, healthy volunteers without evidence of previous infection were randomly assigned to receive a specified dose of oocysts and then monitored for oocyst excretion and clinical illness for 8 weeks (DuPont et al. 1995). Additional clinical studies have provided more information about infection and the immune response for Cryptosporidium (DuPont et al. 1995, Chappell et al. 1996, Dann et al. 2000, Okhuysen et al. 1998, Frost and Craun 1998, Muller et al. 2001) and other waterborne pathogens such as rotavirus and Norwalklike viruses (Ward et al. 1986, Moe et al. 1999). In a population experimental study conducted in the Montreal area, endemic waterborne disease risk was evaluated by monitoring enteric disease in households that had been randomly assigned to one of two groups—one consuming municipal tapwater and the other receiving tapwater further treated by reverse osmosis to remove microbial contaminants (Payment et al. 1991). In a second study, Payment et al. (1997) included for comparison purposes a bottled water group and a group where water was flushed through the household plumbing before use. Observational studies are either descriptive or analytical. In descriptive studies information is available only about the occurrence of disease (from surveillance systems or special surveys) or associations among exposures, demographic characteristics, and disease rates in population groups (ecological studies). Descriptive epidemiology is important for summarizing disease information (e.g., cancer incidence rates) to reveal temporal, demographic, and geographic patterns of occurrence and develop hypotheses about disease etiologies and risk factors, whereas analytical epidemiology is used to test specific hypotheses. Descriptive techniques are also used to detect outbreaks (e.g., an active surveillance system for cryptosporidiosis), but analytical studies are required to evaluate exposure–disease associations and

158

EPIDEMIOLOGIC CONCEPTS

confirm the mode of transmission during the outbreak investigation (Craun 1990, IAMFES 1996). Analytical studies (or population experimental) studies are also required to determine endemic waterborne disease risks. Disease surveillance, no matter how complete the reporting, is inadequate by itself to assess waterborne risks (Craun 1994). 7.6.1

Ecological Studies

Ecological (also called correlational or aggregate) studies are used by epidemiologists to explore associations between health statistics, demographic measures, and other information (e.g., environmental monitoring results) readily available from vital statistics and public records. The ecological study is inexpensive and, as noted in our earlier example, straightforward; however, the interpretation of associations from this analysis is fraught with problems. In ecological studies, health and demographic statistics characterize population groups, rather than the individuals within the groups, and serious errors can result when it is assumed that inferences from a descriptive analysis pertain to the individuals within the group. Often group inferences do not pertain to individual behaviors and exposures, especially waterborne. Neither theoretical nor empirical analyses have offered consistent guidelines for the interpretation of results from ecological studies, and associations found in these studies must be viewed with appropriate caution (Greenland and Robins 1994a, 1994b; Piantadosi 1994). Ecological studies usually provide information to help develop hypotheses for additional study. In some instances, however, the group may be the appropriate unit of study, especially when the disease has a relatively short incubation or latent period and group exposures are shown to be relevant for individuals in the group (Poole 1994; Susser 1994a, 1994b; S. Schwartz 1994, J. Schwartz 1995). Epidemiologists should describe the limitations of their ecological study or why the results are appropriate for assessing risks. Between 1974 and 1982, more than 15 ecological studies associated cancer mortality and incidence with chlorinated drinking water in various locations in the United States (Craun 1993, IARC 1991). Cancer statistics, primarily mortality data for various cancers, were obtained for counties and sometimes census tracts; drinking water exposures for populations in the census tracts or counties were assessed from readily available information about water sources (ground or surface), disinfection practices, and occasionally trihalomethane levels; some demographic information was usually available to describe the population groups (e.g., educational background, nationality, urbanicity). In most studies, cancer mortality, for all and several specific sites, was found to be higher in areas where chlorinated surface water was used than in areas where unchlorinated groundwater was used, but it is difficult to interpret these associations. Are the observed associations due to exposure to chlorinated water, chlorinated byproducts, other contaminants in surface water, or other exposures or characteristics that were not assessed? For example, urban areas generally have higher cancer mortality rates because urban populations have opportunities for other exposures that might also be associated with cancer mortality and may be more likely to be correctly diagnosed with cancer. Urban areas also

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

159

frequently use surface water sources that are almost always chlorinated. Thus, it is difficult to determine with any certainty from these studies whether cancer mortality was causally or merely statistically associated with chlorinated surface water. Chlorinated surface water might serve as a surrogate for another characteristic(s) that is a cause of the observed cancer mortality. In addition, the association was nonspecific, as a variety of cancer sites were implicated. The primary value of these studies was to develop hypotheses for further study by analytical epidemiology. Analytical studies are able to provide both information about possible causal associations and the magnitude of the risk. In contrast to ecological studies, individuals within a population group or geographic area are selected for study. For each study participant, information is obtained about the person’s disease status, exposure to possible risk factors, and other demographic characteristics. Analytical studies can be either longitudinal or cross-sectional (Monson 1980). In a longitudinal study, the time sequence can be inferred between exposure and disease; that is, exposure precedes disease (Monson 1980). In a cross-sectional study, the data on exposure and disease relate to the same time period, making this type of study useful primarily for diseases with a short latent or incubation period. For example, the frequency of serological responses to Cryptosporidium antigens can be measured from sera collected from a cross-sectional survey, and information about sources of drinking water and other recent exposures can be related to the presence of a serological response to the antigens (Frost and Craun 1998; Frost et al. 1998, 2000a, 2000b). Cross-sectional studies can also provide important information for generating hypotheses and for interpreting potential causes of an outbreak. A good example is the cryptosporidiosis outbreak that occurred among residents and visitors to Collingwood, Ontario, during March 1996 (Frost et al. 2000a). The low level of reported diarrheal illness among adult Collingwood residents caused government officials and physicians to question whether an outbreak had occurred. A serological survey found evidence that Collingwood residents were likely to have been infected at the time of the outbreak but did not suffer illness from the infections. A high level of endemic infections prior to the outbreak may have protected Collingwood residents, whereas unprotected nonresidents who drank Collingwood water suffered high attack rates of illness. Longitudinal studies (Monson 1980) are of two distinct and opposite approaches: (1) the cohort study, which begins with an exposure or characteristic of interest and seeks to determine disease consequences of the exposure or characteristic; and (2) the case–control study, which begins with a disease or health condition of interest and seeks information about risk factors and possible exposures. A cohort study is also called a follow-up study. Individuals enter the cohort solely on the presence or absence of certain characteristics, a specific event, or their exposure status (e.g., chlorinated or unchlorinated water; high, moderate, low, no levels of arsenic in water). An advantage of this study is that any number of appropriate health-related outcomes or diseases can be assessed during the follow-up period. Morbidity or mortality incidence rates are determined for various diseases and compared for the exposed and unexposed groups in the cohort. A fundamental requirement is that the investigator not know the disease status of any individual when the cohort is

160

EPIDEMIOLOGIC CONCEPTS

assembled. A cohort can be based on currently defined exposures (e.g., disinfection byproducts or Cryptosporidium levels in water in March 1996) and followed forward in time. Determining possible cancer risks associated with a currently defined exposure (e.g., to disinfection byproducts), however, will require that the cohort be followed for many years into the future. To evaluate Cryptosporidium infection, the cohort can be followed for a much shorter time period. An alternative design for diseases with a long latency period is to assemble a historical cohort based on known exposures at some previous point in time. The follow-up period should be appropriate; that is, a sufficient time should be considered on the basis of the anticipated latency of the disease. For example, if a cohort could be established in terms of known drinking water exposures in 1970 (e.g., to disinfection byproducts), over 25 years of exposure would have already occurred, and the follow-up period would be relatively short. A special kind of cohort study, community intervention studies, can be conducted when a community changes water treatment or sources to improve their water quality. The study is prospective in nature and is conducted during a period before and after the water source or treatment change. Both individual level and community level illness and water exposure information can be collected, either in a longitudinal or cross-sectional study. This type of study has been used to determine enteric disease rates before and after water treatment changes, the relative source contribution of drinking water to community illness rates, and etiologic agents responsible for the observed illness (Calderon and Craun 2000). For comparison purposes, a similar study should also be conducted in a nearby control community that is demographically similar but is not undergoing a change in its water source or treatment. The primary advantage of this type of study is that water quality is improved at all places where persons may consume water (e.g., home, school, work, restaurants) and exposure misclassification is minimized. The household intervention study, as described previously, evaluates the change in water quality only for water consumed at the home, and this may only be a fraction of the water consumed throughout the day by study participants. Other important advantages are that a timeseries analysis of changes in health status can be conducted and that a large number of routinely collected community health surveillance data (e.g., clinical surveillance, hospital admissions, school absences, nursing home illness) can be evaluated in addition to longitudinal and cross-sectional data for diarrheal illness, etiological agents, and serological responses. Illnesses and risk factor data can be collected from participating families by daily diaries and=or telephone surveys. Crosssectional data for illness and risk factors can also be collected by periodic telephone surveys. Also, specific water treatment changes of interest (e.g., health improvements associated with surface water filtration) can be evaluated. A major limitation is that, for most community-level health changes, only relatively large changes can be detected. Since the studies must be conducted in areas considering changes, the areas may not be optimal in terms of water quality or population characteristics. Another limitation is the generalizability of results. Since the studies focus on specific water sources and treatment changes, results may not be generalized to the entire U.S. population; however, the findings may be applied to populations using similar water

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

161

sources and treatment. Furthermore, since people chronically exposed to a water supply may develop some protective immunity to endemic microbes, this type of study may not detect health risks that affect people without prior exposures, such as young children or new residents. In a case–control study (also called a case-comparison or case-referent study), individuals enter the study solely on the basis of disease status without knowledge of their exposure status. A single disease or health outcome (e.g., giardiasis, cryptosporidiosis, lung cancer, bladder cancer, blood lipid levels) is usually selected for study. Individuals with the particular disease or infection are selected during a specified time period within a defined geographic area or from selected hospital(s), clinic(s), or a specified cohort. A comparison group of individuals in which the condition or disease is absent (the controls) is selected, preferably randomly, from the same population in which the cases arise. Existing or past attributes and exposures thought to be relevant in the development of the disease are determined for all study participants (cases and controls). Because previous exposures are determined, a case–control study is sometimes referred to as a retrospective study. Information about any number of individual exposures or behavior (e.g., smoking, use of chlorinated or unchlorinated water, arsenic exposures) can be obtained. The frequency of exposure is compared for individuals with and without the disease to determine possible associations with the disease being studied. Case–control studies have provided a major contribution to our understanding of the causes of many diseases and are frequently used in outbreak investigations. This study design is usually more efficient than the cohort study, requires fewer study participants for adequate statistical power, and is often considered as the first option when studying risk factors. However, information on exposures must usually be obtained by questionnaire (e.g., spouses or parents of cases), and it is often difficult to accurately assess exposures that may have occurred many years ago. Ensuring that the quality and accuracy of information about exposures are similar for cases and controls is difficult.

7.6.2

Time-Series Analyses

A variation of the cohort approach and ecologic study is the time-series analysis. The time-series analysis has been used to relate water quality events (e.g., turbidity) to disease. The study is relatively inexpensive when routinely collected data are used, and it considers the health outcome for the community as a whole, including secondary effects. The community being studied serves as its own control. Both the background level of disease and water quality vary over time, and at least one year of data should be collected for disease and water quality so that seasonal changes can be assessed. The method considers time lags between exposure and the outcome. A time lag is important for waterborne infectious diseases because etiologic agents have different incubation periods ranging from 24 hours to several weeks or more. A major advantage is that the method tends to eliminate many possible confounding factors. For a factor to confound the association, it must vary in a similar way as water quality data, and most important confounders are not likely to vary in the same way as a water quality parameter. However, if higher levels of turbidity occur in the

162

EPIDEMIOLOGIC CONCEPTS

fall and winter, then seasonal risk factors for gastrointestinal disease could confound results from these studies. The incidence of gastroenteritis often varies with season of the year, and thus, the analysis should adjust for possible seasonal effects. Water quality data that can be evaluated include turbidity, coliform analysis, other routinely collected data, or pathogen-specific analyses. Disease measures include emergency room visits, physician visits, routinely collected illness surveillance data, and serological data. The advantage of these studies is that they can, theoretically, detect very small relationships between changes in water quality and illness. This occurs because of the large number of physician visits for gastrointestinal events. A major disadvantage of these studies involves the interpretation of the findings because of uncertainties about the observed associations. In addition, as the measures of exposure tend to be imprecise, it is not yet certain which measures of water quality or health data are most appropriate to include. The studies are also difficult to apply to small systems. Risks will be specific to source waters, type of treatment, susceptibility of distribution systems to contamination, individual susceptibility to disease, and water consumption patterns. It should be recognized that water quality and many of these factors change over time. The methodology has been applied successfully to the evaluation of an outbreak; Morris et al. (1996) evaluated whether previous unreported outbreaks of gastroenteritis had occurred in Milwaukee, the site of a filtered water system in which a large outbreak of cryptosporidiosis had occurred. However, several other studies that used a similar statistical methodology have caused considerable controversy. An study (Aramini et al. 2000) investigated the association between gastrointestinal outcomes in Vancouver, as assessed by hospitalizations, physician visits, and visits to the British Columbia Children’s Hospital emergency room, and water quality parameters, including turbidity and fecal coliform bacteria. Vancouver uses surface water without filtration. An association was found for these gastrointestinal outcomes and water turbidity; relative risks increased as turbidity increased. No association was reported between risk of illness and fecal coliform levels or rainfall. Studies in the United States have also reported an association between drinking water turbidity and gastrointestinal illness in the absence of an outbreak (Schwartz and Levin 1999; Schwartz et al. 1997, 2000). The studies evaluated very low turbidity levels in a filtered surface water supply. When published, the initial Schwartz et al. (1997) study received much criticism. Major concerns included the use of turbidity as a proxy or surrogate measure for risk of microbial contamination, exposure misclassification, and whether the observed turbidity associations are causal. Few epidemiologists accept that these studies have shown, beyond a reasonable doubt, a quantitative, casual association between waterborne diarrheal illness risk and water turbidity. However, the time-series analysis shows promise and will likely continue to be conducted, hopefully with improvements in exposure assessment. 7.6.3

Random and Systematic Error

A study must be of sufficient size and statistical power to detect the expected association. The association observed in each study must be evaluated to determine

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

163

that systematic and random errors are not responsible (Table 7.7). Systematic error or bias affects the validity of a study’s observed association. Random error, a measure of the precision of the risk estimate, is governed by chance. Systematic error can occur in the design and conduct of the study, leading to a false or spurious association or a measure of risk that departs systematically from the true value. Systematic error must be avoided or controlled, and when possible, its likely effect should be assessed. The likelihood that a positive association is due to random error can be estimated by calculating the level of statistical significance ( p value) or confidence interval (CI). In epidemiology, the CI is the preferred measure of random error because it provides a range of possible values for the risk estimate. It should be remembered, however, that random error or chance can never be completely ruled out as the explanation for an observed result and that statistical significance does not imply causality, biological significance, or lack of systematic error. To help interpret a negative association, statistical power calculations should be provided to specify the minimum risk a study was able to detect. Potential sources of systematic error include observation, selection, misclassification, and confounding bias. When the study population is not randomly selected or criteria used to enroll individuals in the study are not comparable for exposed and unexposed individuals or cases and controls, the observed exposure–disease association may be due to selection bias. Selection bias occurs only in study design and must be prevented because it cannot be corrected for in the analysis. To prevent selection bias, the study population must be randomly selected, exposed and unexposed groups must be selected without knowledge of their disease, or cases and controls must be selected without knowledge of their exposure. Observation bias results when disease or exposure information is collected differently from exposed and unexposed groups or cases and controls, respectively. Selective or differential recall of cases or controls about exposure will also result in a biased estimate of risk. For example, people with an illness, especially one that is severe, are more likely than persons without illness to better remember past events and possible exposures. Persons may also provide misleading information, especially when they believe that drinking water or food from a particular restaurant is the source of illness. Blinding study participants and=or investigators about the hypothesis being tested (if possible) and maintaining objectivity in collecting information can help minimize observation

TABLE 7.7 Considerations for Interpreting Epidemiologic Associations Lack of Random Error (Precision) Study size and statistical power

Lack of Systematic Error (Validity) Selection bias Misclassification bias Observation bias Confounding bias

164

EPIDEMIOLOGIC CONCEPTS

bias; however, this is often difficult to do, especially in case–control studies conducted during highly publicized outbreak investigations. Neugebauer and Ng (1990) discuss differential recall as a source of bias in case–control studies and present ways to remedy this problem. Media speculation about a waterborne source of the outbreak may occur shortly after outbreak identification, and this may, at the minimum, prompt cases to more completely recall tapwater consumption habits. Persons may also report illness symptoms because of a perceived risk from drinking water causing selection bias. An objective case definition that includes laboratory confirmation may decrease the total number of cases but will increase study precision. Tapwater consumption patterns may also change as a result of media speculation. Even if cases correctly report tapwater consumption prior to their illness, controls may report less tapwater consumption if they have reduced or eliminated tapwater consumption as a result of the outbreak publicity. The effect of differential recall may be more pronounced when the date of onset of illness was many weeks previous to their interview. For example, the reported association between illness with the use of tapwater compared to the exclusive use of bottled water in Clark County, Nevada, may have been affected by recall bias. Information was obtained from interviews of HIVinfected persons with and without cryptosporidiosis (Craun et al. 2001, Goldstein et al. 1996). Interviews were conducted approximately 100 days after the onset of illnesses. Cases were asked about bottled water consumption for the time prior to their illness, and controls were asked about bottled water consumption for a similar time period. Because of media speculation, it is possible that cryptosporidiosis cases who consumed primarily bottled water for the past year may have been more likely to recall occasional use of tapwater prior to the onset of their illness than would less motived persons who did not have cryptosporidiosis. Recall bias of this type would inflate estimates of waterborne illness risk. To help assess possible recall bias in a situation like this, efforts can be made to verify bottled water consumption of cases and controls, and additional questions should be asked to determine how bottled water users avoided exposure to tapwater. An erroneous diagnosis of disease or erroneous classification of a study participant’s exposure can result in misclassification bias and a poor or incorrect estimate of risk. The probability of misclassification may vary in either a differential or nondifferential manner among the groups under study. Nondifferential misclassification will almost always bias a study toward not observing an association when one may actually be present or underestimating the magnitude of the association. Differential misclassification bias can also result in misleading, incorrect associations that either under- or overestimate the magnitude of risk. In environmental studies where the magnitude of the association is often small, accurate assessment of exposure is critical, as the impact of misclassification can be severe. Exposure must be carefully assessed, so the possibility of nondifferential misclassification bias will not be cited as a general explanation when a small or no association is observed—specifically, ‘‘we observed no association but this does not rule out the possibility of an association; nondifferential misclassification of exposure may have occurred and this will bias a study toward the null hypothesis’’ or ‘‘the magnitude of the association may be

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

165

larger than the one observed because of nondifferential misclassification bias that could not be completely avoided.’’ The imprecise nature of the exposure estimate is a major problem, and more evaluation of routinely collected data is needed to determine if misclassification of exposure or disease is differential or non-differential. Greenland (1988) believes that epidemiologists should not presume that misclassification is nondifferential in environmental studies, and evidence should be provided to support the assertion that misclassification bias is nondifferential. Before initiating an environmental study, investigators should carefully evaluate how exposures will be assessed, and no study should be conducted unless the assessment is expected to be reasonably appropriate and accurate. This simple rule of thumb, if followed, can avoid much of the confusion surrounding the interpretation of results of poorly designed studies. For example, in five case–control studies of bladder and colon cancer risks associated with chlorinated drinking water, exposure was assessed only from information available from a death certificate (IARC 1991). Address at death, birth, or usual address was used to determine previous long-term exposures to chlorinated water. Because of the frequent migration occurring in the past 50 years, it is likely that study results were biased by the misclassification of exposure to chlorinated water. In two of the five studies, an increased risk of bladder cancer was reported, and in three studies an increased risk of colon cancer was reported. Some interpreted these results as evidence for an association between chlorinated drinking water and cancer and cited random misclassification bias as a possible reason an association was not observed in all of the studies. Others interpreted the results of the five studies as not meaningful, citing the possibility of nonrandom misclassification bias (i.e., it could not be determined if bias over- or underestimated the risk). Either interpretation is plausible. However, without additional data, epidemiologists should not assume that the studies found no association because of random misclassification of exposure. Confounding bias may convey the appearance of an association; that is, a confounding characteristic rather than the putative cause or exposure may be responsible for all or much of the observed association. A confounding characteristic can cause or prevent the disease, is not on the causal pathway from exposure to disease, and is associated with the exposure being evaluated. An example of a confounding characteristic is provided by Monson (1980). Cigarette smoking is a cause of lung cancer, and it is also associated with heavy alcohol consumption. In studying lung cancer, an association was observed between heavy drinking and lung cancer; that is, the lung cancer rate was greater in heavy drinkers than in nondrinkers. Alcohol drinking probably does not directly cause lung cancer, and the observed association between drinking and lung cancer is likely caused by confounding bias of cigarette smoking. Confounding bias does not necessarily result from any error of the investigator. It is potentially present in all studies and must always be considered as a possible explanation for any observed association. If a characteristic can be made or demonstrated to have no association with exposure or disease, that characteristic cannot confound the association between exposure and disease. For example, in a study of radon exposure and lung cancer, smoking would not cause confounding bias in a study where smoking habits were similar among radon-

166

EPIDEMIOLOGIC CONCEPTS

exposed and unexposed persons. During study design, confounding bias can be prevented or minimized by matching cases and controls for specific characteristics such as smoking status or randomization of study participants into exposure groups. During data analysis the following are used to assess and control confounding bias: stratification or analysis of risk estimates by various characteristics, such as age, gender, smoking status (e.g., whether risks are greater among the elderly or young, males or females, smokers or nonsmokers) and multivariate techniques, such as regression analysis. Logistic regression that considers the natural logarithm of the odds of disease is often used in studies. In multiple regression analysis, confounding can be controlled; however, it should be remembered that con-founding bias can be controlled or evaluated only for those characteristics for which information is collected. In the example of an observed association between alcohol consumption and lung cancer, a stratified analysis was used to assess the possible confounding bias of cigarette smoking (Monson 1980). Since smoking was the suspected confounding characteristic, the data were analyzed separately in smokers and nonsmokers. Smoking was associated with lung cancer, but no association was seen between alcohol consumption and lung cancer in either smokers or nonsmokers. The observed association between alcohol use and lung cancer was due to confounding bias of cigarette smoking. Not to be confused with a confounding bias, effect modification refers to a change in the magnitude of the effect of a putative cause (Last 1995). A characteristic (e.g., age) can cause confounding bias in one study and modify the risk in another study (Rothman 1996). A classic example of effect modification is the interactive effects of smoking and asbestos exposure. Smoking or asbestos exposure increase the risk of lung cancer. However, exposure to both smoking and asbestos will increase the lung cancer risk much more than would be predicted for each exposure alone. In this case smoking is an effect modifier for asbestos exposure (Kleinbaum et al. 1982). Effect modification is a finding to be reported rather than avoided (Rothman 1996). In an experimental study, randomization is possible—each individual in the study has an equal or random chance of being assigned to an exposed or unexposed group (i.e., tapwater, bottled water, or another water group). Because of this random assignment of exposure, all characteristics, confounding or not, tend to be distributed equally between groups of different exposure. This means that over the long run, if experiments are repeated and similar exposure–disease associations are observed, confounding bias is an unlikely explanation for the observed associations. Selection bias and misclassification of exposure are usually avoided because of the study design, but misclassification of disease and observation bias are still of concern. Reporting bias may be important in experimental studies where disease is self-reported or only symptoms of disease are reported, and if possible, disease status should be confirmed independently by clinical analysis. For example, because of their knowledge of the water source or treatment, persons randomly assigned to a group that receives specially treated bottled water or home filters may report symptoms or disease differently than those persons in a group using tapwater. Reporting bias can be minimized and assessed by including a group that receives bottled tapwater, and then switching the type of water received by each group midway

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

167

through the study in an attempt to blind the participants about their water source. The intent is to eliminate biases or prejudices of the investigators and study participants. In a blinded study either the investigators or study participants or both (double-blinded) do not know to which group, experimental or control, a study participant has been assigned. Laboratory analysts may also be blinded (tripleblind). A study in California (Colford et al. 2002) evaluated the ability to blind participants in household intervention studies. The study was triple-blinded (investigators, the treatment installer, and the study participants), and the sole purpose of the study was to evaluate ability to blind subjects as to whether they had a true treatment device or a sham treatment device installed in their homes. The investigators concluded that study participants could be blinded as to whether they had a true treatment device or a sham device.

7.6.4

Measures of Association

The basic measures of an association in analytical studies are the rate difference (RD) and rate ratio (RR). The rate difference is a measure of the absolute difference between two rates, such as incidence rate of disease for the exposed minus the incidence rate for the unexposed in a study population. The rate ratio is a relative measure of two rates. The ratio of two rates, for example, incidence rate for the exposed divided by the incidence rate for the unexposed in a study population. The rate ratio is also called the relative risk (RR). A RR of unity (1.0) indicates no association or no increased risk; any other ratio signifies either a positive or negative association, provided the association is not subject to systematic error. For example, a RR of 1.8 indicates an 80% increased risk of disease among the exposed; a RR of 0.8 indicates a decreased risk or beneficial effect of 20%. Precision of the risk estimate or random error is assessed by the CI. For example, 95% CI ¼ 1.6–2.0 indicates a precise and statistically significant (the CI is narrow and does not include 1.0) estimate, whereas, CI ¼ 0.8–14.5 indicates the estimate is not statistically significant (the CI includes 1.0) and imprecise (a wide range of values). Because the selection of participants is based on their disease status in a case– control study, the odds ratio (OR) is determined rather than the RR. The OR is the odds or chance of disease among the exposed divided by the odds of disease among the unexposed and is essentially equivalent to the RR, especially if the disease is rare, such as cancer. Another important measure, the attributable risk (AR), is an estimate of the rate of a disease or other outcome in exposed individuals that can be attributed to the exposure in question, provided the association is causal. This measure is obtained by subtracting the rate of disease among the unexposed from the rate among the exposed (see rate difference) (Last 1995). Unfortunately, AR is not always consistently used by epidemiologists, and readers should be careful to understand what the investigator means when AR is reported in a study (Rothman 1996). AR among the exposed or attributable fraction (AF) exposed is used to denote the excess risk among the exposed; population AR or population AF is used to denote the excess

168

EPIDEMIOLOGIC CONCEPTS

risk in the entire population, exposed and unexposed. AR is expressed as either a percent or proportion (Last 1995).

7.6.5

Strength of Association

The magnitude of the RR can help assess if an observed association may be spurious. On the basis of experience (Monson 1980), it is difficult to interpret weak associations, a RR of less than 1.5 (Table 7.8). One or more confounding characteristics can easily lead to a weak association between exposure and disease, and it is usually not possible to identify and adequately measure or control weak confounding bias. On the other hand, a large adjusted RR is unlikely to be completely explained by an unidentified or uncontrolled confounding factor. When the study has a reasonably large number of participants and the adjusted RR is large, random error and confounding bias are less likely to be responsible for an observed association. The magnitude of a RR, however, has no bearing on the possibility that an association is due to observation, selection, or misclassification bias. Any of these biases can lead to a total misrepresentation of an observed association. Since a RR of <1.5 is frequently observed in environmental studies, results of these studies usually cause considerable controversy about whether the observed association represents a real risk or is due to bias. As discussed earlier good-quality exposure assessment measurements and study design are important to interpret risks of this magnitude.

7.6.6

Causality of an Association

Interpretation of data should be made with caution and in the context of all relevant scientific information about the disease and its etiology. No single study, even one with little systematic error, can provide a definitive answer about the exposure– disease association. Results from several studies of different design and different population groups allow a more definitive conclusion, and it may be necessary to consider studies in both the general and special populations. Judging causality in epidemiology is based on guidelines (Beaglehole et al. 1993, Craun 1990, Rothman 1996, Hill 1965), which include  Temporal Association. Exposure must precede the disease, and in most studies this can be inferred. When exposure and disease are measured simultaneously, it is possible that exposure has been modified by the presence of disease.  Strength of Association. The larger the RR or OR the less likely the association is to be spurious or due to confounding bias. However, a causal association should not be ruled out simply because a weak association is observed.  Consistency. Repeated observation of an association under different study conditions supports an inference of causality; however, its absence does not rule it out.

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

169

 Specificity. A putative cause or exposure leads to a specific effect. The presence of specificity argues for causality, but its absence does not rule it out.  Biological Plausibility. When the association is supported by evidence from clinical research or basic sciences (e.g., toxicology, microbiology) about biological behavior or mechanisms, an inference of causality is strengthened.  Dose–Response Relationship. A causal interpretation is more plausible when an gradient is found (e.g., higher risk is associated with larger exposures).  Reversibility. An observed association leads to some preventive action, and removal or reduction of the exposure leads to a reduction in or risk of disease. The inference that an association is causal may be controversial, and epidemiologists have debated how scientific evidence should be evaluated in an attempt to better understand causal inferences. Even when repetitions of an association are observed, questions may remain as to whether these associations really constitute an ‘‘empirical demonstration that serves as a valid platform for (causal) inference’’ or whether ‘‘the process is still steeped in uncertainty’’ (Rothman 1986). Thus, when environmental policymakers and regulators are confronted with associations that suggest the need for action, they must be aware of any uncertainties. Scientific evidence is often conflicting, and the type of evidence or studies that are considered in the evaluation must be given due weight on the basis of issues mentioned previously (i.e., study design, study precision, study validity). Unfortunately for environmental issues, this is a rarely conducted exercise. As the literature grows for many drinkingwater-related studies, more efforts to evaluate causality should be undertaken.

7.6.7

Meta analysis

A meta analysis is sometimes conducted to summarize the results from a group of studies of a related hypothesis (Greenland 1994, Morris et al. 1992, Craun et al. 1993, Anonymous 1992). A frequent application of metaanalysis has been the pooling of results from a number of small randomized controlled clinical trials. By combining studies (usually of identical or similar design), problems associated with the use of small study populations can be overcome and a more precise estimate of risk can potentially be obtained. Generally, a complete, systematic literature review and evaluation of the quality and consistency of findings among the studies TABLE 7.8 Guide to the Strength of an Epidemiologic Association Rate Ratio 1.0 1.0–1.5 1.5–3.0 3.1–10.0 >10.0

Strength of Association None Weak Moderate Strong Infinite

Source: Adapted from Monson (1980).

170

EPIDEMIOLOGIC CONCEPTS

is also conducted. This methodology has been applied to observational studies. For example, a meta analysis of 10 observational studies reported a small increased risk of bladder (RR ¼ 1.21) and rectal (RR ¼ 1.38) cancers associated with chlorinated drinking water (Morris et al. 1992, Craun et al. 1993). This example illustrates the problem of applying meta analysis to observational studies. The analysis failed to evaluate the quality of the individual studies included, consider the differences in study design, or adjust for differing approaches to exposure assessment and confounding and thus contributed very little to our understanding of the observed associations between chlorinated water and cancer (Craun et al. 1993). The utility of a meta analysis to provide a summary risk estimate for observational studies of water disinfection and cancer risk continues to be questionable (Anonymous 1992, Bailar 1995, Poole 1997).

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES Examples of several types of drinking water studies are presented to illustrate study design issues, potential biases, and how to compute measures of association. 7.7.1

Experimental Studies

Microbial Risks. In a population experimental study (Payment et al. 1991) conducted near Montreal, it was estimated that up to 35% (AR exposed) of self-reported, mild, gastrointestinal illness experienced by tapwater consumers over a 15-month period was waterborne; 607 households were randomly assigned to either a group consuming municipal tapwater or a group receiving tapwater further treated by reverse osmosis and monitored for enteric illness (Table 7.9). The water source was a river source contaminated with sewage and treated with predisinfection, coagulation, flocculation, rapid sand filtration, ozone, and chlorine. The tapwater met all current microbiological and physical limits, and no outbreak of illness was reported during the study. Study results reported in Table 7.9 illustrate how measures of association (RD, RR, and AR exposure) between drinking water exposure and gastroenteritis incidence were computed. Highly credible gastroenteritis episodes were defined as those that involved either of the following combination of symptoms: vomiting or liquid diarrhea or nausea, soft diarrhea combined with abdominal cramps, and staying home from work or school or visit to a physician. A similar study was conducted a few years later (Payment et al. 1997) and found an AR of 14% of symptomatic gastroenteritis was related to drinking water. A population experimental study in Australia (Hellard et al. 2001) found no difference between exposed and unexposed families and concluded that drinking water contributed little, if any, to gastroenteritis in the community. Additional studies are currently (2003) being conducted in the United States to better define the magnitude of endemic waterborne disease risk associated with various water sources. It is possible that

171

1.00 0.64

Study Period

3=88–6=88 9=88–6=89 0.65 0.43

Reverse-Osmosis Filtered Tapwater (b) 0.35 0.21

Risk Difference (RD ¼ a 7 b)

a Annual incidence per family adjusted for age, gender, and subregion. Source: Payment et al. (1991).

Tapwater (a) 1.53 2.05

Relative Risk (RR ¼ a=b)

35% 33%

Attributable Risk Exposed (AR ¼ a 7 b=a)

TABLE 7.9 Incidencea of Highly Credible Gastroenteritis Episodes, Experimental Epidemiologic Study

172

EPIDEMIOLOGIC CONCEPTS

the different results observed in Canada and Australia are due to contamination levels of the source water and operation. 7.7.2

Cohort Studies

THMs and Spontaneous Abortions. A large prospective study of pregnant women conducted in California, from 1989 to 1991, revealed statistically significant associations between the spontaneous abortion rate and trihalomethane levels (Waller et al. 1998). The women were recruited from the members from a large managed healthcare organization. The subject’s address was used to determine her residential drinking water utility. The total trihalomethanes water quality reports for the period were obtained directly from the identified utilities. For 77% of the cohort, the trihalomethane levels were estimated by averaging all distribution measurements taken by the subject’s utility within the subject’s first trimester. For the remainder of the cohort who had no other data available, the annual average from the utility’s annual water quality report was used. Similar methods were used to estimate first trimester drinking water levels of individual trihalomethanes. Each woman’s daily tapwater intake at 8 weeks’ gestation was estimated from information taken during a telephone interview. The analysis controlled for the following factors that were independently related to spontaneous abortion: gestational age at interview, maternal age at interview, cigarette smoking, and history of pregnancy loss, maternal race, and employment during pregnancy. Women who drank more than five glasses per day of cold tapwater containing >75 mg=L total trihalomethanes had an adjusted odds ratio of 1.8 for spontaneous abortion (95% CI 1.1–3.0). Of the four individual trihalomethanes, only high bromodichloromethane exposure (consumption of more than five glasses per day of cold tapwater containing >18 mg=L bromodichloromethane) was associated with spontaneous abortion both alone (adjusted OR ¼ 2.0; 95% CI ¼ 1.2–3.5) and after adjustment for the other trihalomethane (adjusted OR ¼ 3.0; 95% CI ¼ 1.4–6.6). Chlorinated Water and Cancer Mortality. A cohort study conducted in Washington County, MD, in 1975 found no statistically significant associations between the incidence of cancer mortality and residence in an area where chlorinated surface water was distributed (Wilkins and Comstock 1981). The cohort was established from a private census during Summer 1963 and followed for 12 years through July 1975. The source of drinking water at home was ascertained and personal and socioeconomic data were collected for each county resident including age, education, smoking history, and number of years residing at the 1963 address. Potential cases of cancer were obtained from death certificate records, the county’s cancer registry, and medical records of the county hospital and a regional medical center. Census data were used to compute age–gender–site-specific cancer mortality rates for 27 causes of death, including 16 cancer sites, cardiovascular disease, vehicular accidents, all causes of death, and pneumonia at the end of the follow-up period in 1975. Three exposure categories were examined: a high exposure group of residents served by chlorinated surface water, a low-exposure group served by unchlorinated

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES

173

deep wells, and a third group served by a combination of chlorinated surface water and groundwater. The average chloroform level from an extensive analysis of chlorinated surface water samples was 107 mg=L. The third group, which likely represented an intermediate exposure, was not used in detailed analyses. In the analysis, confounding bias was controlled and incidence rates were adjusted by multiple regression analysis for age, marital status, education, smoking history, frequency of church attendance, adequacy of housing, and persons per room in the household. Selected cancer mortality rates for males and females are reported in Table 7.10. The RR for liver cancer mortality among females is 1.81 (RR ¼ 19.9=11.0). Although the study was of high quality and well conducted, the associations reported are subject to random error (i.e., all RRs had a CI that included 1.0 and thus were not statistically significant). Even though some 31,000 people were included in the cohort, estimates of the magnitude of bladder cancer risk associated with chlorinated surface water were based on only 29 deaths in females and 51 in males. History of length of residence by each person at the 1963 address was used to estimate his or her duration of exposure to chlorinated and unchlorinated water. For bladder, liver, and lung cancer in females and bladder cancer in males, the association was stronger for persons who had lived in their 1963 domicile for 12 or more years than for those who had been residents for 3 years or less. Among men who had been in their 1963 homes 12 years or more and thus had at least 24 years exposure to chlorinated surface water in 1975, the RR for bladder cancer was 6.46 (95% CI ¼ 1.00–100). Although the estimated magnitude of bladder cancer risk was ‘‘strong,’’ random error is illustrated by the large CI. The estimate ranged from RR ¼ 1.00 (no increased risk) to RR > 100, a very imprecise and statistically unstable estimate. Additional follow-up of the cohort for several more years could possibly have provided a more statistically stable association. Small numbers of

TABLE 7.10

Incidencea of Cancer Mortality in Cohort Study in Maryland Chlorinated Surface Water

Cause of Death

Deaths

Incidence Rate

Liver cancer Kidney cancer Bladder cancer

31 11 27

19.9 7.2 16.6

Liver cancer Kidney cancer Bladder cancer

9 15 46

6.4 10.6 34.6

a

Unchlorinated Groundwater Deaths Females 2 2 2 Males 2 3 5

Adjusted incidence rate per 100,000 person-years. Confidence interval. Source: Wilkins and Comstock.

b

Risk

Incidence Rate

RR

95% CIb

11.0 7.1 10.4

1.81 1.01 1.60

0.64–6.79 0.26–6.01 0.54–6.32

9.0 13.6 19.2

0.71 0.78 1.80

0.19–3.51 0.27–2.69 0.80–4.75

174

EPIDEMIOLOGIC CONCEPTS

cases lead to an imprecise estimate of risk; more years of observation may have yielded more cases.

7.7.3

Case–Control Studies

Water Disinfection and Bladder Cancer. In Massachusetts, a number of towns have used surface water disinfected only with either chlorine or chloramine since 1938, providing an opportunity to compare cancer risks between these two disinfectants. A previous ecological study had revealed bladder cancer mortality to be weakly associated with residence at death in Massachusetts communities using chlorine disinfection (Zierler et al. 1986). Because of likely exposure misclassification bias in this study, a case–control study was conducted to further explore the association (Zierler et al. 1988). Eligible for the study were all persons who were >44 years of age at death and who died during 1978–1984 from either bladder cancer, lung cancer, lymphoma, cardiovascular disease, cerebrovascular disease, or chronic obstructive pulmonary disease while residing in 43 selected communities. Included were 614 persons who died of primary bladder cancer and 1074 individuals who died of other causes. Confounding bias for age, gender, smoking, and occupation was controlled by multiple logistic regression. Analyses included a person’s usual exposure (at least 50% of their residence since 1938 was in a community where surface water was disinfected by only one of the two disinfectants, either chlorine or chloramine) or lifetime exposure to water disinfected with only one of the two disinfectants. The mortality OR for lifetime exposure to chlorinated surface water (OR ¼ 1.6) was higher than the OR for usual exposure (OR ¼ 1.4), and only the OR for lifetime exposure is presented in Table 7.11. After adjusting for various confounding characteristics, a 60% increased risk of bladder cancer mortality was found among lifetime residents of communities that used only chlorinated surface water compared to lifetime residents of communities that used only chloraminated surface water (OR ¼ 1.6 vs. 1.0). The association is statistically significant (i.e., the CI does not include 1.0), and the estimate of risk is precise (i.e., the CI is small). The study is of high quality; systematic bias was evaluated and not felt to be of concern. However, since the magnitude of the association is not large, confounding by unknown, unmeasured characteristics may be present, and it is difficult to interpret this association because it is the only analytical study that compares chlorinated and chloraminated water. TABLE 7.11 Case–Control Study of Bladder Cancer Mortality in Massachusetts Drinking Water Exposure Chlorinated, lifetime exposure Chloraminated, lifetime exposure a

Cases

Controls

ORa

95% CI

251 224

323 387

1.6 1.0

1.2–2.1

Adjusted for age, gender, cigarette pack–years, and residence in community with high-risk occupations. Source: Zierler et al. (1988).

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES

175

Arsenic and Bladder Cancer. Bates et al. (1995) evaluated bladder cancer associations in a U.S. population exposed to relatively low levels of drinking water arsenic. The case–control study of Utah respondents to the National Bladder Cancer Study in 1978 included 117 bladder cancer cases and 266 population-based controls and was conducted in areas where 92% of towns had drinking water arsenic levels less than 10 mg=L; one town had more than 50 mg=L of arsenic. Persons were interviewed, and individual exposures to arsenic in drinking water were estimated by linking residential history information with water sampling information. Two indices of cumulative arsenic exposure were used: total cumulative exposure and intake concentration. Exposures were in the range 0.5–160 mg=L arsenic (mean 5.0 mg=L). No overall increase was reported in bladder cancer risk with increasing exposure to arsenic in drinking water considering either cumulative dose or intake concentration. However, Bates et al. (1995) reported that among ‘‘cigarette smokers there was a non-significant elevation in risk that was not dose related.’’ Among smokers only, positive trends in risk were found for exposures estimated for decade-long time periods, especially in the 30–39-year period prior to diagnosis. Table 7.12 presents information about the magnitude of relative risks, none of which were statistically significant (i.e., the CI included unity), among all participants using cumulative dose of arsenic from water. In a case–control study, the odds ratio (OR) is interpreted as a relative risk; that is, the cancer risk for exposed persons is relative to risk for persons who are unexposed or have low exposure. Persons in the study that had cumulative waterborne arsenic exposures <19 mg were considered as the baseline; persons with exposures of 19–33 mg had 56% higher risks than those with exposures of less than 19 mg. However, persons with exposures of 33–53 mg had 5% less risk, and those with the highest exposure had 41% higher risks. Relative risks did not increase with increased exposure, and none were statistically significant. Kurttio et al. (1999) also considered low drinking water arsenic exposures in a more recent case–control study of bladder and kidney cancer in Finland. Individual exposures and confounding factors were considered for each case of bladder cancer, each case of kidney cancer, and each control. Study participants were selected from a register-based cohort of all Finns who had lived at an address outside the municipal

TABLE 7.12 Case–Control Study of Arsenic and Bladder Cancer in Utah Cumulative Dose of Arsenic (mg) <19 19–33 33–53 >53 a

Bladder Cancer Risk OR (95% CI)a 1 1.56 (0.8–3.2) 0.95 (0.4–2.0) 1.41 (0.7–2.9)

Risks were adjusted for gender, age, smoking, and other possible confounders. Source: Bates et al. (1995).

176

EPIDEMIOLOGIC CONCEPTS

drinking water system during 1967–1980 (n ¼ 144,627). The final study population consisted of 61 bladder cancer cases and 49 kidney cancer cases diagnosed between 1981 and 1995, as well as an age- and sex-balanced random sample of 275 subjects, which were the reference cohort. Kurttio et al. (1999) assessed the levels of arsenic in drilled wells in Finland. Water samples were obtained from the wells used by the study population at least during 1967–1980. Arsenic concentrations in the wells of the reference cohort were low. The reported median concentration was 0.1 mg=L, and the maximum was 64 mg=L; 1% exceeded 10 mg=L. Arsenic exposures that were evaluated included the measured arsenic concentration in the well and the estimated daily dose and cumulative dose of arsenic. In Finland, none of the exposure indicators was statistically significantly associated with the risk of kidney cancer. Bladder cancer tended to be associated with arsenic concentration in the well and with daily dose during years 3–9 prior to the cancer diagnosis (short latency period). Considering a short latency period, the relative risk for bladder cancer for water arsenic concentrations >0.5 mg=L was more than twice the risk for concentrations <0.1 mg=L (Table 7.13). At exposures for 10 years before diagnosis (long latency period), no statistically significant increased risk was observed. Also, no statistically significant increased risks of bladder cancer were seen with cumulative water arsenic dose at either short or long latencies (Table 7.13). Since exposures earlier than 10 years before cancer diagnoses did not show an association with bladder cancer risk, the investigators suggested that relatively recent arsenic exposure might be more relevant for assessing bladder cancer risk. This is consistent with the hypothesis that arsenic may act

TABLE 7.13 in Finland

Case–Control Study of Arsenic and Bladder Cancer Bladder Cancer Risk OR (95% CI)

Exposures Water arsenic (mg=L) <0.1 0.1–0.5 >0.5–64 Daily arsenic dose (mg=day) <0.2 0.2–1.0 >1.0 Cumulative arsenic dose (mg) <0.5 0.5–2.0 >2.0 a

Exposure 3–9 years before diagnosis. Exposure  10 years. Source: Kurttio et al. (1999).

b

Short Latencya

Long Latencyb

1 1.53 (0.8–3.1) 2.44 (1.1–5.4)

1 0.81 (0.4–1.6) 1.51 (0.6–1.6)

1 1.34 (0.7–2.7) 1.84 (0.8–4.0)

1 0.76 (0.4–1.5) 1.07 (0.5–1.5)

1 1.61 (0.7–3.5) 1.50 (0.7–3.2)

1 0.81 (0.4–1.7) 0.53 (0.3–1.1)

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES

TABLE 7.14

177

Smoking, Arsenic in Water, and Bladder Cancer in Finland Bladder Cancer Risk OR (95% CI)

Exposure Water arsenic (mg=L) <0.1 0.1–0.5 >0.5–64

Smoker in the 1970s

Nonsmokera or Ex-smoker in the 1970s

1 1.10 (0.2–6.2) 10.3 (1.2–92.6)

1 0.95 (0.3–3.6) 0.87 (0.3–3.0)

a In this context, one who has never smoked. Source: Kurttio et al. (1999).

as a promoter or cocarcinogen in the late stage of carcinogenesis. Kurttio et al. (1999) also reported a possible synergistic effect between arsenic and smoking (Table 7.14). The association between arsenic exposure and bladder cancer tended to be stronger among persons who smoked in the 1970s. This is consistent with results of experimental studies suggesting that arsenic may promote carcinogenicity of known carcinogens. It is interesting to compare the results of case–control studies in Utah and Finland, since both evaluated latency and low arsenic exposures. Information presented in Table 7.15 shows no statistically significant increased risks of bladder cancer associated with cumulative arsenic doses ranging from <0.5 mg to >10.2 mg. These findings suggest that cumulative exposures to low levels of arsenic may not increase bladder cancer risks. However, it is possible that current levels of arsenic measured in the wells more accurately reflect short-term exposures. Cumulative long-term exposures may be less important than short-term exposures, or these exposures may be more sensitive to misclassification errors when estimating exposure. These

TABLE 7.15 Comparison of Bladder Cancer Risks Reported in Utah (Bates et al. 1995) and Finland (Kurttio et al. 1999) Bladder Cancer Risk OR (95% CI) Cumulative Arsenic Dose (mg)

Short Latency

Long Latency

Finland <0.5 0.5–2.0 >2.0 Utah <4.4 4.4–6.9 6.9–10.2  10.2

(<9 years) 1 1.61 (0.7–3.5) 1.50 (0.7–3.2) (<9 years) 1 1.18 (0.6–2.3) 0.97 (0.5–2.0) 1.11 (0.6–2.2)

(10 years and earlier) 1 0.81 (0.4–1.7) 0.53 (0.3–1.1) (40–49 years) 1 0.52 (0.2–1.8) 0.68 (0.2–2.1) 0.65 (0.2–2.4)

178

EPIDEMIOLOGIC CONCEPTS

two case–control studies have provided provocative findings, and the investigators recommended that further research should be conducted.

7.8

FUTURE TRENDS IN EPIDEMIOLOGY AND DRINKING WATER

Originally concerned with communicable disease outbreaks and epidemics, epidemiology has evolved to include the study of chronic diseases, accidents, and various other health conditions. Its evolution continues as new methods are developed to better study the small risks associated with occupational and environmental exposures (Talbott and Craun 1995). Biochemical techniques are now frequently used in serological and molecular epidemiology and can improve a study’s sensitivity to detect associations (Hulka et al. 1990). Molecular techniques in clinical specimens and environmental samples allow improved detection of pathogens. With the widespread use of improved analytical techniques to measure cellular, biochemical, or molecular alterations in human tissues, cells, or fluids, adverse effects can be studied more carefully in people who do not have overt symptoms of disease. These biological markers may also serve as indicators of exposure or dose. For example, the study of serum antibodies in populations with different water sources and water treatment may help epidemiologists determine endemic waterborne risks of cryptosporidiosis. Descriptive studies are useful to identify emerging problems, to develop specific hypotheses for study by analytical studies, and in some instances to evaluate health conditions and control programs. Experimental and analytical studies are able to provide evidence of a causal association between exposure and disease and estimates of the magnitude of risk, but the studies must be carefully designed and conducted to avoid random and systematic bias. Because small risks are usually observed in environmental studies, it is extremely important to consider the effect of misclassification bias and confounding on the interpretation of a study’s results. When considering the evidence for a specific hypothesis, it is important to critically evaluate each study to determine the amount of information it is able to contribute to the overall evaluation of an association’s causality or magnitude of risk. Few efforts have been made to systematically evaluate drinking water studies either as part of a convened group of epidemiologists or as a review of the literature. An improved systematic approach is needed for an overall evaluation of individual studies and to summarize the information in a more meaningful way than, for example, the meta analysis that attempted to assess risks that may be associated with chlorinated water. As any science, epidemiology has its own vocabulary, and to understand the results of studies, readers must become familiar with the terminology. Readers must also be able to evaluate the quality of the studies. Well-designed and -conducted studies, however, can provide evidence (or the lack thereof) of specific health effects in humans, and estimate the magnitude of risk under real-life exposures or conditions. Epidemiology has been successfully used to reduce uncertainty in risk assessments and regulatory actions and evaluate the beneficial effects of public health control measures. As epidemiologic studies become more precise,

ACKNOWLEDGMENTS

179

bias-free, and improved validity, the effects associated with drinking water contaminants will become better understood. This will also improve the risk assessments associated with those contaminants and allow regulatory agencies to develop policies and regulations with less uncertainty.

ACKNOWLEDGMENTS This chapter was expanded and updated based on the basis of a previously published article, ‘‘An introduction to epidemiology,’’ written by the same authors and published in the public domain in J. Am. Water Works Assoc. 88(9): 54–65, 1996.

REFERENCES Anonymous 1992. Editorial, Chlorinated water and cancer: is a meta-analysis a better analysis? Pediatric Alert 17:91. Aramini, J., J. Wilson, B. Allen, J. Holt, W. Sears, M. McLean, and R. Coper. 2000. Drinking Water Quality and Health Care Utilization for Gastrointestinal Illness in Greater Vancouver. Population and Public Health Branch, Guelph, Ontario: Health Canada. Bailar III, J. C. 1995. The practice of meta-analysis. J. Clin. Epidemiol. 48(1):149–157. Bates, M.N., A.H. Smith, and K. P. Cantor. 1995. Case-control study of bladder cancer and arsenic in drinking water. Am. J. Epidemiol. 141:523–530. Beaglehole, R., R. Bonita, and T. Kjellstrom. 1993. Basic Epidemiology. Geneva, Switzerland.: World Health Organization. Bennett, J. V., S. D. Holmberg, M. F. Rogers, and S. L. Soloman. 1987. Infectious and parasitic diseases. In Closing the Gap: The Burden of Unnecessary Illness. R. W. Amler and H. B. Dull, eds. Oxford, UK: Oxford Univ. Press, pp. 102–114. Calderon, R. L., and G. F. Craun. 2000. Community intervention study for estimation of endemic waterborne disease. Epidemiology 11:S123. Chappell, C. L., P. C. Okhuysen, C. R. Sterling, and H. L. DuPont. 1996. Cryptosporidium parvum: Intensity of infection and oocyst excretion patterns in healthy volunteers. J. Infect. Diseases 173:232–236. Colford, J. M., J. R. Rees, T. J. Wade, A. Khalakdina, J. F. Hilton, I. J. Ergas, S. Burns, A. Benker, C. Ma, C. Bowen, D. C. Mills, D. J. Vugia, D. D. Juranek, and D. A. Levy. 2002. Participant blinding and gastrointestinal illness in a randomized, controlled trial of an inhome drinking water intervention. Emerg. Infect. Diseases 8(1):29–36. Craun, G. F. 1990. Methods for the Investigation and Prevention of Waterborne Disease Outbreaks. EPA=600=1-90=005a. Washington, DC: USEPA. Craun, G. F. 1993. Epidemiologic studies of water disinfectants and disinfection by-products. In Safety of Water Disinfection: Balancing Chemical and Microbial Risks. G. F. Craun, ed. Washington, DC: ILSI Press, pp. 277–302. Craun, G. F. 1994. Chairman, Report of New York City’s Advisory Panel on Waterborne Disease Assessment, New York City Dept. Environmental Protection, Oct. 7.

180

EPIDEMIOLOGIC CONCEPTS

Craun, G. F., F. J. Frost, R. L. Calderon, E. D. Hilbvorn, K. M. Fox, D. J. Reasoner, C. L. Poole, D. J. Rexing, S. A. Hubbs, and A. P. Dufour. 2001. Improving waterborne disease outbreak investigations. Internatl. J. Environ. Health Research 11:229–243. Craun, G. F., R. M. Clark, J. Doull, W. Grabow, G. M. Marsh, D. A. Okun, M. D. Sobsey, and J. M. Symons. 1993. Conference conclusions. In Safety of Water Disinfection: Balancing Chemical and Microbial Risks. G. F. Craun, ed. Washington, DC: ILSI Press, pp. 657–668. Dann S. M., P. C. Okhuysen, B. M. Salameh, H. L. Dupont, and C. L. Chappel. 2000. Fecal antibodies to Cryptosporidium parvum in healthy volunteers. Infect. Immunol. 68:5068– 5074. Doll, R. and A. B. Hill. 1964. Mortality in relation to smoking: ten years’ observations of British doctors. Br. Med. J. i: 1399–1410. DuPont, H. L., C. L. Chappell, C. R. Sterling, P. C. Okhuysen, J. B. Rose, and W. Jakubowski. 1995. The infectivity of Cryptosporidium parvum in healthy volunteers. New Engl. J. Med. 332:855–859. Frost, F. J., A. A. de la Cruz, D. M. Moss, M. Curry, and R. L. Calderon. 1998. Comparisons of ELISA and Western blot assays for detection of Cryptosporidium antibody. Epidemiol. Infect. 121(1):205–211. Frost, F. J. and G. F. Craun. 1998. Serological response to human Cryptosporidium infection (letter). Infect. Immunol. 66(8):4008. Frost, F. J., T. Muller, G. F. Craun, D. Frasier, D. Thompson, R. Notenboom, and R. L. Calderon. 2000a. Serological analysis of a cryptosporidiosis epidemic. Internatl. J. of Epidemiol. 29:376-379. Frost, F. J., T. B. Muller, C. K. Fairley, J. S. Hurley, G. F. Craun, and R. L. Calderon. 2000b. Serological evaluation of Cryptosporidium oocyst findings in the water supply for Sydney, Australia. Internatl. J. Environ. Health Research 10:35–40. Goldstein, S. T., D. D. Juranek, O. Ravenholt, A. W. Hightower, D. G. Martin, J. L. Mesnik, S. D. Griffiths, A. J. Bryant, R. R. Reich, and B. L. Herwaldt. 1996. Cryptosporidiosis: An outbreak associated with drinking water despite state-of-the-art water treatment [published erratum appears in Annals Intl. Med. 125(2):158]. Annals Int. Med. 124(5): 459–468. Greenland, S. 1988. Variance estimation for epidemiological effect estimates under misclassification. Statist. Med. 7:745–757. Greenland, S. 1994. Invited commentary: A critical look at some popular meta-analytic methods. Am. J. Epidemiol. 140:290–296. Greenland, S. and J. Robins. 1994a. Invited commentary: Ecologic studies—biases, misconceptions, and counter examples. Am. J. Epidemiol. 139:747–760. Greenland, S. and J. Robins. 1994b. Accepting the limits of ecologic studies. Am. J. Epidemiol. 139:769–771. Grimm, L. M., M. Goldoft, J. Kobayashi, J. H. Lewis, D. Alfi, A. M. Perdichizzi, P. I. Tarr, J. E. Ongarth, S. L. Moseley, and M. Samadpour. 1995. Molecular epidemiology of a fast-food restuarant-associated outbreak of Escherichia coli O157:H7 in Washington State. J. Clin. Microbiol. 33:2155–2158. Harter, L., F. Frost, and W. Jakubowski. 1982. Giardia prevalence among 1- to 3-year old children in two Washington State Counties. Am. J. Public Health 72:386–388. Hellard, M. E., M. I. Sinclair, A. B. Forbes, and C. K. Fairley. 2001. A randomized blinded controlled trial investigating the gastrointestinal health effects of drinking water quality. Environ. Health Perspect. 109:773–778.

REFERENCES

181

Hill, A. B. 1965. Environment and disease: Association or causation? Proc. Roy. Soc. Med. 58:295–300. Hulka, B. S., T. M. Wilcosky, and J. D. Griffith. 1990. Biological Markers in Epidemiology. Oxford, UK: Oxford Univ. Press. IAMFES. 1996. Procedures to Investigate Waterborne Illness, 2nd ed. Des Moines, IA.: International Association of Milk, Food, and Environmental Sanitarians, Inc. IARC. 1991. International Agency for Research on Cancer, Chlorinated Drinking-Water: Chlorination By-products; Some Other Halogenated Compounds; Cobalt and Cobalt Compounds, Vol. 52. Geneva, Switzerland: World Health Organization. Kent, G. P., J. R. Greenspan, J. L. Herndon, M. S. Lynne, M. Mofenson, J. S. Harris, R. Thomas, and H. A. Waskin. 1988. Epidemic giardiasis caused by a contaminated public water supply. Am. J. Public Health 78:139–143. Kleinbaum, D. G., L. L. Kupper, and H. Morgenstern. 1982. Epidemiologic Research: Principles and Quantitative Methods, Belmont, CA: Lifetime Learning Publications, pp. 403–418. Kurttio, P., E. Pukkala, H. Kahelin, A. Auvinen, and J. Pekkanen. 1999. Arsenic concentrations in well water and risk of bladder and kidney cancer in Finland. Environ. Health Perspect. 107(5):705–710. Last, J. M., ed. 1995. A Dictionary of Epidemiology, 3rd ed. Oxford, UK: Oxford Univ. Press. MacKenzie, W. R., N. J. Hoxie, M. E. Proctor, M. S. Gradus, K. A. Blair, D. E. Peterson, J. J. Kazmierczak, D. G. Addiss, K. R. Fox, J. B. Rose, and J. P. David. 1994. A massive outbreak in Milwaukee of cryptosporidium infection transmitted through the public water supply [published erratum appears in New England Journal of Medicine 331(15):1035 (Oct. 13, 1994)] (see comments). New Engl. J. Med. 331(3):161–167. Moe, C. L., M. D. Sobsey, P. W. Stewart, and D. Crawford-Brown. 1999. Estimating the risk of human calicivirus infection from drinking water. International Workshop on Human Caliciviruses, Atlanta, GA: CDC. Monson, R. R. 1980. Occupational Epidemiology. Boca Raton, FL.: CRC Press. Morris, R. D., E. N. Naumova, R. Levin, and R. L. Munasinghe. 1996. Temporal variation in drinking water turbidity and disguised gastroenteritis in Milwaukee. Am. J. Public Health 86(2):237–239. Morris, R. D., A. M. Audet, I. F. Angelillo, T. C. Chalmers, F. Mosteller. 1992. Chlorination, chlorination by-products and cancer: A meta-analysis. Am. J. Public Health 82:955. Muller T. B., G. F. Craun, and R. L. Calderon. 2001. Serological responses to Cryptosporidium infection. Infect. Immunol. 69(3):1974. Neugebauer, R. and S. Ng. 1990. Differential recall as a source of bias in epidemiological research. J. Clin. Epidemiol. 43(12):1337–1341. Okhuysen P. C., C. L. Chappel, C. R. Sterling, W. Jakubowski, and H. L. DuPont. 1998. Susceptibility and serological response of healthy adults to re-infection with Cryptosporidium parvum. Infect. Immunol. 66:441–443. Oregon Health Division. 1992. A Large Outbreak of Cryptosporidiosis in Jackson County, CD Summary, Portland, OR.: Oregon Health Division. Parsonnet, J., S. C. Trock, C. A. Bopp, C. J. Wood, D. G. Addiss, F. Alai, L. Gorelkin, N. Hargett-Bean, R. A. Gunn, and R. V. Tauxe. 1989. Chronic diarrhea associated with drinking untreated water. Annals Int. Med. 110:985–991.

182

EPIDEMIOLOGIC CONCEPTS

Payment, P., L. Richardson, J. Siemiatycki, R. Dewar, M. Edwards, and E. Frances. 1991. A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiological standards. Am. J. Public Health 81:703– 707. Payment P., J. Siemiatycki, L. Richardson, G. Renaud, E. Franco, and M. Prevost. 1997. A prospective epidemiological study of gastrointestinal health effects due to the consumption of drinking water. Internatl. J. Environ. Health Research 7:5–31. Piantadosi, S. 1994. Invited commentary: Ecologic biases. Am. J. Epidemiol. 139:71–64. Poole, C. 1994. Editorial: Ecologic analysis as outlook and method. Am. J. Public Health 84: 715–716. Poole, C. 1997. Analytical Meta-Analysis of Epidemiologic Studies of Chlorinated Drinking Water and Cancer: Quantitative Review and Re-analysis of the Work Published by Morris et al., Am. J. Public Health 82:955–963. Cincinnati: National Center for Environmental Assessment, USEPA. Rockett, I. R. H. 1994. Population and Health: An Introduction to Epidemiology, Washington, DC: Population Reference Bureau. Rothman, K. J. 1996. Modern Epidemiology, Boston, MA: Little, Brown. Schwartz, J. 1995. Editorial: Is carbon monoxide a risk factor for hospital admission for heart failure? Am. J. Public Health 85: 1343–1345. Schwartz, J. and R. Levin. 1999. Drinking water turbidity and health. Epidemiology 10:86–90. Schwartz, J., R. Levin, and R. Goldstien. 2000. Drinking water turbidity and gastrointestinal illness in the elderly of Philadelphia. J. Epidemiol. Community Health 54:45–51. Schwartz, J., R. Levin, and K. Hodge. 1997. Drinking water turbidity and pediatric hospital use for gastrointestinal illness in Philadelphia. Epidemiology 8:615–620. Schwartz, S. 1994. The fallacy of the ecological fallacy: The potential misuse of a concept and the consequences. Am. J. Public Health 84: 819–823. Snow, J. 1965. Snow on Cholera. New York: Hafner. Susser, M. 1994a. The logic in ecological: I. The logic of analysis. Am. J. Public Health 84: 825–829. Susser, M. 1994b. The logic in ecological: II. The logic of design. Am. J. Public Health 84:830–835. Talbott, E. and G. F. Craun. 1995. Introduction to Environmental Epidemiology, Boca Raton, FL: CRC Press. Waller, K., S. H. Swan, G. DeLorenze, and B. Hopkins. 1998. Trihalomethanes in drinking water and spontaneous abortion. Epidemiology 9:134–140. Ward, R. L., D. I. Bernstein, E. C. Young, J. R. Sherwood, D. R. Knowlton, and G. M. Schiff. 1986. Human rotavirus studies in volunteers: Determination of infectious dose and serological response to infection. J. Infect. Diseases 154:871–880. Wilkins, J. R. and G. W. Comstock. 1981. Source of drinking water at home and site-specific cancer incidence in Washington County, Maryland. Am. J. Epidemiol. 114:178. Zierler, S., R. A. Danley, and L. Feingold. 1986. Type of disinfectant in drinking water and patterns of mortality in Massachusetts. Environ. Health Perspect. 69:275. Zierler, S., L. Feingold, R. A. Danley, and G. Craun. 1988. Bladder cancer in Massachusetts related to chlorinated and chloraminated drinking water: A case–control study. Arch. Environ. Health 43:195–199.

8 APPLICATION OF RISK ASSESSMENTS IN CRAFTING DRINKING WATER REGULATIONS BRUCE A. MACLER, Ph.D. Toxicologist, U.S. Environmental Protection Agency, San Francisco, California

8.1

INTRODUCTION

Assessments of adverse health outcomes and estimations of health risks from contaminants of drinking water are used at several points in the overall development of U.S. drinking water regulations. They are applied during the extensive discussions surrounding consideration of a contaminant’s risk and approaches to risk mitigation, and in the required elements of National Primary Drinking Water Regulation (NPDWR) proposals, support documents, and final rules. Risk assessments are of several types with substantially different intents and characteristics. No one approach fits all applications. The type of assessment chosen depends on the application and the nature and availability of relevant data. The regulatory process starts with the identification of a constituent as either a possible contaminant of drinking water or as having some adverse toxicological or pathogenic properties. Available health data are then collected and analyzed, leading to the possible development of a toxicological profile of adverse health outcomes and dose–response characteristics. This is then coupled with occurrence data and Disclaimer : The views expressed in this chapter are those of the author and do not necessarily represent those of the USEPA. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

183

184

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

exposure estimates to provide an estimate of national public health impact. If a decision is made to move forward on a NPDWR, more refined assessments are then used to develop maximum contaminant level goals (MCLGs). Separate assessments are used to help identify possible maximum contaminant levels (MCLs) or treatment techniques consistent with public policies. In some cases, risk assessments may comprise regulatory elements themselves, such as sanitary surveys or comprehensive performance evaluations. Finally, risk assessments are necessary as formal components of regulatory supporting documents to help estimate risk reductions and quantify public health benefits. Some of these assessments, particularly those associated with public health policies and legal requirements, are highly constrained in form and content. Additionally, while the art and science of environmental risk assessment continue to evolve, we still have substantial uncertainty and imprecision in these estimates. Effective interpretation of these risk products requires some understanding of their purpose and form. This chapter details these applications and presents some examples.

8.2 RISK ASSESSMENT APPROACHES FOR DRINKING WATER REGULATIONS In risk management activities to ensure the safety of drinking water, assessments of health risk are performed to answer questions posed in the management process (Fig. 8.1). Examples include broad questions such as ‘‘What is the nature and magnitude of waterborne disease in the United States?’’—answers to which can help define where the safety of drinking water fits into the overall considerations of public health, or identify particular situations (e.g., undisinfected wells, crossconnections) that are associated with disease. More familiarly, risk assessments can address narrower questions such as ‘‘What is the risk that oral ingestion of hexavalent chromium will cause lung cancer?’’ or ‘‘What is the likelihood of Cryptosporidium illnesses from a turbidity spike in an unfiltered surface water system?’’ where the answers might help define regulatory actions. Specific questions, such as ‘‘What is the likely differential number of cancers prevented between a MCL for arsenic at 3 mg=L and one at 5 mg=L?’’ or ‘‘What are failure modes in the operation of an upflow clarifier?’’ may need to be answered to compare or determine possible solutions to the identified problems. Risk assessment approaches for drinking water health and regulatory questions fall into three main types: those based on epidemiologic data, those calculated using mathematical risk models, and those based on analysis of systems and components. Epidemiologic and risk model approaches have been described more completely in earlier chapters, so only their applicability to regulatory development is addressed here. For this purpose, it is important to remember that there are substantial and important limitations to these approaches and that the conclusions from such assessments must be considered with due caution by those using the information. In addition, management decisions are often required in the face of uncertainty and

8.2 RISK ASSESSMENT APPROACHES FOR DRINKING WATER REGULATIONS

185

Figure 8.1 Integration of data collection, risk assessment, and risk management [source: Farland (2000)].

lack of information. Risk assessments can provide useful information, but seldom give clear, unambiguous answers. With respect to epidemiologic data, information from waterborne disease outbreaks, intervention studies, or controlled experiments on humans can be used directly to both quantitatively and qualitatively describe risks. However, the precision and accuracy of epidemiologic data typically limit these assessments to situations where excess risks (risks above background levels) are about 1% or greater. More often, epidemiologic studies of large groups or populations can describe only effects greater than 10–100% or more. In general, causal associations between contaminants and effects are considered significant only when effects are severalfold higher than background levels. Therefore, risk assessments done directly from these data can generally describe risks and answer risk questions only to this level of resolution. Because public health questions for drinking water often involve situations where risks are substantially below epidemiologic resolution, this approach is frequently inappropriate. Perhaps the most familiar form of risk assessment is based on mathematical models used to extrapolate existing data to make quantitative estimates relevant to other situations. This well-known approach, first described by the National Academy of Sciences (NAS) in 1983 (NRC 1983), organizes the process of human health

186

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

Figure 8.2 NAS risk assessment paradigm (left circle) and risk management (right circle) [source: USEPA 1999c)].

risk assessment into four steps: hazard identification, dose–response assessment, exposure assessment, and risk characterization (Fig. 8.2). Hazard identification involves an evaluation of whether exposure to a substance would produce an adverse or otherwise undesirable effect. The data used to make such a determination usually come from animal studies. In some instances, human data may be available for the contaminant of interest. Dose–response assessment involves a more quantitative evaluation of the empirical evidence relating a specific exposure dose to the effect of interest. In particular, the available data are examined to determine the relationship between the magnitude of the exposure and the probability of the observed effect. Exposure assessments involve an evaluation of contaminant occurrence data, characterization of the environmental fate and transport of the contaminant from the source to the exposed population by different media (e.g., air, water, food), and physiological considerations of different exposure routes (e.g., ingestion, inhalation, dermal contact). These toxicity and exposure products are combined in the risk characterization. The risk characterization describes the overall nature and magnitude of risk posed to human populations exposed to a particular contaminant. Included in the description is a discussion of what is known and not known about the hazards posed by the substance, what models were used to quantify the risk and why they were selected, assumptions and uncertainties associated with the qualitative and quantitative aspects of the assessment, and general level of confidence in the assessment. A caution with these is that, while the calculated results are often presented as point value expressions of risk, it must be recognized that the farther the extrapolation from the original data, the more uncertainty and less precision in the results. The animal data themselves result from what are essentially small, highly controlled epidemiologic studies, and thus have the resolution limitations noted above. In addition, experimental variation may be 10% or greater. Because these studies are

8.2 RISK ASSESSMENT APPROACHES FOR DRINKING WATER REGULATIONS

187

most often conducted with high exposures to the contaminants of interest and generate high risks, models are used to extrapolate the data to answer questions about lower environmental exposures and=or risks. These models have inherent limitations that magnify uncertainties. Exposure estimates likewise have substantial variation and uncertainty. The net result is that quantitative estimates of risk cannot describe a defined point risk for a certain exposure, but instead a range of possible risks. These ranges tend to increase with increasing model complexity to frequently span orders of magnitude. Most often, the range includes zero. A third approach to risk assessment is based on analysis of an entire system or operation to identify vulnerabilities that could allow contaminants to reach the consumer. It is based on standard engineering design assessment approaches used to identify failure modes, judge probabilities of occurrence, and describe consequences. These have been adapted to focus on vulnerabilities in systems or operations that could result in human exposure to contaminants. This approach begins with a full description of the system or flow diagram of the process. Points where contamination can occur are identified. The likelihood and consequences of contamination at these points are described. From such an assessment, management actions can focus on controlling high-impact situations (Fig. 8.3). In the food

Figure 8.3 Hazard assessment critical control point (HACCP) risk assessment approach.

188

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

industry, hazard assessment critical control point (HACCP) programs have been put into place to protect the public from, among other things, contaminated shellfish and other food items. Sanitary surveys, source water vulnerability assessments, and comprehensive performance evaluations (CPEs) are examples of this type seen in the drinking water industry. This is also the approach used for water system security and counterterrorism assessments. As currently practiced, these are qualitative, rather than quantitative in nature. Once the system is described and vulnerabilities identified, probabilities and consequences are typically rated on the basis of best professional judgment and using categories such as ‘‘high, medium, or low’’ or ‘‘minor, significant, or catastrophic.’’ The end result may be a list of vulnerabilities with some rankings for risks and consequences. Managers can use this information to identify problem areas and prioritize activities. Although not commonly practiced, it should be noted that such system analysis approaches are open to quantification of risks (Buchanan and Whiting 1998).

8.3

RISK MANDATES FROM THE SAFE DRINKING WATER ACT

The Safe Drinking Water Act (SDWA), as amended in 1986 and 1996 [Title XIV, Sec. 1412(b)] has language that directs the U.S. Environmental Protection Agency (USEPA) to establish MCLGs for contaminants of public health concern for drinking water: ‘‘Each maximum contaminant level goal established under this subsection shall be set at the level at which no known or anticipated adverse effects on the health of persons occur and which allows an adequate margin of safety.’’ These goals, which are not enforceable themselves, are to be used to set the enforceable NPDWRs [Title XIV, Sec. 1412(b)]: Each national primary drinking water regulation for a contaminant for which a maximum contaminant level goal is established under this subsection shall specify a maximum contaminant level for such contaminant which is as close to the maximum contaminant level goal as is feasible . . . . For the purposes of this subsection, the term ‘‘feasible’’ means feasible with the use of the best technology, treatment techniques and other means which the Administrator finds, after examination for efficacy under field conditions and not solely under laboratory conditions, are available (taking cost into consideration).

In addition, a provision added to the Act in 1996 specifies priorities for selecting contaminants for rulemaking to take into consideration, among other factors of public health concern, the effect of such contaminants upon subgroups that comprise a meaningful portion of the general population (such as infants, children, pregnant women, the elderly, individuals with a history of serious illness, or other subpopulations) that are identifiable as being a greater risk of adverse health effects due to exposure to contaminants in drinking water than the general population.

8.4 DEVELOPING MCLS AND TREATMENT TECHNIQUES

189

In addition, the Act specifies that USEPA shall document health risks, opportunities for risk reductions, and benefits and costs of mandating these reductions and make these available for public comment. This health risk reduction and cost analysis (HRRCA) is a required component of the regulatory process. By this and other language, Congress described both the necessary characteristics of risk assessments for drinking water regulations and the broad principles for managing these risks. It can be seen that by their words, Congress established a precautionary policy with regard to drinking water safety. The MCLG was to be set conservatively with respect to risk to more vulnerable individuals. The MCL was to be set to reflect the MCLG, with the additional considerations for technical feasibilities and costs. These broadly stated goals set directions, but were not sufficiently described to be used for specific regulatory decisions. USEPA evolved operational interpretations of Congressional intentions for both the MCLG and for the acceptable public health risks associated with MCLs and treatment techniques following the 1986 amendments. These have been used consistently for NPDWRs from that time. Following the 1996 amendments, USEPA developed additional approaches for the benefit and cost analyses for the HRRCA.

8.4

DEVELOPING MCLS AND TREATMENT TECHNIQUES

The SDWA grants the USEPA Administrator the authority to publish a MCLG and promulgate a NPDWR for a contaminant if the contaminant may have an adverse effect on the health of persons, the contaminant is known or likely to occur in drinking water with a frequency or level of health concern, and there is a meaningful opportunity for health risk reduction. Development of a NPDWR normally begins with the identification of a drinking water contaminant. As provided in the SDWA, USEPA must list candidate contaminants for regulation on a periodic basis (USEPA 1998a). Following listing, more detailed health, occurrence, exposure, and treatment technology information is gathered. When adequate information becomes available, a determination is made on whether to go forward with development of a NPDWR proposal (USEPA 2002a). This determination uses a protocol developed and recommended to USEPA by the National Drinking Water Advisory Council (NDWAC). The health risk information is combined with occurrence data (levels, frequency, national distribution, persistence, etc.) and exposure estimates to predict the national number of individuals exposed above advisory levels. This risk assessment forms the basis for USEPA’s determination if regulation would provide a meaningful opportunity for health risk reduction. 8.4.1

Maximum Contaminant Level Goals

Once the decision is made to move forward, the MCLG is determined. MCLGs are risk assessment products developed by USEPA Office of Water and Office

190

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

of Research and Development. MCLGs are strictly health-based levels. They are developed and set at contaminant levels believed to be without appreciable health risk to individuals, to be consistent with the provision ‘‘set at the level at which no known or anticipated adverse effects on the health of persons occur.’’ Additionally, they must address the concern for the protection of sensitive subpopulations. Therefore, risk assessments used to develop MCLGs must at a minimum provide an estimate for a zero-risk exposure level for humans who may be more sensitive to the contaminant. While these risk assessments are caveated to be upper bounds for estimated risks, such that the true risks may be less or even zero, the assessments do not have to, nor are they designed to, estimate the full range of true risks to the average individual. MCLG risk assessments use the available toxicologic and epidemiologic health study data. The data are evaluated with respect to the nature of the adverse health effects from the contaminant, the strength of evidence for causal relationships, and their quality. Depending on the outcome, a dose–response estimate is made. As noted above, these data are almost always limited in quantity and quality, yielding substantial ranges for uncertainty. USEPA has chosen as a matter of policy to work to risks at the more conservative end of these ranges, in order to comply with the provision that the MCLG ‘‘allows an adequate margin of safety’’ in the face of these uncertainties. Therefore, poorer-quality data will lead to more stringent quantitative descriptions of risk. For chemicals that produce adverse health effects and are not considered to be carcinogenic (noncarcinogens), the MCLG is based on the reference dose (RfD), which is defined as an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. In practice, the RfD is set at a plausible zero-risk level. USEPA assumes that a physiological threshold exists for noncancer health effects from chemical contaminants, below which the effect will not occur. Thus the MCLG will be a nonzero number. Depending on the quality of the available toxicity data, the RfD is usually derived from an experimental no-observed-adverse-effect level (NOAEL), identified as the highest dose in the most relevant study that did not result in a known adverse effect. The adverse effects chosen may themselves be mild and without clinical significance, but typically represent early stages in progression to more serious disease. In order to extrapolate from the data to human exposures protective of sensitive subpopulations, the NOAEL is divided by various uncertainty factors to derive the RfD. These uncertainty factors conservatively account for the variation in human response to the contaminant, extrapolation to human responses if animal data were used, the nature of the studies, data quality, and relevance. The result of this is that the RfD may differ from the NOAEL by as little as a factor of 3 (e.g., nitrate, arsenic) or as much as 1000 (e.g., methyl bromide, chlorobenzene). The RfD takes the form of dose ingested per unit body weight per day (mg kg 1 day 1). RfDs that have been reviewed and agreed on by consensus within USEPA are listed in USEPA’s Integrated Risk Information System (IRIS).

8.4 DEVELOPING MCLS AND TREATMENT TECHNIQUES

191

The RfD, which is based on the total daily amount of contaminant taken up by a person on a body weight basis, is converted to a drinking water equivalent level (DWEL) concentration and adjusted for the percentage contribution of other sources [relative source contribution (RSC)] of the contaminant besides drinking water (air, food, etc.) to arrive at the MCLG. This calculation traditionally assumes a lifetime consumption of 2 L of drinking water per day by a 70-kg adult, which is about the upper 90th percentile consumption level. More recent USEPA regulatory risk assessments, such as those for arsenic, have considered different drinking water consumption rates in addition to this default value to represent specific populations (infants, agricultural workers) and situations (USEPA 2001a). A different approach is taken for contaminants that may be carcinogenic. USEPA assumes as a default position that no toxicity threshold exists for induction of cancer and thus, there is no absolutely safe level of exposure. Until relatively recently, once it was determined that a contaminant is a known or probable human carcinogen, the MCLG was automatically set at zero. USEPA has now revised its guidelines for cancer risk assessments to reflect the increasing understanding of the several steps in the progression of cancer (USEPA 1999b, 2001b). Some contaminants have been reevaluated for their carcinogenicity using the current draft version of these guidelines. A good example is chloroform (USEPA 1998a), which is now considered a carcinogen only at very high exposures associated with tissue damage, such that a threshold is indicated. This allows calculation of a nonzero MCLG using a margin of exposure (MoE) approach. As with RfDs, these determinations of carcinogenicity and their associated dose–response assessments that have been reviewed and agreed on by consensus within USEPA are listed in IRIS. An alternative approach is used for situations where the data on carcinogenicity of a contaminant are equivocal or too scanty to make a clear judgment. These contaminants are termed ‘‘possible human carcinogens.’’ For these, the MCLG may be derived from their relevant noncancer health effects as described above. The resulting RfD is divided by an additional uncertainty factor of 10 as a margin of safety for the possible carcinogenicity. In a vein similar to that for carcinogens, microbial pathogens and indicator organisms are also assigned a MCLG of zero as a matter of policy, from the consideration that one infective unit (oocyst, cyst, virus particle, bacterium) could be sufficient to cause an infection. The available data on infectivity are supportive for this assumption for the pathogenic viruses studied and for the protozoa, Giardia and Cryptosporidium. While it is less clear that this is so for pathogenic bacteria, the data cannot exclude this possibility. 8.4.2

Identifying Candidate MCLs

Once the MCLG is established, it is combined with information on contaminant occurrence, treatment technologies, and analytical methods to suggest and evaluate possible regulatory criteria for further discussion. Although the MCLG is a regulatory value, it is not enforceable; the NPDWR is the enforceable regulatory element.

192

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

One element in the standard setting process is to determine whether the NPDWR should be a MCL or a treatment technique. An enforceable MCL can be established to control exposure to a contaminant when appropriate analytical methods exist to quantify the contaminant and determine compliance at the MCL. When methods are not available, as for certain microbial exposure situations, treatment techniques may be established that do not directly measure exposure, but use other indicators for compliance. Two risk considerations come into play in the process of setting a MCL to reduce a contaminant as close as feasible to the MCLG. The first has to do with MCLGs of zero. It is impossible to quantify a contaminant or confirm treatment to a zero exposure. Therefore, MCLs must be above zero, and thus have some risk. A risk benchmark is used to identify appropriate safe drinking water exposures. These guide selection of analytical techniques, treatment approaches, and, ultimately, the MCL choices. This benchmark is based on a consideration of de minimus or ‘‘acceptable’’ risk. As a matter of policy, USEPA Office of Water has used an acceptable risk range for chemical carcinogens from one additional cancer per million people to one additional cancer per 10,000 people exposed to the contaminant over a lifetime (USEPA 1998d, 2001a). For pathogenic microorganisms, an acceptable risk of one additional infection per 10,000 people exposed per year has been used (USEPA 1989a, 1998c). These allowable exposures are estimated from the associated dose–response curves. For carcinogens, this dose–response assessment uses models to extrapolate from the available data to zero exposure–zero risk, defining a curve that is essentially linear at low exposures. The resulting ‘‘cancer slope factor’’ allows for a convenient probability analysis of risks to individuals associated with different exposures. The exposures for the acceptable risk range are taken directly from the curve. However, as noted above, the uncertainties increase substantially in these extrapolations. USEPA traditionally uses the 90th percentile upper bound of the modeled results to minimize the possibility that risks in this exposure range are not greater than estimated. The end result of this is that the risk from a lifetime consumption of water at a given level is unlikely to be greater than estimated, is more likely substantially less, and may be zero. A similar probabilistic risk approach is used for microbial contaminants. The dose–response models used for microbial risk assessments are somewhat more complex in that they must be selected to account for the particulate nature of the infective material in the environment. The modeled exposures are then used for further risk management. For contaminants with noncancer health risks, the MCLG is used as the starting point for determining a MCL. Because this uses a ‘‘bright line’’ reference point (the DWEL), exposures need only be estimated and compared to the MCLG. However, since this is nonprobabilistic, alternative MCLs cannot be considered on the basis of risks. For most noncarcinogens, the MCL is set equal to the MCLG. The estimated risks associated with different exposures are then matched against the existing environmental exposure levels to determine the magnitude of the public health problem to be solved. As discussed in other chapters, the additional con-

8.4 DEVELOPING MCLS AND TREATMENT TECHNIQUES

193

siderations for treatment technologies and their feasibility and the availability of appropriate analytical methods are factored in at this point. 8.4.3

Health Risk Reduction and Cost Analysis

The second application of risk assessment in the development of a MCL or treatment technique is in the estimation of public health benefits to be gained by regulation. This is used in the management discussions leading up to a regulatory proposal. Under the SDWA of 1986, it was necessary to consider only analytical and treatment feasibility in establishing the MCL. However, a cost benefit assessment was produced as an element of the regulatory impact analysis (RIA). The 1996 revisions to the SDWA required USEPA to explicitly consider costs and benefits in determining the MCL. Therefore, an expanded HRRCA is now required as part of a regulatory proposal. To estimate benefits for different MCLs or treatment requirements, the risks to an individual at the resulting exposure levels must be multiplied by estimates of the number of individuals in the United States exposed to the different levels. In practice, a relationship between the number of individuals exposed versus exposure level is first produced, then further manipulated to account for existing and proposed treatment controls. This relationship may be calculated stepwise or by using Monte Carlo simulations based on exposure distributions. From this, the number of cancers or microbial illnesses avoided at a given regulatory level can be estimated. These ‘‘body counts’’ can be matched with information on the costs of treating the associated diseased and the dollar value of avoiding illness to give quantitative information on the monetary benefits of the regulation. These benefits are then matched against the implementation and compliance costs to utilities and oversight agencies (USEPA 1999a). From a risk assessment perspective, it must be remembered that carcinogen dose–response curves represent upper-bound risks; thus estimates of the number of cancers based on these curves are also upperbound values for any given exposure level. This approach is most useful for benefits from reducing cancer or microbial illness risks. This is both because these disease endpoints are recognizable and definitive and because the impacts are quantifiable from their probabilities. This is not so for noncancer risks from chemicals. Because these are described by nonprobabilistic, zero-risk DWELs associated with subclinical health effects of indeterminate public health importance, it is difficult to quantify or assign monetary value to the benefits of reducing exposures to these chemicals. 8.4.4

Risk Assessments as Regulations

Risk assessments can comprise regulatory elements themselves. These may apply directly to utilities or secondarily through requirements on primacy agencies. The type of risk assessments currently required in NPDWRs are all qualitative system analyses. These include treatment system sanitary surveys found in the Total Coliform Rule (USEPA 1989b) and Interim Enhanced Surface Water Treatment Rule (IESWTR) (USEPA 1998b), the watershed sanitary surveys in the Surface

194

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

TABLE 8.1 Australian Framework for Management of Drinking Water Quality First element Commitment to drinking water quality management System analysis and management Assessment of the drinking water supply system Planning preventive strategies for water quality management Implementation of operational procedures and process control Verification of drinking water quality Incident and emergency response Supporting requirements Employee awareness and training Community involvement and awareness Research and development Documentation and reporting Review Evaluation and audit Review and continual improvement

Water Treatment Rule (SWTR) (USEPA 1989a), and the CPEs, also in the IESWTR. These all require on-site evaluations to determine sources of contamination and vulnerabilities to failures that could compromise water quality. These analyses are similar to the HACCP process used by the National Aeronautics and Space Administration (NASA) and U.S. Department of Agriculture (USDA) to protect foods. Australia has taken a more formal HACCP approach to protect their drinking water. Termed the ‘‘framework for management of drinking water quality,’’ it is to be incorporated into the Australian Drinking Water Guidelines (Australian Department of Health and Ageing 2001). The approach includes a systematic and comprehensive analysis of the entire route of water from source to tap (Table 8.1). This analysis leads to identification of hazards, sources of hazards, and associated risks. These are addressed in the subsequent institution of protective measures to yield multiple barriers to contamination. The emphasis of the HACCP approach is on proactive protection, rather than reactive responses. 8.4.5

Regulatory Reviews of NPDWRs

The 1996 SDWA Amendments required that USEPA review all existing NPDWRs every 6 years to determine if information available subsequent to promulgation would support regulatory revision. Congress stipulated that all such revisions must maintain, or provide for greater, protection of the health of persons. USEPA, working with the NDWAC, developed a protocol for these reviews, driven largely by reevaluations of health risk information (USEPA 2002a). USEPA principally considered whether any new evaluation of oral ingestion risks could lead to revision

REFERENCES

195

of the MCLG. This approach allows the Agency to use up-to-date risk assessment approaches in its reevaluations, which will include a wider range of possible adverse health outcomes (reproductive and developmental) and better characterizations of carcinogenicity. The potential here is that some carcinogens would be reclassified in such a way that their MCLs could be relaxed while maintaining the same level of health protection to the public. The reclassification of chloroform as a threshold carcinogen is an example. This would recognize that improved risk assessments could reduce the scientific uncertainties that led to excessively stringent NPDWRs.

8.5

FUTURE OUTLOOK

Risk assessments are used both formally and informally at several points within the regulatory process. They serve specific purposes and mandates and may have severe constraints on their representations of risk. While assessments may be qualitative or quantitative in nature, they are always limited by the available information and our current abilities to understand disease processes and predict outcomes. They will always be inexact. Recognizing and accepting their limitations is important for making sound regulatory decisions. An increased awareness of how risk assessments may be used will improve future regulatory discussions.

ACKNOWLEDGMENTS The author wishes to acknowledge the patient help of Fred Pontius, without which this chapter would have been the poorer.

REFERENCES Australian Department of Health and Ageing. 2001. Framework for Management of Drinking Water Quality: A Preventative Strategy from Catchment to Consumer. Canberra, Australia: National Health and Medical Research Council. Buchanan, R.L. and Whiting, R.C. 1998. Risk Assessment: A means for linking HACCP plans and public health. J. Food Protect. 61:11:1531–1534. Farland, W. H. 2000. Current and Proposed Approaches to Assessing Children’s Cancer Risk. USEPA=National Institute of Environmental Health Sciences Workshop: Information Needs to Address Children’s Cancer Risk. March 30 and 31. Washington, DC: USEPA Office of Research and Development. National Academy of Sciences. 1983. Risk Assessment in the Federal Government: Managing the Process. National Academy Press, Washington, DC. Title XIV—Safety of Public Water Systems; Sec. 1412(b), National Drinking Water Regulations: Standards; 42 USC Sec. 300g.

196

RISK ASSESSMENTS IN DRINKING WATER REGULATIONS

USEPA. 1986. The Risk Assessment Guidelines of 1986. EPA=600=8-87=045. Washington, DC: USEPA. USEPA. 1989a. Drinking Water; National Primary Drinking Water Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionella, and Heterotrophic Bacteria; Final Rule. Fed. Reg. 54:27486–27541. USEPA. 1989b. Drinking Water; National Primary Drinking Water Regulations; Total Coliforms (including fecal coliforms and E. coli); Final Rule. Fed. Reg. 54:27544–27568. USEPA. 1998a. Announcement of the Drinking Water Candidate List; Notice. Fed. Reg. 63:10273–10287. USEPA. 1998b. National Primary Drinking Water Regulations: Disinfectants and Disinfection Byproducts Notice of Data Availability; Proposed Rule. Fed. Reg. 63:15606–15692. USEPA. 1998c. National Primary Drinking Water Regulations: Interim Enhanced Surface Water Treatment; Final Rule. Fed. Reg. 63:69478–69521. USEPA. 1998d. National Primary Drinking Water Regulations: Disinfectants and Disinfection Byproducts; Final Rule. Fed. Reg. 63:69390–69476. USEPA. 1999a. Health Risk Reduction and Cost Analysis for Radon in Drinking Water: Notice. Fed. Reg., 64:9560–9599. USEPA. 1999b. Draft Guidelines for Carcinogen Risk Assessment. NCEA-F-0644. Washington, DC: USEPA. USEPA. 1999c. Research and Development: Fiscal Years 1997–1998 Research Accomplishments. Washington, DC: USEPA Office of Research and Development. USEPA. 2001a. National Primary Drinking Water Regulations; Arsenic and Clarifications to Compliance and New Source Contaminants Monitoring; Final Rule. Fed. Reg. 66:6976–7066. USEPA. 2001b. Notice of Opportunity to Provide Additional Information and Comment. Draft Revised Guidelines for Carcinogen Risk Assessment. Fed. Reg. 66:59593–59594. USEPA. 2002a. National Primary Drinking Water Regulations; Announcement of the Results of EPA’s Review of Existing Drinking Water Standards and Request for Public Comment. Fed. Reg. 67:19029–19090. USEPA. 2002b. Announcement of Preliminary Regulatory Determination of Priority Contaminants on the Drinking Water Contaminant Candidate List. Fed. Reg. 67: 38222–38244.

9 ‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

9.1

INTRODUCTION

Since the earliest days of public water supply, science and scientific advances have formed the foundation regulatory agencies use to define safe drinking water. More recently, the adequacy or ‘‘soundness’’ of scientific and technical studies for regulatory decisionmaking and policy has received increasing attention. During the regulatory process, stakeholders frequently call into question the adequacy of the science that underlie proposed U.S. Environmental Protection Agency (USEPA) regulations and regulatory policy decisions. Traditionally, science has been granted a special status within developed Western society. In 1938, Albert Einstein protested plans to require Italy’s intelligentsia to pledge loyalty to the fascist regime. He wrote (Einstein 1990) the pursuit of scientific truth, detached from the practical interests of everyday life, ought to be treated as sacred by every Government, and it is in the highest interests of all that honest servants of truth should be left in peace.

But far from ‘‘being left in peace,’’ science and scientists are often engulfed in the controversy that can surround a regulatory decision or agency rulemaking, as stakeholders contend with an agency and with each other over whether a particular study represents ‘‘sound’’ science. Indeed, even defining what science is has been Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

197

198

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

difficult as there is no universally applicable definition that could be applied to all fields of inquiry (Lastrucci 1963). Laypeople, scholars, and scientists themselves define the term in varying ways and employ it in a variety of contexts. Indeed, defining what science is not is more easily done than defining what science is (Taylor 1996). Nevertheless, the attributes of ‘‘sound’’ science begin to emerge as past definitions various writers have put forth for science itself are considered. Pickering (1992) construes science simply as ‘‘a way of being in, getting on with, making sense of, and finding out about the world.’’ Ackoff et al. (1962) states that Science may be considered a process of inquiry; that is, as a procedure for answering questions, solving problems, and developing more effective procedures for answering questions and solving problems. Science is also frequently taken to mean a body of knowledge.

Lastrucci (1963) proposes that ‘‘Science may be defined quite accurately and functionally as: an objective, logical, and systematic method of analysis of phenomena, devised to permit the accumulation of reliable knowledge.’’ More recently, Shermer (1997) defined science as ‘‘a set of methods designed to describe and interpret observed or inferred phenomena, past or present, and aimed at building a testable body of knowledge open to rejection or confirmation.’’ Today, the content of science is often confused with the methods of science. Much of the content of science is changing. What may be scientific (i.e., accepted as true, or ‘‘sound’’) today may become unscientific (i.e., regarded as untrue, or ‘‘unsound’’) in the future. In addition, the demarcation between science and nonscience is not always clear. In practice, it is not a clearly visible line but an area that is both shifting and subject to debate—depending on the authorities one accepts. With this backdrop, ‘‘sound’’ science might be defined as reliable knowledge obtained and accumulated by objective, logical, and accepted systematic methods of analysis appropriate for the phenomena, subject, or field of study. But even with such a simple definition, agreement regarding ‘‘soundness’’ of the science behind a specific study, regulation, or regulatory policy may still be elusive. The purpose of this chapter is to examine the dynamics behind these disagreements and discuss principles to guide discussion of scientific issues within the context of the Safe Drinking Water Act (SDWA).

9.2

ELEMENTS OF ‘‘SOUND’’ SCIENCE

‘‘Sound’’ science in drinking water regulation is no different from good science in general. For each field of scientific inquiry, textbooks and professional reference books are usually available addressing topics such as experimental design, procedures, data analysis, research methodologies, and other elements necessary for a study to be considered ‘‘sound’’ within that field. In practice, however, no study is perfect and techniques used in every study could be improved on. Clearly, a study that failed to meet minimal criteria for a key element—say, for example, too few subjects with a given exposure level, or some obvious fatal flaw in experimental design— would not be considered ‘‘sound.’’ But even when no fatal flaw is evident, disagree-

9.2 ELEMENTS OF ‘‘SOUND’’ SCIENCE

199

ments regarding the soundness of a study may stem from issues regarding lack of objectivity, poor reasoning, ambiguity, and=or an absence of fair-mindedness. 9.2.1

Objectivity

Objectivity in science refers to attitudes devoid of personal whim, bias, or prejudice. It focuses on methods for ascertaining publicly demonstrable qualities of the phenomenon or subject being studied (Lastrucci 1963). Evidence in science is to be factual, not conjectural, although at times a scientist must make attempts to extend understanding beyond known facts by proposing theories, offering opinions, and making educated guesses. Even so, demonstration of evidential proof is needed to achieve acceptable scientific truth. Science is a subjective enterprise insofar as it is practiced by imperfect people. In reality, much of research is trial and error, depending on factors other than scientific laws and method (Kneller 1978). The scientist usually does not spend much time thinking about scientific laws or principles. Other things such as trying to get an experimental apparatus to work, finding a way of measuring something more accurately, or keeping a treatment system operating, occupy the time. At times a scientist may hardly know what he or she is trying to prove. By its very nature, research is typically a continual feeling out into the dark. German physicist Max Planck (1932) once observed: Anybody who has been seriously engaged in scientific work of any kind realizes that over the entrance to the gates of the temple of science are written the words: Ye must have faith. It is a quality which the scientist cannot dispense with.

When pressed to say what is being done, a scientist may present a picture of uncertainty or doubt, or even confusion. But the scientific method (discussed below) encourages a rigorous, impersonal mode of procedure dictated by the demands of logic and objective procedures. Indeed, authority in science is achieved by the accumulation of publicly ascertainable evidence supporting an argument. It should not be a consequence of mere opinion (no matter how strongly felt), not faith alone (although an element of faith may be involved), nor of mere assumed verity (which would make it a presupposition). The ideal of objectivity, in effect, recognizes that ‘‘sound’’ science must be publicly verified according to the consensus of suitably trained observers. Should a study have a bias or limitation unknown to the scientist, it will likely be uncovered by objective peer observers. 9.2.2

Reason and Truth Claims

Scientific reasoning may take several forms. Two of the most common are deductive and inductive, discussed below. Kneller (1978) identified the following forms of scientific reasoning:  Retroduction—a scientist encounters an anomaly and then seeks a hypothesis from which the existence of the anomally can be deduced. She reasons backward from the anomally to a hypothesis that will explain it.

200

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

 Hypotheticodeduction—rather than reason to a hypothesis from data, the scientist begins with a hypothesis and deduces conclusions, general statements, or particular predictions from it.  Inductive reasoning—used to infer a general regularity from statements of particular instances.  Analogical reasoning—employed when the scientist arrives at a hypothesis by seeing an analogy between apparently unrelated phenomena. The pursuit of sound science is ultimately a quest for truth within a scientific context. Zeno (ca. 475 B.C.) made one of the earliest attempts to define truth, by reducing alternative positions to the absurd, known as the reductio ad absurdum argument. No position or statement that generates contradictions can be considered true. This law of noncontradiction is one of the fundamental principles of logical thought. At best, however, it is a negative test for truth, demonstrating that some positions are false, but failing to determine which ones are true. Deductive Reasoning Aristotle (384–322 B.C.) was the first Western philosopher to elaborate rules for deductive reasoning, although he also accepted inductive reasoning, discussed below. Deductive reasoning is simply arguing from the general to the particular, using formal rules of logic. For example, consider the premise ‘‘drinking water meeting USEPA regulations is safe to drink’’ (the general). And the premise ‘‘the drinking water of town x meets USEPA regulations’’ (the particular). Then it follows that ‘‘town x’s drinking water is safe to drink’’ (the conclusion). Deductive reasoning involves formal rules of logic in a series of propositions, called a ‘‘syllogism.’’ A valid syllogism has a particular structural form, the subject of introductory textbooks on philosophy and logic. But not all valid syllogisms are true. The principal limitation of deductive reasoning is that, for a syllogism to be considered ‘‘sound,’’ the premises must be true. Often in scientific and regulatory policy debates the premises underlying the arguments presented are actually assertions or presumptions based on unstated assumptions, the truth of which are assumed but unknown. In the example above, if town x’s water also contained 1000 mg=L of perchlorate, a contaminant not regulated by USEPA as of this writing (2002), the soundness of this argument would be in question. The Inductive Method Francis Bacon (1561–1626) is credited with advancing the inductive method of discovering scientific truth. He formulated the basic rules of induction, which became the forerunner of Cannons of Inductive Logic, by John Stuart Mill (1806–1873). Mill’s inductive method is summarized by these rules: 1. The method of agreement—the one factor common to all antecedent situations where an effect occurs is probably the cause of the effect.

9.2 ELEMENTS OF ‘‘SOUND’’ SCIENCE

201

2. The method of difference—whenever an effect occurs when A is present but not when it is absent, then A is probably the cause of the effect. 3. The joint method—the first two methods are combined when one method alone does not yield a definite result. 4. The method of concomitant variations—when an antecedent factor varies concomitantly with a consequent factor, then the former is probably the cause of the latter. The principal limitation of the inductive method is that one cannot be absolutely sure of the conclusion without complete or universal observation or knowledge, which is impossible to achieve. Nevertheless, the inductive method has played an important role in the development of public water supply. For example, in one of the first epidemiologic investigations, Dr. John Snow plotted the location of deaths from cholera on a map of central London in 1854 (see Chapter 1). The area’s 11 water pumps were also located on the map. Snow observed a correlation: that cholera occurred almost entirely among those who lived near and drank from the Broad Street water pump. He believed that the disease was caused by an unidentified microorganism in the drinking water contaminated by fecal material from diseased persons. This was a radical belief in that day (an assumption for Snow), and Snow had to argue his findings and theory before the local authorities. He was persuasive and the handle of the pump was removed, ending the neighborhood epidemic that had taken more than 500 lives. But it was not until almost thirty years later, in 1883, that Koch finally isolated Vibrio cholerae as the organism responsible for causing this disease. The Scientific Method In essence, the scientific method involves both deductive and inductive reasoning. Truth claims are determined by appeal to factual evidence, based on observation and experience. Beliefs are treated as hypotheses subject to repeat testing and open to public confirmation or refutation. Hypotheses must be tested by deducing its implications in the form of predictions and comparing them with the results of observations or experiments. The ideal hypothesis is precise and testable, accounts for the known facts, and predicts at least one new fact (Kneller 1978). In general, the scientific method involves four basic steps: 1. Formulating a statement carefully, clearly 2. Predicting the implications of such a belief 3. Performing controlled experiments to confirm or refute these implications and observing the consequences 4. Accepting, rejecting, or modifying the statement as a result The principal limitation of the scientific method is that experimentation is not always possible. Also, it is not always possible to formulate a problem in a clear and concise way that could be subjected to experimentation and observation.

202

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

Observations do not interpret themselves, but are interpreted by a human mind, through the observer’s biases, or even by where the research dollars originated. However, a number of techniques can be used to limit the influence of bias in research (Wildavsky 1995):  Experimental Design. Experiments should be structured in advance to limit or eliminate bias to the degree possible. The experimental design should be such that the number of factors that could affect the outcome are limited and should include appropriate controls.  Replication. Scientists should disclose not only what was found in a research study but also how it was found. Experiments should be described in detail, along with a description of the controls used. If other scientists can replicate the results, it provides further evidence that the results are real and objective.  Peer Review. Scientists submit findings to scientific journals, which are reviewed by peers for possible errors or leaps in logic. The methods of experimentation should be reviewed to determine whether the experimental designs were valid and show what the author claims. Only when scientists from different areas agree that the research is valid should it be published. Properly conducted, peer review minimizes bias, and over time accumulated data tend to overthrow erroneous theories and expose fraud. Peer review is one form of ‘‘peer involvement,’’ discussed below.  Falsifiability. Well-conceived hypotheses are falsifiable. Within the constraints of the scientific method, a claim could never be considered true unless it could be proved or disproved. Indeed, a distinguishing feature of a ‘‘sound’’ science is the formation and testing of a hypothesis. Davies (1992) emphasizes the importance of hypothesis testing: A powerful theory is one that is highly vulnerable to falsification, and so it can be tested in many detailed and specific ways. If the theory passes those tests, our confidence in the theory is reinforced. A theory that is too vague or general, or makes predictions concerning only circumstances beyond our ability to test, is of little value.

Postmodern Science The modern worldview dominating Western culture for most of the twentieth century according to the presupposition that the sciences are rational. Logic is used to outline problems, interpret observations, and formulate and study hypotheses. The postmodern worldview, first articulated in 1979 by Lyotard (1997), became increasingly influential and pervasive in Western culture, especially since the early 1990s. Postmodernism argues that there is no fixed vantage point beyond our own structuring of the world. Hence, there is no thing as objective reality (i.e., something that is true regardless of whether someone believed it or not). Objective reason is considered a myth to the postmodernist. Objective understanding and the power of reason are rejected, making all scientific truth relative. What is thought to be scientific knowledge, what is considered to be a firm grasp of

9.2 ELEMENTS OF ‘‘SOUND’’ SCIENCE

203

truth and reality, is viewed only as a subjective opinion. Hence, a scientific statement can be true for one person and not true for someone else. Scientific truth becomes whatever anyone decides to believe regardless of external evidence, and the common ground necessary of meaningful exchange of ideas is eroded. The power of persuasion is diminished and assertion of power over others to believe in a particular way or to control their behavior is the natural next step. A key premise of scientific relativism is that one theory is simply replaced by another because of paradigm shifts, but neither is closer to any objective truth. The scientific method is undermined or even irrelevant, as objective truth is believed not to exist, even in what is known as the ‘‘hard’’ sciences (chemistry, biology, etc.). Proponents of one scientific theory develop their own language and scientific viewpoint, with no common language between proponents of other scientific theories. What is true or rational in one scientific branch is not necessarily so in another. For example, a concept considered true by an epidemiologist would not necessarily be considered true by a toxicologist or engineer. Furthermore, no common ground on which to exchange ideas would be thought to exist between these disciplines. Each discipline develops its own worldview and language apart from the other. Hence, nothing is true and false in an absolute sense, nor is there a need for discussion between scientists or disciplines to integrate differing disciplines into a unified scientific theory or construct to describe objective reality. Something is considered true scientifically simply because it is believed, regardless of whether it is antithetical to other scientific or observable ‘‘fact.’’ An important danger in scientific relativism is that, once the ties between belief and objective reality become severed, irrational beliefs in any discipline may become firmly entrenched and difficult to change. Wynn and Wiggins (2001) note that Once people acquire a belief, they tend to adhere to that belief, even in the face of contradictory evidence. Explanations developed to explain phenomena become fixed, even when those explanations are shown to be irrational or based on wrong evidence.

The disconnections between disciplines and the divergence of views regarding the role of rational thought described above can be the source of extreme conflict over the soundness of a scientific study. Indeed, inherent in the postmodern worldview and scientific relativism is a serous contradiction—observation and logical inference are used to argue that observation and logical inference tell us nothing. By using the tools of science to argue its views, the postmodern view demonstrates the belief that these tools actually work, rather than that these tools are worthless. Arguing that scientific relativism is universally true is internally inconsistent and contradictory. The postmodern view also has serious ramifications for public water supply. Consider the waterborne disease outbreak in Walkerton, Ontario, in which there were 11 deaths and more than 1000 people infected in a population of 5000 (O’Conner 2002). Sincere citizens believed that the water was safe to drink, which governed their behavior to drink the water in the first place, but did not invalidate the objective reality external to the victims that disease-causing microorganisms were present in their drinking water.

204

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

Scientific relativism within the postmodern view undermines the very concept of a universally applicable ‘‘sound’’ science. Although imperfect, science attains the goal of discovering, to the degree possible, an objective, rational truth much of the time. Reflecting on this nature of science, Ziman (2000) concludes To put it simply . . . scientists still formulate and try to solve practical and conceptual problems on the basis of their shared belief in an intelligibly regular, not disjointed, world outside themselves. They still go on theorizing, and testing their theories by observation and experiment. They still try as best they can to eliminate personal bias from their own findings and are extremely canny in their acceptance of the claims of others. To that extent at least, we, the public at large, have just as good grounds as we ever did for believing (or doubting) the amazing things that science tells us about the world in which we live.

9.2.3

Clarity

Lack of clarity can originate from areas already discussed: lack of objectivity and poor reasoning. However, it can also originate from doublespeak, unclear definitions, and unstated presumptions. ‘‘Doublespeak’’ is language that pretends to communicate but really does not. Basic to doublespeak is incongruity between what is said or left unsaid, and what really is (Lutz 1989): It is the incongruity between the word and the referent, between seem and be, between essential function of language—communication—and what doublespeak does— mislead, distort, deceive, inflate, circumvent, obfuscate.

There are at least four kinds of doublespeak:  Euphemism—an inoffensive or positive word or phrase used to avoid a harsh, unpleasant, or distasteful reality. Not all euphemisms are doublespeak, but become so when used to mislead or deceive.  Jargon—the specialized language of a trade, profession, or similar group. Within a group, jargon functions as a verbal short hand allowing members of the group to communicate with each other clearly, efficiently, and quickly. But jargon can also be doublespeak when it is pretentious, obscure, and esoteric terminology used to project an air of profundity, authority, and prestige. Jargon doublespeak makes the simple appear complex, the ordinary profound, the obvious insightful. When a member of a group uses its jargon to communicate with a person outside the group, and uses it knowing that the nonmember does not understand the language, it is doublespeak.  Gobbledygook or bureaucratese—overwhelming the hearer with words, the bigger the words and the longer the sentences, the better. Gobbledygook sounds impressive, but when examined in print, usually does not make sense.

9.2 ELEMENTS OF ‘‘SOUND’’ SCIENCE

205

 Inflated language—language designed to make the ordinary seem extraordinary, to make everyday things seem impressive, to give an air of importance to people, situations, or things that would not normally be considered important, to make the simple seem complex. Ambiguous definitions can also result in confusion and, in some cases, doublespeak. Many words have more than one definition. Meaning can be hidden or confused when definitions of words change in a statement or study, or have inserted presuppositions. Lastrucci (1963) notes that a good definition can be neither true nor false. It is not a factual position. A definition is simply an explicit declarative statement or resolution; it is a contention or an agreement that a given term will refer to a specific object. It cannot be logically tested for ‘‘truth.’’ Its ‘‘truth’’ is established by declaration; it is what the definer says it is. Lack of clarity results when a scientist or regulator interjects a presumption within a definition or statement, or changes the meaning of a word in midsentence. For example, how should the term ‘‘safe drinking water’’ be defined? An operational definition might be ‘‘Safe drinking water is water that meets current drinking water regulations.’’ But this statement contains a presumption that current drinking water regulations are sufficient to protect public health. We might then define ‘‘safe drinking water’’ as water that poses no significant risk to public health. What is ‘‘significant’’? How do we define ‘‘public’’? What about sensitive people within the general population? And what about contaminants that are not regulated? Hence, a definition of ‘‘safe drinking water’’ could mean many things. Indeed, suppose a customer calls their water company asking ‘‘Is my water safe to drink? And the water company representative responds by saying ‘‘Yes, it meets all current regulations.’’ Both customer and water company representative may likely feel satisfied, but should they be? All rational disciplines, including the scientific method, take certain things for granted. These beginning points are called presuppositions, things considered true even though they cannot be proved within that particular discipline, or proved at all. All knowledge has certain beginning points that are simply to be accepted, and that are impervious to scientific methods. Indeed, differences regarding whether a particular study constitutes sound science can arise simply because presuppositions are unstated, or because investigators view the study through a different perspective based on differing presuppositions and basic assumptions. 9.2.4

Critical Thinking

A prerequisite to being able to fairly evaluate scientific studies is the development of good thinking habits, or, in short, learning to think critically. In this context, critical thinking is the disciplined art of ensuring that the best thinking one is capable of is used to evaluate the scientific study in question. Paul and Elder (2001) note that a well-disciplined critical thinker  Raises vital questions and problems, formulating them clearly and precisely  Gathers and assesses relevant information, and can effectively interpret it

206

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

 Comes to well-reasoned conclusions and solutions, testing them against relevant criteria and standards  Thinks open-mindedly within alternative systems of thought, recognizing and assessing, as need be, their assumptions, implications, and practical consequences  Communicates effectively with others in figuring out solutions to complex problems A basic component of critical thinking is fair-mindedness. To be fair-minded is to treat every viewpoint relevant to a study in an unbiased, unprejudiced way. By nature, humans tend to prejudge the views of others, placing them into ‘‘favorable’’ (agrees with us) and ‘‘unfavorable’’ (disagrees with us) categories. Less weight tends to be given to contrary views, especially when political motives are involved. When recognizing mistakes in reasoning, commonly called ‘‘fallacies’’ (Table 9.1), most people will tend to see the mistakes in the reasoning they already disapprove of rather than in their own reasoning. Although they can develop some proficiency in making their opponents thinking look bad, subjecting their own thinking to similar scrutiny might yield similar limitations. Fair-mindedness entails treating all viewpoints alike (evaluating all views or aspects of a study according to the same criteria), uninfluenced by a particular person or group’s advantage. This does not mean that all viewpoints are equally true or equally acceptable, but it does mean that they are evaluated fairly with the same criteria and thinking processes before acceptance, rejection, or modification.

9.3

PEER INVOLVEMENT

The quality and credibility of scientific studies forming the basis of regulatory policies can be enhanced through peer involvement. Peer involvement is generally considered to be the process whereby an investigator or agency staff involve subject matter experts from outside their own program in one or more aspects of a study or development of a work product. It is an active outreach to and participation by the broad scientific, engineering, and economics communities beyond the investigator’s organization (external) as well as within the investigator’s organization (internal). In general, peer involvement takes two forms: peer input and peer review. Peer input, sometimes called peer consultation, generally refers to an interaction during development of a study, policy, or work product, providing an open exchange of data, insights, and ideas. It may be characterized by a continual iterative interaction with scientific experts during the study or work product development. Peer input provides contributions to the development of the study or work, but does not substitute for peer review, which is more rigorous. There is no written definition of peer review that applies across the federal government (USGAO 1999). In general, the goal of peer review is to obtain an independent, third-party review of the product from experts who have not substan-

9.3 PEER INVOLVEMENT

TABLE 9.1

207

Directory of Fallacies

Ad hominum: attacking the person rather than his or her arguments Ad ignorantiam (appeal to ignorance): arguing that a claim is true just because it has not been shown to be false; arguing from incomplete information Ad misericordiam (appeal to pity): appealing to pity as an argument for special treatment Ad populum: appealing to the emotions of a crowd; appealing to a person to ‘‘go along’’ with the crowd, without presenting any reasons to show that the crowd is an informed or impartial source Affirming the consequent: a deductive fallacy of the following form: If p then q q Therefore, p Both premises could be true and the conclusion still false; overlooks alternative explanations Begging the question: implicitly using your conclusion as a premise; assuming the issue in question in your response Complex question: posing a question of issue in such a way that one cannot agree or disagree with you without committing oneself to some other claim you wish to promote Composition: assuming that a whole must have the properties of its parts; opposite of ‘‘division’’ Denying the antecedent: a deductive fallacy of the form: If p then q Not p Therefore, not q Both premises could be true and the conclusion still false; overlooks alternative explanations Division: assuming that the parts of a whole must have the properties of the whole; opposite of ‘‘composition’’ Equivocation: using a word in more than one sense False cause: generic term for a questionable conclusion about a cause and effect False dilemma: reducing the options you consider to just two, often sharply opposed and unfair to the person against whom the dilemma is posed (arguing from a false dilemma is sometimes a way of not playing fair, and it overlooks alternatives) Loaded language: using language whose only function is to sway the emotions of the readers or hearers, either for or against the view you are discussing; making your argument look good by caricaturing the opposite side Nonsequitur: drawing a conclusion that ‘‘does not follow’’; a conclusion that is not a reasonable inference from the evidence; this is a very general term for a bad argument— try to determine specifically what is (supposed to be) wrong with the argument The ‘‘person who’’ fallacy: allowing one vivid example to outweigh a careful examination of data Persuasive definition: defining a term in a way that appears to be straightforward but that in fact is subtly loaded (positively or negatively) Petitio principii: Latin for ‘‘begging the question’’ Poisoning the well: using loaded language to disparage an argument before even mentioning it (continued )

208

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

TABLE 9.1

(Continued)

Post hoc, ergo propter hoc (literally, ‘‘after this, therefore because of this’’): assuming causation too readily on the basis of mere succession in time Provincialism: mistaking a local fact for a universal one Red herring: introducing an irrelevant or secondary subject and thereby diverting attention from the main subject; usually the red herring is an issue about which people have strong opinions, so that no one notices how their attention is being diverted Straw man: caricaturing an opposing view so that it is easy to refute Suppressed evidence: presenting only the part of a piece of evidence that supports your claim while ignoring the parts that contradict your claim Weasel word: changing the meaning of a word in the middle of your argument, so that your conclusion can be maintained, though its meaning may have shifted radically

tially contributed to its development. When experts have a material stake in the outcome of the peer review (such as a regulated entity) or have participated substantially in the development of the product, their reviews are more appropriately referred to as peer input. Peer review involves an indepth assessment of the assumptions, calculations, extrapolations, alternative interpretations, methodology, acceptance criteria, and conclusions of the study or work product and of the supporting documentation. USEPA (1998) and other government agencies (USGAO 1999) have formal peer review guidelines and policies. Scientists, regulators, and other stakeholders—no matter how brilliant—simply cannot think out every thing for themselves all of the time, nor do they have expertise in all areas involved in drinking water regulation. Much of the time, no alternative exists except to trust the experts in a particular area. But can they be trusted? Experts carefully selected will certainly be knowledgeable in their subject area, but they also may have an interest in furthering a view that is in their own interests, delivering what is popularly known as a ‘‘snow job.’’ The late physicist Richard Feynman, a recognized leader of modern science, described most appropriately qualities of a trustworthy expert and the importance of scientific integrity in a 1974 commencement speech at the California Institute of Technology. Feynman (1989) told the graduating class to cultivate a kind of scientific integrity, a principle of scientific thought that corresponds to the kind of utter honesty—a kind of leaning over backwards. For example, if you’re doing an experiment, you should report everything that you think might make it invalid—not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you’ve eliminated by some other experiment, and how they worked—to make sure the other fellow can tell they have been eliminated . . . . In summary, the idea is to try to give all the information to help others to judge the value of your contribution; not just the information that leads to judgment in one particular direction or another. The first principle is that you must not fool yourself—and you are the easiest person to fool. So you have to be very careful about that. After you’ve not fooled yourself, it’s

9.4 SCIENTIFIC DISAGREEMENT

209

easy not to fool other scientists. You just have to be honest in a conventional way after that. I would like to add something that’s not essential to the science, but something I kind of believe, which is that you should not fool the layman when you’re talking as a scientist. I’m talking about a specific extra type of integrity that is [more than] not lying, but bending over backward to show how you’re maybe wrong, that you ought to have when acting as a scientist. And this is our responsibility as scientists, certainly to other scientists, and I think to laymen.

In sum, trustworthy experts are those who  Understand their responsibility to provide their expertise without claiming to know more than they really do  Seek to conduct scientific studies of the highest quality  Don’t try to evade questions or difficulties  Pursue objectivity, reason, and clarity  Apply fair-minded, critical thinking Over time, experts in a field of study develop a reputation on the basis of a track record of the quality of their work, actions, and public statements. They may become generally regarded as a reliable, trustworthy resource—a process generally referred to as peer regard. Reliance on trustworthy experts is essential for peer involvement—peer input and peer review—to be effective.

9.4

SCIENTIFIC DISAGREEMENT

Scientists often disagree. The same can be said for engineers, architects, and all other scientifically based disciplines. Even the use of peer-reviewed science will not, unfortunately, guarantee agreement among all parties. Scientists often weigh differing strands of evidence and supporting data differently. But disagreement among scientists does not by itself mean that a particular scientific view is unsound; that depends on the available evidence. In its report, Setting Priorities for Drinking Water Contaminants, a committee of the National Research Council (NRC 1999) addressed the issue of scientific disagreement as it applies to evaluating contaminants in drinking water for possible regulation: [The committee] takes the position that scientific disagreements are the norm and do not signal a deviation from sound science. These disagreements may be based on values other than strictly scientific ones, however, this does not mean that the sides of the debate are not based on sound science. Indeed, it is not unusual for scientists to disagree on the application of sound science to public policy issues. Any scheme that affects the provision of public water is likely to engender legitimate scientific disagreement. The

210

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

report also recognizes that identifying and agreeing on what is sound science is itself a difficult and error-prone enterprise. It therefore makes no recommendations on what ‘‘soundness’’ entails, letting the accepted mechanisms of peer regard, peer review, and scientists’ habits of critical thinking continue to serve as the ultimate arbiters.

In an atmosphere of high mistrust, disagreements can cause breakdown in communication and fracture relationships between scientists and=or stakeholders. But when trust is maintained, disagreement can stimulate discussion and lead to better solutions. In a sense, the more vigorous the discussion between trustworthy experts and=or stakeholders, often the better the resulting solution. In some cases, a neutral third-party facilitator is necessary for stakeholders or groups of scientists to communicate with each other when disagreement is high or emotionally charged and trust is low. (See Chapter 11 for a discussion of public involvement.)

9.5

‘‘JUNK’’ SCIENCE

In recent years the idea has arisen that certain studies are so poor as to warrant being called ‘‘junk science’’ (Milloy 2001). Junk science is argued to be the source of many health scares and scams. Specifically, ‘‘junk science’’ is defined as (Milloy 2001) the manipulation of statistics to promote special policy agendas that have nothing to do with public health and safety. It can be disseminated by special interest groups, social and political activists, businesses seeking to hurt rival companies, and politicians. Unfortunately, many gullible journalists pass on the bad information, alarming the public and causing much harm.

It is well known that measurements and statistics can be manipulated to deliberately deceive (Brignell 2000, Lutz 1989, Huff and Geis 1993), and studies must be carefully evaluated if this is suspected. But the fact that measurements and statistics are used in a study does not mean that the weaknesses of a study are deliberately intended. Technical studies and research typically involve data collection, analysis, and interpretation (Bendat and Piersol 2000). Furthermore, scientific techniques have limitations. Even so, all studies have strengths, weaknesses, limitations, and appropriate applications. Indeed, there is no generally accepted, uniform criterion or set of criteria by which to identify a particular study as ‘‘junk,’’ although Milloy (2001) has proposed such criteria particularly with regard to interpreting epidemiologic data. Evaluation of a study according to such criteria may serve a useful purpose in sharpening the interpretation and application of a study. But simply labeling a study as ‘‘junk,’’ is generally not helpful to fairly assess the strengths and weaknesses of a particular study and often serves more as a colorful political device to further an agenda. In general, data should not be categorically discarded, even from weak studies. The merits and applicability of each study must be considered in light of its strengths and weaknesses. Distortion or misinterpretation of scientific findings by news media can cause unnecessary alarm and cloud the proper interpretation of a study (Murray et al. 2001, Best 2001).

9.6 CAUSATION AND CAUSAL INFERENCE

9.6

211

CAUSATION AND CAUSAL INFERENCE

Vigorous debate over the ‘‘soundness’’ of scientific studies is perhaps most intense in the area of assessing causation, especially causation in relation to health effects of drinking water contaminants. Causation and causal inference have been discussed in Chapters 6 and 7, and are reviewed in detail by Klaassen (2001) and Rothman (1998). Scientific studies are performed by imperfect scientific people using imperfect scientific methods and techniques. Indeed, they must be properly evaluated applying fair-minded, critical thinking, and appropriate criteria (i.e., Tables 9.1 and 9.2). Toxicologists and epidemiologists (and engineers as well as water plant operators) are trained in their respective disciplines to think differently from one another, usually start with differing fundamental assumptions, and have differing analytical approaches. Hence, debates over the interpretation and meaning of toxicologic and epidemiologic studies can be sharp. Toxicology and epidemiology are discussed further in Chapters 6 and 7, respectively. An excellent case study of the difficulties in assessing causation and disputes between disciplines is provided in USEPA’s regulation of chloroform in drinking water. Chloroform is regulated as one of the total trihalomethanes (TTHMs). In 1994, USEPA proposed a maximum contaminant level goal (MCLG) for chloroform. At that time, the agency assumed, because of a lack of data to the contrary, that there was no safe threshold for chloroform’s potential carcinogenic effects. On the basis of this assumption, the agency initially proposed an MCLG of zero for chloroform. Subsequent to this proposal, extensive new science became available concerning the carcinogenicity of chloroform. USEPA considered these data, and concluded that a genotoxic mechanism was not likely to be the predominant influence of chloroform on the carcinogenic process, and that there was a reasonable scientific basis to conclude that the chemical caused cancer through cytotoxicity. This conclusion was counter to the agency’s longstanding policy of setting MCLGs for potential carcinogens at zero. At issue was the strength of the large body of experimental evidence for cytotoxicity, as opposed to a genotoxic mechanism, and how that evidence should be applied to regulate chloroform. USEPA’s proposal to set a nonzero MCLG for chloroform was vigorously opposed by environmental groups and many epidemiologists, who were skeptical of the experimental evidence. On the other hand, toxicologists were more likely to be convinced of the experimental evidence, even going as far as to take a popular vote in strong support of a nonzero MCLG for chloroform at a session held at the 2000 Society of Toxicology Conference. How much experimental or other evidence is necessary to justify a regulatory decision concerning the MCLG (or MCL, for that matter) that is counter to a longstanding policy assumption, as in the case of chloroform, or political pressure? There is no simple answer to this question. Rothman (1998) notes that Even when they are possible, experiments (including randomized trials) do not provide anything approaching proof and in fact may be controversial, contradictory, or irreproducible . . . . Laboratory studies often involve a degree of observer control that cannot be approached in epidemiology; it is only this control, not the level of observa-

212

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

TABLE 9.2

Potential Errors in Environmental Studies

Were appropriate controls used? The control group in health effect studies must be as similar as possible to the study group except for the exposure of interest If the study group has a characteristic that the control group does not have and that could cause the effect of concern, then a ‘‘false-positive’’ result may be derived from the experiment A ‘‘false-positive’’ is an erroneous correlation between exposure and health effect, which leads to a wrong conclusion Important factors to consider in establishing a control group in a human epidemiological study are the effects of age, smoking, gender, and preexisting conditions Controls are also important in nonhuman, scientific studies to isolate the variable(s) or phenomenon(a) of interest and minimize or eliminate confounding factors Was an appropriate baseline established? Increases and decreases are observed partly because of the existing data and convenience and partly because of the point the researcher has in mind In some cases, knowing where the baseline begins is necessary to properly evaluate the study results Was the baseline varied to determine whether the conclusion is robust? Would the conclusions be affected if a different baseline were chosen If the conclusion is the same no matter what the baseline, the study is on firmer ground Are conclusions extrapolated from the ‘‘parts’’ to the ‘‘whole’’? Observing effects on a micro scale (e.g., biochemical reaction) may not indicate effects on a macro scale (e.g., organism) Conclusions drawn from experiments on separate components do not necessarily apply to the whole unit Were direct measures used? Direct measures should be used for the important factors, characteristics, effects, etc. Indirect measurements may be useful, but they may not necessarily indicate directly the effect or phenomena under investigation Are the stated trends consistent across the measured data? Following trends in data decreases reliance on appearances and guesses Do the measured data consistently support the claims and conclusions made Was the normal range for the phenomenon in question established as a standpoint from which to judge trends? Establishing the normal or typical variation of a phenomena or measurement is important, if possible Effects within the normal range may not be attributable to a cause Excursions from the normal range are the phenomena of interest, not variations within the normal range Were the same types of measurements used consistently? Measurements must be accurate and correct Measurements must also be compatible if they are to be compared to historical data or other measures Actual measurements are generally preferable to model estimates: Accuracy and reliability depend on actual measurements Model estimates must be validated by actual measurements Models are not reality, but descriptions of reality or of a phenomena

9.6 CAUSATION AND CAUSAL INFERENCE

TABLE 9.2

213

(Continued)

Rely as little as possible on model extrapolations that are difficult or impossible to verify and as much as possible on actual measurements Be aware of recall bias in conducting surveys and assessing exposure: Memory and recall have been proved to be unreliable. Exposure assessments should rely on actual measurements whenever possible Consider the duration of exposure: Short-term exposures will differ in effects from chronic exposures Evaluate separate effects to determine if there is really something to be concerned about: Clusters of effects or use of surrogate measures can be misleading Specific effects must be identified and measured Be aware of extrapolation from high to low doses: Establishing an effect at a high exposure level is easier than at low exposure levels It is usually assumed that there is no safe dose, however low, for a chemical that has been shown to be a hazard at however high a dose In toxicology, ‘‘the dose makes the poison;’’ this means that the larger the dose, the greater the effect However, there may be no scientific basis for assuming that a chemical is dangerous at low doses just because it is dangerous at high doses; such an assumption is usually made by regulatory agencies as a policy assumption Many substances are beneficial at low doses but toxic at higher doses Has a mechanism been firmly established? A cause must precede its effect A cause and an effect must be connected in order for definitive conclusions to be reached Knowing that a plausible mechanism has not yet been identified or tried out should provoke skepticism Have conditions of applicability been established? Under what conditions do the conclusions or observations apply? Do not accept residual explanations: Do not accept the reasoning that if it is not X or Y, it must be Z, when other possibilities exist Be on the lookout for logical fallacies (see Table 9.1) Don’t draw final conclusions from one study: If a chemical is truly responsible for an effect, then similar studies of that chemical should show similar effects Scientists must be able to replicate results before they can confirm or deny the causal implications of a study that shows a correlation between exposure and illness Look to other studies for confirming or contradicting results Be skeptical: Environmental studies may be flawed because of one or more errors Whether the claims are of excessive harm or of no harm at all, they should be approached on a ‘‘show me’’ basis On complicated issues, keep score, and think carefully: Track the supporting evidence and conflicting expert opinions Do not discard data, even from weak studies; consider the merits of each study in light of its strengths and weaknesses Separate out rival arguments so that the evidence pro and con can be accumulated (continued )

214

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

TABLE 9.2

(Continued)

Apply fair-minded thinking to evaluate divergent views: Welcome differences of opinion, because they may bring new insights By comparing evidence supporting and opposing rival views, new insights are likely to emerge Often, the more vigourous the fair-minded discussion of opposing views that occurs, better solutions and conclusions are fashioned Evaluate data analytical techniques and results carefully Data analysis techniques and statistics must be appropriate to the data sets analyzed Statistical tests must be properly interpreted Alternative interpretations of data should be considered

tion, that can strengthen the inferences from laboratory studies. And again, such control is no guarantee against error.

As a body of scientific knowledge is developed over time, discoveries and conclusions might challenge conventional thinking or longstanding policy. In the case of the chloroform MCLG, court action was necessary for resolution of the issue, discussed below.

9.7

SCIENCE AND SDWA REGULATIONS

The U.S. Environmental Protection Agency’s (USEPA’s) Office of Ground Water and Drinking Water (OGWDW) has set forth the objective to implement a balanced public health protection program based on ‘‘sound science and adequate data’’ (Dougherty 1996) Indeed, the 1996 Safe Drinking Water Act (SDWA) amendments specifically requires USEPA to the degree that an agency action is based on science, to use the best available, peer-reviewed science and supporting studies conducted in accordance with sound and objective scientific practices, and use data collected by accepted methods or best available methods (if the reliability of the method and the nature of the decision justifies use of the data). Given the added importance of science to the SDWA rulemaking endeavor, the sharp disagreements that can often exist as to what constitutes ‘‘good’’ or ‘‘sound’’ science become more significant. At various times within the regulatory process, stakeholders should take full opportunity to engage in discussions with USEPA and other stakeholders regarding the adequacy of science forming the basis of regulations and regulatory policy. The regulatory process is helped, not harmed, by the formulation and defense of different hypotheses that extend beyond known facts (Wildavsky 1995). But the criteria for what counts as knowledge, the epistemology of science, must remain high. Emphasis is needed on the quality of science, not who performed or funded the study. Wildavsky (1995) argues that replacing a scientific hierarchy of values and the honoring of common scientific standards with inchoate feelings shifts the focus from

9.8 SCIENCE AND THE COURTS

215

searching for universal truths on which to base decisions to the expression of personal testimonies. What is true becomes far less important or even irrelevant, as in the case of the postmodern view, compared with what is personally authentic. Not all regulatory decisions are based on science. Although science cannot answer every question, replacing science with sincere feelings undermines the scientific underpinnings of good decisionmaking. Otherwise sound science becomes defined by an eminent authority, or by whomever is in power, or by whomever can overpower everyone else. Throughout the drinking water regulatory development process, stakeholders and USEPA present and evaluate expert opinions, scientific and engineering data, modeling results, and other forms of information. Deliberations occur over scientific data with arguments for and against positions, policies, and regulatory options. Disagreement among experts as to what is true can sometimes leave stakeholders at a loss as to the best way to approach a regulatory issue. Scientific truth is not infallible, and can change over time. Disagreement regarding sound science is to be expected as experts move beyond facts that can be proved using a scientific protocol or logical reasoning to express beliefs that cannot be scientifically documented. Scientific integrity depends on institutions that maintain competition between scientists, and scientific groups who are numerous, dispersed, and independent. Scientific claims should be subject to replication and verification. Informed opinions of scientific and engineering experts are needed throughout the regulatory process, in terms of what is scientifically factual as well as what may be possible, probable, and improbable.

9.8

SCIENCE AND THE COURTS

The courts have an important oversight role regarding USEPA’s application of science. Chapter 22 reviews the role of the courts in assessing science in legal proceedings related to toxic tort cases. SDWA provisions allow a petitioner to challenge the science behind a USEPA regulatory action, as discussed below.

9.8.1

Judicial Review

The SDWA gives the federal courts jurisdiction to review USEPA actions. Section 1448 of the SDWA allows for judicial review of actions regarding primary drinking water regulations (including MCLGs). A petitioner must file in the U.S. Court of Appeals for the District of Columbia Circuit. The judicial review process is discussed in the following sections, and additional details are available elsewhere (Mintz and Miller 1991, Prillaman and Rubin 1992). Petitions must be filed not later than 45 days after a regulation has been promulgated or an order has been issued, with one exception. A petition can be filed later if it is based solely on grounds arising after the expiration of such period. Filing a petition within this time limit is critical. Any USEPA action not petitioned against

216

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

within the period will not be subject to judicial review in any civil or criminal proceeding for enforcement or in any civil action to enjoin enforcement. A key question in any court proceeding is whether a petitioner has ‘‘standing’’ to bring the lawsuit. Standing is a concept derived from several court decisions interpreting provisions of the U.S. Constitution that authorize the judiciary to exercise its authority only when alleged personal injury is fairly traceable to the purportedly unlawful conduct and is likely to be redressed by the requested relief. Harm can be actual or threatened, but it must be distinct and palpable. Harm to an ideological interest is insufficient. Only an organization, group, or individual that has standing could challenge the soundness of the science underlying a USEPA rule. The principles of standing were explained in a 1991 decision of the U.S. Court of Appeals for the DC (District of Columbia) Circuit (Court), which reviewed a challenge to several MCLs promulgated by USEPA in January 1991 (Court of Appeals 1992). The petitioners in that case included the International Fabricare Institute (IFI), an organization of dry cleaners that use perchloroethylene, and several major manufacturers, including Dow Chemical Company and Shell Oil Company. In response to USEPA’s challenge to the petitioners’ standing, the Court of Appeals (1992) ruled: An organization such as IFI may have standing, even when the organization itself has not suffered an injury, if it can show that (1) its members would otherwise have standing to sue in their own right; (2) the interest it seeks to protect are germane to the organization’s purpose; and (3) neither the claim asserted nor the relief requested requires the participation of individual members in the lawsuit.

The court also stated that there must be some indication, even slight, that the litigant before the court was intended to be protected, benefited, or regulated by the statute under which the suit was brought. The court agreed that there was sufficient basis to grant the dry cleaners’ standing to challenge USEPA actions because MCLGs are used by USEPA to establish liability under the Superfund. As for the manufacturers, the court was satisfied that they operate public water systems and therefore would be directly regulated by the MCLs.

9.8.2

The Judicial Review Process

An overview of the judicial review process is provided in Figure 9.1. A lawsuit commences with the filing of a brief statement by the petitioner in the U.S. Court of Appeals after the final regulation is published in the Federal Register. This ‘‘petition for review’’ identifies the formal citation to the regulation, but need not and usually does not identify the particular issue(s) the petitioner finds objectionable. The petition is usually short, about one page in length. Multiple parties may file petitions on any given rule. After a petition is filed, there is a 30-day period within which another party may file to intervene. An intervenor may join on the side of USEPA to help defend a regulation, or on the side of the petitioner to provide support in challenging a regulation. All petitions are initially handled as separate lawsuits. The court eventually consolidates the cases, that is,

9.8 SCIENCE AND THE COURTS

217

Figure 9.1 Judicial review process.

groups them together in one case before the court. A petitioner may request that the court isolate its petition into a separate case if reason exists to do so. The Supreme Court has ruled that challengers to a regulation litigate on their own time; that is, a regulation will remain in effect during the course of the challenge, unless the USEPA or the court issues a ‘‘stay’’ (postponement) of the regulation pending appeal. SDWA regulations are complex, and often have thousands of pages of documentation in the administrative record. Several months are usually required for both the petitioner and USEPA to identify and consider all the pertinent facts and legal

218

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

arguments relating to the challenge. Typically, USEPA or a petitioner will request the court delay by setting up a formal schedule for submitting briefs and presenting oral arguments to the court. A request for a delay is presented in the form of a motion. A motion seeks to establish a schedule within which the parties can consider settlement and provides that, should settlement discussions fail, the parties will agree to return to the court to establish a schedule to submit briefs. In case out-of-court negotiations fail, the Court of Appeals sets a briefing schedule. Petitioners are given a certain period of time within which to file briefs that present the arguments for overturning the regulation. USEPA is then given a period of time to respond. Occasionally, a party may intervene on the side of USEPA or on the side of the petitioner. That party, referred to as an ‘‘intervenor,’’ files a supportive brief at the same time according to the briefing schedule. Petitioners are then given a short period of time, usually 14 days, to file a reply brief. The judicial review is limited to the administrative record for the rule in dispute. Briefs seldom contain new information, but instead focus the court’s attention on deficiencies in the rulemaking record. Typically, the petitioner will identify pertinent portions of the statute and case law in support of the petitioner’s contention that the court should act to overturn (or uphold) a regulation. The court will schedule the case for oral arguments, which in some cases may be years after filing of the initial petition. Each side is given 15 minutes, sometimes up to an hour, to address a panel of three judges (Court of Appeals 1984). The primary purpose of oral argument is not to repeat what has already been stated in the briefs, but to respond to questions the judges will raise about the arguments and facts in dispute. The law establishes no specific timeframe within which the court must issue a written decision. The court of appeals typically does one of three things. It may affirm a regulation in its entirety; it may overturn a regulation and remand it for further consideration by the agency; or it may direct USEPA to take specific action. The court may choose to affirm one portion of the regulation, overturn another portion, and direct USEPA to take specific action on yet another portion of the regulation. If the court rules entirely in favor of the agency, the regulation remains in place, as is. If the court rules in favor of the petitioner, to the extent that the court has overturned the regulation, ordinarily that portion of the regulation objected to is no longer in effect.

9.8.3

Deference

The Court of Appeals has construed the applicable principles of the Administrative Procedure Act to require it to give substantial deference to USEPA when reviewing its rules. The Court of Appeals (1992) has stated We will reverse [a] USEPA action only if it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law . . .. This highly deferential standard of review presumes Agency action to be valid . . . . The rationale for deference is particularly strong when USEPA is evaluating scientific data within its technical expertise; In an area characterized by scientific and technological uncertainty, . . . this

9.8 SCIENCE AND THE COURTS

219

court must proceed with particular caution, avoiding all temptation to direct the Agency in its choice between rational alternatives. Despite this deferential standard, we must [e]nsure that USEPA has examined the relevant data and has articulated an adequate explanation for its action . . . . The USEPA is required to give reasonable responses to all significant comments in a rulemaking proceeding . . . . We will therefore overturn a rulemaking as arbitrary and capricious where USEPA has failed to respond to specific challenges that are sufficiently central to its decision.

In short, the court has decided that USEPA’s interpretation of its own regulation will be accepted unless it is plainly wrong. Of course, a rule may not be wrong, but still be based on marginal science. Even so, the court’s job is not to determine whether USEPA’s judgments are correct, but whether they are properly reasoned. Overall, USEPA has had remarkable success in defending against challenges to drinking water regulations issued under the SDWA. Given the great deference granted to the agency by the court, a rulemaking or portion thereof must be clearly incorrect or fatally flawed to justify an appeals court ruling that it is arbitrary and capricious. In cases where the science is unclear, or equally qualified experts disagree, the court will defer to the scientific judgments made by the agency. This critical point must be carefully considered by anyone contemplating a legal challenge to a USEPA drinking water rulemaking on scientific grounds—the court will presume the agency to have made correct judgments of science, and it will be up to the petitioner to prove that the agency was clearly wrong. In general, disagreements between reasonable scientists and engineers within the agency and the regulated community during the course of development of a rulemaking will generally not be sufficient to justify an arbitrary and capricious ruling. The burden of proof is on the petitioner to show that the agency is incorrect. 9.8.4

Example: Chloroform MCLG

In some respects, ‘‘sound’’ science is what the courts say it is, at least in regard to scientific disputes that are elevated through court action. As noted above, the courts typically defer to agency judgments regarding science. However, Pontius (2000) and Quill and Fischer (2000) have reviewed the notable case of chloroform, mentioned above, where the court ruled that USEPA had indeed violated the SDWA by failing to use the best peer-reviewed science in setting the MCLG (Court of Appeals 2000). There are several aspects of this case that greatly influence the agency’s application of science in rulemakings. As mentioned previously, USEPA proposed an MCLG of zero for chloroform in 1994. At that time, the agency assumed, because of a lack of data to the contrary, that there was no safe threshold for chloroform’s potential carcinogenic effects. Subsequent to this proposal, extensive new science became available concerning chloroform’s carcinogenicity. USEPA considered these data in special Notices of Data Availability (NODAs) in 1997 and 1998, to seek comment on the new data and analysis on chloroform. The 1998 NODA concluded that a genotoxic

220

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

mechanism was not likely to be the predominant influence of chloroform on the carcinogenic process, and that there was a reasonable scientific basis to conclude that the chemical caused cancer through cytotoxicity. In December 1998, USEPA published its final rule establishing the MCLG for chloroform. Although the agency reaffirmed the findings in the 1998 NODA, it did not apply those findings to the final rule, and set the MCLG for chloroform at zero. According to press reports, the Clinton administration bowed to pressure from environmentalists in making this decision (Inside Washington Publishers 1998). The Chlorine Chemistry Council (CCC), a business council of the American Chemistry Council, along with other industry petitioners, filed a petition for review, arguing that the agency violated the SDWA requirement to ‘‘use the best available, peer reviewed science’’ when it promulgated the zero MCLG for chloroform. Typically, petitioners will challenge a rulemaking by questioning the agency’s scientific determinations. In this case, CCC and others sought to compel USEPA to apply its own scientific conclusions, rather than the scientific conclusions of others. This lawsuit drew much attention, prompting amicus (i.e., Friend of the Court) briefs by a group of 13 reputable scientists and by Rep. Thomas Bliley (R–VA), chair of the House Commerce Committee. Both amicus briefs supported CCC’s petition. In a rare action, USEPA asked the Court of Appeals on Feb. 24, 2000 to vacate or invalidate the chloroform MCLG it issued Dec. 16, 1998 (BNA 2000a, 2000b). Hence, whether the MCLG for chloroform should have been set at zero was no longer at issue. Even so, an important issue before the court was whether the agency had violated the law by not proceeding in 1998 with the best science at hand, but delaying the decision for Science Advisory Board (SAB) review. On March 31, 2000, the U.S. District Court issued a ruling vacating the MCLG for chloroform, finding that USEPA had indeed violated the SDWA by failing to use the best available peer-reviewed science in setting the MCLG (Court of Appeals 2000). The court found the zero MCLG to be arbitrary and capricious and in excess of statutory authority. The fact that the agency arrived at a novel, politically charged outcome, was deemed of no significance. The DC Circuit Court ruled in favor of the petitioners, concluding that ‘‘[i]n promulgating a zero MCLG for chloroform, EPA openly overrode the ‘‘best available’’ scientific evidence, which suggested that chloroform is a threshold carcinogen.’’ The court also rejected the agency’s rationale that it adopted a zero MCLG because it wished to consult further with its SAB about chloroform’s carcinogenicity. The court found that ‘‘however desirable it may be for EPA to consult an SAB and even to revise its conclusion in the future, that is no reason for acting against its own science findings in the meantime.’’ Significantly, the court concluded that ‘‘EPA cannot reject the ‘best available’ evidence simply because of the possibility of contradiction in the future by evidence unavailable at the time of the action—a possibility that will always be present.’’ The DC Circuit’s findings have two important implications: (1) an agency may be acting illegally when it relies on default assumptions when the best available science supports less conservative approaches for assessing risk and (2) the best

9.9 FUTURE DEVELOPMENTS AND TRENDS

221

available science is the scientific evidence that is available at the time of a rulemaking. The possibility of contradiction based on further scientific data or peer review is not a legitimate basis for rejecting the science that currently exists. 9.9

FUTURE DEVELOPMENTS AND TRENDS

The topic of ‘‘sound’’ science and drinking water regulation is multifaceted. Alternative views and truth claims in the ‘‘marketplace’’ of ideas must be carefully evaluated when formulating regulations and regulatory policy. Disagreements over science and scientific issues are to be expected and will certainly continue. The issues and dynamics reviewed in this chapter demonstrate the difficulties faced in assessing the soundness of a particular scientific study. Typically, a spectrum of beliefs and data must be deciphered and evaluated to extract those scientific studies, facts, assumptions, and beliefs that will stand the test of time. Determining what constitutes sound science requires time to gather factual information, consideration of opposing views and alternative hypotheses, application of good cross-discipline, fair-minded, critical thinking habits and a clear objective. For informed public policy and regulatory policy discussions, context, source, presumptions, and bias in scientific studies all must be evaluated in order to determine when science is truly sound. USEPA actions and use of science under the SDWA are subject to scrutiny by the federal courts. The judiciary gives the agency much leeway and USEPA’s administrative determinations are usually upheld by the courts. Even so, the availability of judicial review provides a constant reminder to the agency to be attentive to all procedural safeguards and to carefully consider the arguments and scientific data presented by those affected by its regulations. To make effective use of the right to judicial review, two things are essential: (1) legal arguments and factual data must be presented to USEPA prior to promulgation of the final rule and (2) petitions to challenge a final rule must be filed within 45 days after the rule is published in the Federal Register. The decision to legally challenge a USEPA rule on scientific grounds must be made by carefully evaluating the potential gains if the petition is successful and the potential risks and costs if it is not. Under the SDWA, reliance on default assumptions may be problematic, or even illegal, when the best available science supports less conservative approaches for assessing risk. The ‘‘best available science’’ under the SDWA is the scientific evidence that is available at the time of a rulemaking. The possibility of contradiction based on further scientific data or peer review is not a legitimate basis for rejecting the science that currently exists. REFERENCES Ackoff, R. L., S. K. Gupta, and J. S. Minas. 1962. Scientific Method: Optimizing Applied Research Decisions. New York: Wiley.

222

‘‘SOUND’’ SCIENCE AND DRINKING WATER REGULATION

Beagle, R., R. Bonita, and T. Kjellstrom. 1993. Basic Epidemiology. Geneva: World Health Organization. Bendat, J. S. and A. G. Piersol. 2000. Random Data, New York: John Wiley & Sons, Inc. Best, J. 2001. Damned Lies and Statistics: Untangling Numbers from the Media, Politicians, and Activists. Berkeley: University of California Press. BNA. 2000a. EPA Asks federal court to vacate standard setting chloroform MCLG. Daily Environment Report. Bureau of National Affairs, Feb. 28. BNA. 2000b. Industry wants court to issue opinion—chloroform. Daily Environment Report. Bureau of National Affairs, March 10. Brignell, J. 2000. Sorry, Wrong Number! The Abuse of Measurement. Broughton, UK: J. E. Brignell and the European Science and Environment Forum. Court of Appeals. 1984. Information Pamphlet Regarding Oral Argument Procedures. U.S. District Court of Appeals, D.C. Circuit, Washington, DC. Court of Appeals. 1992. International Fabricare Institute for Itself and on Behalf of its Members v. USEPA. U.S. District Court of Appeals for the District of Columbia, Case 91-1838 (Dec. 6, 1994). Court of Appeals. 2000. Chlorine Chemistry Council and Chemical Manufacturers Association v. EPA. U.S. Court of Appeals, District of Columbia Circuit, Case 99-1627 (March 31, 2000). Davies, P. 1992. The Mind of God: The Scientific Basis for a Rational World. New York: Simon & Schuster. Dougherty, C. C. 1996. USEPA Memorandum Regarding OGWDW Reorganization, April 11. Washington, DC: Office of Ground Water and Drinking Water. Einstein, A. 1990. The World as I See It: Out of My Later Years. New York: Quality Paperback Books. Feynman. R. 1989. Cargo Cult Science. Surely You’re Joking, Mr. Feynman: Adventures of a Curious Character. New York: Bantam. Hill, A. B. 1965. The environment and disease: association or causation? Proc. Roy. Soc. Med. 58:295–300. Huff, D. and I. Geis. 1993. How to Lie with Statistics. New York: W. W. Norton and Company. Inside Washington Publishers. 1998. Political pressure may result in ‘‘zero’’ standard for chloroform. Risk Policy Report 5(11):8–9. Klaassen, C. D., ed. 2001. Casarett & Doull’s Toxicology: The Basic Science of Poisons. New York: McGraw-Hill. Kneller, G. F. 1978. Science as a Human Endeavor. New York: Columbia Univ. Press. Lastrucci, C. L. 1963. The Scientific Approach; Basic Principles of the Scientific Method. Cambridge, MA: Schenkman. Lutz, W. 1989. Doublespeak. New York: Harper & Row. Lyotard, J.-F. 1997. The Postmodern Condition: A Report on Knowledge, trans. (from French) G. Bennington and B. Massumi. Minneapolis: Univ. Minnesota Press. Milloy, S. J. 2001. Junk Science Judo. Washington, DC: Cato Institute. Mintz, B. W. and N. G. Miller. 1991. A Guide to Federal Agency Rulemaking, 2nd ed. Washington, DC: Administrative Conference of the United States. Murray, D., J. Schwartz and S. R. Lichter. 2001. It Ain’t Necessarily So. Lanham, MD: Rowman and Littlefield.

REFERENCES

223

NRC. 1995. Classifying Drinking Water Contaminants for Regulatory Consideration. Washington, DC: National Academy Press. NRC. 1999. Setting Priorities for Drinking Water Contaminants. Washington, DC: National Academy Press. O’Connor, D. R. 2002. Report of the Walkerton Inquiry: The Events of May 2000 and Related Issues. Toronto: Ontario Ministry of the Attorney General. Paul, R., and L. Elder. 2001. Critical Thinking. Upper Saddle River, NJ: Prentice Hall. Pickering, A. 1992. From science as knowledge to science as practice. In Science as Practice and Culture. A. Pickering, ed. Chicago: Univ. Chicago Press. Planck, M. 1932. Where is Science Going? New York: Norton. Pontius, F. W. 2000. Chloroform: Science, policy, and politics. J. Am. Water Works Assoc. 92:12. Prillaman, J. and K. Rubin. 1992. Safe Drinking Water Act. Environmental Law Practice Guide. New York: Matthew Bender. Quill, T. F. and D. B. Fischer. 2000. D.C. Circuit elevates science over politics in its chloroform decision. Risk Policy Rept. 7:12:36–38. Rothman, K. J. 1998. Causal inference in epidemiology. In Modern Epidemiology, 2nd ed. K. J. Rothman and S. Greenland, eds. Philadelphia: Lippincott Williams & Wilkins. Shermer, M. 2002. Why People Believe Weird Things: Pseudoscience, Superstition, and Other Confusions of Our Time. New York: H. Holt. Taylor, C. A. 1996. Defining Science: A Rhetoric of Demarcation. Madison: Univ. Wisconsin Press. USEPA. 1998. Science Policy Council Peer Review Handbook. EPA 100-B-98-00. Washington, DC: Office of Research and Development. USGAO. 1999. Federal Research: Peer Review Practices at Federal Science Agencies Vary. GAO=RCED-99-99. Washington, DC: U.S. General Accounting Office. Wildavsky, A. 1995. But Is It True? A Citizen’s Guide to Environmental Health and Safety Issues. Cambridge, MA: Harvard Univ. Press. Wynn, C. M. and A. W. Wiggins. 2001. Quantum Leaps in the Wrong Direction. Washington, DC: Joseph Henry Press. Ziman, J. 2000. Real Science: What It Is, and What It Means. Cambridge, UK: Cambridge Univ. Press.

10 BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION ROBERT S. RAUCHER, Ph.D. Executive Vice President, Stratus Consulting Inc., Boulder, Colorado

10.1

INTRODUCTION

The Safe Drinking Water Act (SDWA) as amended in 1996 requires that the U.S. Environmental Protection Agency (USEPA) consider benefits and costs when setting drinking water regulations. The USEPA Administrator is required to issue a formal ‘‘determination’’ that the benefits of each standard ‘‘justify’’ the costs. Further, the Administrator is authorized to set maximum contaminant levels (MCLs) at levels less stringent than what is technologically feasible if the benefits are found not to justify the costs. Although the term ‘‘justify’’ is not defined in the statute, the objective under standard economic principles is to identify the MCL at which the benefits exceed the costs by the widest margin—the point where the ‘‘net benefits’’ are the greatest. This chapter provides an overview of the methodology and criteria used for evaluating the benefits and costs of drinking water regulations. Key questions addressed are 1. How are benefits and costs assessed? In other words, how are the estimates derived, and how reliable or controversial are they likely to be? 2. How should benefits be compared to costs? This question raises issues related to how one should interpret a benefit-cost comparison in setting a public health-oriented regulatory level, and includes the issue of how nonquantifiable benefits and costs should be considered. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

225

226

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

Neither of these questions can be answered with simple responses, because many complex factors and uncertainties must be considered. Nonetheless, this chapter examines the conceptual foundation and key principles for how benefit-cost analysis (BCA) should be performed and interpreted for setting MCLs. It is intended to provide readers with a clear overview of the issues and techniques, and an understanding of the key points of contention in the debates about how BCA should be conducted and interpreted for setting standards.

10.2

BENEFIT-COST ANALYSIS (BCA) UNDER THE SDWA

Common sense suggests that the benefits and costs of various options should be considered (in some manner) when making regulatory policy or other decisions that are aimed at protecting public health. Before the 1996 Amendments, however, standard setting under the SDWA could not take into consideration what the quantified health benefits of a regulation might be, or how those benefits compared to costs. Instead, the statute prior to 1996 required that the USEPA establish technologybased standards in which the MCLs were to be set as close to the ‘‘risk free’’ levels (MCL goal) as ‘‘feasible,’’ where feasibility pertained to technologically achievable contaminant removals and practical limits of quantitation. Public health risk reduction benefits were typically examined, but these benefits were rarely quantified in any meaningful or systematic manner, nor could they be taken into account in standard setting. Under the 1996 amendments to the SDWA, new statutory language (1) required USEPA to conduct and publish of a benefit-cost analysis with every rulemaking effort and (2) enabled the Agency to use benefit-cost information in selecting how stringently to set the standard. These two features are noteworthy, especially the latter provision, which enables the Administrator to set enforceable standards that may be less stringent than what is deemed technically feasible, if the BCA indicates the less stringent MCL is justified. More specifically, the 1996 amendments now require that USEPA publish a report describing the public health risk reduction benefits and national compliance costs for every standard that it proposes or promulgates. The SDWA [Sec. 1412 (b)(3)(C)] requires that the mandated benefit-cost analysis, which is referred to as a health risk reduction and cost analysis (HRRCA), include the following:
Quantifiable and nonquantifiable health risk reduction benefits from reductions in the contaminant of concern
Quantifiable and nonquantifiable health risk reduction benefits from reductions in co-occurring contaminants
Quantifiable and nonquantifiable costs, including monitoring, treatment, and other costs
Incremental costs and benefits associated with each alternative MCL

10.3 HISTORICAL APPLICATION OF BCA

227


The effects of the contaminant on the general population and on sensitive subpopulations
Any increased health risk that may occur
Other relevant factors, including the quality and extent of information, the uncertainties in the analysis, and factors related to the degree and nature of the risk In other words, the SDWA now requires that USEPA conduct a benefit-cost analysis (viz., HRRCA) that contains both quantitative and nonquantitative information, compares incremental benefits to incremental costs, and indicates the presence and impacts of uncertainties in the analysis. The HRRCA must be available for public review and comment as part of every rulemaking action. On the basis of the HRRCA, the Administrator is required to issue a formal ‘‘determination’’ that the benefits of each standard ‘‘justify’’ the costs. Further, the Administrator is authorized to set MCLs at levels other than what is technologically feasible if the benefits are found not to ‘‘justify’’ the costs. In other words, the Amendments enable the Administrator to set the standard ‘‘that maximizes health risk reduction benefits at a cost that is justified by the benefits.’’ Hence, statutory requirements now formally mandate that public health risk reduction benefits be systematically estimated, communicated to decision-makers and the public, and then evaluated vis-a-vis costs in making regulatory decisions. It is worth noting that neither the statutory language or the legislative history specify what ‘‘justify’’ means, leaving the term open to interpretation.

10.3

HISTORICAL APPLICATION OF BCA

Since the dawn of time, humans have weighed the pros and cons of their options before acting. As a formal policy evaluation tool, however, benefit-cost analysis was spawned by language in the Flood Control Act of 1936, which mandated its use by the U.S. Army Corps of Engineers in evaluating water resource projects. The concepts underlying BCA are well grounded in economic theory, but early applications were generally unsophisticated and often politically skewed to promoting specific water projects. Over subsequent decades, the BCA maturation process has refined its conceptual foundation, empirical methodologies, and policy interpretations. Nonetheless, widely recognized limitations remain. In the environmental, health, and safety areas in particular, significant challenges to the application of BCA include uncertainties, gaps in the available data, controversies surrounding methods for quantifying physical effects or placing monetary values on nonmarket outcomes (such as change in risks to health), and issues of equity. These unresolved issues suggest that BCA should not be used as a strict decision rule for defining which policy options can be considered and that must be selected. Rather, with better use of sound scientific and policy approaches to address uncertainties and other problems, BCA can be used, as now intended under the

228

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

amended SDWA, as a practical, objective, and valuable tool that contributes to a more informed decisionmaking process. Over the past few decades, BCA has been applied successfully and constructively to several important public health, safety, and environmental issues. For example, USEPA conducted an outstanding BCA of lead exposures from vehicle exhaust associated with use of leaded fuels. This BCA was instrumental in accelerating the phasedown of lead concentrations in motor fuels (absent the BCA, society would have continued to bear the costs of higher lead exposures). Useful benefit-cost applications to drinking water issues include a study that demonstrated that the pre-1996 SDWA statutory requirement that USEPA regulate 25 additional contaminants every 3 years was diverting scarce resources away from addressing more critical drinking water health risks. The analysis showed that some regulations cost over $1 billion for each cancer case avoided, whereas MCLs for other contaminants could achieve the same level of protection at less than $1 million per case avoided. As a consequence, close to 99% of the regulatory program’s carcinogenic risk reductions could be achieved at approximately 60% of the cost if there were greater flexibility in selecting which contaminants to regulate. Stated in another manner, the same monetary investment could have yielded far greater public health benefits if the regulations had been established on a benefit-cost basis (Raucher et al. 1994). These and other applications illustrate that, when pursued with due care (1) practical solutions to the inherent limitations of BCA can be found; (2) BCA can be a feasible, objective, and valuable tool for decisionmakers; and (3) BCA can be used to promote as well as criticize environmental programs (i.e., it is not a device intended solely to undermine the fabric of the nation’s health, safety, and environmental regulations). The balance of this chapter discusses issues related to how BCA should be conducted and interpreted to best ensure that the nation’s investments in drinking water yield the greatest public health benefits possible.

10.4

USEPA POLICIES AND PRACTICES

Before the 1996 Amendments to the SDWA, USEPA was not allowed to consider BCA issues in setting MCLs, nor was it required under the statute to estimate benefits. Nonetheless, since 1981 a series of Executive Orders—coupled with regulatory reviews by the Office of Management and Budget (OMB)—required that the Agency make some attempts at estimating benefits and comparing them to costs. Typically, these pre-1996 BCAs were very simple and qualitative, relying on very general information such as listing the potential adverse health effects associated with a contaminant of interest, and in some instances indicating the number of water systems and people possibly exposed to levels of regulatory interest. One key exception was the MCL for lead, for which a fairly extensive benefits analysis was developed, consisting of health risk assessments and economic valuations for many types of adverse health effects that were likely to be reduced by the rule. The BCA for this rulemaking was similar to some of the better analyses that the

10.5 COMPARING BENEFITS TO COSTS

229

Agency had conducted in its nondrinking water programs. For example, between 1981 and 1996, some reasonably sophisticated BCAs had been conducted by USEPA in the context of some of its air quality and wastewater regulatory activities. With passage of the 1996 Amendments to the SDWA, the USEPA drinking water program has attempted to step forward with better BCAs. For example, a benefits workgroup was convened as part of National Drinking Water Advisory Council (NDWAC) activities, and this panel established several guiding principles for using BCA in a manner consistent with good economic theory and public policy. In addition, USEPA’s analyses for SDWA-related rulemakings issued since 1996 have provided more quantitative and comprehensive BCAs than were typically seen before the Amendments. Nonetheless, some post-1996 BCAs (e.g., for MCLs more recently proposed or promulgated, such as for disinfection byproducts, radon, and arsenic) have raised concerns amongst some reviewers and stakeholders regarding shortcomings in how benefits were estimated and portrayed, how benefits were compared to costs, and how the Agency interpreted the BCAs. On a broader level (i.e., beyond the drinking water program), USEPA has been revisiting key BCA issues. The USEPA Office of the Administrator has published Guidelines for Preparing Economic Analyses (USEPA 2000) that address the core issues of how to conduct a BCA. In addition, the Agency’s Science Advisory Board (SAB), Environmental Economics Advisory Committee (EEAC), issued a report on the key issues of how USEPA should assign monetary values to regulations that reduce the risk of premature fatality (June 2000). Both reports provide practical and sound guidance on how BCAs should be conducted, but to date the HRRCAs from the USEPA’s drinking water office have not fully adhered to the spirit or letter of the guidance.

10.5

COMPARING BENEFITS TO COSTS

The goal of a BCA should be to help guide decisionmaking toward options that lead to the highest level of well-being for society as a whole (economists refer to this as ‘‘maximizing social welfare’’). This means looking for the MCL at which the benefits exceed the costs by the widest margin.

10.5.1

Maximizing Net Benefits

The point at which the benefits of an MCL exceed the costs by the widest margin occurs where the ‘‘net benefits’’ are the greatest. Net benefits equals the benefits minus costs. Sometimes the ratio of benefits to costs (B=C) is used as a way to evaluate a policy. If the B=C ratio is greater than 1, then the benefits outweigh costs. The use of a B=C ratio is valid for considering alternative investments, because the option yielding the greatest rate of return (highest B=C ratio) can be identified. However, for regulatory policy decisions such as setting an MCL, the net benefits concept (benefits minus costs) is more appropriate rather than the B=C ratio. This is

230

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

because the objective is to identify the MCL that maximizes social welfare (net benefits), and not to develop an investment strategy that maximizes rates of return. To identify the MCL that yields the greatest net benefit, economists look for the point where marginal benefits are equal to marginal costs. These ‘‘marginal’’ concepts refer to the change in benefits and the change in costs for each possible increased stringency of an MCL (i.e., what the additional benefits and costs would be if an MCL were to be made 1 mg=L more stringent than the last option considered).

10.5.2

Incremental Benefits and Costs

In reality, there are rarely data to examine how marginal benefits and costs change at such a tiny (i.e., 1 mg=L) change in an MCL, so instead the terms incremental benefits and incremental costs are used instead of marginal benefits or marginal costs. These ‘‘incremental’’ terms refer to examining how benefits and costs change from one MCL option to the next (e.g., moving from a 50 mg=L arsenic standard to a 20-mg=L MCL option, and then examining the incremental benefits and costs of moving from 20 mg=L to 10 mg=L, and so forth). The incremental perspective allows one to view regulatory options one step at a time, and to identify the point (MCL option) at which moving to the next, more stringent option would add more costs than benefits (where incremental benefits become outweighed by incremental costs). By selecting the lowest MCL for which the incremental benefits still outweigh the incremental costs, the policy will yield the greatest possible net benefits to society. This is why the SDWA statute now specifies that the HRRCA reveal the incremental costs and incremental benefits of each MCL option—it is the comparison of these incremental benefits and costs that enables one to maximize social welfare. This concept of maximizing net benefits by comparing incremental benefits to incremental costs is also how most economists would interpret the intent of the SDWA BCA provisions. One limitation of some drinking water BCAs to date has been the omission of an incremental benefit-cost comparison. Instead, some BCAs have shown total benefits and total costs, and justify MCL selections on the basis of the assumption that benefits are roughly the same size as costs. This type of benefit-cost comparison only indicates that the regulation may be a ‘‘breakeven’’ proposition at best (net benefits of zero)—there has not been any attempt to consider using the analysis to maximize net benefits, only to show that net benefits may be positive. Tables 10.1 and 10.2 illustrate the distinction between a total and an incremental BCA perspective, drawing on an analysis of hypothetical MCL options for MTBE (Raucher et al. 2002). Table 10.1 reveals the total (or average) benefits for each MCL option, and compares them to the hypothetical total cost estimates for each MCL. For each MCL option, the total benefits outweigh the total hypothetical costs. If one were choosing the MCL solely on the basis of the criterion that net social benefits (benefits minus costs) are positive, then any of the MCL options would pass the test. If a 10-mg=L standard were technically feasible, the depicted comparison of total benefits to total costs would ‘‘justify’’ that option.

10.5 COMPARING BENEFITS TO COSTS

231

TABLE 10.1 Total Benefits and Costs for Hypothetical MTBE Examplea Regulatory Scenario

Total Average Benefits ($)

Total Costs (Hypothetical) ($)

Net Average Benefits ($)

192 230 273

15 75 195

177 155 78

MCL ¼ 20 mg=L MCL ¼ 14 mg=L MCL ¼ 10 mg=L

a All values are expressed in 1999 dollars rounded to the nearest million. Total costs are based on a hypothetical example. Results are for fatal cancers only and assume a linear dose–response model and low-end hypothetical costs, and use the mean values from the benefits distribution.

Source: Raucher et al. (2002).

A more relevant and informative comparison would examine the incremental benefit and cost estimates. This is shown in Table 10.2, in which the total amounts from Table 10.1 are converted to their incremental counterparts (using simple subtraction). This incremental comparison tells a story different from that of the preceding one—namely, that the first regulatory increment (baseline to 20 mg=L) provides expected net improvements to welfare, but that the more stringent options each add more in costs than they add in benefits. This reveals the value of the incremental perspective, because it identifies that going from 20 to 14 mg=L would cost more to society than it would pay back in benefits—the more stringent options creates a net loss in social welfare relative to the 20-mg=L option.

10.5.3

Accounting for System Size

An additional point to consider in comparing benefits and costs is the need to break the analysis down by system size category. This is because drinking water treatment costs (the costs of compliance) tend to be relatively high on a per unit basis in small (rural) systems. Therefore, households served by small and=or rural water systems TABLE 10.2

Incremental Benefits and Costs for Hypothetical MTBE Examplea

Regulatory Scenario Baseline to MCL ¼ 20 MCL ¼ 20 to MCL ¼ 14 MCL ¼ 14 to MCL ¼ 10 a

Incremental Mean Benefits ð$Þ

Incremental Costs (Hypothetical) ð$Þ

Incremental Net Benefits ð$Þ

192 38 43

15 60 120

177 (22) (77)

All values are expressed in 1999 dollars rounded to the nearest million. Cost estimates are based on a hypothetical example. The results assume a linear dose–response model and low-end hypothetical costs, and use the mean values from the benefits distribution. Fatal cancers only. Source: Raucher et al. (2002).

232

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

typically bear a disproportionately high share of the regulatory costs relative to customers in larger systems that enjoy economies of scale in treatment. While a uniform MCL for small and large systems does provide households with roughly ‘‘equal health protection’’ in terms of exposure, regardless of where they live or the size of the community water system (CWS) that serves them, the cost burden that each bears can be significantly different. In essence, any given MCL is likely to impose much higher costs per unit of risk reduction benefit received by households served in small systems relative to the costs per risk reduction borne in larger communities. This raises a fundamental issue of fairness—should families served by small systems be forced through regulations to pay much higher costs for their risk reduction benefits than do households in larger, more urban settings? When BCA results are cast solely on the basis of national aggregates of benefits and costs, small system benefit-cost impacts are often obscured under the preponderant share of total costs and benefits borne by larger systems. Therefore, BCA results should also be made available on a system size basis, so that the benefit-cost tradeoffs borne by small system customers can be readily evaluated. These may reveal that the cost per unit of risk reduction (e.g., the cost per cancer fatality avoided) may be unreasonably large for households served by small systems. This type of finding can then be used as a basis for considering alternative MCLs, for considering providing supportive compliance funding to small systems, or both. For example, using USEPA’s results for the proposed radon rule, it is evident that the largest impact in terms of systems bearing compliance costs will be in the smallest size categories. These small CWS bear a disproportionate share of the costs, due predominantly to economies of scale in radon removal treatment technologies. And, because there are relatively few individuals exposed in small systems,

Figure 10.1 Percent of CWSs impacted by the proposed radon MCL [source: derived by Stratus Consulting Inc. from USEPA HRRCA (USEPA 1999)].

10.6 MEASURES OF RISK REDUCTION BENEFITS

233

the benefits tend to be disproportionately low. Figure 10.1 reveals the percentages of national costs and benefits borne by CWS according to size categories (populations served), as derived by USEPA in the HRRCA (USEPA 1999). In the two smallest size categories (systems serving between 25 and 500 persons), the customers will bear over 40% of the total nationwide compliance costs, yet realize only 5% of the national benefits. Hence, even if the benefits were roughly equal to costs at the national level, the small system residents would nonetheless be paying a disproportionately high cost per unit of risk reduction relative to their large system counterparts. The issue of environmental justice for small system customers is especially critical when considering how many of the regulation-impacted systems are in the small system size categories. In Figure 10.1, a line is superimposed on the benefitcost shares to indicate the percentage of CWS that will be impacted by the rule. Approximately 37% of the systems affected by the rule are in the smallest size category, bearing 17% of the nation’s compliance costs, but receiving only 1% (or less) of the benefits. When aggregating up to somewhat larger systems, an estimated 75% of the systems directly incurring compliance costs at an MCL of 300 pCi=L are in the two smallest size categories; and 94% of the regulation-impacted CWSs are in the three smallest size categories. Equity concerns need to be taken into explicit consideration when 94% of the impacted systems accrue only about 20% of the national benefits, but bear over 60% of the nation’s compliance costs.

10.6

MEASURES OF RISK REDUCTION BENEFITS

Drinking water regulations generate benefits that are in the form of reduced risks that exposed people will suffer adverse health impacts. In other words, the benefits are thought of as a reduced number of illnesses (morbidity) or deaths (mortality). This raises the question of how one predicts the number of such adverse health effects avoided, and also how one places a dollar value on ‘‘saving lives’’ or ‘‘protecting good health.’’ In other words, the key issues that arise are (1) how to quantify the benefits and (2) how (or whether) to assign monetary values to these benefits, so that benefits can be compared directly to costs. 10.6.1

Quantifying Risk Reduction Benefits

Drinking water regulatory benefits typically are quantified in terms of the number of illnesses or premature fatalities avoided as a result of reducing exposures to contaminants. As such, the quantified estimates are based on a series of analyses that constitute a ‘‘risk assessment.’’ These start with contaminant occurrence (How many systems have this contaminant, and at what concentrations?), exposure assessment (Given a concentration in drinking water, how much does the person actually ingest—what is their ‘‘dose?’’), and dose–response (What level of health risk is associated with the estimated level of exposure or dose?).

234

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

Ultimately, the risk is characterized in some quantitative fashion (to the degree feasible), such as the excess lifetime risk of cancer for a person exposed to a given concentration for a specified duration of time. For example, there may be a 1.0  106 lifetime risk (i.e., a one in a million chance) that someone exposed to contaminant X at a concentration of 10 mg=L would suffer a cancer as a consequence. This result implies that for every million people exposed at this level, there would be one expected excess case of cancer in that population due to the presence of that compound in drinking water. Such results typically are based on several key assumptions, such as having the person consume 2 L of water per day in a 70þ year lifetime of exposure. Whether underlying assumptions are reasonable or representative is often a point of contention between USEPA and stakeholders. Using the preceding example, if a regulation eliminated this risk 106 entirely for one million people, then the benefits would typically be quantified as one cancer case avoided per lifetime (or 0.014 cancers avoided per year, using the typical but questionable assumption that the risk is spread equally over a 70-year lifetime). If the type of cancer is predominantly fatal (low survival=remission rate), then one might portray the result as one life saved. Technically, the 5-year survival rate for a cancer should be used to split the estimated number of cases into those that probably would be fatal and those that would be nonfatal (but serious) health effects. While ‘‘lives saved’’ is a convenient and readily understood metric, it is important to recognize several important facts: 1. There is no identifiable individual whose life is saved. Instead, the outcome reflects what is referred to as a ‘‘statistical life’’ because what the regulation has really accomplished is to reduce a low-level risk that was borne across a large population. In effect, the MCL reduced the risk levels borne by 1 million people, and the estimate of one ‘‘life saved’’ is just a convenient metric for portraying that risk reduction. 2. Because every person is mortal, no regulatory action truly ‘‘saves lives.’’ Instead, MCLs may reduce a number of ‘‘premature fatalities’’ or, put another way, ‘‘increase the life expectancy’’ of someone who might otherwise have suffered a fatal illness. For example, by avoiding a fatal cancer that typically strikes late in life, a rule may add 10–20 years of life expectancy of every predicted statistical life saved. This is typical of lung cancer and bladder cancer, where the typical age of onset is in the mid-60s and mid-70s, respectively. If the typical life expectancy for someone spared one of these cancers would otherwise be 85 years, then each fatal cancer case avoided yields a benefit of 10–20 ‘‘life years saved’’ (LYS). Alternatively, if the fatal cancer avoided was a childhood leukemia, then the expected benefit would be closer to 70–80 LYS per case avoided. Thus, the LYS metric is a useful to way to portray and compare benefits.

10.6.2

Quality-Adjusted Life Years

Quality-adjusted life years (QALYs) provide a measure of composite (or overall) risk, taking into account multiple effects that might be produced by a substance or

10.6 MEASURES OF RISK REDUCTION BENEFITS

235

mixture. The QALY approach starts with a LYS calculation, and then assigns subjective weights to each remaining year of life expectancy to reflect the quality of life under the health state projected for each year; that is, the QALY approach calculates the excess probability of death from a substance and converts it to years of life lost. These years lost are subtracted from the mean lifespan in the population. The remaining years of life are multiplied by a fraction that represents the ‘‘desirability’’ of life under the conditions of chronic, nonfatal, effects produced by the substance. The weights are assigned within the range of zero to one, with a higher score reflecting a higher anticipated quality of life. In essence, the LYS approach is a simplified QALY application, with each year given an equal weight of one regardless of projected health status during that year. In QALY approaches, each health effect contributes in one of two ways: (1) it can adjust the lifespan, decreasing it through premature death, or (2) it can reduce the quality of the remaining years of life. The result is a single number characterizing the equivalent number of effect-free years of life left to an individual as a result of exposures. A lower QALY value represents a greater individual risk of lower health-related well-being (in economic parlance, a lower utility for the individual). The lower the product of the mean QALY and the size of the impacted population, the greater the welfare loss for the population. The QALY approach has been suggested by OMB as a useful way for federal agencies to portray benefits. Although the QALY approach to developing a composite measure of risk is feasible and has been widely applied in the health care field, it remains highly controversial because of conceptual and empirical hurdles. The most notable drawback is the subjectivity in having analysts determine the appropriate measure of ‘‘desirability’’ for a year of life characterized by the presence of a specific, nonfatal, effect. The weights assigned to the desirability of a given year and health status are subjectively assigned by experts, and do not reflect any systematic revealed or stated preference elicited from the affected public. On the other hand, proponents of the QALY approach stress the desirability of having a single metric that can be applied to a wide range of possible changes in health states, including both cancer and noncancer effects. Advocates also point out that while the weights assigned may be subjective, they may be less arbitrary and more informative than the LYS approach, in which each year of longevity gained or lost is assigned an equally value (of one). 10.6.3

Valuing Risk Reduction Benefits

Once the change in risk of adverse health effects has been estimated (e.g., 10 excess statistical cancer fatalities would be avoided annually), a range of valuation techniques can be applied to estimate benefits. The two common measures of monetary value for human health are cost of illness (COI) and willingness to pay (WTP). Assigning monetary values to quantified changes in health risks raises numerous normative and positive issues. For example, some arguments have been aired that life is ‘‘priceless’’ and that public health and environmental policies should not be subject to benefit-cost analyses. However, regardless of how one philosophically views the issue of assigning monetary values to changes in health status, sound

236

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

policy development requires that such a step be taken. At a minimum, an implicit valuation of changes in health risk should be performed, such as in a cost-effectiveness context of the ‘‘cost per change in health risk.’’ 10.6.4

Willingness to Pay: The Value of a Statistical Life

An important observation helps define how to address the thorny ethical, conceptual, and empirical issues associated with ‘‘valuing good health’’ or ‘‘placing a dollar value on lives saved.’’ The key is to recall that MCLs reduce risks, they do not ‘‘save lives’’ or ‘‘improve health’’ per se. Therefore, the key to valuation is to examine how people respond to and reveal their values (or preferences) about risks. Every day, people face a wide range of risks to their health and safety. Some of these risks are borne involuntarily (e.g., exposure to pollutants, or a genetic predisposition to certain types of disease); some risks are confronted by choice (e.g., choosing whether to use tobacco, install smoke detectors, wear a seat belt, ride a motorcycle, or accept a job in a risky occupation). By observing how individuals make choices about the level and types of risks they bear, and the level of cost they incur to reduce risks (or the rewards they receive when accepting increased risks), economists have been able to make clear inferences about the monetary worth people place on risk reductions. There are over 26 published research papers in which economists have developed estimates of the ‘‘value of a statistical life’’ (VSL) by looking at how people state or reveal their willingness to pay (or to accept compensation) for lower (or elevated) risks. Typical studies examine wage rate premiums in risky professions, or consumer behavior in purchasing risk-reducing items. Willingness-to-pay estimates represent monetary measures of the value individuals place on the change in quality of life achieved as a result of a risk reduction. The WTP-based measures are the conceptually appropriate approach, in accordance with well-established and broadly accepted principles of welfare economics. USEPA has reviewed the WTP literature and found a midrange value for VSL of $6.1 million (in 1999 dollars), and the range spans roughly from $1 million to $20 million (USEPA 1997). This USEPA finding is generally accepted as a reasonable interpretation of the literature, although there are important controversial issues how the VSL estimate should be adjusted when applied to drinking water standards (as discussed in several sections below). The VSL concept is suitable for application to MCLs (or other environmental, public health, or safety programs) because the value concept corresponds to the risk reduction context—MCLs reduce low level risks across a large population, and the VSL estimates reflect how people value small changes in low level risks that also are spread across a large population. For example, a VSL estimate may be based on a $610=year wage premium per worker in an occupation where the risk of fatal accident is 1 in 10,000 per year. This means that for every 10,000 exposed workers, one statistical premature fatality is expected each year. Collectively, these workers enjoy a combined wage premium of $6.1 million annually ($610 times 10,000). Thus, the VSL estimate would be $6.1

10.6 MEASURES OF RISK REDUCTION BENEFITS

237

million. This is directly parallel to the risk reduction benefits of an MCL in the sense that the number of fatalities, and the overall dollar value, reflect risk-based ‘‘statistical lives’’ over a large population, and not identifiable individuals. However, other issues are important in how VSL estimates should be applied in the MCL context, such as the timing of the risks, the amount of life extension generated per case, and other attributes of the risks and the impacted populations (as addressed below).

10.6.5

Cost of Illness

Nonfatal risks (i.e., for illnesses) can be more difficult to monetize using a WTP approach than premature fatalities. The literature on WTP values for avoiding morbidity is not well developed, hence most empirical work relies instead on the COI approach. The COI approach estimates medical expenses and lost income due to premature death or illness. Categories of costs usually estimated include hospitalization, emergency room, physician costs, drug costs, the value of time spent being sick (i.e., the value of work lost as a result of illness, and the value of nonwork time spent being sick). This approach places no value on the time or the lives of those who are not in the labor market, nor does it recognize the lost utility to those who suffer pain (i.e., the lost value of being healthy). The attraction of the COI approach is that it can be used, with various modifications, for a wide range of health endpoints, including mortality and morbidity from either chronic or acute exposure. Another attraction is that it is readily understood by policymakers and the general public, and it is a relatively straightforward exercise to obtain defensible estimates of medical costs, lost wages, and so forth. USEPA has developed a report on the COI of specific adverse health endpoints (e.g., bladder cancer), and it is readily available for direct application in benefit studies (USEPA 2001). The principal weakness of the COI approach is that it underestimates the full benefits of a risk reduction. It is an ex poste measure of cost, not the conceptually appropriate ex ante measure of value. COI will understate the value that individuals place on being healthy and alive, and will understate costs for those who are not in a position to secure appropriate medical attention and those who are not part of the paid labor force. It is possible to attempt to correct for this deficiency by crudely adjusting COI estimates to reflect limited empirical evidence on the relationships between COI and WTP estimates. For example, where both types of estimates have been derived for the same health endpoint, WTP has been twice (or more) the COI estimate. Given the limitations of the COI approach, the use of COI is generally recommended only where no WTP estimates are available (e.g., for most nonfatal illnesses), and that analysts clearly identify the results as lower bound estimates of benefits (i.e., that observed costs are being used as a proxy for unobserved values). COI estimates may also be presented side by side with WTP-based estimates (e.g., for valuing changes in mortality risks) to create a lower bound or as part of a sensitivity analysis.

238

10.7

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

BENEFITS TRANSFER TO DRINKING WATER

Benefits transfer (BT) refers to the practice of using empirical results derived from one or more primary research studies, and applying these results to another context or policy. For example, the published literature that develops estimates of VSL from data from occupational settings (based on wage–risk tradeoffs) is usually used to assign monetary values to reductions in fatal risks that accompany the setting of an MCL. The key question to consider is whether (or how) VSL estimates should be transferred to drinking water applications. BT is often used in evaluating public policies because there is rarely the time or budget necessary for conducting primary research efforts that directly apply to the policy in question. For example, it would be very time-consuming and expensive to design and conduct a credible study of the value people place on the risk reductions associated with a potential change in a specific drinking water parameter. For example, a highly credible and applicable research effort using state-of-the art survey or hedonic methods to collect and analyze data from a large sample might take more than $1 million and a year to complete. Because the use of BT is often a practical necessity, there are standard practices and procedures about how BT should be conducted and evaluated. The standards come from the peer-reviewed economics literature [e.g., see Desvousges et al. (1992)] and are embedded in some federal programs for valuing changes in environmental quality [e.g., National Oceanic and Atmospheric Administration’s regulations for natural resource damage assessment (U.S. Department of Commerce 1996)]. While the quality (scientific validity and robustness) of the various pieces of empirical VSL literature is an important matter, the large number of well-regarded studies (e.g., the 26 used in USEPA’s assessment) provides a convincing weight of evidence that the body of work provides a reasonably credible range of values. Instead, in the VSL context, the primary focus is on issues of the applicability of the estimates—which are derived predominantly from occupational and other accidental death contexts—to the issue of risks posed in by drinking water. In terms of the applicability of VSL estimates for the drinking water context, the key issues are that the VSL estimates derived from the published literature are often based on types of risk (e.g., accidental deaths in occupational settings) that affect groups of individuals (e.g., blue-collar employees in their working years) that differ from the types of risks and impacted populations that apply in a drinking water context. Thus, there is a need to consider whether the attributes of the risks and the attributes of the affected populations are similar between the original empirical research effort and the policy application (and if not, whether some adjustments might be feasible and appropriate). For example, changes in drinking-water-related risks may occur many years or decades after a regulation changes exposure levels (e.g., for carcinogens), whereas most VSL studies pertain to accidental deaths that occur immediately. The VSL studies reflect observed values for an average aged person (between the ages of 35 and 40) who might immediately lose all their remaining life expectancy (40–50

10.7 BENEFITS TRANSFER TO DRINKING WATER

239

years). In contrast, the benefits of an MCL for a carcinogen to an average aged person should reflect the values for losing perhaps 15 years of life expectancy, and that potential loss would occur perhaps 30 years in the future. There is good reason to suspect that a 38-year-old reflecting on dying 30 years from today (instead of in 45 years) would assign a lower value to reducing that risk than she or he would to dying immediately. On the other hand, some have pointed out that VSL estimates may be understated if they reflect values for people who have less than average risk aversion or below average education and incomes, and=or if people dread involuntary cancer risks more than other types of risks they can better control personally.

10.7.1

Adjusting VSL

The issues described in the preceding section illustrate why it may be misleading to apply literature-based estimates of VSL directly to the context of changes in the quality of drinking water. This raises the question of whether some adjustment to the VSL estimates may be appropriate. This section discusses the issue of whether and how such adjustments may be necessary. This question was addressed by the USEPA SAB Environmental Economics Advisory Committee (EEAC), and the approach described here draws on (and conforms with) the report issued on the topic (SAB EEAC 2000). The report recommends including latency periods for delayed onset health effects, and discounting the results to present values (using the same discount rate applied to all other benefits and costs). The SAB EEAC also recommends accounting for income growth over the relevant period, and applying a range of income elasticities to reflect the increased WTP for risk reduction that accompanies income growth. The SAB EEAC does not endorse further empirical adjustments to VSL estimates for application to environmental contexts (e.g., to reflect a potential dread factor for cancer), but does suggest conducting sensitivity analyses for potential age-based adjustments to VSLs.

10.7.2

Accounting for Latencies

To date, USEPA analyses of drinking-water-related risk reductions have used the simplifying assumption that risk reductions accrue immediately after reduction in exposure. This assumption skews the benefit analysis significantly by counting cancer cases that are avoided far into the future as cases that are avoided immediately. By assuming ‘‘immediacy’’ of benefits realization, USEPA implicitly assigns a latency period of zero years for cancer risk reductions. Although it may be true that the exact length of the latency period is variable and uncertain, it is widely recognized in the scientific community that latency periods are not zero for most carcinogens. Rather, the latency period is likely to be many years (e.g., 10, 20, or 30 or more years) for typical carcinogenic modes of action. Instead, multiyear latency scenarios

240

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

should be used, with a range of multiyear latency scenarios applied as a mechanism to reflect the degree of uncertainty regarding the latency period. Evidence for accommodating nonzero latency periods in the benefit-cost analysis can be found by examining the age of onset for specific cancer endpoints of relevance. For example, National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) Program data indicate that the mean age of onset for bladder cancer in the United States is after 71, average age at death is 77, and less than 1% of fatal cases arise in people under 45 (less than 5% of the cases occur in people under 55, and 15% in people under 65) (NCI 1998). This type of readily available information can be used, in concert with sound scientific evidence, to develop plausible latency period scenarios for the analysis. 10.7.3

Discounting Costs and Benefits

Introducing scientifically justified latency period scenarios for cancer cases avoided in turn raises the important issue of discounting the future health effects avoided by the proposed regulation. Basic economic welfare theory is quite clear about the appropriateness of discounting future benefits and costs, and the need to treat future benefits and costs in a consistent manner (i.e., applying the same discount rate to all benefits and costs). The SAB EEAC fully agreed with this principle, and stated that future cancer fatalities need to be valued by taking the latency periods into account and discounting the associated VSL estimates back to present value, using the same discount rate as applied to all other costs and benefits (SAB EEAC 2000). There may be some exceptions to this principle in cases where intergenerational equity impacts are under consideration, but this is not the case for most drinking water quality issues. Risk reductions associated with changes in water quality are realized by the same generation of individuals for whom exposures are reduced via the costs incurred (e.g., for additional treatment). OMB has issued guidance on discounting based on a review of the economics literature, and requires that federal agencies discount costs and benefits at a real rate of 7% (OMB 1992). Under the OMB guidance, other discount rates can also be applied in the form of sensitivity analyses, if accompanied by suitable justification. As one alternative scenario, analysts may want to consider that an expert panel assembled by the U.S. Public Health Service (Gold et al. 1996) directly addressed the issue of discounting future health risk reductions, and clearly articulated that it is the conceptually appropriate approach to apply a positive discount rate to such future health risk reduction outcomes. This panel advised using 3%. For another feasible discount rate scenario, analysts should consider the empirical evidence available in the peer reviewed literature on how much people discount future fatal risks (Moore and Viscusi 1988, 1990; Cropper et al. 1994). These studies suggest that, on average, individuals actually apply a discount rate that may be considerably higher than 7% to future fatal risk reductions (e.g., empirical evidence suggests rates of 10 to 12% or higher—perhaps as high as rates charged to credit card balances).

10.7 BENEFITS TRANSFER TO DRINKING WATER

241

10.7.4 Adjusted VSLs to Reflect Latency, Discounting, and Income Growth As noted above, USEPA has been using $6.1 million (1999 U.S. dollars) as the most likely benefits value for each premature fatality avoided, based on the mean estimate the Agency developed from its review of the VSL literature. The $6.1 million estimate does not reflect the latency, discounting, or other factors that would affect this valuation. Such adjustments can be made in accordance with the SAB EEAC report (SAB EEAC 2000). The SAB report states that two adjustments should be made to the VSL estimates: (1) latency and discounting need to be factored into the estimate, as noted above, and (2) VSLs should be adjusted to reflect the impact that projected real income growth might have on the VSLs of future (wealthier) Americans. Age-adjusted VSLs are also suggested by the SAB, but in the form of a sensitivity analysis. In an AWWA Research Foundation (AWWARF) project examining this issue (Raucher et al. 2002), the research team developed Monte Carlo simulations to make both sets of SAB-recommended adjustments, using the USEPA’s Weibull distribution with an estimated mean of $6.1 million for the unadjusted VSL as the starting point. The distribution of adjusted VSL estimates is then developed by applying a Monte Carlo simulation that draws a latency period at random from a uniform distribution, bounded by 10 and 30 years. For each latency period, future income-adjusted VSLs are generated based on estimates of real income growth over that period (U.S. Bureau of Economic Analysis 2000), and then an income elasticity drawn at random from the values recommended by SAB (income elasticities of 0.08, 0.4, and 1.0). Elasticities here refer to the percent change in VSL per percent change in income. Hammitt (2000) suggests that VSL is not very sensitive to income, noting that ‘‘typical estimates suggest that a 1% change in income yields less than a 1% change in VSL—often 0.5% or less.’’ This suggests an income elasticity of 0.5 or less. Finally, for the relevant latency period, the income-adjusted VSL is discounted back to the present value using a discount rate drawn at random from a uniform distribution, bounded by discount rates of 3% and 7%. The adjusted VSLs thus reflect latency, income growth, and discounting. Table 10.3 reveals the adjusted VSL results for three cancer endpoints, based on alternative types of adjustment. The mean value generated for the discounted latency- and income-adjusted VSL using this procedure is $2.7 million per premature fatality (1999 dollars). The median estimate is $2.1 million, and the 25th and 75th percentiles are approximately $1.3 million and $3.4 million, respectively. The standard deviation is $2.2 million. Because the simulations relied on uniform distributions, this mean estimate for discounted latency- and income-adjusted VSL corresponds to a mean discount rate of 5%, applied over a mean latency period of 20 years, with a mean income elasticity of 0.49. If the applicable latency periods tend to be longer and=or if the applicable discount rates tend to be higher, the mean adjusted VSL estimate would be lower than $2.7 million (and vice versa). But these results reveal that use of unadjusted VSL estimates can significantly overstate the benefits of MCLs (and in the case of some cancers of interest, the degree of overstatement is a factor of over 225%).

242

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

TABLE 10.3 Adjusted VSL Estimates (Means, in Millions 1999 Dollars)

Mean EPA-derived estimate (unadjusted VSL from Weibull distribution) SAB EEAC-endorsed adjustments (latency, with income growth and discounting) Age-adjusted VSL sensitivity analysis (age, income growth, latency, discounting) Using VSLY and life years saved (age of onset, latency, discounting)

Bladder Cancer

Lung Cancer

Kidney Cancer

6.1

6.1

6.1

2.7

2.7

2.7

1.3

1.9

2.2

1.2

1.5

1.6

Source: Raucher et al. (2002).

10.8

UNCERTAINTY AND VARIABILITY

One of the largest challenges in conducting and interpreting BCAs for drinking water standards (and many other environmental and public health programs) is that many uncertainties and data gaps inevitably arise in the analysis. This is especially true for the benefits side of the analysis, but also in cost analyses as well. 10.8.1

What are Uncertainty and Variability?

The terms uncertainty and variability have distinct meanings in risk assessment and benefit-cost analysis, and each raises its own issues and can be addressed in different manners. Uncertainty refers to a lack of knowledge, reflecting a gap or unknown in the applicable field of science. A classic uncertainty involves developing dose–response relationships for humans based on laboratory studies using rodents. Toxicologists develop data on dose–response findings from laboratory studies in which rodents are exposed for short periods to high doses of a contaminant. These high-dose experimental results in mice or rats then need to be translated into risk levels in humans who are exposed to relatively low doses in drinking water over a long time period. Because scientific knowledge is far from complete on these matters, these cross-species and high-dose–low-dose extrapolations are not fully understood. Because there are no known answers to these problems, these issues of how to develop or interpret dose–response functions for drinking water exposures reflect considerable uncertainty. Variability reflects differences that exist as a basic state of the world — it pertains to the fundamental reality that each person is different from the next person in relevant ways, and that conditions may differ from place to place or from time to time. Exposure assessments tend to have several important variabilities, including the amounts of water that different people consume (i.e., the mean level of per capita CWS tapwater consumption is approximately 1 L=day, but data suggest that roughly 10% of the population consumes more than 2 L=day).

10.8 UNCERTAINTY AND VARIABILITY

243

Other exposure-related variabilities of interest include the fact that different people will have different durations of exposure (some will reside in the same CWS for 70 years, some for only a few months), different body weights, or different activity patterns (e.g., time away at an office or school with a different water supply). Concentrations of a contaminant in their water may vary over time (e.g., as a result of seasonal variations in source water quality, changes in treatment regimes, distribution system flushing events, a customer’s use of bottled water, or in-home treatment units). Risk assessment and benefit-cost analyses need to reflect these variabilities in some fashion. For example, different people consume different amounts of tapwater on an average day; these differences or variabilities in drinking water consumption are a recognized and unavoidable fact of life. These variabilities can be observed and measured by collecting data from a sample of individuals (whereas uncertainties cannot be addressed by simply observing the world and collecting data). Analysts can thus understand and measure variability by collecting good data and using them in their studies. As data are collected, a probability distribution of values can be derived (e.g., that the mean level of tapwater consumption is approximately 1 L=day, and the 90th percentile is roughly 2 L=day).

10.8.2

Addressing Uncertainties and Variabilities

Uncertainties in components of a BCA typically are addressed through the use of standard assumptions or uncertainty factors. These assumptions and adjustment factors typically are somewhat arbitrary and are intended to be conservative (i.e., erring on the side of being overly protective), because the risk assessor’s primary duty is to determine the level that poses ‘‘zero risk.’’ For example, in determining the MCL for uranium, USEPA relied on kidney toxicity data derived from laboratory animal experiments. A ‘‘no-effect level’’ was observed in the laboratory studies of 60 mg kg1 day1. To translate this rodent-based finding to humans, Agency risk assessors applied an uncertainty factor of 100 when converting the rodent results into the human-oriented safe daily dose of 0.6 mg=kg (i.e., 60 divided by the uncertainty factor of 100). This adjusted no-effect exposure level is called the ‘‘oral reference dose’’ and reflects the dose at which no risks are anticipated in humans, including an ample safety margin. This is how uncertainty is typically addressed for noncarcinogens posing risks from chronic exposure. Variability in this application arises in translating this safe ‘‘oral reference dose’’ for humans (a dose per body weight) into a drinking water concentration (measured in micrograms per liter). Typically, risk assessors remain conservative at this stage as well, and apply a set of high-end assumptions to ensure that the most sensitive and most highly exposed persons are protected. For the uranium example, which is typical, USEPA assumes that a 70-kg person who drinks 2 L of CWS tapwater daily and obtains 80% of his or her total uranium exposure through drinking water. In reality, these standard assumptions reflect upper ends of the variability distributions rather than average or typical conditions. But they are adopted by

244

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

risk assessors to assure safety to even highly exposed and highly sensitive persons in the overall population. On the basis of these exposure assumptions, the uranium ‘‘safe’’ oral reference dose (0.6 mg kg1 day1) is translated into a ‘‘safe concentration in drinking water’’ of 16.8 mg=L [(0.6 mg  80%)=(2 L=day  70 kg)]. The Agency then rounds to 17 mg=L and defines this as the drinking water equivalent level (DWEL) for uranium. In this typical risk assessment application for a systemic (noncarcinogenic) compound, variability is addressed through a simple set of assumptions that apply the high-end values observed from the underlying variability distributions.

10.9 PRECAUTIONARY ASSUMPTIONS VERSUS CENTRAL TENDENCIES The preceding discussion described how uncertainty and variability typically are addressed in risk assessments, wherein standard conservative assumptions are used to address both uncertainty and variability. This use of conservative assumptions and uncertainty factors is related to the ‘‘precautionary principle’’ and is suitably used in the risk assessment context to ensure that the designated ‘‘safe’’ concentration (DWEL) is indeed risk free for even the most highly sensitive and highly exposed individuals, with an ample margin of safety. Although the use of precautionary assumptions is well established and accepted in the context of a risk assessor’s mission to define a concentration that is ‘‘safe,’’ the use of these same assumptions is not suitable for a BCA. This is because the BCA should reflect the most accurate or typical conditions. In other words, the risk assessment process is designed to err on the side of safety, but the BCA is intended to reflect an evaluation of what health risk reduction benefits can be expected from a potential MCL. When conservative assumptions and adjustment factors are applied throughout a BCA, the effects become compounded, and this results in risk reduction benefits estimates that will be greatly exaggerated. For example, the chance that the persons consuming the 90th percentile volume of water per day (i.e., roughly 2 L) are also the same persons that are exposed for the 99th percentile duration (e.g., assume 70 years) is only 1 in 1000 (10.%  1.0% ¼ 0.1%), assuming that the events are statistically independent. If one used just these two conservative assumptions alone, one would have a benefit estimate that reflected the 99.9th percentile of the likely benefits distribution. Adding in uncertainty factors for the dose–response relationship would drive the results further toward extremely low probability outcomes. Several important entities have recognized the critical distinction between using the precautionary assumptions in risk assessment and using more realistic central tendencies (or whole distributions) when conducting BCAs as part of a risk management process. USEPA has acknowledged this in several recent rulemaking packages, and the U.S. General Accounting Office (USGAO) issued an excellent report on this issue as well (USGAO 2000).

10.9 PRECAUTIONARY ASSUMPTIONS VERSUS CENTRAL TENDENCIES

245

TABLE 10.4 Impact of Exposure-Related Assumptions on Benefit-Cost Analysis Factor

Impact Relative to Central Estimate

(a) Daily tapwater consumption (2 L=day) (b) Duration of exposure (70þ years) Occurs in 1% of systems Occurs in 5% of systems Occurs in 10% of systems (c) Combined impact in lifetime exposure estimate Occurs in 1% of systems Occurs in 5% of systems Occurs in 10% of systems

1.8

12.4 8.7 6.1 22.4 15.7 11.0

Source: Raucher (2001).

USEPA has made some efforts to use averages and plausible ranges in its most recent BCAs, especially for drinking water consumption. However, precautionary elements still persist throughout the analyses, including the use of overly conservative dose–response information and lifetime durations of exposure. This problem can be addressed using some standard methods, including simple sensitivity analyses and more elaborate probabilistic assessments (such as applying Monte Carlo techniques). An AWWARF project provides a detailed illustration of these issues (Raucher et al. 2002). Tables 10.4 and 10.5 also indicate the extent to which various precautionary assumptions can individually and collectively impact a BCA. TABLE 10.5 Impact of Cancer Risk Assessment Assumptions on Benefit-Cost Analysisa (a) Use of linear dose–response function (relative to suitable nonlinear alternative) MTBE illustration (at mean) Arsenic illustration (repair model) (b) Use of 95th upper confidence limit (relative to maximum likelihood) (c) Combined illustrative impact [if both (a) and (b) are relevant] (d) Impact when combined with exposure illustration (Table 10.4) a

12.8 3–5 2–3 6–38.4 66–860

Note that results are case-specific, depending (for example) on degree and type of nonlinearity over relevant exposure range, and difference between high-dose data points and low doses of regulatory relevance. Source: Raucher (2001).

246

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

10.10

OMITTED OR UNQUANTIFIED BENEFITS AND COSTS

Another challenge in developing and interpreting BCAs arises when an important benefit or cost cannot be readily quantified or expressed in monetary terms. For example, the principal health risk benefit underlying the December 2000 uranium standard is kidney toxicity. The level of renal toxicity risk is highly uncertain and therefore cannot be quantified (i.e., there is no way to estimate a projected number of disease cases avoided). In such a circumstance, benefits cannot be directly compared to costs. When potentially important benefits (or costs) cannot be directly included in a quantitative BCA, the analyst has several options. An unsatisfactory option is to ignore the omitted benefits or costs, and base the decision only on those benefits and costs that can be included. This is undesirable because if important benefits are left out, then an MCL will not be set as stringently as it should. Likewise, if important costs are omitted, then the BCA would suggest an MCL that is overly stringent. Another unsatisfactory option is to use the existence of an unquantified benefit (or cost) as an excuse to set an MCL at a level that probably does not maximize net benefits. Even though an unquantified benefit may be important and should not be overlooked, it should not be used ‘‘carte blanche’’ as an excuse to set an overly stringent MCL (and vice versa, for an omitted cost). While a potentially significant unquantified (or unmonetized) cost or benefit should not be ignored, nor should it be afforded undue weight and influence. To determine how much weight should be given to considering an unquantified benefit or cost, several informative and appropriate options can be explored to try to include the omitted (nonmonetized) benefits or costs within the BCA framework in as useful and objective a manner as possible. In some cases, this will simply entail providing a good qualitative discussion of the unquantified outcome so that decisionmakers can take it into account along with the numeric BCA findings. If benefits already exceed costs, then a qualitative discussion of nonmonetized benefits only helps reinforce the obvious outcome (and the same is true if the omitted component is a cost and the monetized net benefits are already negative). Where the omitted element might alter the net benefit result (e.g., an important benefit is omitted where the monetized BCA components yield a negative net benefit), a ‘‘breakeven’’ analysis may be useful. This is a semiquantitative approach in which the analyst backcalculates from the estimated net benefit how large the value of the omitted benefit (or cost) would need to be for benefits and costs to be equal (net benefits are zero). This is sometimes referred to as ‘‘implicit valuation.’’ For example, if monetized benefits exceeded costs by $100 million, then the nonmonetized benefit would need to be worth at least $100 million for the BCA to ‘‘break even.’’ It may be quite obvious that the omitted benefit is (or is not) likely to be worth this amount of money. In the uranium example, the implicit valuation outcome for the unquantified benefit was that the ‘‘cost per person exposed’’ (but not necessarily having any adverse health effect) would have to be worth at least $100,000 for the incremental benefits to be at least as great as the incremental

10.11 UNCERTAIN COSTS

247

costs of the MCL. A policymaker or stakeholder can then judge whether it seems likely that such an expense is warranted.

10.11

UNCERTAIN COSTS

While most of the discussion in this chapter has been about issues and uncertainties in estimating benefits, the estimated cost of compliance can also be highly uncertain and controversial. Costs can be highly uncertain for a variety of reasons, including differences in whether or how treatment plant retrofit issues are addressed or how estimates account for difficulties or lack of economies in treatment at a small scale. Other areas that cause differences in cost estimates include how to account for the potential scarcity of skilled operators, reflect limited space and access for expanded treatment facilities (especially at decentralized wells), or consider residuals management issues associated with water treatment-generated wastestreams. Given these and other issues, it is not uncommon for cost estimates developed by some parties to be to be far lower (e.g., one-tenth the level) or higher than estimates predicted by other stakeholders. In this case, it is probably prudent to show BCA results using two alternative cost scenarios. This is a simple sensitivity analysis of how (or whether) the policy implications drawn from the BCA are altered, depending on which cost estimates are applied.

10.12

FUTURE OUTLOOK

BCA is now mandated under the SDWA, and the use of BCA enables standard setting for MCLs to move away from technical feasibility considerations alone. The BCA provisions enable policymakers to consider whether the mandated investment in compliance expenditures will produce suitable returns in the form of public health benefits. This is an important step forward. Since BCAs now play an important role in MCL determinations, it is vitally important that they be performed and interpreted properly. BCAs should adhere to established economic principles by trying to maximize net social benefits. This implies that the analyses should focus on comparing incremental benefits to incremental costs, benefits should be valued according to willingness to pay (as feasible), benefits transfer needs to be done carefully, and VSLs need to be adjusted to reflect latency periods and discounting where applicable. It also is important that uncertainties and variabilities be recognized and considered (e.g., through the use of sensitivity analyses or other, more sophisticated techniques such as Monte Carlo analysis). BCAs should be based on average or typical parameter values and not a compilation of exceedingly conservative assumptions (such as traditionally used when the precautionary principle is applied to risk assessment). Unquantified benefits and costs also need to be considered in a fair and systematic manner (and neither be ignored nor given undue weight). BCA results

248

BENEFIT-COST ANALYSIS AND DRINKING WATER REGULATION

should also be portrayed according to system size, so that small system impacts can be more fully discerned. There are important uncertainties and challenges associated with applying BCA to the drinking water context. BCAs will often be inexact and must be performed and interpreted with due caution. Accordingly, BCAs generally should not be used as a strict decision rule. Nonetheless, BCA is a very useful tool for helping ensure that America’s investment in MCL compliance costs will yield suitable public health returns. When conducted and interpreted in a sound and objective manner, BCAs will be very informative in guiding the nation’s drinking water investments so that they generate the greatest public health returns possible. BCA as applied to drinking water standard setting is a new and evolving practice, and it probably will take time and debate over alternative policy interpretations before this tool becomes more fully integrated into the regulatory policy framework.

ACKNOWLEDGMENTS An earlier and more extensive version of this chapter was prepared as a ‘‘White Paper’’ with the generous support of the National Rural Water Association. All errors and omissions are those of the author.

REFERENCES Cropper, M. L., S. K. Ayded, and P. R. Portney. 1994. Preferences for lifesaving programs: How the public discounts time and age. J. Risk Uncert. 8:243–265. Desvousges, W. H., M. C. Naughton, and G.P. Parsons. 1992. Benefit transfer: Conceptual problems in estimating water quality benefits using existing studies. Water Resources Research 28(3):675–683. Gold, M. R., J. E. Siegel, L. B. Russell, and M.C. Weinstein. eds. 1996. Cost-Effectiveness in Health and Medicine. New York: Oxford Univ. Press. Hammitt, J. K. 2000. Valuing Lifesaving: Is Contingent Valuation Useful? Harvard Center for Risk Analysis. Risk in Perspective, 8(3). Moore, M. J. and W. K. Viscusi. 1988. The value of changes in life expectancy. J. Risk Uncert. 1:285–304. Moore, M. J. and W. K. Viscusi. 1990. Models for estimating discount rates for long-term health risks using labor market data. J. Risk Uncert. 3:381–401. NCI. 1998. What You Need to Know about Bladder Cancer. National Cancer Institute (http:==cancernet.nci.nih.gov=wyntk_pubs=bladder.htm; accessed July 21, 2000). OMB. 1992. Guidelines and Discount Rates for Benefit-Cost Analysis of Federal Programs. Circular A-94. Washington, DC: Office of Management and Budget (http:==www.whitehouse.gov=omb=circulars=a094=a094.html; accessed May 24, 2001). Raucher, R. 2001. Blending Science with Policy: Precautionary Assumption and Their Impact on Benefit-Cost Analyses and Drinking Water Standards. White Paper. Duncan, OK: National Rural Water Association.

REFERENCES

249

Raucher, R., A. M. Dixon, J. A. Drago, and E. Trabka. 1994. Evaluating the cost-effectiveness of the federal drinking water regulatory program. J. Am. Water Works Assoc. 86(8):28–36. Raucher, R., D. Burmaster, D. Crawford-Brown, and A. Cox. 2002. Quantifying Public Health Risk Reduction Benefits. Denver: American Water Works Association Research Foundation. SAB EEAC. 2000. An SAB Report on EPA’s White Paper Valuing the Benefits of Fatal Cancer Risk Reductions. EPA-SAB-EEAC-00013. Washington, DC: U.S. Environmental Protection Agency, Science Advisory Board, Environmental Economics Advisory Committee. U.S. Bureau of Economic Analysis. 2000. Regional Accounts Data—Projections to 2045. (http:==www.bea.doc.gov=bea=regional=project=projlist.htm; accessed Aug. 30, 2000). U.S. Department of Commerce. 1996. Natural Resource Damage Assessments, Final Rule. 15 CFR Part 990. Fed. Reg. 61(4) (Jan. 5, 1996). Washington, DC: National Oceanic and Atmospheric Administration. USEPA. 1997. Benefits and Costs of the Clean Air Act 1970–1990. Report to Congress. Washington, DC: USEPA Office of Air and Radiation. USEPA. 1999. Health Risk Reduction and Cost Analysis for Radon in Drinking Water: Notice, Request for Comments, and Announcement of Stakeholder Meeting. Pre-publication copy of Federal Register Notice for Review and Public Comment. EPA-815-2-99-002. Washington, DC: U.S. Environmental Protection Agency. USEPA. 2000. Guidelines for Preparing Economic Analyses. EPA 240-R-00-003. Washington, DC: U.S. Environmental Protection Agency. USEPA. 2001. Cost of Illness Handbook. Prepared by Abt Associates, Inc. Washington, DC: USEPA Office of Pollution Prevention and Toxics (http:==www.epa. gov=opptintr=coi_handbook). USGAO. 2000. Use of Precautionary Assumptions in Health Risk Assessments and Benefits Estimates. Report to Congressional Requesters. GAO-01-55. Washington, DC: U.S. General Accounting Office.

11 PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

11.1

INTRODUCTION

The Safe Drinking Water Act (SDWA) mandate an ambitious agenda for program implementation and issuance of drinking water regulations. Public participation in the development and implementation of the act is essential for the partnership ethic embodied in the act to be realized. To facilitate public involvement, the U.S. Environmental Protection Agency (USEPA) Office of Ground Water and Drinking Water (OGWDW) has committed to an open process of stakeholder meetings to inform and involve interested parties in various rulemakings. Since enactment of the 1996 SDWA Amendments, the number of public stakeholder meetings on various regulatory issues has increased dramatically (Fig. 11.1). Stakeholder meetings are expected to remain at a high level throughout implementation of the 1996 SDWA Amendments. This chapter reviews the mechanics and opportunities for stakeholders to become involved in the development of drinking water regulations.

11.2

WHO IS THE PUBLIC?

The concept of public participation can be misleading because the erroneous impression is given that the ‘‘public’’ is a single homogeneous, identifiable entity. In reality, however, the ‘‘public’’ is a diverse mixture of individuals, groups, and organizations, Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

251

252

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

Figure 11.1 Trend in public stakeholder meetings on drinking water issues announced in the Federal Register.

each with different interests and differing points of view. The term ‘‘public’’ or ‘‘publics’’ is used here as a generic term referring broadly to groups, individuals, organizations, and entities with an interest in drinking water. ‘‘Stakeholders’’ generally refer to those people, organizations, and=or groups that are directly affected by an agency rulemaking or that have a direct interest in the outcome of a rulemaking. USEPA’s Science Advisory Board (SAB) has noted that ‘‘the definitions of the terms ‘stakeholder’ and ‘stakeholder processes’ have become highly elastic’’ (SAB 2001). These terms are used to refer to any interaction by USEPA with groups outside the agency, or even to involvement of people within the agency. ‘‘Stakeholders’’ may be experts, nonexperts, or semiexperts from government agencies, nongovernmental organizations, corporations, citizens, and other private parties. Diverse ‘‘stakeholders’’ desire to have input regarding how new USEPA drinking water regulations are fashioned. Certain procedures are necessary to ensure that all who would like to participate and provide input to the agency can participate on an equal basis, or nearly equal basis. Although mechanisms exist by which stakeholders can participate in regulation development, the influence or effect will vary, depending on the public involvement opportunities provided by USEPA and the resources available to the particular individual or organization. 11.3

OBJECTIVES DETERMINE INVOLVEMENT LEVEL

A number of mechanisms may be used by USEPA and other governmental agencies to encourage and allow public involvement. The mechanism used is determined by the agency’s objective and the degree of public involvement desired. Table 11.1 summarizes the principal opportunities and their characteristics, advantages, and limitations. Stakeholders must understand and take advantage of the various mechanisms if their efforts to shape future rules is to be effective. Providing a fair and equitable means for everyone with an interest in a regulation to have an opportunity to participate in its development is one of the purposes of the

11.4 INVOLVEMENT DURING THE RULEMAKING PROCESS

253

TABLE 11.1 Levels of Public Involvement in Regulation Development Level

Interaction

1. Stakeholder power

Stakeholders act without communicating with government agency; government agency may react

2. Shared power

Stakeholders and government agency solve problems together on equal basis Government agency asks stakeholders for meaningful input and intends to listen and act on input received but retains power to reject input Government agency asks stakeholders for limited input but would prefer not to listen at all; government agency retains power to reject stakeholder input and may have little incentive to accept it Government agency presents information and decisions to stakeholders; stakeholders listen and react Government agency acts without communicating at all with stakeholders

3. Stakeholders consult

4. Stakeholders talk

5. Government talks

6. Government power

Example Citizen investigations, public protests over government agency action or inaction, industry chooses to meet standards more or less stringent than the government’s Cooperative projects, joint problem solving, negotiated rulemaking Some advisory committees, informal meetings, ongoing dialog, some public hearings

Typical of public hearings, request for comments on some proposed rules, pro forma meetings, and advisory committees

Some public meetings, promulgated rules, press releases and government publications, some legal and enforcement actions Some investigations, some rulemakings, some legal and enforcement actions

Administrative Procedure Act (APA), which defines minimum procedural requirements for federal agency rulemakings. Public participation in programs under the SDWA, including minimum requirements and suggested program elements are defined in the Code of Federal Regulations, Title 40, Part 25 (CFR 2002). 11.4

INVOLVEMENT DURING THE RULEMAKING PROCESS

There are several points during the development of drinking water regulation whereby the public may participate and provide input to regulatory decisions (see Table 11.2). The key points are prior to rule proposal, during rule proposal, following rule proposal, and after rule promulgation.

254

Private meetings with USEPA

Comments submitted to USEPA

Event=Action

Stakeholders may submit information, comments, and requests on regulatory issues to USEPA at anytime; most effective if information or comments are presented to key agency staff within the window of time that the agency is considering the particular issue, prior to internal agency closure Private meetings of one or more groups with USEPA may be formal or informal, and may or may not involve dialog or consensus building; most effective if the issues and agenda are clearly defined, with adequate preparation and documentation by all parties involved

Characteristics

TABLE 11.2 Public Involvement Opportunities Limitations USEPA may not be able to address the issue(s) raised immediately because of timing, resource limitations, or statutory requirements

Potential for dialog or consensus may be hindered by the absence of other stakeholders; no opportunity to develop support of stakeholders not present; usually limited following rule proposal; may be difficult to arrange or not possible because of scheduling conflicts or perceived favoritism from other stakeholders

Opportunities Stakeholders need not feel constrained to approach the agency with regulatory concerns or issues; can be used to alert USEPA to new issues or problems needing attention or resolution

Opportunity to clearly present to USEPA key issues without the potential for disruption from other stakeholders; potential for dialog, exploration of alternatives and solutions, and development of consensus on possible approaches or solutions; most effective as part of an open stakeholder process prior to formal rule proposal

255

Oral testimony at a public hearing

Proposed rule comment period

The traditional opportunity for the public to present written views and supporting technical information on a proposed rule; most effective if data collection performed in advance so that comments and recommendations have the strongest technical foundation possible; comments must be submitted to the designated address by the specified deadline The traditional opportunity for the public to present formal oral comments on a proposal; most effective if key points are presented succinctly, with supporting data; effectiveness depends on how well the statement is delivered as well as its content; registration for the public hearing is usually required Travel required to the location of the public hearing; advance notice may be relatively short; preparation required to ensure that the oral statement presented is effective; poor oral presentation or inadequate documentation may detract from the persuasiveness of the statement Opportunity to present views to the agency and other stakeholders in attendance; key points can be reinforced without getting lost in voluminous written comments submitted on a proposal; potential for interaction with other stakeholders to explore alternatives and develop consensus

(continued )

Poor written explanation or inadequate documentation may detract from the persuasiveness of the comments

Thoughtful consideration of the issues is possible out of the public eye; more issues can be addressed and more documentation provided compared to oral testimony; opportunity to place key information and documentation into the rulemaking record

256

(Continued)

NDWAC meeting

FACA advisory committee

Stakeholder meeting participation

Event=Action

TABLE 11.2

Open public meeting held by the agency to brief and receive input on a regulation or regulatory issues; meeting typically facilitated by a neutral third party; stakeholders may provide oral comments, written comments, or both Formal advisory committee under FACA to address a regulatory issue or need; meetings are generally announced and open to the public; meetings generally include time for public comments during which stakeholders may present information and recommendations to the committee for consideration Formal NDWAC meetings and=or NDWAC work groups are held regularly; meetings are generally announced and open to the public; meetings generally include time for public

Characteristics

Issues will be considered by the advisory committee whose membership is balanced to represent all relevant interests; some opportunity for consensus between diverse stakeholders

Potential for dialog or consensus may be hindered by the presence of other stakeholders

Opportunity for clear presentation to USEPA regarding key issues; potential exists for dialog, exploration of alternatives and solutions, and development of consensus on possible approaches or solutions Issues will be considered by the advisory committee whose membership is balanced to represent all relevant interests; opportunity for consensus between diverse stakeholders

In general, the agency is not bound to accept the advisory committee’s recommendations

In general, the agency is not bound to accept the advisory committee’s recommendations

Limitations

Opportunities

257

Regulatory negotiation (reg-neg)

SAB Drinking Water Committee Meeting

comments during which stakeholders may present information and recommendations to the committee for consideration Formal SAB DWC meetings and SAB executive committee meetings are held regularly; meetings are generally announced and open to the public; meetings generally include time for public comments during which stakeholders may present information and recommendations to the committee for consideration Formal FACA advisory committee charged to negotiate the provisions of a proposed rule; meetings are generally announced and open to the public; meetings generally include time for public comments during which stakeholders may present information and recommendations to the committee for consideration In general, the agency is not bound to accept the SAB’s recommendations

Not all regulatory negotiations succeed, depending on the issues involved and the stakeholders involved

Issues will be considered by the SAB; some opportunity for consensus between diverse stakeholders

Issues will be considered by the reg-neg committee whose membership is balanced to represent all relevant interests, including USEPA; opportunity for consensus between diverse stakeholders

258

11.4.1

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

Involvement Prior to Rule Proposal

Prior to issuance of a formal proposed rule, the agency has considerable flexibility in meeting with the public and discussing elements of an anticipated regulation. If time allows, USEPA may seek public comment prior to a formal proposal through one of several mechanisms. The agency could issue an Advance Notice of Proposed Rulemaking (ANPRM) in the Federal Register. The ANPRM provides an early warning that a regulation is being considered or will be developed. Public comments on particular aspects of the new rule are typically requested by the agency. Another mechanism used to solicit public input early in the rulemaking process is the issuance of draft discussion documents or a draft rule. These documents are usually not published in the Federal Register, although a notice of their existence is usually published on the USEPA OGWDW Website (www.epa.gov/safewater/ new.html). The draft may also include preliminary regulatory language and related issue papers that discuss various aspects of the draft rule. Stakeholder meetings, workshops, conferences, and public meetings sponsored by USEPA and others provide opportunities for the agency to discuss issues regarding new regulations with its publics. Stakeholder meetings facilitated by a neutral third party are frequently held on drinking water regulatory issues. Private informal meetings can also be held between USEPA and individuals or representatives from organizations to discuss technical data and regulatory options. USEPA may also issue status reports on a rule as a means to initiate public discussions. USEPA has total discretion as to whether to engage the public or any particular group in discussions prior to issuance of a proposed rule. Issuance of an ANPRM or a draft rule, meetings with specific groups, and participation in workshops and conferences are optional. The Agency has the freedom to not discuss the substance of a new rule with anyone and to develop a proposed rule totally on its own. The Agency could decide to meet with representatives of a specific organization but choose not to discuss any substantive issues regarding a specific future proposed rule. The Agency has discretion in determining the degree to which it will interact with stakeholders, and any specific organization, prior to rule proposal. No legal restrictions exist that prevent the Agency from meeting with specific organizations to discuss the elements of an anticipated rulemaking. However, certain perceptions will influence how the Agency interacts with interest groups and professional organizations. Not being perceived as showing favoritism to a particular stakeholder, for example, is considered particularly important when publics have irreconcilable differences on issues related to a rulemaking. The tendency in these cases, however, is for the Agency to become isolated, allowing it to be the final decisionmaker on issues of public dispute but at the same time hindering public involvement. This can foster an adversarial relationship between the Agency and various publics. The public can submit information, data, or ideas by writing to the Agency at any time. Consideration given to such letters depends mostly on the agency’s internal schedule. If time allows, information submitted to the agency, either from a meeting or by letter, is usually considered. If the agency has completed its internal delibera-

11.4 INVOLVEMENT DURING THE RULEMAKING PROCESS

259

tions on a proposal, referred to as ‘‘closure,’’ then the information may not be considered.

11.4.2

Involvement during Rule Proposal

The formal public comment period required by the APA is initiated by publication of the proposed rule in the Federal Register. The proposal will state the length of the public comment period, which ‘‘must be sufficient to fairly appraise interested parties of the issues involved, so that they may present responsive data or argument’’ (APA 1946). USEPA must inform the public of the data and assumptions on which the Agency’s proposal is based. Those interested must have an opportunity to ‘‘challenge the factual assumptions on which [the Agency] is proceeding and to show in what respect such assumptions are erroneous’’ (Congress 1974). In practice, the public will not have access to USEPA supporting documentation until the proposed rule is published in the Federal Register. Comments submitted by the public must be postmarked by the last day of the public comment period designated in the notice to be considered as part of the record for the rulemaking. The comment period is typically 30, 60, 90, or 120 days, depending on the complexity of the rulemaking. Comments received are placed in the drinking water docket, where they are available by appointment for public inspection and reproduction. Comments received after the close of the comment period are also placed in the docket and may be considered by the agency if time allows but are not considered as part of the rulemaking record that would be the subject of judicial review. Following the close of the comment period preceding promulgation of a new rule, the agency considers the comments received on the proposal and prepares a response. Technical support documents are updated and revised on the basis of comments received, and a final rule is prepared and issued.

11.4.3

Involvement after Rule Proposal

Open discussion with stakeholders is usually limited by USEPA after a proposed rule is signed by the administrator, at which time the agency is considered to have an official position. The primary concern regarding interaction with publics after administrator signature has to do with the possibility of the agency being influenced by ex parte communications. This refers to ‘‘off-the-record’’ communications between agency decisionmakers and other parties during informal rulemaking. Examples of ex parte communications might include special meetings of USEPA staff with an environmental group or water utility representatives, individual contacts, and even private telephone conversations. In the context of informal rulemaking, a communication is ex parte when it is an off-the-record, private discussion between agency decisionmakers and others concerning the substance of the agency’s proposed rule. As a result, public hearings held during proposed rule comment periods are typically very formal, with USEPA attorneys present, and a transcript is prepared for the rulemaking record.

260

11.4.4

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

Ex Parte Communications

The APA places no restrictions on ex parte communications made in informal rulemakings (ACUS 1991). In fact, the concept of ‘‘parties’’ and ‘‘off-the-record’’ contacts do not really apply in informal rulemaking under APA Section 553, because participation in informal rulemakings is not limited to named parties. Ex parte communications also are not restricted by the SDWA. Early court cases held that ex parte communications in informal rulemakings were improper (Home Box Office 1977). Later cases clarified the legality of ex parte communications in informal rulemakings (ACUS 1991, Hercules 1978, Sierra Club 1981). The importance to effective regulation of continuing contact with stakeholders cannot be underestimated. Informal contacts may enable the agency to win needed support for its program, reduce future enforcement requirements by helping those regulated to anticipate and shape their plans for the future, and spur the provision of information which the agency needs. Although court decisions allow ex parte communications in SDWA rulemaking, USEPA is very sensitive to unfavorable criticism that could occur if it were to base its decisions on private discussions with individuals outside the agency. Critics note that decisionmakers may be influenced by communications that may be made privately, which is counter to the need for accountable government. This is a particular concern when certain groups or individuals are allowed privileged access to upper-level management and decisionmakers at USEPA. Ex parte communications are feared to reduce the incentive to submit thoughtful comments or carefully prepared supporting data for people who believe that other interests have privileged access to decisionmakers. On the other hand, nonpublic candid contacts between the agency and interested parties can be useful in working out tentative and compromise solutions. Comments and supporting data must be presented to the agency at the time the rule is proposed in order to be considered as part of the rulemaking record. Comments and data submitted after the close of the comment period may be considered by the agency but will not necessarily be included as part of the rulemaking record that would be the subject of judicial review if the final rule were challenged in court. If a party was successful at persuading the agency to adopt a certain decision, the basis for the decision must be in the rulemaking record. Otherwise, the agency must open the record to make available for comment any new information on which its decision may be based. If the decision resulted in a final rule that differed substantially from the proposal, an additional notice and comment period would be needed under most circumstances (ACUS 1991). A USEPA administrator’s memorandum of May 31, 1985, instructs agency employees to ‘‘be certain (1) that all written comments received from persons outside the Agency (whether during or after the comment period) are entered in the public record for the rulemaking [and] (2) that a memorandum summarizing any significant new factual data or information likely to affect the final decisions received during a meeting or other conversations is placed in the public record’’ (ACUS 1991). This practice has been reaffirmed since that time and continues to be agency policy.

11.5 FEDERAL AGENCY ADVISORY COMMITTEES

261

In general, the OGWDW does not release information to one group outside the agency without making it available to other groups. Selected technical regulatory information might be released early at the appropriate time, such as technical information on cost, occurrence, and treatment technology. However, internal management recommendations and policy issues and options are generally not released until an agency management decision has been reached. A distinction is generally made between discussion of technical information and policy issues. The agency is generally willing to work with stakeholders to develop the best factual, technical information base for regulatory decisions but does not include the public in formulation of agency policy. When new information or arguments are received, issuance of a Notice of Data Availability (NODA) and reopening the comment period is the only mechanism available to ensure that everyone with the desire to has an opportunity to comment. If the agency is under a court-ordered deadline, flexibility may not exist to reconsider a proposal unless the new data or arguments presented are compelling enough for USEPA to convince the court to extend the schedule so that the record can be reopened.

11.5

FEDERAL AGENCY ADVISORY COMMITTEES

When formal public involvement is desired in the development of a particular regulation or guidance, USEPA may form an advisory committee. In general, an agency of the federal government is required to comply with the requirements of the Federal Advisory Committee Act (FACA) (FACA 1996) (see also Table 11.3) when it establishes or uses a group that includes nonfederal members as a source of advice. Under FACA, an advisory committee is established only after both consultation with the General Services Administration (GSA) and receipt of a charter from the agency forming the advisory committee. The primary function of an advisory committee is to assist elected or appointed officials by making recommendations to them on issues that the decisionmaking body considers relevant. These issues may include policy development, project alternatives, financial assistance applications, work plans, major contracts, interagency agreements, and budget submissions, among others. Advisory groups can provide a forum for addressing issues, promote constructive dialog among the various interests represented on the group, and enhance community understanding of the Agency’s action. When USEPA establishes an advisory committee, provisions of the FACA and GSA Regulations on Federal Advisory Committee Management must be followed. These requirements are:  The development of a charter that has been approved by the GSA and Office of Management and Budget (OMB). It must contain the committee’s objectives and the scope of its activities, the period of time necessary for the committee to carry out its objectives, the agency responsible for providing the necessary

262

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

TABLE 11.3 Administrative Statutes and Executive Orders Affecting Public Participation in USEPA Rulemaking Federal Advisory Committee Act (FACA) Governs the establishment of and procedures for advisory committees that provide advice or recommendations to the federal government. Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act (SBREFA) Generally requires agencies to assess the impacts on small entities, including small businesses, small governmental jurisdictions, and small organizations, of rules subject to notice and comment rulemaking requirements. For rules that may impose significant economic impacts on a substantial number of small entities (SISNOSE), agencies must prepare a regulatory flexibility analysis of the potential adverse economic impacts of small entities, participate in a Small Business Advocacy Review Panel (propose rule stage), and prepare a Small Entity Compliance Guide (final rule stage). For rules that impose a SISNOSE, public participation requirements include: opportunity for public comment on the agency’s initial regulatory flexibility analysis; opportunity for participation by small entities through the reasonable use of techniques including, among other things, open conferences, public hearings, and solicitation and receipt of comments over computer networks; and solicitation of advice and recommendations from small entity representatives identified by the agency after consultation with the Chief Counsel for Advocacy of the Small Business Administration. Unfunded Mandates Reform Act of 1995 (UMRA) Generally requires agencies to assess the effects on state, local, and tribal governments and the private sector of rules subject to notice and comment rulemaking requirements. Public participation requirements include: for rules containing significant federal intergovernmental mandates, agencies must develop an effective process to allow elected officials of state, local, and tribal governments (or their designated, authorized employees) to provide meaningful and timely input in the development of the regulatory proposal; and for rules that may significantly or uniquely affect small governments, agencies must develop a small government agency plan that provides for notifying potentially affected small governments, enabling officials of small governments to have meaningful and timely input in the development of regulatory proposals with significant federal intergovernmental mandates, and informing educating, and advising small governments on compliance with regulatory requirements. Executive Order 12898, ‘‘Federal Actions to Address Environmental Justice in Minority Populations and Low-Income Populations’’ Generally requires each federal agency, to the greatest extent practicable and permitted by law, to make achieving environmental justice part of its mission by ensuring meaningful public participation of minority and low-income populations, including identifying potential effects and mitigation measures, and improving accessibility of public meetings, documents, and notices to affected communities. Executive Order 13175, ‘‘Consultation and Coordination with Indian Tribal Governments’’ Requires most federal agencies to develop and utilize an effective process that allows elected officials and other representatives of Indian tribal governments to provide meaningful and timely input on regulations, legislative comments, proposed legislation, and policies that have substantial direct effects on one or more Indian tribes, and to appoint a federal official to oversee the implementation of that process.

11.5 FEDERAL AGENCY ADVISORY COMMITTEES

TABLE 11.3

263

(Continued)

Executive Order 12866, ‘‘Regulatory Planning and Review’’ Whenever feasible, agencies must seek views of appropriate state, local, and tribal officials before imposing regulatory requirements that might significantly or uniquely affect those governmental entities. Each agency must assess the effects of federal regulations on state, local, and tribal governments, including specifically the availability of resources to carry out those mandates, and seek to minimize those burdens that uniquely or significantly affect such governmental entities, consistent with achieving regulatory objectives. In addition, as appropriate, agencies must seek to harmonize federal regulatory actions with related state, local, and tribal regulatory and other governmental functions. Executive Order 13166, ‘‘Improving Access to Services for Persons with Limited English Proficiency’’ Requires each federal agency to examine the services it provides, and then identify, develop and implement a system by which limited-English-proficient persons can meaningfully access those services consistent with, and without unduly burdening, the fundamental mission of the agency. The order also requires that each federal agency draft guidance pursuant to Title VI of the Civil Rights Act of 1964, as amended, to ensure that recipients of federal financial assistance take reasonable steps to provide meaningful access to their programs and activities. Administrative Procedure Act (APA) Standardizes administrative procedures for all governmental agencies. For actions subjected to the APA’s formal rulemaking requirements (most USEPA rulemakings), the APA generally requires agencies to publish a general notice of proposed rulemaking in the federal register, and to give interested persons an opportunity to participate through submission of written data, views, or arguments. For actions subjected to the APA’s formal rulemaking or formal adjudication requirements, the APA prescribes additional procedures for agency hearings, which include, among other things, requirements for notice and an opportunity for interested parties to submit facts and arguments, proposed findings and conclusions, or exceptions to agency decisions. Negotiate Rulemaking Act of 1990 Authorizes federal agencies to use the formal regulatory-negotiation (reg-neg) process to formulate proposed rules. If consensus is reached, then the resulting rule is likely to be easier to implement and the likelihood of subsequent litigation is diminished. Even if consensus on a draft rule is not achieved, the process may be valuable as a means of better informing the regulatory agency of the issues and concerns of affected stakeholders. Although the agency is permitted to publish as its own the consensus proposal adopted by the negotiating committee, the agency is not required to publish either a proposed or final rule merely because a negotiating committee proposed it. Source: USEPA (2000a).

support for the committee, and a description of the duties for which the committee is responsible. The charter must be renewed every 2 years.  The Establishment Federal Register Notice. At least 15 days before the charter is filed for a new committee, USEPA is required to publish an establishment notice in the Federal Register. This notice describes the nature and purpose of

264

     



PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

the committee, the agency’s plan to attain fairly balanced membership, and a statement that the committee is necessary and in the public interest. Balanced membership. Advisory committees must be ‘‘fairly balanced’’ in points of view represented. The Meeting Federal Register Notice. Each advisory committee meeting must be noticed in the Federal Register at least 15 days prior to the meeting. To close a meeting to the public, approval must be obtained both from the USEPA Administrator and the USEPA General Counsel. Detailed minutes must be kept of all advisory committee meetings. Open meetings. Interested persons may file written statements with any advisory committee, attend any advisory committee meeting (unless properly closed), and appear before any advisory committee. DFO attendance. Each meeting must be attended by a Designated Federal Official (DFO), a full-time federal employee who is authorized to adjourn the meeting and approve the agenda. Documents available to the public. All advisory committee documents (including drafts) shall be available to the public on request.

An important characteristic of advisory committees formed under FACA is that membership on the committee is to be balanced so that all relevant stakeholder interests are represented. This does not mean that anyone with an interest in the issue may be a member of the advisory committee, but that committee members are specifically selected so that all necessary points of view are represented. When an advisory committee is formed under FACA to advise the agency on a regulatory issue, committee formation and all meetings are announced to the public by notice in the Federal Register. USEPA may provide advice and administrative support for FACA advisory committees, but in general, the committees function independently from the agency. A notable exception occurs when agency staff sit at the table and participate like any other committee member. In such cases, agency staff may steer the committee, either deliberately or unknowingly, toward the agency’s view or away from the view of other stakeholders. Two important permanent advisory committees formed under FACA used by USEPA for advice on drinking water regulations are the National Drinking Water Advisory Council (NDWAC) and the USEPA Science Advisory Board Drinking Water Committee (SAB DWC). In particular, since enactment of the 1996 SDWA amendments, USEPA has utilized the NDWAC and its working groups as a principal mechanism for obtaining public input and recommendations on a variety of regulatory issues.

11.5.1

National Drinking Water Advisory Council (NDWAC)

The NDWAC was established in the 1974 SDWA to ‘‘advise, consult with, and make recommendations to the administrator on matters relating to activities, functions, and

11.5 FEDERAL AGENCY ADVISORY COMMITTEES

265

policies’’ of USEPA under the SDWA. The 15 members of the NDWAC are appointed by the administrator. Five members each are appointed from the general public, appropriate state and local agencies, and representatives of private organizations or groups with an active interest in water hygiene and public water supply. Two members must represent small, rural public water systems. Nominations for appointment to the NDWAC are sought from the public each fall through a Federal Register notice. Members of the NDWAC receive compensation for the time they are engaged in the business of the council. The NDWAC usually meets several times each year. Meeting times and locations are announced in the Federal Register and usually include time for public statements. Meetings of the NDWAC can provide opportunity for public discussion of issues related to new regulations. Although USEPA’s OGWDW provides administrative support for the council, the council functions independently from USEPA. The agency is not obligated to seek the council’s advice on every matter, and it is not obligated to follow the council’s recommendations. 11.5.2

USEPA Science Advisory Board

The SDWA requires USEPA to request comments from the USEPA SAB before proposing any maximum contaminant level goal (MCLG) or National Primary Drinking Water Regulation. The SAB was first established administratively in 1973 as a part of USEPA’s Office of Research and Development. In 1976, the SAB was moved to the administrator’s office to provide it with more independence and to expand its scope. Congress provided statutory authorization for the SAB in 1978 under the Environmental Research, Demonstration, and Development Authorization Act (ERDA 1978a). One of the SAB’s statutory functions is to review and provide advice and comments on the scientific adequacy of ‘‘any proposed criteria document, standard, limitation, or regulation’’ issued by USEPA (ERDA 1978b). The importance of the SAB’s role was noted by Congress (1977a) in the legislative history of the SAB’s enabling legislation: In recognition that political and temporal pressures often weaken the credibility of scientific and technical data which support EPA’s regulatory program . . . [t]his legislation is intended to enhance the status, scope and responsibilities of the Board, leading to more vigorous and complete independent scientific review of data and analyses used for reaching decisions.

Congress (1977b) later commented [M]uch of the criticism of [EPA] might be avoided if the decisions of the Administrator were fully supported by technical information which had been reviewed by independent, competent scientific authorities.

The OGWDW typically initiates the SAB Drinking Water Committee (DWC) review process by submitting documents to the SAB for review or by requesting comment on specific issues. Documents released by USEPA to the SAB must also be made available to the public. The SAB DWC considers the issues and forwards its report to

266

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

the SAB Executive Committee. Following SAB Executive Committee approval, the SAB DWC report is forwarded to the agency. The SAB can respond any time prior to promulgation of the regulation. USEPA does not have to postpone promulgation if no comments are received. Meetings of the SAB are announced in the Federal Register and usually include an opportunity for public statements. Agency representatives may or may not be present during public statements.

11.6

REGULATORY NEGOTIATION

The traditional APA rulemaking procedure can discourage affected stakeholders from communicating with one another. Stakeholders with different interests typically assume conflicting and antagonistic positions. The end result can be expensive and time-consuming litigation. Adversarial rulemaking deprives affected parties of the benefits of shared information, knowledge, expertise, and technical abilities. Cooperation in developing and reaching agreement on the provisions of a rule is also discouraged. The regulatory negotiation (reg-neg) process is conducted under the authority of the Negotiated Rulemaking Act of 1990 (NRA 1990). If used to develop a proposed rule, negotiations are held between members of a FACA advisory committee whose membership includes a representative of each affected interest. Meetings are facilitated by a neutral third party, are open to the public, and information is freely exchanged. New ideas and comments from the public can be offered to the advisory committee for consideration in its negotiations. A key feature of the reg-neg (regulatory negotiation) process is that a representative from USEPA functions as a member of the committee, negotiating with other committee members. The goal of committee deliberations is to reach consensus on the language or issues involved in the draft rule. Not all rulemakings, however, are appropriate for the reg-neg process. The process is instituted at the agency’s discretion, and not all reg-negs succeed. A successful reg-neg results in shared power and consensus building of regulatory solutions. If consensus is reached, then the resulting rule is likely to be easier to implement and the likelihood of subsequent litigation is diminished. Even if consensus on a draft rule is not achieved, the process may be valuable as a means of better informing the regulatory agency of the issues and concerns of affected stakeholders. Although the Agency (USEPA) is permitted to publish as its own the consensus proposal adopted by the negotiating committee, the Agency is not required to publish either a proposed or final rule merely because a negotiating committee proposed it. In the negotiative rulemaking process, a neutral, third-party facilitator is necessary to keep the process moving and provide mediation between negotiators. The facilitator becomes involved early in the process, even before a decision is made to use the negotiative process for rule development. The facilitator is critical to the success of each of the following phases involved in regulatory negotiation (Pritzker and Dalton 1990):  Evaluation Phase. In this first phase, the facilitator works with USEPA to identify the key stakeholders. The facilitator interviews each stakeholder and

11.6 REGULATORY NEGOTIATION

267

other knowledgeable sources to identify the critical issues, and provides a report to USEPA to assist the agency in deciding whether a negotiation is likely to succeed. Not all negotiations are successful. If stakeholders are too far apart in their issues and=or beliefs, or are unwilling to work toward solutions within a negotiative process with other stakeholders who may think differently than they do, then the agency may choose to use the traditional rulemaking approach, rather than risk expending resources on a negotiative process that is unlikely to succeed.  Convening Phase. Following the decision to use the negotiative process to develop a rule, the facilitator role changes to convener. The facilitator works with the stakeholders to set meeting dates and organizes and hosts the meetings. Ground rules are typically discussed and drafted at the first formal meeting, and ratified by the group at the beginning of the second meeting. Ground rules define the purpose of the discussions, participation, meetings, decisionmaking, agreement, schedule, and safeguards for the negotiators. Issues for future discussion are identified and a meeting schedule developed. Typically, several meetings may be needed to develop a baseline of shared knowledge, consisting of presentations of data by informed experts invited in advance by the negotiators. Consultants and researchers may be invited to present data or otherwise discuss issues and present opinions to the negotiators. All meetings are announced in advance and are open to the public.  Negotiation Phase. Once ground rules have been established, issues identified, and all relevant information presented and discussed, the facilitator guides the negotiators in identifying and evaluating regulatory options. As options are narrowed, the facilitator’s role is principally mediation.  Rulemaking Phase. If an agreement is reached, USEPA proceeds as agreed. A formal agreement is signed, whereby stakeholders agree to not contest or litigate issues for which agreement is reached. However, for issues where agreement was not reached, stakeholders reserve the right to comment and act, as they deem appropriate. Stakeholders reserve the right to comment and act, as they deem appropriate, on issues for which agreement was not reached. The agency develops a proposed rule based on the signed agreement in principle. Organizations are selected to participate as stakeholders during formal negotiations so that all identifiable interests or points of view are represented. Not every organization or individual that wants to participate will necessarily be represented at the negotiating table. For a negotiative regulatory process to succeed, there must be a limited number of identifiable stakeholders. On the other hand, representation at the negotiating table must be balanced in terms of points of view, and organizations must select representatives who can adequately represent the organization’s interests during negotiations. Organizations and negotiators involved must be willing to negotiate in good faith to reach ‘‘consensus.’’ There must be a reasonable likelihood ‘‘consensus’’ will be reached within the time limit that may be imposed by the SDWA and=or USEPA.

268

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

The representatives at the negotiating table constitute the negotiating committee, or FACA committee, named after the federal law (FACA) under which the negotiations take place, discussed previously. Technical support is typically required during the negotiative rulemaking processes. All participants can provide their own technical support, or in some cases a separate technical working group is formed to provide analytical support and technical assistance to the entire negotiating committee. The primary purpose of the technical working group is to conduct scientific data analyses for the negotiating committee so that committee decisions can be based on the best available science. Needed analyses typically include the development of baseline estimates (treatment practices, risk estimates) and rigorous evaluation of regulatory options formulated by the negotiating committee (treatment impacts, risk reduction, benefits, and costs). The principal drawback to a regulatory negotiative process for rule development is that it typically requires significant resources and time. Deliberations can take many months or years, depending on the issues to be addressed. Regulatory deadlines may make stakeholder negotiations impractical. In some cases, only selected issues may be addressed to meet time constraints. To be successful, the creation of a negotiating committee to formulate a draft rule or principles of agreement involves elements of what might be termed ‘‘community.’’ Creation of a FACA committee necessarily involves the formation of an artificial community that includes the negotiators as well as their supporting organizations and other interested stakeholders. Individually, the negotiators and the organizations represented must have realistic expectations regarding what can be achieved working within this community of stakeholders. In some cases, certain organizations and=or individuals may have only fought with each other as enemies either politically or via the news media, never having entered into constructive dialog. The negotiative process provides an opportunity for such a dialog to occur, if the participants choose to work together toward common solutions, rather than simply fight. In other cases, organizations and=or individuals begin the process virtually unknown to each other, and a period of time is needed to develop rapport and effective working relationships. Stakeholders involved in negotiative deliberations reflect great diversity in their views and the expertise that they bring to the negotiating table. Individuals and=or organizations with fundamentally different worldviews regarding drinking water must learn to respect differences, and through the negotiative process shift from fighting each other to fighting the problem together. Although core values of individuals and=or organizations may differ, these differences help define boundaries, not necessarily barriers to productive discussions and formulation of solutions. An attitude of learning from others and seeking to build on common ground helps keep discussions moving in a positive direction.

11.7

JUDICIAL REVIEW

The final opportunity available to stakeholders to influence a USEPA rulemaking is judicial review of a newly promulgated rule. Rubin and Pontius (1993) have

11.8 USEPA’S PUBLIC INVOLVEMENT POLICY

269

reviewed the basics of the judicial review process. Challenging a final agency rule in court can be costly and time-consuming. More importantly, litigation of a promulgated rule can impair or irrepairably damage informal working relationships between the agency and the petitioner. The risk involved must be evaluated carefully before a decision is made to legally challenge a final rule. Judicial review is discussed further in Chapter 9. 11.8

USEPA’S PUBLIC INVOLVEMENT POLICY

On Jan. 19, 1981, USEPA published its first agencywide Public Participation Policy (USEPA 1981): to ensure that managers plan in advance needed public involvement in their programs, that they consult with the public on issues where public comment can be truly helpful, that they use methods of consultation that will be effective both for program purposes and for the members of the public who take part, and finally that they are able to apply what they have learned from the public in their final program decisions.

The 1981 Policy complemented regulations on ‘‘Public Participation in Programs under the Resource Conservation and Recovery Act, the Safe Drinking Water Act, and the Clean Water Act,’’ 40 CFR Part 25, promulgated in 1979 (CFR 2002). Part 25 covers procedures that the Agency (or state, tribe, etc.) should or must follow. Like the 1981 Policy, these procedures address matters associated with information, notification, consultation responsibilities, public hearings, public meetings, advisory committees, responsiveness summaries, permit enforcement, rulemakings, and work elements in financial assistance agreements. Since issuance of the 1981 Policy, public involvement techniques have expanded. USEPA has also developed and extended its methods of ensuring compliance with environmental regulations through partnerships, technical assistance, information and data access, and public involvement under the laws it implements. In addition, public involvement has increasingly become an important part of agency decisionmaking at all levels, ranging from advisory committees for national rules to local involvement in permitting, cleanups, and other initiatives. Further, USEPA developed tools to assist in conducting public involvement and consultation. These include  RCRA Public Involvement Manual (EPA530-R–96–007, Sept. 1996)  Public Involvement in Environmental Permits: A Reference Guide (EPA599R00–007, Aug. 2000)  The Model Plan for Public Participation (EPA300-K–96–003, Nov. 1996)  Environmental Justice in the Permitting Process (EPA=300-R–00–004, Dec. 1999)  The Office of Pesticide Program’s How to Participate in EPA Decision-making (63 FR 58038, Oct. 1998).

270

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

In July 1999, USEPA committed to evaluate and update USEPA’s public involvement requirements and assess how well its regulations and policies ensure public involvement in decision-making (USEPA 1999a). In November 1999, USEPA sought public comment opinion on whether the 1981 Policy needed to be revised and updated (USEPA 1999b). USEPA collected, analyzed, and posted public comments on the Internet at http://www.epa.gov/stakeholders. The formal review of USEPA’s public participation policy and regulations culminated in a December 2000 report (USEPA 2000a). USEPA concluded that the 1981 Policy was basically sound and workable, but required updating. On Dec. 28, 2000, (USEPA 2000b) the agency issued a Draft Public Involvement Policy (hereafter called the Draft Policy) updating and strengthening (but not fundamentally changing) the 1981 Policy. This Draft Policy applies to all USEPA programs conducted under the laws and Executive Orders that USEPA implements. The Draft Policy addresses public involvement in all of the USEPA decisionmaking, rulemaking, and program implementation activities. It states that The fundamental premise of this Draft Policy is that, in all its programs, USEPA should provide for meaningful public involvement. This requires that everyone at USEPA remain open to receive all points of view and extend every effort to solicit input from those who will be affected by decisions. This openness to the public furthers our mission to protect public health and safeguard the natural environment by increasing our credibility and improving our decision-making. Our willingness to remain open to new ideas from our constituents, and to incorporate them where appropriate, is absolutely essential to the execution of our mission. At the same time, we should not accord privileged status to any special interest, nor accept any recommendation or proposal without careful, critical examination.

The term the public is used in the Draft Policy in the broadest sense, meaning the general population of the United States. Many segments of ‘‘the public’’ may have a particular interest or may be affected by Agency programs and decisions. In addition to private individuals, ‘‘the public’’ includes, but is not limited to, representatives of consumer, environmental and other advocacy groups; environmental justice groups; indigenous people; minority and ethnic groups; business and industrial interests, including small businesses; elected and appointed public officials; the media; trade, industrial, agricultural, and labor organizations; public health, scientific, and professional representatives and societies; civic and community associations; faithbased organizations; research, university, education, and governmental organizations and associations; and governments and agencies at all levels. Public agencies that serve as coregulators may have a dual role; they can be beneficiaries of public involvement in their decisionmaking processes as well as stakeholders who provide input into USEPA’s decisions. The term public involvement is used in the draft policy to encompass the full range of actions and processes that USEPA uses to engage the public in the Agency’s work, and means that the Agency considers public concerns, values, and preferences when making decisions. Public involvement is intended to enable the public to work with the Agency and hold it accountable for its decisions.

Public involvement at USEPA relies heavily on discretion by Agency officials and a separate strategy for implementing the Draft Policy has been prepared (USEPA 2002). The Draft Policy states an intended bias in favor of public involvement. All reasonable efforts should be made to ensure that the public is informed and given appropriate opportunities for involvement. Those opportunities should not be judged solely by their quantity, but also by whether they are designed to improve the quality of USEPA’s decisions. Opportunity for public involvement in rulemaking that requires public notice and comment will always be provided, but not every document or decision requires public involvement. Every involvement opportunity does not call for the inclusion of all potentially interested persons; including representatives of various interests may be sufficient. Agency officials have the flexibility to determine appropriate public involvement, and recognize that agreement among all parties, while valuable, is not always needed. In addition, USEPA retains the discretion to make decisions or take actions to preserve and protect the environment and public health. The Draft Policy is not a rule, is not legally enforceable, and does not confer legal rights or impose legal obligations on any member of the public, USEPA, or any other agency. However, it is USEPA’s statement of a strong commitment to full and meaningful public involvement in Agency activities. As a policy, the Draft Policy is not binding on states, tribes, and local governments that implement federally delegated, authorized, or approved programs.

11.9

THE FUTURE OF PUBLIC PARTICIPATION

Public participation is an important aspect in the development of new regulations. The principal opportunity for public involvement in traditional rulemaking is the proposed rule comment period. Opportunities may exist prior to proposal of a rule for the public to discuss issues of concern with the agency and in public forums such as stakeholder meetings and through the NDWAC and SAB. Since enactment of the 1996 SDWA amendments, USEPA has increased the frequency of stakeholder meetings to provide opportunities to inform and engage stakeholders in meaningful discussions on a variety of anticipated rulemakings. USEPA has the discretion to discuss, share information, and work with the public, individuals, and specific organizations on issues related to a new rulemaking both before and after formal proposal. Stakeholders should take full advantage of opportunities to discuss with the agency the technical issues and facts bearing on a rulemaking prior to proposal. Stakeholder interaction is usually limited by the agency after proposal because of concerns regarding ex parte communications, perception of favoritism, or other reasons. Traditional rulemaking allows an opportunity for public input but rarely allows for meaningful interaction and consensus building. In contrast, stakeholder meetings and a formal FACA advisory committee or regulatory negotiation (reg-neg) process provide an alternative to the traditional rulemaking process, utilizing an open forum and consensus process to integrate various public concerns in the development of a proposed rule. An emerging chal271

272

PUBLIC INVOLVEMENT IN REGULATION DEVELOPMENT

lenge will be to engage the public and stakeholders in regulatory discussions without releasing security-sensitive information or compromising the security of public water systems. Security concerns may limit the amount and=or type of data made available by USEPA for public review in the future.

REFERENCES APA. 1946. Legislative History of the Administrative Procedure Act. S. Document 248. 79th Congress, 2nd Session, 200. ACUS. 1991. A Guide to Federal Agency Rulemaking, 2nd ed. Washington, DC: Administrative Conference of the United States. CFR. 2002. Public Participation in Programs under the Resource Conservation and Recovery Act, the Safe Drinking Water Act, and the Clean Water Act. Code of Federal Regulations, Title 40, Part 25. Congress. 1974. House of Representatives Rept. 93-1606, 93rd Congress, 2nd Session, 33. Congress. 1977a. H.R. 95-157. House of Representatives Rept. 157, 95th Congress, 1st Session, 38. Congress. 1977b. H.R. 95-722. House of Representatives Rept. 722, 95th Congress, 1st Session, 16. ERDA. 1978a. Environmental Research, Demonstration, and Development Authorization Act. 42 USC Sec. 4365. ERDA. 1978b. Environmental Research, Demonstration, and Development Authorization Act. 42 USC Sec. 4365(e). FACA. 1996. Federal Advisory Committee Act. 5 USC App. 2, Secs. 1–15. Hercules. 1978. U.S. Court of Appeals 1978. Hercules Inc. v. EPA, 598 F.2d 91, D.C. Circuit. Home Box Office. 1977. U.S. Court of Appeals 1977. Home Box Office Inc. v. FCC, 567 F.2d 9 (D.C. Circuit), certification denied, 434 U.S. 829. NRA. 1990. Negotiated Rulemaking Act of 1990. P.L. (Public Law) 101-648. Pontius, F. W. 1993. Involving the public in developing regulations. J. Am. Water Works Assoc. 85:20. Pritzker, D. M. and D. S. Dalton. 1990. Negotiated Rulemaking Sourcebook. Washington, DC: Administrative Conference of the United States. Rubin, K.A. and F.W. Pontius. 1993. Influencing regulation through litigation. J. Am. Water Works Assoc. 85:19. SAB 2001. Improved Science-Based Environmental Stakeholder Processes. EPA-SAB-ECCOM-01-006. Washington, DC: USEPA Science Advisory Board. Sierra Club. 1981. U.S. Court of Appeals 1981. Sierra Club v. Costle, 657 F.2d 298, D.C. Circuit. USEPA. 1981. EPA Policy on Public Participation. Fed. Reg. 46:5736. USEPA. 1999a. Aiming for Excellence: Actions to Encourage Stewardship and Accelerate Environmental Progress. EPA 100-R-99-006. Washington, DC: USEPA. USEPA. 1999b. Review of Environmental Protection Agency Public Participation Policies. Fed. Reg. 64:66906–66913.

REFERENCES

273

USEPA. 2000a. Engaging the American People: A Review of EPA’s Public Participation Policy and Regulations with Recommendations for Action. EPA 240-R00-005. Washington, DC: Office of Policy, Economics and Innovation. USEPA. 2000b. Draft Public Involvement Policy; Proposed Policy. Fed. Reg. 65:82335–82345. USEPA. 2002. Draft Recommendations for Implementing EPA’s Public Involvement Policy. January 10. Washington, DC: Office of Policy, Economics and Innovation.

PART III CONTAMINANT REGULATION AND TREATMENT

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

12 CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS STIG E. REGLI Environmental Engineer, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

PAUL S. BERGER, Ph.D. Microbiologist, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

THOMAS R. GRUBBS, P.E. Environmental Engineer, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

12.1

INTRODUCTION

The purpose of this chapter is to review how the U.S. Environmental Protection Agency (USEPA) has developed federal drinking water regulations that control for pathogens and disinfection byproducts (DBPs). Pathogens and DBPs are considered together because control for either one of these groups of contaminants has a direct influence on the other. This chapter is organized chronologically to inform how perception of public health concern from pathogens and DBPs has changed, how this has influenced changes in the Safe Drinking Water Act (SDWA), and how the SDWA has influenced regulation development by USEPA. 12.2

CONTROL OF WATERBORNE PATHOGENS BEFORE THE 1970s

Early development of U.S. drinking water standards is discussed in Chapter 1, but is reviewed briefly here as related to waterborne pathogens. Federal legislation to Disclaimer: The views expressed in this chapter are those of the authors and do not necessarily reflect the views or policies of the USEPA. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

277

278

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

control infectious disease in the United States began with the National Quarantine Act of 1878. This law was designed to prevent the introduction of infectious diseases into the United States. The Act was modified in 1890 and 1893, giving the Treasury Department authority to prevent the introduction of contagious diseases across state lines. In 1912, as a result of severe outbreaks of intestinal disease among steamship passengers on the Great Lakes, the Treasury Department issued the first waterrelated regulation, which prohibited the use of a common drinking water cup aboard interstate carriers such as ships and trains (McDermott 1973). The 1912 regulations led to the 1914 Treasury Department standards, which prescribed mandatory limits for bacteria in interstate carrier supplies. Under these standards, the level of Bacteria coli (i.e., coliform bacteria) were limited to less than 2.2 coliforms per 100 mL, and the total bacterial count was not to exceed 100 mL. Later revisions of the drinking water standards by the U.S. Public Health Service (USPHS) in 1925, 1942, 1946, and 1962 modified the coliform standard and added standards for several inorganic chemicals. In 1925, the total bacterial count requirement was dropped when a standard was established for turbidity, which is a rough measure of the cloudiness of the water (created by light scatter due to particulate matter). The recommended monitoring frequency was based on expert judgment (NRC 1977). Table 12.1 provides a history of the early coliform standards from 1914 through 1962. Thus, up through the early 1970s, the only federal mandate, the U.S. Public Health Service Act, covered interstate carrier supplies (train stations, ship terminals, etc.), but not other drinking water supplies. The nation’s drinking water supplies were primarily protected by state and community standards, although many states lacked legal authority to require water systems to comply with existing standards (Congress 1976). Why were coliform bacteria used during this period as an index of water pollution? The reason is that these microbes were common in the gut of warm-blooded animals and could be measured easily. Coliforms are a group of closely related bacteria (family Enterobacteriaceae) that are nonsporeforming, facultatively anaerobic, Gram-negative staining rods that ferment the sugar lactose, producing acid and=or gas. Few other bacteria ferment lactose. Thus, by using culture media that contained lactose as the only carbohydrate, and a means for detecting acid or gas production, coliforms could be differentiated from other bacteria without difficulty. Any bacterium that grew in the culture, producing acid and=or gas, was identified as a coliform, regardless of its genus or species; thus coliforms were identified operationally rather than taxonomically. In practice, most coliforms are Escherichia coli or species of Klebsiella, Enterobacter, and Citrobacter. Of the many coliform strains, only a tiny minority are pathogenic. The use of the coliform group as a fecal indicator was controversial from the beginning. The difficulty is that most coliforms are also widespread in natural water and soil. Thus a coliform-positive sample does not necessarily indicate fecal contamination. One coliform, Escherichia coli, seldom survives outside the gut for long and thus was considered a satisfactory fecal indicator, but the absence at that time of a suitable culture medium that could distinguish E. coli in a mixture of many other

12.2 CONTROL OF WATERBORNE PATHOGENS BEFORE THE 1970

TABLE 12.1

Federal Microbial Drinking Water Standards for Interstate Carriers

Standard 1914 (Treasury)

1925 (USPHS)

1942 (USPHS)

1946 (USPHS) 1962 (USPHS)

279

Requirements=Limits Total number of bacteria on standard agar plate (24 h, 37
C) 100=mL Bacterium coli present in no more than one of five tubes, each with 10-mL lactose-containing broth (B. coli same as coliform group) B. coli present in 3 of 5 tubes, each with 10-mL lactosecontaining broth, in more than 5% of samples when 20 samples=month One sample when <20 samples=month B. coli present in 10% of the tubes If 5 tubes, each with 10-mL lactose-containing broth, are used, then limits are same as the 1925 limits (except that term ‘‘coliform group’’ is used instead of B. coli) If 5 containers, each with 100-mL lactose-containing broth, are used, then coliform group cannot be present in all 5 containers in more than 20% of the tubes when 5 samples=month One sample when <5 samples=month If 5 containers, each with 100-mL lactose-containing broth, are used, then coliform group cannot be present in 60% of the containers=month Check samples—if coliforms present in 3 tubes (10-mL portions) or all 5 containers (100-mL portions), then daily check samples taken until results satisfactory Turbidity: 10 ppm (silica scale) Same as 1942 standards, except use of check samples clarified Either lactose-containing broth (now called fermentation tube method ) or membrane filter technique may be used If lactose-containing broth is used, limits are same as 1942 standard If membrane filter is used, coliform colonies shall not exceed 3=50 mL, 4=100 mL, 7=200 mL, or 13=500 mL in 2 consecutive samples: >1 sample when <20 samples=month >5% of samples when 20 samples=month If membrane filter is used, the arithmetic mean coliform density of all samples collected per month shall not exceed 1 colony=100 mL Monitoring frequency based on population served, ranging from 2 to 500 per month, as agreed to by the state and USPHS Turbidity: 5 units

waterborne bacteria precluded its use. Other organisms were proposed as fecal indicators during this period, but only the enterococci and fecal coliforms were given serious consideration (Geldreich 1966, Kabler and Clark 1960, Committee on Public Health Activities 1961). Fecal coliforms are a subgroup of coliform

280

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

bacteria that are distinguished by their ability to grow at elevated temperatures (now defined as 44.5
C) and primarily consist of E. coli. To clarify terminology, the terms ‘‘coliforms,’’ ‘‘coliform group,’’ ‘‘coliform bacteria,’’ and ‘‘total coliforms’’ are interchangeable. Although coliform standards changed over the years, they were all based on the density of coliform bacteria in water. The technical basis for the various coliform density standards during the early years is not obvious, but apparently was the result of expert judgment based on experience. The minimum required monitoring frequency was not based on any statistical review, but rather on what was considered attainable for systems that were under careful control (Woodward 1959). Even before the 1962 standards were published, the public health community was finding increasing reason to question the effectiveness of total coliforms as an index of safety in drinking waters containing the hepatitis A virus and the protozoan pathogen, Entamoeba histolytica (Committee on Public Health Activities 1961).

12.3 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1970s During the early 1970s, the 1962 USPHS drinking water standards were in effect, at least for interstate carriers. However, the results of several investigations were published in the early 1970s that further challenged the effectiveness of the USPHS standards. For example, in 1969=70, the USPHS conducted a national survey of 969 public water supplies to determine whether they were meeting the current standards. Of the 969 supplies, including those of 22 cities, only 59% met the USPHS coliform limits, and only 10% met both the coliform limits and the monitoring frequency recommended by the USPHS standards (McCabe et al. 1970). In addition, as part of another study, 36 synthetic organic chemicals were found in the drinking water of the system serving New Orleans. The USPHS standards did not address organic contaminants. These national survey findings were among the most important in prompting Congress, in 1974, to pass the SDWA. This legislation broadened coverage to all water systems that regularly provided drinking water to at least 25 people or had at least 15 service connections (SDWA 1974). This legislation (and subsequent amendments) provides the legal basis for all USEPA drinking water regulations. The 1974 SDWA required USEPA to publish (the legal term is ‘‘promulgate’’) regulations using a phased approach. First, USEPA was directed to publish National Interim Primary Drinking Water Regulations (NIPDWRs), based on the 1962 USPHS standards. Then the Agency was directed to contract with the National Academy of Sciences (NAS) for a 2-year study to identify and describe all potentially harmful contaminants in drinking water, both biological and chemical. Finally, USEPA was required to publish more comprehensive drinking water standards, based on the NAS study and other data found in the literature or the results of ongoing research. The SDWA also provided USEPA with direction on developing drinking water regulations. USEPA was to publish regulations for any contaminant that might pose a

12.3 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1970s

281

health risk. Setting regulations was to be a two-step process. First, USEPA was to publish a recommended maximum contaminant level [RMCL; later known as a maximum contaminant level goal (MCLG)] for each contaminant to be regulated. The RMCL=MCLG, which was a federally nonenforceable health goal, was to be set at a level such that no adverse health effect would occur, with an adequate margin of safety. The sole reason for publishing the RMCL=MCLG was to inform the public what the Agency considered a safe concentration. Second, USEPA was to set the maximum contaminant level (MCL), which was federally enforceable, as close to the RMCL=MCLG as feasible, taking state and utility implementation costs into account. The SDWA allowed USEPA to set a treatment technique requirement for a contaminant, in lieu of an MCL, if it could not be measured accurately at the level of health concern (because the analytical method was not sufficiently sensitive or was too costly). 12.3.1

Total Coliform Rule (TCR)

Prompted by the SDWA requirements, USEPA published the NIPDWRs in 1975, extending the 1962 USPHS standards, coliforms included, to all public water systems. The TCR consisted of two types of maximum contaminant levels (MCLs): the single-sample MCL and the monthly average. Both MCLs were based on the density of total coliforms and were complicated. They varied with the analytical method, the sample volume, and the number of samples collected per month. For example, if a system collected fewer than 20 samples per month and used the membrane filter technique to test for total coliforms, the single-sample MCL was no more than 4 coliforms=100 mL in more than one sample per month. Under the TCR, systems were required to monitor for total coliforms in the distribution system at a frequency based on the number of people a system served. The required minimum monitoring frequency ranged from 1 sample per month for systems serving 1000 or fewer people to 500 samples per month for systems serving more than 4,690,000 people. States were allowed to reduce monitoring for small community water systems (CWSs) that used protected ground waters and for noncommunity water systems (NCWSs). (CWSs are public water systems that regularly serve at least 25 year-round residents, while NCWSs are public water systems that regularly serve an average of at least 25 individuals daily at least 60 days out of the year, but not 25 residents year-round.) If the coliform density in a single sample exceeded a particular value (depending on the method and sample volume), the system was required to collect daily check samples from the same site until two consecutive samples were coliform-negative. With approval from the state, a system could substitute chlorine residual monitoring for up to 75% of the required number of coliform samples, if that system maintained at least 0.2 mg=L of free chlorine throughout the distribution system and met other chlorine monitoring requirements. Provisions of the 1975 TCR are summarized in Table 12.2. One SDWA provision is the public notification requirement. Unlike the USPHS standards, if any MCL is exceeded, including those for total coliforms, the system is required to notify the public within a specified time and manner. Other than public

282

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

TABLE 12.2

USEPA’s 1975 Total Coliform Rule

Rule Component Limits

Monitoring location Monitoring frequency Action after positive sample Analytical methods Other requirements

Requirements=Limits ‘‘Single’’ sample limit—based on coliform density and varied with the analytical method used (fermentation tube technique vs. membrane filter), sampling frequency per month (<20 vs. 20þ), and sample volume Monthly average limit—based on coliform density and varied with the analytical method used and sample volume Representative points within distribution system Based on population served, ranging from quarterly to 500 per month Check samples daily, until 2 consecutive samples satisfactory Fermentation tube technique and membrane filter technique Use best source water available Systems allowed to monitor chlorine residual as substitute for up to 75% of required coliform samples, at rate of 4 chlorine samples for each required coliform sample. Required chlorine residual  0:2 mg=L under this provision

notice and check samples, the 1975 TCR did not require remedial action after either type of MCL violation (although a state, at its discretion, could set more stringent requirements). The rationale for using total coliforms continued to evolve from being a conservative indicator of fecal contamination to a convenient indicator of treatment efficiency and distribution system integrity. The premise was that treatment that controlled coliforms would also minimize the likelihood of pathogen occurrence. Thus, total coliforms began to be used to assess the vulnerability of a system to fecal contamination, not whether the system was fecally contaminated. 12.3.2

Turbidity and Heterotrophic Bacteria

Among the other USEPA standards set in 1975 was one for turbidity. The USEPA regulation set a MCL for turbidity of 5 nephelometric turbidity units (NTUs) for systems using surface waters, and required such systems to monitor turbidity levels daily. The Agency’s primary reason in 1975 for regulating turbidity was to minimize its interference with chlorine disinfection. The Agency also recommended (but did not require) that systems monitor bacterial plate counts [also called standard plate count, heterotrophic plate count (HPC), total bacterial counts, etc.]. The bacterial plate count is the total number of bacterial colonies per sample volume growing on a

12.3 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1970s

283

specified culture medium under defined incubation conditions; it is used as a rough index of drinking water quality. The premise is that a high HPC, or a sudden increase in HPC, reflects inadequate water treatment. In 1977, the use of a coliform standard and HPC was supported by the National Research Council (NRC 1977). 12.3.3

Trihalomethanes (THMs)

In the early 1970s, researchers in the Netherlands and the United States demonstrated that THMs are formed as a result of drinking water chlorination (Rook 1974, Bellar et al. 1974). This class of disinfection byproducts (DBPs) is generated when the disinfectant chlorine reacts with naturally occurring material (called precursors) in the water. USEPA subsequently conducted the National Organics Reconnaissance Survey (NORS) of halogenated organics (Symons et al. 1975) to measure the concentrations of four THMs (chloroform, bromodichloromethane, dibromochloromethane, and bromoform) in the raw water and treated water of 80 communities nationwide and to determine what effect the raw water source and water treatment practices had on the formation of these compounds. NORS indicated that chloroform occurred invariably in water that had been chlorinated, while it was absent or at much lower concentrations in raw water. Of the four THMs [collectively known as total THMs or (TTHMs)], chloroform represented about 75% of the total concentration. Among the 80 communities surveyed, chloroform concentrations were found to be as high as 0.311 mg=L and concentrations for TTHMs were as high as 0.482 mg=L, with a mean TTHM concentration of 0.067 mg=L. A subsequent USEPA study, the National Organics Monitoring Survey (NOMS) (USEPA 1978), found results similar to those of the NORS. The data collected under NORS and NOMS, as well as other research, indicated that TTHM levels in drinking water vary depending on the season, chlorine contact time, water temperature, water pH, type of water (i.e., groundwater vs. surface water), chemical composition of the raw water (especially bromide concentration), and treatment methodology. During this time, toxicologic studies conducted on rats and mice indicated that chloroform was carcinogenic (NRC 1977). While cancer bioassay studies were only available for chloroform, USEPA was concerned that other THMs, as well as other DBPs, that could not be measured, might pose health risks. In 1979, to address the above-mentioned concerns, USEPA published a regulation that set an MCL of 0.10 mg=L for TTHMs. This value was based on the feasibility of measuring and controlling for TTHMs, and the need to balance the requirement for continued disinfection to control pathogens while simultaneously lowering exposure to TTHMs (USEPA 1979). Systems serving 10,000 people or more were required to collect four distribution system samples per quarter for each water treatment plant, and determine MCL compliance based on a running annual average (i.e., an average of all samples collected during the most recent four quarters). The standard also allowed a state to reduce the monitoring frequency for systems that could demonstrate that TTHM levels were consistently below the MCL. USEPA limited the TTHM standard to larger systems because the majority of small systems use

284

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

groundwaters that are low in THM precursor content and because of concern that smaller systems might lack the technical expertise needed to control TTHMs properly without undermining pathogen control.

12.4 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1980s During the early 1980s, USEPA’s major focus in drinking water was on toxic and carcinogenic chemicals. One reason for this focus was the conventional wisdom among many that waterborne pathogen control was a mature field, and that the technology for controlling pathogens was well understood. Thus additional regulatory effort in this area was of subordinate interest. Subsequent findings and events slowly shifted USEPA’s direction on this point. The periodic surveillance reports of waterborne disease outbreaks published by the Centers for Disease Control and Prevention (CDC) suggested that incidence rates were still high, especially given the premise that few outbreaks are recognized and reported. One USEPA-funded study in Colorado, for example, found that only about one-quarter of these outbreaks were being recognized and reported (Harter et al. 1985), and other studies suggested a higher rate of unreported waterborne disease (Bennett et al. 1987). Moreover, newly recognized waterborne pathogens were being identified, including the protozoa Giardia and Cryptosporidium, the bacterium Legionella, and the caliciviruses (including the Norwalk agent). Some of these pathogens were found to be more resistant to disinfection than the total coliforms, undermining the belief that pathogen control was mature. In addition, traditional assumptions on filtration and disinfection efficacy began to be questioned. In one study, enteroviruses were found in the filtered water of a large water system that employed full treatment (Payment 1981). One epidemiology study indicated that drinking water was linked to nosocomial (i.e., hospital-acquired) infections caused by the opportunistic bacterial pathogen group, Mycobacterium avium complex (Du Moulin and Stottmeier 1986). These concerns led USEPA to evaluate the need for regulatory activity to control these newly recognized pathogens and reduce outbreaks (USEPA 1983, 1985). This activity was strongly bolstered by the 1986 SDWA reauthorization. In this reauthorization, Congress, frustrated by USEPA’s slow progress in publishing regulations (Pontius and Clark 1999), required USEPA to regulate 83 specified contaminants by 1989, including total coliforms, turbidity, viruses, Giardia lamblia, Legionella, and heterotrophic plate count (HPC). Congress also directed USEPA to publish regulations requiring all water systems to disinfect, with appropriate criteria for waivers (called variances). In addition, the reauthorization also required USEPA to establish criteria under which filtration would be required by systems using surface water sources. These Congressional mandates affirmed the direction that USEPA was already proceeding in controlling pathogens. In contrast to USEPA’s strategy for controlling harmful chemicals in drinking water by establishing an MCL for each chemical, the

12.4 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1980s

285

Agency approach for controlling pathogens was to set a treatment technique requirement instead. More specifically, USEPA’s approach was to (1) assess, by use of water type (i.e., groundwater vs. surface water) and water quality indicators, the extent to which treatment might be needed; (2) set treatment requirements based on a specified pathogen removal=inactivation level; and (3) use water quality and engineering design and operating indicators to ensure that any required treatment was sufficient and reliable. This approach assumed that routine pathogen monitoring was not economically or technologically feasible, and thus setting MCLs were inappropriate. In response to Congressional direction, in 1989 USEPA published a revised TCR (USEPA 1989b) and the Surface Water Treatment Rule (SWTR) (USEPA 1989a). These rules were meant to complement each other and are described below. 12.4.1

Revised Total Coliform Rule

USEPA embarked upon a revision of the 1975 TCR to address perceived shortcomings of the rule. Among the shortcomings were (1) an inadequately documented technical basis for the MCLs, (2) complexity of the MCLs, (3) concern that the monitoring frequency for small systems was not sufficient to determine water quality, (4) concern that coliform-positive samples were being invalidated or disregarded too readily as merely being a local plumbing system problem or improper sample collection and handling, (5) concern that systems were not required to determine whether a total coliform-positive sample was the result of fecal contamination (i.e., contained E. coli) that would have necessitated an urgent response, and (6) concern that some major parts of the distribution system were not being monitored. One shortcoming of the TCR, the previously mentioned deficiency of total coliforms as an indicator of Giardia and Cryptosporidium, was to be addressed by the SWTR and by upcoming rules pertaining to groundwater disinfection. The revised TCR, published in 1989, based the MCL on the presence or absence of total coliforms in a 100-mL sample, rather than on coliform density, as before. This was done for several reasons: (1) data in the literature do not demonstrate a quantitative relationship between coliform densities and either pathogen density or the potential for a waterborne disease outbreak, (2) the revised MCL is easier to understand, (3) it is simple to determine the presence or absence of coliforms and not have to consider the uncertainties associated with estimates of coliform density, and (4) there is less concern about coliform die-off during the time between sample collection and analysis because any decrease in coliform density will seldom result in a complete die-off of all coliforms in that sample. The technical basis of the MCL was based on a study of small systems that found that coliforms are distributed very unevenly in the distribution system; the pattern follows a lognormal distribution, and the variance is large (Pipes and Christian 1982, Christian and Pipes 1983). According to the study, if 60 or more samples are collected, and no more than 5% are coliform-positive, then a 95% confidence exists that the fraction of water with coliforms present is less than 10%. USEPA defined a reasonably safe drinking water as one that contained coliforms in no more than 10% of its volume. This led USEPA to set a MCL of 5.0%; thus, no more than

286

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

5.0% of the total coliform samples collected during a month could be positive. The revised TCR, however, allows small systems to have one coliform-positive sample per month without being in violation of the MCL, because an infrequent coliformpositive sample probably is not a health risk. The nonuniform coliform distribution also is problematic with regard to the required monitoring frequency. On the basis of study data, even when the arithmetic mean coliform density in distribution system water was greater than 1=100 mL, the probability that a 100-mL sample would not contain even a single coliform cell was high. Consequently, a few samples per month would not adequately represent water quality. Thus, USEPA proposed to increase the monitoring frequency for small systems, compared to the 1975 rule. Heavy and passionate opposition from states and small systems concerned about implementation costs led USEPA to conclude that the proposed approach was not feasible. The revised TCR, as finally published, represented a compromise. Under the revised TCR, the monitoring frequency for a small system was similar to the previous rule; however, a system collecting fewer than 5 samples per month (generally systems serving 4100 people or fewer) must have an on-site sanitary survey (i.e., inspection) every 5 years, with some exceptions. The combination of sampling and a periodic sanitary survey allows a small system to have an adequate grasp of its drinking water quality. The term ‘‘sanitary survey’’ was defined as an on-site review of the water source, facilities, equipment, operation, and maintenance of a system to determine whether it could deliver safe water. Conceptually, the revised TCR assumes that small systems are providing the public with safe water. However, a total coliform-positive sample brings this assumption into question. Given the uneven coliform distribution, two coliformfree ‘‘check’’ samples after a total coliform-positive sample cannot be used to invalidate the positive sample. The TCR assumes that all coliform-positive samples are valid; although it allows a state to invalidate a total coliform-positive sample under certain situations, the criteria for invalidation are narrow and the procedures precisely defined. Under the revised TCR, after a total coliform-positive sample, the system must collect a set of repeat samples (3 or 4) within 24 hours and then at least 5 routine samples during the next month of operation. If all repeat samples and remaining routine samples the same month and all routine samples the next month are coliform-negative, then some assurance is provided that the contamination is not extensive or has been eliminated. Also every total coliform-positive sample must be tested for the presence of either fecal coliforms or E. coli. If either is present, the drinking water is fecally contaminated, an acute health risk may exist, and the state must be alerted quickly. The revised TCR eliminated the chlorine substitution policy of the 1975 TCR because no utility was using this provision. The more important provisions of the revised TCR are summarized in Table 12.3. 12.4.2

Surface Water Treatment Rule (SWTR)

To respond to the SDWA mandate to control the pathogens mentioned in the previous section (Giardia, viruses, and Legionella) in surface water and to meet

12.4 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1980s

TABLE 12.3

287

Revised Total Coliform Rule

Rule Component MCL

Monitoring location

Monitoring frequency

Action after positive sample

Analytical methods Sanitary surveys Sample invalidation

Requirements=Limits If system collects 40þ samples per month, no more than 5.0% of the monthly samples (routine þ repeat) can be total coliform-positive If system collects < 40 samples per month, no more than one sample per month can be total coliform-positive If a routine and repeat sample from a site are total coliform-positive, and one is also E. coli- or fecal coliform-positive (acute MCL violation) Samples collected throughout distribution system according to a sample siting plan subject to state approval CWS: 1–480 samples=month, based on population served CWS: 1 sample=quarter if system serves 25–1000 people, uses a protected groundwater, and passes a sanitary survey NCWS: If system uses surface water or groundwater under the direct influence of surface water, or if it serves >1000 people, it must monitor at same rate as CWS of comparable size Other NCWSs: quarterly (less in some circumstances) Test total coliform-positive culture for presence of either fecal coliforms or E. coli Collect a set of repeat samples (3 or 4) within 24 h; one must be at the same site as the original positive, the others must be upstream and downstream within 5 service connections of the positive Collect at least 5 routine samples the next month of operation (some exceptions) Both routine and repeat sample results count toward MCL determination If MCL violation, notify state and public Only EPA-approved methods used for compliance samples If system collects <5 samples=month, it must have on-site sanitary survey every 5 years (some exceptions) State may invalidate positive sample if (1) laboratory admits analysis error, (2) repeat sample is positive at original positive site and negative at other sites within 5 service connections (i.e., plumbing system problem), (3) substantial grounds exist that positive result does not reflect water quality and state documents rationale in writing State must invalidate negative sample if coliform test indicates that high levels of heterotrophic bacteria may have interfered with coliform analysis System must resample within 24 h

288

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

the 1986 SDWA mandate regarding the filtration of systems using surface water sources, USEPA published the SWTR in 1989. The SWTR also met the SDWA mandates to control HPC and turbidity and, at least for systems using surface water, to require these systems to disinfect. The SWTR set MCLGs of zero for Giardia, viruses, and Legionella, because USEPA was not aware of any pathogen concentration other than zero at which it was safe to drink the water. The Agency set a treatment technique requirement rather than MCLs for these pathogens because pathogen monitoring was not considered economically or technologically feasible. The SWTR applies to all systems using surface water or groundwater under the direct influence of surface water (such as springs and shallow wells). Such groundwaters have been associated with waterborne giardiasis. The SWTR was structured to require adequate treatment effectiveness, ample treatment reliability, and protection from pathogen intrusion into the distribution systems. Each of these elements had been implicated as causing a growing number of waterborne disease outbreaks during the 1970s and 1980s (Craun 1988). Of special concern were unfiltered surface water systems, which had been associated with a waterborne disease outbreak rate that was 8 times that of filtered systems (USEPA 1989a). A central issue in developing the SWTR was what minimum level of treatment should be set. Two perspectives were considered in addressing this issue. One perspective was to define the level of treatment that could be attained by a welloperated system using conventional treatment (coagulation, flocculation, sedimentation, rapid granular filtration, chlorine disinfection). Such systems rarely had been implicated in a waterborne disease outbreak. Research on filtration and disinfection indicated that such systems could achieve at least a 3log (99.9%) and 4log (99.99%) reduction through the removal and=or inactivation of Giardia and viruses, respectively. The second perspective was to establish an acceptable health risk value for Giardia and determine the level of treatment needed to achieve this value. Giardia was used as the target organism to evaluate risk because this organism is substantially more resistant than viruses to disinfection, and USEPA assumed that if Giardia were treated to acceptable levels, viruses would also likely be treated to acceptable levels. In addition, more source water occurrence data were available for evaluating risk from Giardia than for viruses. On the basis of these data, USEPA estimated that a 3log removal=inactivation level for Giardia would provide, for most systems, an annual risk of infection rate of less than one in 10,000 (104). This risk level was considered comparable to other risk guidelines (Regli et al. 1988, Macler and Regli 1993). USEPA recognized that systems with poor source water quality might have a higher risk, even after a 3log treatment, but decided not to base the treatment level on the density of Giardia in the source water, given the absence of an adequate analytical method to recover and enumerate this pathogen in a system’s source water. However, USEPA did recommend that systems with poor-quality source water consider additional treatment.

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

289

Both perspectives supported the treatment criterion that provide the central requirement of the SWTR; specifically, all systems are required to use a level of treatment sufficient to achieve a 3log removal=inactivation of Giardia and a 4log removal=inactivation of viruses. Systems practicing filtration could meet these criteria by using standard design and operating conditions. To ensure that systems were meeting the specified removal=inactivation criteria, the SWTR requires filtered systems to measure finished water turbidity (i.e., the turbidity level of the filtered water before the water enters the distribution system) and the disinfection residual both entering and within the distribution system. The SWTR turbidity limits and monitoring requirements, which depend on the type of filtration process used and the number of people the system serves, and the disinfection criteria are provided in Table 12.4. To ensure public health, an unfiltered system has to meet stringent source water quality criteria, maintain an effective watershed control program, and meet the TTHM standard. If all of these criteria are met, the SWTR allows an unfiltered system to remain unfiltered if it can meet the 3log=4log treatment through disinfection alone. To ensure that the unfiltered system provides sufficient disinfection, the SWTR and associated USEPA guidance provide tables of CT values for disinfection inactivation [CT is the product of disinfectant residual concentration (C) in milligrams per liter (mg=L) and disinfectant contact time (T ) in minutes]. CT values are provided for 3log inactivation of Giardia and 4log inactivation of enteric viruses by disinfectant type (e.g., chlorine, chloramines, ozone, chlorine dioxide), water pH, and water temperature. To control Legionella and HPC, as required by the 1986 SDWA, USEPA assumed that the SWTR removal=inactivation requirements for Giardia and viruses would minimize the concentration of bacteria. Some bacterial pathogens such as Legionella, however, may flow from the distribution system to residential or office hot-water plumbing systems, and proliferate in this warm environment. On the basis of a legal interpretation, USEPA decided that it did not have the legal authority to control Legionella within plumbing systems, which are generally outside the authority of the water system. However, the Agency has issued guidance on the control of Legionella in plumbing systems (USEPA 1988).

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s AND BEYOND In the late 1980s, USEPA had identified a number of disinfection byproducts (DBPs) other than the TTHMs. Several animal studies and epidemiology studies suggested that some disinfectants and DBPs might pose a public health risk. These included DBPs from chlorine (e.g., haloacetic acids, haloacetonitriles), from ozone (bromate, aldehydes), from chlorine dioxide (chlorite, chlorate), and from chloramines (cyanogen chloride). A major dilemma for the Agency was how to control DBP formation without compromising pathogen control. This balance became even more

290

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

TABLE 12.4

Surface Water Treatment Rule

Turbidity

Disinfection Unfiltered Systemsa

Monitor source water turbidity at least every 4 h; samples must not exceed 5 NTU

Monitor disinfection conditions daily to calculate inactivation through the CT principle and confirm that the system is achieving at least a 3log inactivation of Giardia and a 4log inactivation of viruses Maintain a detectable disinfectant residual throughout the distribution system Have redundant disinfection components or automatic shutoff to prevent untreated water from entering the distribution system

Filtered Systems Monitor combined filter effluent turbidity at least every 4 h;b for conventional filtration and direct filtration, 95% of monthly samples must not exceed 0.5 NTU and no individual sample may exceed 5 NTU; for diatomaceous earth and slow sand filtration, 95% of monthly samples must not exceed 1 NTU and no individual sample may exceed 5 NTU; for alternative filtration technologies such as membranes or bag filters, 95% of monthly samples must not exceed 1 NTU and no individual sample may exceed 5 NTU (after the system has demonstrated that filtration and disinfection together achieve a 3log inactivation and removal of Giardia and a 4log inactivation and removal of viruses)

Monitor disinfection conditions specified by the state daily to confirm that filtration and disinfection together achieve a 3log inactivation and removal of Giardia and a 4log inactivation and removal of viruses Maintain a detectable disinfectant residual throughout the distribution system

a In addition to the requirements in this table, unfiltered systems must meet certain criteria on an ongoing basis to remain unfiltered. These include monitoring of the source water for total coliform or fecal coliform at least 5 times per week, with at least 90% of samples taken in the previous 6 months below 100=100 mL or 20=100 mL, respectively; having no more than two events in any 1-year period or five events in a 10-year period when the source water turbidity exceeds 5 NTU; developing and implementing a watershed control program to minimize the potential for source water contamination by Giardia or viruses, including a yearly on-site state inspection; and continued compliance with MCLs for total coliform and TTHM. b Small systems and certain filtration technologies qualify for less frequent monitoring.

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

291

crucial in 1993, when a massive outbreak of waterborne cryptosporidiosis in Milwaukee became a major national news story. About 400,000 people became ill and 50 people died (MacKenzie et al. 1994). This outbreak put considerable pressure on USEPA to accelerate the development of regulations to control Cryptosporidium. In addition, the results of a survey (LeChevallier et al. 1991a) indicated significantly higher levels of Giardia in source waters used by drinking water supplies than had previously been reported. Both the outbreak and the Giardia survey data raised questions about the adequacy of the SWTR limits. Because of the technical complexity posed by the need to balance pathogen and DBP risk, and the lack of sufficient DBP occurrence, treatment, and health effects data, USEPA initiated a negotiated rulemaking in 1992 to control both types of risk. A negotiated rulemaking involves the direct participation of individuals representing different organizations with different points of view that would be affected by the rule (stakeholders). USEPA sometimes uses this rulemaking approach for the development of a complicated rule where the regulatory strategy is uncertain or where competing goals need to be met. The negotiation is conducted under the provisions of the Federal Advisory Committee Act (FACA), and the different parties attempt to reach a consensus on the provisions of the rule to be proposed. The FACA process that USEPA initiated in 1992 was the first of three related negotiated rulemakings (the two others are discussed below). In all three, the negotiators included representatives of USEPA, state, and local health and regulatory agencies, public water systems, elected state officials, consumer groups, and environmental groups. The Advisory Committee, as the negotiating team was called, met from November 1992 through June 1993. A few of the tough issues with which the Advisory Committee struggled were whether to:  Regulate DBPs through MCLs or through a treatment technique  Minimize formation of DBPs by establishing a regulatory limit for their naturally occurring organic precursors (e.g., total organic carbon) in the water before disinfection  Provide greater protection against pathogens, in conjunction with new DBP limits, by tightening the SWTR  Develop a second round of DBP controls along with the first, or to wait until better scientific information became available. Following months of intensive discussions and technical analysis, the committee recommended the development of three sets of rules: (1) a two-staged Disinfectants=Disinfection Byproduct Rule (Stage 1 DBPR and Stage 2 DBPR), (2) an ‘‘interim’’ ESWTR (IESWTR), and (3) an Information Collection Rule (ICR). The intent was to use the monitoring and treatment data that would result from the ICR to develop the IESWTR and Stage 2 DBPR. The IESWTR would apply to systems serving 10,000 people or more. The Committee also agreed that a ‘‘long-term’’ ESWTR (LT-ESWTR) would be needed for systems serving fewer

292

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

than 10,000 people and possibly to refine requirements for larger systems based on new information. The Advisory Committee agreed that the implementation schedules for IESWTR and LTESWTR should be ‘‘linked’’ to the implementation schedule for the Stage 1 DBPR to assure simultaneous compliance. This linkage would balance the control of pathogen and DBP risk. The Committee also concluded that additional data on health risk, occurrence, and the relative efficiencies of various technologies were needed to better understand the risk tradeoff, determine the most appropriate means for accomplishing an overall reduction in risk, and support development of the IESWTR, LT-ESWTR, and Stage 2 DBPR. To meet this objective the Committee agreed that utilities would collect additional occurrence and treatment information under the ICR, and USEPA and others would conduct additional research. The three first-phase rules were proposed in 1994 to solicit public comment. The major goals and features of these three rules, as finalized, are discussed below.

12.5.1

1996 SDWA Amendments for Pathogen and DBP Control

In August 1996, before any of the new rules mentioned in the previous section (except for the ICR) could be finalized, Congress reauthorized and amended the SDWA. Among the new provisions, the 1996 amendments set a schedule for USEPA to publish the IESWTR (1998), Stage 1 DBPR (1998), LTESWTR (2001), and the Stage 2 DBPR (2002), in line with the FACA Advisory Committee recommendations. The 1996 amendments also required USEPA to publish a regulation that ‘‘governs’’ the recycle of filter backwash water within a treatment plant. This requirement was probably prompted because several outbreaks of cryptosporidiosis had occurred at drinking water systems where recycled filter backwash was identified as a possible cause (Craun 1998). Recycled filter backwash is discussed later. In addition to specific regulatory provisions, the 1996 SDWA Amendments required USEPA and CDC to conduct epidemiology studies in several communities that would provide a basis for estimating the number of cases of waterborne disease occurring each year in the United States. This requirement may have been prompted by the size of the Milwaukee outbreak of cryptosporidiosis as well as the results of two epidemiology studies in Canada. One epidemiology study indicated that, for a well-treated system using a poor-quality source water, about one-third of the cases of gastrointestinal illness were drinking-water-related (Payment et al. 1991). A follow-up epidemiology study at this system, using a different approach, resulted in similar findings (Payment et al. 1997). The 1996 SDWA amendments also required USEPA to publish a list of unregulated contaminants every 5 years that are known or anticipated to occur in water systems and that may need to be regulated. The first list, known as the Contaminant Candidate List (CCL), was published on March 2, 1998 (USEPA 1998a). The purpose of the CCL is to determine which unregulated drinking water contaminants should be given high priority for research and regulatory consideration. According to the amendments, USEPA is to select at least five contaminants from the CCL

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

TABLE 12.5

293

Pathogens on the Contaminant Candidate List

Protozoa Acanthamoeba Microsporidium [Enterocytozoon, Encephalitozoon (formerly Septata)] Viruses Caliciviruses Adenoviruses Coxsackieviruses Echoviruses Bacteria Helicobacter pylori Mycobacterium avium complex Aeromonas hydrophila Algae Algae and their toxins

every 5 years and determine whether to regulate them. Ten pathogens were included on the 1998 CCL (see Table 12.5). The FACA recommendations mentioned previously were modified somewhat by a major technical difficulty. The original intent was to use the ICR monitoring and treatment data to revise the SWTR (IESWTR and LTESWTR) and to develop the Stage 2 DBPR. However, because the analytical method for Giardia and Cryptosporidium performed below expectations, publication and implementation of the ICR stalled while a concerted effort was made to improve method performance. The delay complicated the planned regulatory phasing approach, and led to USEPA separating the LTESWTR into two parts: the LT1ESWTR, which would not be based on ICR data, and LT2ESWTR, which would be based on these data. The regulations are described below.

12.5.2

Information Collection Rule (ICR)

The purpose of the ICR (USEPA 1996), published in May 1996, was to collect occurrence and treatment information needed to develop the LT2ESWTR and Stage 2 DBPR. The ICR generally covered systems serving at least 100,000 people, although a more limited set of ICR requirements applied to groundwater systems serving between 50,000 and 100,000 people. About 300 systems operating 500 treatment plants were involved with the extensive ICR data collection. The ICR required surface water systems to collect source water samples, and in some cases finished water samples, monthly for 18 months, and test them for the following organisms: Giardia, Cryptosporidium, viruses, total coliforms, and fecal coliforms or E. coli. The ICR also required systems serving at least 100,000 people to determine the concentrations of a host of disinfectant and DBP concentrations in water from different parts of the system. Besides valuable occurrence data, the rule

294

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

required systems to provide specified system operating and engineering data to USEPA. In addition, depending on the TOC levels in their source or finished water, systems were required to conduct bench- or pilot-scale studies with granular activated carbon or membranes to evaluate their effectiveness in controlling TTHMs and haloacetic acids. The data collected under the ICR, with new information that became available through other surveys and research, allowed USEPA to identify strategies for minimizing DBP formation, determine appropriate treatment levels to protect against pathogens, and estimate how much it would cost utilities nationwide to implement various regulatory options. ICR monitoring began in July 1997 and ended in early 1999. 12.5.3

Stage 1 Disinfection Byproducts Rule (DBPR)

USEPA established a second FACA Advisory Committee in March 1997 to collect, share, and analyze information and data developed after the 1994 proposal, as well as build consensus on the implications of this new information. In July 1997, the Advisory Committee reached agreement on the following major issues related to DBPs: (1) set MCLs for TTHM, five haloacetic acids (HAA5), chlorite, and bromate; (2) specified enhanced coagulation requirements as part of DBP control (Advisory Committee suggested modifications in the proposed requirement, based on new data on coagulation effectiveness and the prevalence and multiple uses of predisinfection); and (3) included a methodology by which a surface water system that was likely to modify its disinfection practices, after consulting with the state, could assure that pathogen control would not be significantly degraded as the result of modifying disinfection practices to meet MCLs for TTHM and HAA5 (this requirement appeared in the IESWTR and LT1ESWTR). USEPA published a summary of the Advisory Committee’s recommendations (USEPA 1997a). In 1998, USEPA published the Stage 1 DBPR (USEPA 1998c), which applies to all CWSs and nontransient NCWSs that add a disinfectant during any part of the treatment process (including as a residual). In addition, transient NCWSs that use chlorine dioxide were required to control the level of chlorine dioxide entering the distribution system because of concerns about short-term health effects. The Stage 1 DBPR included a total of 11 nonenforceable MCLGs and seven enforceable MCLs or maximum residual disinfectant levels (MRDLs). Among the provisions, the rule reduced the MCL for TTHMs from the 1979 standard of 0.10 to 0.080 mg=L. Table 12.6 provides a summary of the MCLGs and MCLs in the Stage 1 DBPR. In addition to the MCLs and MRDLs, the Stage 1 DBPR set a treatment technique to reduce the formation of unregulated DBPs in systems that were most likely to have higher levels of these DBPs. This treatment technique required conventionally treated surface water systems to remove specified amounts of organic materials [measured as total organic carbon (TOC)], using enhanced coagulation or enhanced softening. Table 12.7 contains the basic TOC removal requirements, with compliance based on a running annual average. Systems could avoid this TOC removal requirement if (1) the source water had low TOC levels, (2) enhanced

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

295

TABLE 12.6 MRDLGs, MRDLs, MCLGs, and MCLs in Stage 1 Disinfection Byproduct Rule Disinfectant

MRDLG (mg=L) MRDL (mg=L)

Chlorine

4 (as Cl2)

4.0 (as Cl2)

Chloramines

4 (as Cl2)

4.0 (as Cl2)

0.8 (as ClO2)

0.8 (as ClO2)

Chlorine dioxide (at treatment plants using ClO2) Disinfection Byproduct Total trihalomethanes (TTHM) consisting of Chloroform Bromodichloromethane Dibromochloromethane Bromoform Haloacetic acids (five) (HAA5) consisting of Monochloroacetic acid Dichloroacetic acid Trichloroacetic acid Monobromoacetic acid Dibromoacetic acid Bromate (at treatment plants using ozone) Chlorite (at treatment plants using ClO2)

MCLG (mg=L)

MCL (mg=L)

NA

0.080

—a 0 0.06 0 NA

NA NA NA NA 0.060

— 0 0.3 — — 0

NA NA NA NA NA 0.010

0.8

1.0

Compliance Based on Running annual average of distribution system samples Running annual average of distribution system samples Individual samples taken at the treatment plant or in the distribution system Compliance Based on Running annual average of all distribution system samples; sample value is the sum of the four individual THMs Running annual average of all distribution system samples; sample value is the sum of the five individual HAAs Running annual average of samples at treatment plant Average of distribution system three-sample sets taken at least monthly

a

USEPA proposed and finalized an MCLG of zero for chloroform. The U.S. Court of Appeals for the District of Columbia vacated the MCLG and remanded it to USEPA for action.

coagulation was ineffective in removing the TOC, or (3) the TOC was relatively nonreactive with chlorine. The TOC removal requirement was limited to conventionally treated surface water systems because (1) such systems generally had higher precursor levels and DBP levels than did systems using other technologies and (2) the process was relatively inexpensive for conventionally treated systems but expensive and technologically less practical for other types of systems (e.g., in a direct filtration system, which does not use the sedimentation step, the higher coagulant

296

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

TABLE 12.7 Required TOC Removal by Enhanced Coagulation=Enhanced Softening for Surface Water Systems Using Conventional Treatmenta,b Source Water Alkalinity, mg=L as CaCO3 (%) Source Water TOC (mg=L)

0–60

>60–120

>120

>2.0–4.0 >4.0–8.0 >8.0

35.0 45.0 50.0

25.0 35.0 40.0

15.0 25.0 30.0

a b

Softening systems must meet the TOC removal requirements in the right-hand column. Also applies to utilities that treat groundwater under the influence of surface water.

doses needed for enhanced filtration would overload the filter). By using the enhanced coagulation or enhanced softening process, a system would have less precursor material to react with chlorine and thus would have less reason to switch from chlorine to an alternative disinfectant. Switching disinfectant was a concern because even less was known about the health effects associated with DBPs from these alternative disinfectants than was known about chlorinated DBPs. As stated previously, the SDWA requires USEPA to take implementation costs into account in setting an MCL. This means that the Agency cannot set an MCL that a system is not able to consistently meet using the ‘‘best available technology’’ (BAT) that is affordable. In setting an MCL, USEPA must identify an affordable BAT. For the Stage 1 DBPR, the BAT for each MCL (or MRDL) is indicated in Table 12.8. USEPA was aware of technologies that, if implemented, would produce lower levels of DBPs; however, their costs (especially for small systems without economies of scale) were not considered affordable. Because of the complexity of the rule and the need to balance risks and achieve simultaneous compliance with multiple rules, USEPA developed guidance for both systems and states for implementation of the rule (as was also done with the IESWTR). The guidance documents included ones addressing enhanced coagulation and softening, alternative disinfectants, and simultaneous compliance.

12.5.4 Strengthening the SWTR: The IESWTR, LT1ESWTR, and Filter Backwash Recycling Rule The same FACA Advisory Committee that helped develop the Stage 1 DBPR also assisted in the development of a revision to the SWTR, known as the IESWTR. In July 1997, the Advisory Committee reached agreement on IESWTR criteria. The final rule was published in 1998 (USEPA 1998d). The IESWTR applies to systems covered by the SWTR (systems using surface water or groundwaters under the direct influence of surface water) that serve 10,000 people or more. The centerpiece of the revision was to require a 2log Cryptosporidium removal. For systems using rapid granular filtration (i.e., filtration as used in conventional treatment or direct filtration), the turbidity standard was

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

TABLE 12.8

297

BAT for Disinfectants and Disinfection Byproducts

Disinfectant=DBP

Best Available Technology Disinfectants

Chlorine residual

Chloramine residual

Chlorine dioxide residual

Control of treatment processes to reduce disinfectant demand and control of disinfection treatment processes to reduce disinfectant levels Control of treatment processes to reduce disinfectant demand and control of disinfection treatment processes to reduce disinfectant levels Control of treatment processes to reduce disinfectant demand and control of disinfection treatment processes to reduce disinfectant levels Disinfection Byproducts

Total trihalomethanes Total haloacetic acids Chlorite

Bromate

Enhanced coagulation or GAC10,a with chlorine as the primary and residual disinfectant Enhanced coagulation or GAC10, with chlorine as the primary and residual disinfectant Control of treatment processes to reduce disinfectant demand and control of disinfection treatment processes to reduce disinfectant levels Control of ozone treatment process to reduce production of bromate

a GAC10 means granular activated carbon with an empty-bed contact time of 10 minutes and reactivation frequency for GAC within at least 6 months.

tightened, and continuous turbidity monitoring was required for individual filters. Under the IESWTR, if the turbidity of an individual filter exceeds USEPA-specified performance levels, the system is required to evaluate and correct the deficiency of the filter. The major features of the IESWTR are indicated in Table 12.9. The IESWTR supplemented, but did not replace, the SWTR. Some systems, especially smaller ones using relatively clean source water, use other filter processes such as slow sand filtration or diatomaceous earth filtration. The IESWTR did not change the existing turbidity performance criteria for these systems because data indicated that these filter processes, as prescribed under the SWTR, already achieve at least a 2log Cryptosporidium removal. The IESWTR also modified the criteria for avoiding filtration by requiring the watershed control program to control for Cryptosporidium (but did not change the existing SWTR criteria for avoiding filtration), because available occurrence data suggested that Cryptosporidium densities in the raw water of unfiltered systems were similar to those in the finished water of many filtered systems. The IESWTR did not require, or provide an option for, Cryptosporidium monitoring in the source water,

298

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

TABLE 12.9 Interim Enhanced Surface Water Treatment Rule (IESWTR)a Profiling and Benchmarking Systems judged to be most likely to modify their disinfection practices to comply with the Stage 1 DBPR (defined as those with TTHM or HAA5 running annual averages exceeding 0.064 and 0.048 mg=L, respectively) were required to determine the level of Giardia (and, in some cases, virus) inactivation daily for one year; before modifying disinfection practices, systems were required to consult with the state to ensure that modifications would result in no significant increase in microbial risk Combined Filter Effluent Turbidity For systems using conventional filtration or direct filtration, 95% of samples taken during the month cannot exceed 0.3 NTU, and no individual sample can exceed 1 NTU; diatomaceous earth and slow sand filtration must continue to meet the SWTR standards (Table 12.4); systems using alternative filtration technologies must demonstrate that the technology achieves a 2log removal of Cryptosporidium and meet state-specified turbidity limits Individual Filter Monitoring Systems using conventional filtration or direct filtration must monitor the turbidity of each individual filter every 15 min; if certain triggers indicating poor or degraded filter performance are exceeded, systems must conduct actions designed to improve performance a

With the exception of the combined filter effluent turbidity provisions for conventional filtration and direct filtration, requirements in this table are in addition to SWTR requirements listed in Table 12.3.

because analytical methods were not yet available for making adequately accurate determinations. At the beginning of the rule development process to control Cryptosporidium, it was thought that this microbe was exceptionally resistant to disinfectants, with the possible exception of ozone. Later, new animal study data suggested that this view was too pessimistic. Nevertheless, the IESWTR required that filtration alone provide the 2log Cryptosporidium reduction, rather than a combination of filtration and disinfection. At the time the IESWTR was finalized, USEPA did not have criteria (such as CT values) for estimating inactivation of Cryptosporidium by effective technologies such as ozone and ultraviolet light. The IESWTR also set constraints on systems that might want to cut back the existing disinfection process to comply with DBPR limits. The system must first evaluate the effect on Giardia and viral control of any proposed change to a disinfection practice (referred to as disinfection profiling and benchmarking), and satisfy the state that the proposed changes in treatment would not significantly undermine existing pathogen control.

12.5 CONTROL OF WATERBORNE PATHOGENS AND DBPs IN THE 1990s

299

The IESWTR included another provision requiring states to conduct on-site sanitary surveys for all surface water systems. The requirement, which was more specific than the sanitary survey requirement in the TCR, stipulated that sanitary surveys had to include an evaluation of source water quality, treatment, distribution system, finished water storage, pumping facilities, monitoring and reporting data, system management and operation, and operator compliance with state requirements. While most states have had a sanitary survey program in place for many years, these programs often differed significantly among the states with regard to content or frequency. The overall purpose of the IESWTR provisions was to tighten pathogen control from source to tap, thereby minimizing the potential for waterborne disease. Subsequent to publishing the IESWTR, USEPA published the Long Term 1 Enhanced SWTR (LT1ESWTR) (USEPA 2002). The LT1ESWTR extended coverage of the IESWTR to surface water systems serving fewer than 10,000 people. The LT1ESWTR has basically the same requirements as the IESWTR, except that the turbidity of individual filters must be analyzed less often and less monitoring is required to establish the disinfection benchmark. Another modification to the SWTR is the Filter Backwash Recycling Rule (FBRR) (USEPA 2001). As part of normal maintenance of the filtration process, a filter is periodically taken off line and cleaned by reversing the water flow, or backwashed. The backwash process usually involves a large volume of water flowing through the filter. Typically, the backwashed water is directed to the plant influent water or to a specific location in the treatment process. Thus, any pathogens in the backwashed water may be returned to the system. If the point of backwash recycling is not at the beginning of the treatment process, pathogen removal may be compromised. Moreover, when the large volume of backwash water enters a point in the treatment process other than at the initial treatment stage, where it can be distributed as part of the entire plant influent, the combined normal influent water and the recycling backwash water may create a hydraulically overloaded condition, particularly for small systems, that could overwhelm existing treatment. These two situations could degrade Cryptosporidium removal efficiency. To prevent this problem, USEPA published the FBRR, as mandated by the 1996 SDWA amendments. The FBRR requires (with some exceptions) a system using either conventional or direct filtration to recycle backwash water so that it passes through the entire existing treatment process, rather than at an intermediate point where the backwash water would receive incomplete treatment. The rule also ensures that systems and states will have the recycle flow information necessary to evaluate whether a site-specific backwash recycle practice might compromise a system’s ability to achieve the required 2log Cryptosporidium removal. 12.5.5

Ground Water Rule

USEPA is currently developing a regulation that would require an undisinfected groundwater system to assess whether its source water is at high risk for fecal

300

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

contamination and, if so, to undertake corrective action. In 2000, (USEPA 2000a), USEPA proposed a regulation known as the Ground Water Rule (GWR) that included the following provisions: (1) a periodic state-conducted sanitary survey for every system and a requirement to correct each significant deficiency identified, (2) an initial phase of source water monitoring by undisinfected systems using a fecal indicator (such as E. coli, enterococci, or coliphage), and (3) for each undisinfected system, an assessment of the hydrogeological characteristics to gauge whether the source water might be vulnerable to fecal contamination. A high-risk system would be required to disinfect or provide some other means of protecting the public. Under the GWR, as proposed, a high-risk system that disinfects would be required to consistently achieve a 4log (99.99%) virus inactivation. The GWR is being developed in response to the 1996 SDWA amendments, which directed USEPA to publish regulations requiring all groundwater systems to determine whether they needed to disinfect. The SWTR already requires systems using surface water or groundwaters under the direct influence of surface water to disinfect. 12.5.6

LT2ESWTR and Stage 2 DBPR

In March 1999, USEPA reconvened the FACA Advisory Committee to develop recommendations for the LT2ESWTR and Stage 2 DBPR. The intent is to implement the two rules simultaneously for the same reason that the IESWTR was linked with the Stage 1 DBPR, i.e., to prevent a system from compromising pathogen control to meet new DBPR requirements. USEPA expects to propose these recommendations during 2003 and finalize them in 2004. The primary goal of the LT2SWTR is to modify the SWTR such that the public health risk of waterborne cryptosporidiosis would be equivalent, that is, that the required treatment level would depend on the density of Cryptosporidium in the source water. This goal is supported by data from the ICR and supplemental surveys that indicated that mean pathogen levels in different source waters of filtered systems could vary by several orders of magnitude and that minimum treatment levels prescribed by the IESWTR and LT1ESWTR would seldom result in comparable levels of protection across systems. The new occurrence data also indicated that Cryptosporidium densities in the raw water of unfiltered systems were substantially higher than the finished water of most filtered systems. To address these issues, a filtered system using a poor-quality source water would need more treatment than those using a good-quality source water. Also, unfiltered systems, in general, while continuing to meet the filtration avoidance criteria, would need some minimum level of inactivation for Cryptosporidium to achieve comparable protection to filtered systems. The Advisory Committee recommended the following provisions for LT2ESWTR: (1) an initial round of Cryptosporidium monitoring in source waters over a period of 1 year (small systems) or 2 years (large systems) to determine appropriate levels of treatment; (2) a ‘‘toolbox’’ of technologies from which systems could choose to provide additional treatment for Cryptosporidium; (3) an option for small systems to monitor the E. coli density in source water, in lieu of Cryptosporidium, to determine whether Cryptosporidium monitoring and possible

12.6 A VIEW TOWARD THE FUTURE

301

additional treatment was needed; (4) a minimum 2log Cryptosporidium inactivation requirement for unfiltered systems; and (5) specific operational and design criteria for ultraviolet light that would allow its use as an option for providing additional treatment for Cryptosporidium control. In addition, the Advisory Committee recommended that 6 years following the completion of the initial source water monitoring, systems be required to conduct a second round of monitoring to determine possible changes in the source water quality and associated treatment. The goal of the Stage 2 DBPR is to more tightly control the formation of DBPs by setting concentration limits at specific points in the distribution system, rather than more general limits as is the case with the Stage 1 DBPR. This goal is supported by ICR data indicating that TTHM and HAA5 levels could greatly exceed the Stage 1 MCL at various locations within the distribution system and that the highest levels of TTHM and HAA5 rarely occur at the same locations or at the maximum residence times (as had been assumed under the Stage 1 MCL). The Advisory Committee thus recommended the following provisions for the Stage 2 DBPR: (1) retain the MCLs for TTHMs and HAA5 at 0.080 and 0.060 mg=L, respectively, but revise the compliance calculation from a running annual average of measurements throughout the distribution system to a running annual average at each site within the distribution system; and (2) identify the monitoring sites within the distribution system that have the highest TTHM and HAA5 concentrations, and then retain these sites as monitoring locations. USEPA believes that these rule modifications will reduce cancer risks and adverse developmental and reproductive effects, and minimize differences in protection among the population served, regardless of where in the distribution system water is obtained.

12.6

A VIEW TOWARD THE FUTURE

As a new century begins, U.S. drinking water is much safer than ever before. Since the early 1990s, USEPA has published new regulations to control pathogens more tightly and to minimize the formation of harmful DBPs. These regulations are the SWTR, as revised, TCR, and Stage 1 DBPRs. The emphasis for pathogen control has been on setting treatment requirements and using easily measured indicators of water quality and other measures (e.g., periodic sanitary surveys) to assess the adequacy and reliability of that treatment. The regulatory focus of DBP control has been through MCLs. Yet challenges remain. The quality of source waters may degrade with population increases, complicating water treatment and prompting direct water reuse. Newly discovered deficiencies in the distribution system are complicating control of distribution system integrity. The adequacy of current and impending regulations for the control of emerging pathogens is an issue. Beyond the technical challenges, media attention on actual or perceived deficiencies in public water supplies, along with highly visible waterborne disease outbreaks, have resulted in some public anxiety about the quality of our drinking water. The use of bottled water has increased markedly over a short period of time.

302

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

The challenge is to allay this concern, not through overregulation, but by using a systems management strategy for providing microbially safe drinking water. This strategy encompasses the multibarrier approach for pathogen control: watershed control to minimize pathogen introduction to surface waters and groundwaters used by water supplies, adequate and reliable water treatment, and proper operation and maintenance of the distribution system. It also includes operator training, periodic on-site inspections by the state, and laboratory accreditation. The complexity and expanse of this challenge is requiring USEPA to work closely with states and the regulated community in developing new drinking water regulations. Research is an important element in developing a scientifically defensible, costeffective regulation. For an emerging pathogen, the specific research data needed varies with the pathogen (CCL pathogens are presented in Table 12.5), but generally data are needed to address the following questions: To what extent will existing or forthcoming regulations address an emerging pathogen? To what extent does the pathogen occur in source waters or grow within the distribution system? What is the health risk associated with a pathogen, especially in sensitive subpopulations? How effective are various water treatment processes in controlling a pathogen? How effective are various operational and maintenance procedures associated with the distribution system for controlling a pathogen? For DBPs, issues include the relative significance of developmental and reproductive effects associated with DBPs and, if these effects are significant, which DBPs are of most concern and how they can best be controlled. Much of this research is under way.

REFERENCES Bellar, T. A., J. J. Lichtenberg, and R. C. Kroner. 1974. The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66(12):703–706. Bennett, J. V., S. D. Holmberg, M. F. Rogers, and S. L. Solomon. 1987. Infectious and parasitic diseases. In Closing the Gap: The Burden of Unnecessary Illness. R. W. Amler and H. B. Dull, eds. New York: Oxford Univ. Press. Christian, R. and W. Pipes. 1983. Frequency distribution of coliforms in water distribution systems. Appl. Environ. Microbiol. 45:603–609. Committee on Public Health Activities. 1961. Coliform organisms as an index of water safety. J. Sanitary Eng. Div. ASCE 87(SA6):41–58. Congress. 1976. House of Representatives Report 93-1185 (from 93rd Congress, 2nd session, on Safe Drinking Water Act). Washington, DC: U.S. Government Printing Office. Craun, G. F. 1988. Surface water supplies and health. J. Am. Water Works Assoc. 80(2):40–52.

REFERENCES

303

Craun, G. F. 1998. Memorandum from G. Craun to U.S. Environmental Protection Agency (M. Negro), dated 10=26=98. Waterborne outbreak data 1971–1996, community and noncommunity water systems. Du Moulin, G. C., and K. D. Stottmeier. 1986. Waterborne mycobacteria: An increasing threat to health. ASM News 52:525–529. Geldreich, E. E. 1966. Sanitary significance of fecal coliforms in the environment. Water Pollution Control Research Series Publication WP-20-3. Cincinnati: U.S. Dept. Interior, Federal Water Pollution Control Administration. Harter, L. et al. 1985. A three-state study of waterborne disease surveillance techniques. Am. J. Public Health 75:1327–1328. Kabler, P. W. and H. F. Clark. 1960. Coliform group and fecal coliform organisms as indicators of pollution in drinking water. J. Am. Water Works Assoc. 52:1577–1579. LeChevallier, M. W., D. N. Norton, and R. G. Lee. 1991a. Occurrence of Giardia and Cryptosporidium spp. in surface water supplies. Appl. Environ. Microbiol. 57:2610–2616. LeChevallier, M. W., D. N. Norton, and R. G. Lee. 1991b. Giardia and Cryptosporidium spp. in filtered drinking water supplies. Appl. Environ. Microbiol. 57(9):2617–2621. MacKenzie, W. R., N. J. Hoxie, M. E. Proctor, M. S. Gradus, K. A. Blair, D. E. Peterson, J. J. Kazmierczak, D. A. Addiss, K. R. Fox, J. B. Rose, and J. P. Davis. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. New Engl. J. Med. 331(3):161–167. Macler, B. and S. Regli. 1993. Use of microbial risk assessment in setting US drinking water standards. Internat. J. Food Microbiol. 18:245–256. McCabe, L. J., J. M. Symons, R. G. Lee, and G. G. Robeck. 1970. Survey of community water supply systems. J. Am. Water Works Assoc. 62:670–687. McDermott, J. H. 1973. Federal drinking water standards—past, present and future. J. Environ. Eng. Div. ASCE (Proc. Paper 9924), 99(EE4):469–478. Morris, R. D. et al. 1992. Chlorination, chlorination by-products, and cancer: a meta-analysis. Am. J. Public Health 82(7):955–963. NRC. 1977. Drinking Water and Health. Washington, DC: National Academy Press. Payment, P. 1981. Isolation of viruses from drinking water at the Pont-Viau water treatment plant. Can. J. Microbiol. 27:417–420. Payment, P., L. Richardson, J. Siemiatycki, R. Dewar, M. Edwardes, and E. Franco. 1991. A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiological standards. Am. J. Public Health 81:703–708. Payment, P., J. Siemiatycki, L. Richardson, G. Renaud, E. Franco, and M. Prevost. 1997. A prospective epidemiological study of gastrointestinal health effects due to the consumption of drinking water. Internat. J. Environ. Health Research 7:5–31. Pipes, W., and R. Christian. 1982. Sampling Frequency—Microbiological Drinking Water Regulation. EPA 570=9-82-001. Washington, DC: USEPA. Pontius, F. W. and S. W. Clark. 1999. Drinking water quality standards, regulations, and goals. In Water Quality and Treatment, 5th ed. R. D. Letterman, ed. New York: McGraw-Hill. Regli, S., A. Amirtharajah, B. Borup, C. Hibler, J. Hoff, and R. Tobin. 1988. Panel discussion on implications of regulatory changes for water treatment in the United States. In Advances in Giardia Research. P. M. Wallace and B. R. Hammond, eds. Calgary: Univ. Calgary Press.

304

CONTROL OF DRINKING WATER PATHOGENS AND DISINFECTION BYPRODUCTS

Reif, J. S. et al. 1996. Reproductive and developmental effects of disinfection by-products in drinking water. Environ. Health Perspect. 104(10):1056–1061. Rook, J. J. 1974. Formation of haloforms during chlorination of natural waters. Water Treat. Exam. 23:234. SDWA 1974. Safe Drinking Water Act (Public Law 93–523). Symons, J. M., T. A. Bellar, J. K. Carswell, J. Demarco, K. L. Kropp, G. G. Robeck, D. R. Seeger, C. L. Slocum, B. Smith, and A. A. Stevens. 1975. National Organics Reconnaissance Survey for Halogenated Organics. J. Am. Water Works Assoc. 67(11):634–647. USEPA. 1975. National Interim Primary Drinking Water Regulations. Fed. Reg. 40:59566–59588. USEPA. 1978. National Organics Monitoring Survey (NOMS). Cincinnati: Office of Ground Water and Drinking Water, Technical Support Division. USEPA. 1979. National Interim Primary Drinking Water Regulations; Control of Trihalomethanes in Drinking Water. Fed. Reg. 44:68624. USEPA. 1983. National Primary Drinking Water Regulations: Advanced Notice of Proposed Rulemaking. Fed. Reg. 48:45502–45521. USEPA. 1985. National Primary Drinking Water Regulations; Synthetic Organic Chemicals, Inorganic Chemicals and Microorganisms; Proposed Rule. Fed. Reg. 50:46936–47022. USEPA 1988. Control of Legionella in plumbing systems. In Reviews of Environmental Contamination and Toxicology 107:79–92. G. W. Ware (ed.), New York: Springer-Verlag. USEPA. 1989a. National Primary Drinking Water Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionella, and Heterotrophic Bacteria; Final Rule. Part II. Fed. Reg. 54:27486. USEPA. 1989b. National Primary Drinking Water Regulations; Total Coliforms (Including Fecal Coliform and E. coli); Final Rule. Fed. Reg. 54:27544. USEPA. 1994a. National Primary Drinking Water Regulations; Disinfectants and Disinfection Byproducts; Proposed Rule. Fed. Reg. 59:38668. USEPA. 1994b. National Primary Drinking Water Regulations; Enhanced Surface Water Treatment Requirements; Proposed Rule. Fed. Reg. 59(145):38832. USEPA. 1994c. National Primary Drinking Water Regulations; Monitoring Requirements for Public Drinking Water Supplies; Proposed Rule. Fed. Reg. 59(28):6332. USEPA. 1996. National Primary Drinking Water Regulations: Monitoring Requirements for Public Drinking Water Supplies; Final Rule. Fed. Reg. 61:24354. USEPA. 1997a. National Primary Drinking Water Regulations; Disinfectants and Disinfection Byproducts; Notice of Data Availability; Proposed Rule. Fed. Reg. 62:59388–59484. USEPA. 1997b. National Primary Drinking Water Regulations: Interim Enhanced Surface Water Treatment Rule Notice of Data Availability. Fed. Reg. 62:59486. USEPA. 1998a. Announcement of the Drinking Water Contaminant Candidate List; Notice. Fed. Reg. 63:10273–10287. USEPA. 1998b. National Primary Drinking Water Regulations; Disinfectants and Disinfection Byproducts; Notice of Data Availability; Proposed Rule. Fed. Reg. 63:15606–15692. USEPA. 1998c. National Primary Drinking Water Regulations. Disinfectants and Disinfection Byproducts. Final Rule. Fed. Reg. 63:69390–69476 (http:==www.epa.gov=safewater= mdbp=dpbfr.pdf).

REFERENCES

305

USEPA. 1998d. National Primary Drinking Water Regulations. Interim Enhanced Surface Water Treatment Rule. Final Rule. Fed. Reg. 63:38832–38858 (http:==www.epa.gov= safewater=mdbp=ieswtrfr.pdf). USEPA. 2000a. National Primary Drinking Water Regulations: Ground Water Rule. Proposed Rule. Fed. Reg. 65:30193–30274. USEPA. 2000b. Stage 2 Microbial and Disinfection Byproducts Federal Advisory Committee Agreement in Principle. Fed. Reg. 65:83015–83024 (http:==www.epa.gov=fedrgstr= epa-water=2000=december=day-29=w3306=htm). USEPA. 2001. National Primary Drinking Water; Filter Backwash Recycling Rule; Final Rule. Fed. Reg. 66:31085–31105 (http:==www.epa.gov=safewater=mdbp=fr-fbr.html). USEPA. 2002. National Primary Drinking Water Regulations: Long Term 1 Enhanced Surface Water Treatment Rule. Final Rule. Fed. Reg. 67:1812–1844. (http:==www.epa.gov= safewater=mdbp=lt1eswtr.html) USPHS. 1962. Part 72B Interstate quarantine. Fed. Reg. 27:2152–2155. USPHS. U.S. Public Health Service. 1943. Public Health Service drinking water standards. J. Am. Water Works Assoc. 35:93–103. Woodward, R. 1959. Sampling Frequency. Report of the first meeting, Advisory Committee on Revision of the Public Health Service Drinking Water Standards (met March 24–25, 1959, Washington, DC), pp. 29–37. Bureau of State Services, Public Health Service, U.S. Dept. of Health, Education, and Welfare.

13 REGULATING RADIONUCLIDES IN DRINKING WATER DAVID R. HUBER Regulation Manager, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

13.1

INTRODUCTION

Radiation is an ominous word to most people, one that evokes a wide spectrum of thoughts and emotions, both rational, and not so rational. A multitude of different, sometimes conflicting images born of the times, of ignorance, of fact, of history, may come to mind: missiles and mushroom clouds; the Manhattan Project; mutual assured destruction (MAD) policy; Yucca Mountain waste disposal; civil defense; cold war; clandestine technology; cancer cause, cancer cure; cosmic radiation; cheap energy; Chernobyl and Three Mile Island crises; X rays; fallout; and radium wristwatch dials. As a result, regulating radioactivity in drinking water is not a simple, purely rational, linear process, involving, as it does, new, changing scientific data, power politics, differing philosophical opinions and perceptions of risk, competing interagency views and policies, cost considerations, and statutory mandates. All people are exposed to radiation, both consciously and unconsciously, voluntarily or involuntarily. Some is natural, background radiation from cosmic rays, radon, and other naturally occurring radionuclides in food, air, or water, and some is created by human technology. Limitations to radioactivity exposure by the public have been established by regulating drinking water, limiting medical exposure, and waste disposal. Informing the public of risks from natural sources enables people to make reasonable choices concerning their health. But because of the significance of unavoidable natural exposures to radiation, some ask whether health risk reduction dollars could be spent more wisely in other arenas (such as highway safety or vascular disease research) where there is higher fatality, than by governmental limits in certain media such as drinking water. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

307

308

REGULATING RADIONUCLIDES IN DRINKING WATER

However, simply rank ordering and informing the public of the relative magnitudes of the risks to which they are exposed does not always add value to the debate on limiting today’s contaminant of concern. Comparisons between familiar risks and unfamiliar risks using quantitative statistical expressions can be misleading (Powell 1996). It can reduce risk to a single dimension, usually death, and trivialize the individual components of risk perception. While rank ordering risks, benefits, and costs can place hazards in a useful context for comparison, this approach is unsatisfactory because risk and responses to it are multifaceted. Some people object to smoking limitations or seat belt and helmet laws as infringing on individual liberties, while others point out that some choices affect others’ welfare and wallets. Congress has mandated protection from contaminants in public drinking water systems but not from non-public systems, such as private wells. Public drinking water systems are those which serve at least 25 people or 15 service connections for at least 60 days per year and may be publicly or privately owned. Still, limiting risk to reasonable levels in this population will be viewed as a benefit by some systems and individuals, and perceived as an unnecessary expense by others. Radium is perceived as benign by some simply because it is natural, yet it evokes profound concern in others. Relative to other accepted everyday risks we face from transportation accidents, beestings, storms, or acts of violence, sedentary lifestyles and fatty foods, smoking, and recreational sports, radiation may seem inherently more mysterious and ominous and therefore riskier. The commonness of certain events or activities may lull us into complacently accepting their risks, or else we are informed and inspired to make changes to avoid their risks. Whatever is outside of our own control, we may want ‘‘fixed’’ by someone else (i.e. the government). The choice or control over risk exposure sometimes defines the difference in concern. A risk imposed without consent (contamination of drinking water) can be of more concern than a voluntarily chosen risk (eating hamburgers and fries). The author has spoken to two pack-a-day smokers very concerned about the possibility of exposure to household radon. Perhaps it was that added risk that was of concern, but the fact that it was radiation made it seem a more real and potent threat. Exposure to diagnostic X rays may cause a modicum of concern, but it is viewed as conscious, voluntary, and necessary, while radiation risks from choosing to live a mile high in Denver, Colorado, may simply go unnoticed due to lack of knowledge, or accepted as a rational tradeoff considering other benefits. Exposure to natural radiation from a house built of brick with the latest granite countertops or floor tiles and rock fireplace may not register as being any risk at all. Radon gas in the house, though, may evoke a high degree of concern, whereas outdoor exposure to radon or cosmic rays may not be even considered. In fact, psychological research has uncovered some 47 elements influencing risk perception (Covello 1992, 1983). Slovic (1986) has identified fright, outrage, or dread factors that color perceptions of risk and make it less acceptable. For example, anxiety rises if risk is involuntary, seen as inequitable, inescapable, is novel or unfamiliar, is man-made rather than natural, is hidden and causes irreversible

13.1 INTRODUCTION

309

damage, is a danger to children and future generations, arouses dread, produces identifiable victims (catastrophic group disaster e.g. airline crash), appears poorly understood by science, and elicits contradictory statements from responsible sources. Covello and Merkhofer (1994) also add to the above such factors as institutional trust, media attention, accident history, clarity of benefits, reversibility, personal controllability, personal stake and attributability to the list. In the area of risk communication, it is axiomatic that perception is reality. The message sent is not necessarily the message received. Personal filters from nature and nurture color perceptions of a person’s environment. Explaining the government’s best intentions during a rulemaking is fraught with communication risks. The government operates largely under the precautionary principle: better safe than sorry. But the question is how much risk insurance are people willing to pay for. An ounce of prevention is worth a ton of cure, yet to some people, for some contaminants, the expression is turned on its head: an ounce of perceived risk is not worth a ton of prevention dollars. Sandman (1987) states that the public sees risk as more than hazard. It includes other factors such as dread, outrage, fairness and control. ‘‘The public’’, he says, ‘‘pays too little attention to hazard, the experts pay absolutely no attention to outrage’’. While a scientific assessment of risk as described by NRC (1983) entailing exposure evaluation, hazard identification, dose response and finally a risk characterization, may not take into account outrage, regulators and risk managers must pay attention to both. As discussed in this chapter, USEPA had to balance the desire for absolute minimal risk from some citizens and groups with the concerns by others for being able to afford compliance costs of the rule. USEPA decided on one hand not to raise MCLs across the board as proposed in 1991 in order to maintain existing protections. On the other hand, the agency decided not to lower the radium standard for Ra-228 in response to newer risk information, or to set the lowest possible uranium standard, because of unjustifiable costs of doing so. Scientists have differing opinions regarding data interpretation and uncertainty factors for newly emerging contaminants, much more so than radionuclides with a known mode of action, resulting in large differences in risk assessment factors such as the reference dose. Even with radionuclides, the question of absence or existence of a threshold of effect was the subject of Court action because scientists dispute the meaning of certain data. As Covello and Merkhofer (1994) point out, The current state of the art of risk assessment does not permit questions of science to be clearly separated from questions of policy. In practice, assumptions that have potential policy implications enter into risk assessment at virtually every stage of the process. The ideal of a risk assessment that is free, or nearly free, of policy considerations is beyond the realm of possibility.

Clearly, a reasonable function of government is to protect general public welfare, including protection of public health from known risks. Reducing public health risk, where it can be feasibly done, is wise public policy. At the same time, many choices can only be left up to the individual, recognizing the inevitability of risks in life, and

310

REGULATING RADIONUCLIDES IN DRINKING WATER

government or the private sector may educate people to help them make wise choices to reduce their own risks. Where Congress decides the risk to the citizen is reducible and worth federal intervention it passes a law and directs the executive agencies to write regulations to implement the law. In a time of changing public awareness regarding pollution and radiation, Congress enacted the Safe Drinking Water Act (SDWA) of 1974 (see Chapters 1 and 4). This was followed by amendments in 1986 and 1996. The first interim drinking water rules set by the U.S. Environmental Protection Agency (USEPA) in 1976 included limits on radioactivity. These regulations were revised in 2000 (USEPA 2000b). This chapter discusses key aspects of the development of these radionuclide regulations. USEPA follows the mandates of the SDWA as amended in 1996 to set limits for contaminants in drinking water as close as feasible to a level at which there is known or anticipated adverse effects to the health of persons allowing an adequate margin of safety. In writing new regulations for public drinking water, the Agency has the leeway to establish limits above where they might be set according to the formula of the law in order to maximize health risk reduction benefits at a cost that is justified by the benefits [Sec. 1412(b)(6)]. After the final regulation was published in 2000, four groups ultimately decided to file a petition for review in the US District Court to force a review of the rule. The City of Waukesha WI, along with several interveners were concerned with the radium MCL. The Nuclear Energy Institute’s primary focus was on the man-made beta and photon emitters produced by power plants and the Department of Energy; the National Mining Institute’s interest was largely in uranium, while Radiation Science and Health, an advocacy group, sought review of the zero MCLG for ionizing radiation. The Court heard arguments that USEPA 1) did not conduct a proper benefit-cost analyses pursuant to the SDWA and Administrative Procedure Act, 2) did not use best available science to determine MCLs and MCLGs, and 3) did not adequately respond to comments. The Court ruled in USEPA’s favor on all counts providing clear explanations and interpretations of portions of the SDWA (City of Waukesha v. EPA 2003).

13.2

RADIATION BASICS

A background on radiation terminology and simplified explanation of ionizing radiation is provided in this section for nontechnical readers. Additional detail is available in health physics texts, and the reader is encouraged to further explore this fascinating field. An atom consists of a nucleus of positively charged protons and neutral neutrons surrounded by shells of negative electrons. An atom normally has an equal number of electrons and protons; more of one kind or the other produces an ion. An element is defined by the number of protons, its atomic number. Radium has 88 protons, or it is no longer radium. The atomic mass is the number of protons plus neutrons. An isotope of an element has the fixed number of protons but different number of

13.2 RADIATION BASICS

311

neutrons. For example, radium 228 (228Ra), 226Ra, 224Ra, and 223Ra are isotopes of radium. The combination of 88 protons and 140 neutrons yields radium 228. An isotope of radium (with the same number of protons) yet two less neutrons is radium 226. Collectively, radioactive isotopes of various elements are commonly called ‘‘radionuclides.’’ A radioactive element ‘‘decays’’ in order to reach a more stable energy configuration. The result, if it is a different element, is called a ‘‘progeny’’ or ‘‘daughter.’’ If not stable, it also decays to another daughter, resulting in a decay chain of progeny until a stable configuration (i.e., the element lead) is reached. There are a variety of modes of decay involving different particles including alpha, beta, gamma, as well as positron and gamma waves. Some types of radiation are emitted from the nucleus and others from the electron shells. Alpha particles consist of two neutrons and two protons, the nucleus of a helium atom. When a radionuclide emits an alpha particle it loses two protons and becomes a new element. A negative beta particle is an electron, and results from the decay of a neutron into an electron and a proton. That additional proton results in the progeny being a different element. A positive beta particle is called a positron and can result from the transformation of a proton into a neutron and a positive particle resulting in a new element (with one less proton). Gamma emissions involve a loss of energy, yet do not add or subtract protons to change the atom to a new element, or change the number of neutrons to result in a new isotope. Gamma radiation, known by a broader term ‘‘photon radiation’’ in USEPA rules, is a form of high-energy short-wavelength electromagnetic radiation. Other radiation along the electromagnetic spectrum are X rays, ultraviolet, visible light, and radio waves. A gamma ray is billions of times more energetic than an ordinary photon and can be stopped by thick concrete or lead shielding. It is of concern as external radiation because of its penetrating power. A beta particle can be stopped by clothing or aluminum foil. An alpha particle, which is far more massive (a proton or neutron is about 2000 times more massive than an electron), is stopped by skin or paper. However, once ingested and inside the body, damage is more easily done to unprotected internal organs. When a heavy nuclide such as uranium is bombarded by neutrons, it can split into fission fragments, releasing large amounts of energy and producing new radionuclides. Many synthetic (human-made), elements and isotopes are produced by nuclear power plant reactors, or other controlled reactions. Approximately 3700 plus natural and synthetic isotopes are known to date. Radionuclides decay at a fixed rate specific to each element, and the amount of time for half of the atoms element to decay is termed the half-life. It ranges from fractions of a second to billions of years. Different isotopes of an element can have far different half-lives; it is 1600 years for 226Ra but 3.6 days for 224Ra. Radioactivity is described in terms of the number of transformations occurring in a given time in units called curies or becquerels. It may also be described as a dose of energy imparted to cells in units of rads (radiation absorbed doses) or rems (‘‘roentgen equivalent in man’’ units). The emissions from one gram of radium is called a curie, in honor of the co-discoverers. One curie is 3.7
1010 transforma-

312

REGULATING RADIONUCLIDES IN DRINKING WATER

tions per second. Because this value is so large, a picocurie is typically used, which is one-trillionth of a curie. This translates to 0.037 transformations per second or approximately one emission every 27 seconds. The International System (SI) unit of activity is the Becquerel, which is one emission every second. One Becquerel equals 27 picocuries. The energy imparted to cells by the emissions is called the absorbed dose and is measured in units called the rad (acronym defined above). In SI units, one gray equals 100 rads. The damage caused by radiation may affect chromosomes in the nucleus or organelles in the cell. The high energy particle or wave strips off electrons (ionizes) molecules in its path, so the particles or waves of radiation are called ionizing radiation. Physiologically based biokinetic models are used to describe the distribution and retention of radionuclides in the body, and this information is used to calculate the absorbed dose at different locations in the body. Because of differences in mass and charge, 1 rad of energy deposited by alpha particles creates more damage in a short distance than a beta particle or a single neutron, proton, or photon. The difference in damage is considered in a different unit of dose equivalent called the rem (acronym defined above) or more commonly one-thousandth of a rem, the millirem (mrem). A quality factor of 20 is used to account for the relative differences in harm between an electron or gamma or X rays, and a alpha particle. The dose equivalent in rems of one rad of alpha particles is 20 rems. The SI unit equivalent to the rem is the sievert (Sv), where 1 Sv equals 100 rem. A unit of effective dose equivalent (EDE) uses weighting factors for various organs to allow for the differences in sensitivity to the effects of radiation. Thus, 4 mrem EDE will generally have more tightly grouped range of risks for a suite of radionuclides than 4 mrem (without the weighting factors). However, size matters, and units need to be understood to communicate the protection afforded by a maximum contaminant level (MCL). Any change in units used needs to reflect the same quantity of radiation. Accordingly, care must be taken to change the amount along with the units for equivalency. Sometimes units are preferred because they look more favorable when reported. Certainly the impression given to the public by reporting 1 Becquerel of radiation is better than 27 pCi. Other times conformity with international convention is simply preferred.

13.3

SDWA REQUIREMENTS FOR RADIONUCLIDE STANDARDS

The 1974 SDWA provided for a health-based recommended maximum contaminant level (RMCL), and an enforceable standard, the MCL. The amendments of 1986 stipulated that USEPA simultaneously establish the two numbers when writing a National Primary Drinking Water Regulation (NPDWR). The RMCL was changed to a maximum contaminant level goal (MCLG), a health-based numerical target number, and is the level at which no known or anticipated adverse effect on the health of persons occur and that allows an adequate margin of safety. The other is the enforceable standard, the MCL. The MCL is to be set as close as feasible to the

13.3 SDWA REQUIREMENTS FOR RADIONUCLIDE STANDARDS

313

MCLG using the best available treatment (BAT) technology in field (not laboratory) use, taking costs into consideration. 13.3.1

Linear No-Threshold Assumption

Ionizing radiation causes cancer, an established fact observed in humans at high radiation doses (the Japanese bomb survivors, radium dial painters, worker exposure to radioactivity) and in animal experimentation. Ionizing radiation, in stripping electrons from molecules in vital places (in DNA), can initiate or promote a cancer, although a small amount of radiation produces a statistically tiny probability of that happening. However, the greater the number of particles emitted the greater the risk that one particle will cause a serious malfunction leading to uncontrolled cell growth. USEPA has adopted by policy an MCLG of zero for most carcinogens. This policy that all exposure to carcinogens be prevented where possible because any exposure can conceivably result in malignancy, has its origins in the Eisenhower administration. Detection of carcinogens at lower and lower concentrations is now possible, representing smaller and smaller risks. The Agency generally sets MCLs so that the resulting cancer risk falls within a band of 1
104 to 1
106 (1 in 10,000 to 1 in 1,000,000). USEPA utilizes the linear nonthreshold (LNT) hypothesis, which presumes that a single insult to a cell is enough to initiate cancer (that there is not starting point or threshold of effect) since experimentally, a single alpha particle traversing a cell nucleus is sufficient to induce a mutation caused by un- or misrepaired DNA lesions and chromosomal damage. It presumes that a straight-line relationship exists between radiation dose and cancer risk. This straight line links the damage from a single ionizing event to the observed cancers in larger doses. Whether the relationship between the radiation and the effect is linear or quadratic or something else, the severity of the resulting effect (cancer) is not related to the number of particles that cause the cancer, which would be a ‘‘deterministic’’ effect (at high doses, the severity of acute radiation sickness is related to dose), but the probability of cancer is proportional to the number of particles being fired at the cells by the radioactive element, a ‘‘stochastic’’ effect. A degree of scientific controversy exists regarding the existence of a threshold of effect at the low environmental levels of radiation. Is there a point for some or all radionuclides or some types of cells where cells are not affected, or are able to adaptively respond at low levels to the insult to protect the cell at higher levels? Is there a stimulating or beneficial ‘‘hormetic’’ effect of radiation? Or is the cellular response in defense a sure sign of the adverse nature of the threat? This is a continuing area of research and debate, but without convincing proof to the contrary, the LNT model is utilized for regulatory purposes. 13.3.2

Non-cancer Effects

The MCL for uranium is based on chemical toxicity, not cancer, although cancer is also a concern. For noncancer effects, the MCLG is determined by several factors. The reference dose is calculated by dividing the lowest, or no-observed-adverse-

314

REGULATING RADIONUCLIDES IN DRINKING WATER

effect level (LOAEL or NOAEL) by appropriate safety factors (from 3- to 10-fold to account for inter- and intraspecies variability, short versus chronic study, LOAEL versus NOAEL to arrive at a reference dose (RfD) in units of milligrams per kilogram per day (mg kg1 day1). Multiplying by 70 kg adult weight and dividing by 2 L=day consumption yields the drinking water equivalent level (DWEL) number in milligrams per liter (mg=L). This number is multiplied by the relative source contribution (RSC), to determine the MCLG. The RSC is the proportion of the total contaminant consumed via drinking water versus food, dermal absorption, or inhalation. The higher the RSC, the correspondingly higher the MCLG for non-carcinogens. But, simply taking drinking water as a percentage of food plus other exposures may show it to be a high percent or a low percent with no relationship to the total reference dose allowed before effects are seen. For example, if water is only 20% of the total exposure from all sources, multiplying the RfD (converted to the DWEL) by 20% would yield a small MCLG. Yet if that total from all exposures is only 50% of the RfD, the MCLG might be set artificially low because it leaves much of the reference dose ‘‘unused’’. Even though water was 20% of the total exposure, the total was only half of the RfD allowance. Alternatively, an arguably better method is to subtract the actual amounts from food or other exposures leaving the rest of the RfD as the limit for water. The MCLG may be more correctly set higher, because after exposures from food etc. are subtracted from the RfD, there is still much leeway to what is an effect level. It leaves that portion of the RfD (or less for an additional safety factor), to become the MCLG. In the above case if the DWEL were 100, the first MCLG (by percentage) would be 20, and the second (by subtraction) could be 60 (80% of the 50% is food, etc. ¼ 40, leaving 60 for water). Additional safety factors to account for error may be applied to lower that level. Better data can change any one of these factors and may result in a change to the calculated MCLG when USEPA reviews regulations and revises them as required every 6 years. For carcinogens with no proven threshold of effect, the MCLG remains at zero because the adverse cellular effect is seen in concentrations above zero.

13.4

1976 RADIONUCLIDE REGULATIONS

There are several possible approaches for setting radionuclide limits: limiting either the concentration, the dose, or the risk, and doing so either for individual nuclides or for a group of nuclides. Considering the relationships between concentration, dose, and risk may be helpful. The concentration in picocuries of emissions in water result in a dose of energy imparted to the body, which then results in a certain risk of disease or death. A different concentration of each nuclide is needed to produce a given dose, and that dose results in different risks depending on the organs affected. Thus, whichever expression, concentration, dose or risk is selected to be the MCL the other two will vary considerably. Set the limit as concentration, and the dose and the risk vary widely. Set the dose as a limit, and risk and concentrations vary. Set risk as a constant, and concentrations and dose differ.

13.4 1976 RADIONUCLIDE REGULATIONS

315

USEPA usually sets limits for individual contaminants. But, having several contaminants present at their limits increases the total risk in the drinking water. Therefore, limiting a group of contaminants in aggregate allows individual latitude as long as the mixture does not exceed the standard. In 1976, USEPA regulated radium, alpha-particle emitters, and beta-particle and photon emitters under the 1974 SDWA. USEPA expected new data, and to change the rule within a short timeframe as new models became available and termed the rule national interim primary drinking water regulations. The 1976 rule set an MCL of 5 pCi=L as combined standard for 226Ra þ 228Ra. This limit provided protection for the most people at a reasonable cost given the data at the time. The risk was believed to be between 5
105 (1 excess death in 20,000) and 2
104 (1 in 5000). Alpha emitters were set at a total limit of 15 pCi=L and termed gross alpha. ‘‘Gross,’’ however, did not include either uranium, or radon, both alpha emitters, because they would be included in future rules. The uranium limit was finally established 24 years later, in the 2000 rule, but regulation of radon is still pending. There is no single risk number associated with the 15-pCi=L limit for alpha emitters because the concentration limit applies regardless of the nuclide. Because each nuclide emits alpha radiation of different energies and may accumulate in different organs of varying sensitivity to radiation, 15 pCi=L yields greatly different risks depending on the nuclide in question. The beta particle and photon emitters’ MCL was established as a dose to the whole body or to the critical (most affected) organ, accompanied by a table of the picocurie concentration limits for each nuclide at 4 mrem. The nuclides and limits were calculated on the basis of a 2-L=day drinking water intake using the 168-hour data listed in Maximum Permissible Body Burdens and Maximum Permissible Concentrations of Radionuclides in Air or Water for Occupational Exposure (NBS Handbook 69), which was incorporated by reference into the rule. The 1976 rule attempted to limit maximum risk to a reasonably conservative level: one excess death per million people per year from synthetic radiation in drinking water. One a year translates to 70 times that in 70 years, and thus, 7
105 risk. USEPA selected a dose limit of 4 mrem as being closest to the risk limit for whole-body dosers, although other nuclides would represent a much smaller risk at that same dose. (The origin of 4 mrem is less certain, reportedly a compromise between a log number of 3 mrem and an average of 5 mrem.) Regardless, in performing an analysis of risk of the selected 4 mrem, the geometric mean of the relative and absolute models was 0.8 deaths per million persons per year rather than one death. Therefore the lifetime risk (0.8
70 years) was 5.6
105. This was meant to be a ceiling or maximum risk represented by a whole-body doser, not a target. It was not intended to be a goal like an MCLG to which all of the nuclides should be as close to as possible. Rather, it was to be a ceiling, the highest risk, represented by the whole-body dosers. (Later, in 2000, an MCLG of zero would be published for ionizing radiation.) In setting a dose limit USEPA understood that risks would vary widely. The greatest risk could be estimated at 4 mrem, but having both a constant risk and a constant dose was not possible. The Court ruling explained that USEPA is not required to set or revise numbers to achieve a constant risk (City of Waukesha v. EPA 2003).

316

REGULATING RADIONUCLIDES IN DRINKING WATER

This approach of setting a dose limit makes each nuclide equal to the overarching limit of dose or risk that serves as the MCL for the group. For example, the beta particle and photon emitters have a limit equal to the dose of 4 mrem. There is of necessity a table of individual concentration values for each radionuclide equal to that 4 mrem dose for monitoring purposes. Concentration has to be specified because water systems monitor the concentration, not the dose or risk. Any combination of nuclides would be allowed as long as all together (aggregately) they did not exceed the assigned dose limit, in this case 4 mrem. This is the ‘‘sum of the fractions’’ approach and could work for a limit based on risk as well. Each nuclide’s individual concentration limit is set equal to the whole dose or risk selected as the MCL. The quantity of a nuclide actually present would be divided by its picocurie allowance resulting in a fraction of what is allowed. If 14 pCi of nuclide Q equals the MCL of, say, 4 mrem, and 14 pCi is present, the limit of 5 mrem is reached. If any other nuclide is present, it will cause the limit to be exceeded. However, if two nuclides are present at half their respective picocurie allowances, that also means that each is at half of the 4 mrem allowance as well. Both halves added bring the water to the limit. If there are multiple nuclides, the sum of all the fractions of individual 1 þ 14 ¼ 43 Þ: Since this exceeds one concentration limits present (e.g. 16 þ 23 þ 12 (unity) the MCL is exceeded. Compliance with the MCL would mean that one or more of the constituents could be treated until the sum of the fractions is one or less. A third approach is to limit total risk from any combination of nuclides. Risk as the MCL approach would not have been feasible in 1976 because of the inability to specify a risk for a given concentration for each nuclide. By 2000, USEPA could calculate risks at a given concentration using the Federal Guidance Report 13, published in September 1999 (USEPA 1999). A caveat is in order with respect to risk. If a risk limit were set for a group or an individual nuclide in drinking water such as 226Ra, the concentrations equal to the given risk would change as risk models changed. The limiting concentrations would rise or fall based on model results. This could pose great difficulties for public water systems by putting them in or out of compliance with the MCL especially if risk were done on an individual basis. A total group risk ceiling allows variations to constituents as long as the total risk cap was not exceeded, but a calculation of risk via modeling carries its own risks! USEPA could have set a limit on the total concentration of beta=photon emitters just as it did for alpha emitters. The risk would then vary considerably depending on the individual nuclides present. Alternatively, each nuclide could also have separate concentration limits as individual MCLs where the total dose, and the total risk, would be driven by the combination of nuclides actually present in the water. However, when USEPA developed limits for beta and photon emitters in 1976, dose was selected as the limit for a very practical reason: to retain the ability to revise concentration numbers as dose conversion methodologies matured, without having to change the overall MCL. USEPA expected the numbers to change in a few years, and published the rule as interim. However, 10 years later, the numbers had not been changed in any way, and the SDWA in 1986 declared that all interim rules were final national primary drinking water regulations (NPDWRs). Regulation writers reasoned in 1976 that if the concentration was changed periodically, but still limited

13.5 1991 PROPOSED RADIONUCLIDES RULE

317

to a 4-mrem dose, only a new handbook need be published correcting the table, rather than a rulemaking, with its attendant proposal, public comment and final rulemaking. With an MCL as dose, an underlying number could change while leaving the MCL itself, the 4-mrem dose, alone. What seemed like a good idea in 1976, in later years, did not seem as wise. In the late 1990s, the USEPA drinking water program decided against an effort to change the beta and photon emitters to higher limits and to allow future concentrations to float by tracking on periodic guidance documents. USEPA believed that changing the numbers to which water systems would be accountable rightly deserved a period of public comment. Even if the MCL did not change, if the concentrations of individual constituents tied to that dose limit change, it affects treatment at water systems and at sites where remedial actions use the limit for that constituent as a cleanup target. Furthermore, changing concentration, changes the risk, and the 1986 Amendments specified that MCLs be revised whenever treatment technology would permit greater protection to the health of persons.

13.5

1991 PROPOSED RADIONUCLIDES RULE

In 1986, USEPA had published a list of 83 possible contaminants for future regulation for public comment in an Advanced Notice of Proposed Rulemaking (ANPRM). However, later that year, the SDWA reauthorization passed, and Congress responded by requiring USEPA to regulate all 83 contaminants, including uranium and radon. The 1986 SDWA prescribed a review of the rules every 3 years. In addition, MCLGs for radionuclides were needed. In 1991, USEPA published a proposal increasing both radium 226 and 228 from 5 pCi=L together, to 20 pCi=L each (USEPA 1991). In addition, 226Ra would have been removed from the count of the gross alpha limit of 15. The new limit, termed adjusted gross alpha, would allow greater amounts of other contaminants to take 226 Ra’s vacated place within the limit. The 1991 proposal included a limit of 20 mg=L (according to the rule, equivalent to 30 pCi=L) for uranium. It also changed the MCL for beta and photon emitters from 4 mrem to 4 mrem EDE. The latter change appeared as though the MCL had remained the same, but used newer units. However, the change in units changed the associated concentration limits dramatically. On average for the synthetic (human-made) nuclides, 4 mrem is actually equivalent to about 0.9 mrem EDE for the aggregate of nuclides. Regardless of the dose units, the actual radiation needed to remain the same for the effect on the body to remain the same. To illustrate, in units of Fahrenheit, 72 F is comfortable. But in switching to a new brand of thermostat using Celsius units, the temperature should not be set again to 72 C but to 22 C. At 72 C the room would be decidedly hotter, the equivalent of about 162 F. Likewise changing from 4 mrem to 4 mrem EDE made the allowance of the radionuclides’ concentration in picocuries ‘‘hotter.’’ At a minimum, the change in units should have been accompanied by a change in the MCL number from 4 mrem to 0.9 mrem EDE. However, because the rule covered many individual nuclides and not an average, it still would have resulted in many nuclides being at higher picocurie concentrations than they had been before.

318

REGULATING RADIONUCLIDES IN DRINKING WATER

The 1991 proposed limit for radon was controversial, and a series of Congressional actions delayed final action on the 1991 radionuclides proposal. Furthermore, the regulatory approach was novel if not unorthodox. The Agency attempted to apply an overall risk bubble concept of costs and benefits for radium and radon to avoid disproportionate costs of controlling radium versus radon. Radon is less expensive to treat because it is volatile. A greater risk reduction ‘‘bang for the buck’’ could be obtained by treating radon than radium. Therefore, the risks and rewards of both contaminants were considered together in the proposal. Radium was separated into two different MCLs: the beta-emitting 228Ra and the alpha-emitting 226 Ra each at 20 pCi=L. By adding in the benefit of regulating radon, these contaminants could be given higher limits, which were thought to be less risky than before. However, a rule to allow more radium, while removing some radon, might have a positive overall impact in the northeast radon corridor, but would have negative consequences in the upper Midwest, where radium is prevalent. Nationally the risks may have been lower overall, but regionally they were not. The concept of protecting the ‘‘health of persons’’ in SDWA is rooted in consideration of individual not collective risks. In essence, the 1991 proposal suggested relaxing the standards on all fronts. Unfortunately, this had long-term negative consequences for the water systems in compliance with the 1976 rule, and for those systems not in compliance. Some water systems were hesitant to install expensive treatment or even to explore alternative water sources if USEPA intended to declare their water to be safe by increasing the MCL. USEPA had given the states’ ample discretion in enforcement, but this was a tough policy call. On one hand, the 1986 Amendments declared interim rules to be final rules. Some systems would fall back on the ‘‘interim’’ status even in the year 2000, as an excuse for noncompliance since 1976. At least one city would be sued by citizens and lose. On the other hand, a proposal to increase limits gave plausible pause to both systems and enforcers. On one hand, the 1976 rule was still the law and enforceable until changed. The 1991 proposal did not legally count in decisionmaking. On the other hand, it did telegraph the Agency’s direction of thought. Unfortunately, 9 years ensued to a final rule, perpetuating the dilemma to those systems out of compliance, and causing a hardening of the attitudes for those with varying points of view. Fortunately, the 2000 final rule is more scientifically defensible than the 1991 proposal. In particular, the 2000 final rule was crafted with the spirit and letter of the law on an important point: it ensures that protection obtained is protection maintained. Significantly for this rule and future rules, a later court ruling discussed below affirmed that revisions to MCLs should not backslide in protection.

13.6

1996 SDWA AMENDMENTS AND RULE REVISIONS

A change in an MCL has an effect on public health and the operation of the water treatment systems. In the 1974 SDWA, Section 1412(b)(4) stated that ‘‘revised national primary drinking water regulations shall be amended whenever changes

13.6 1996 SDWA AMENDMENTS AND RULE REVISIONS

319

in technology, treatment techniques, and other means permit greater protection of the health of persons, but in any event such regulations shall be reviewed at least once every 3 years’’. Again in 1986 the SDWA amendments addressed this issue in Section 1412 (b)(9) with very similar yet even stronger language. In the 1996 Amendments, what became known as ‘‘antibacksliding’’ was debated in Congress and the intent of the 1986 Amendments was retained in slightly different language. Revision to the 1976 rule was subject to the following provision [Sec. 1412(b)(9)]: The Administrator shall, not less often than every 6 years, review and revise, as appropriate, each national primary drinking water regulation promulgated under this title. Any revision of a national primary drinking water regulation shall be promulgated in accordance with this section, except that each revision shall maintain, or provide for greater, protection of the health of persons.

A critical factor in writing a defensible rule is ensuring that it complies with the mandates of the enabling legislation. A reasonable person reading the plain sense of the law can understand the intent of Congress. The House of Representatives and Senate legislative histories illuminate possible uncertainties. The reasonable reading, the simple, plain meaning is the first choice. Very rational sounding, but wrongheaded arguments can be put forward by overcomplicating the terms and introducing concepts that are not present in the legislation. For example, the word ‘‘protection’’ became a source of some misunderstanding arising from using ‘‘risk’’ and ‘‘protection’’ interchangeably. That issue arose in terms of the radium and alpha standards, and in a different form for the beta and photon emitters. ‘‘Maintaining protection’’ is not synonymous with ‘‘maintaining risk’’ because risk is calculated on the basis of changing models. Protection relates to deterrence or shielding from harmful agents, in this case, ionizing radiation. Therefore, the numerical limit, the drinking water MCL, guards against excess consumption of radionuclide-contaminated water. Risk is the exposure to the chance of harm. It is expressed as an estimation of danger or harm arising from the lack of protection from a harmful contaminant. The concentration that still remains in the water at or below the MCL is what causes the risk. Increasing the numerical limit causes an automatic increase in risk, given an upward dose–response relationship. It means more of a bad thing . . . is a worse thing. But, can an MCL ever increase above its current level if feasibility is questioned, or if the MCLG increases, or if its risk has declined? With respect to changing the limit based on feasibility=cost considerations the SDWA exempts pre-1986 MCLs from its cost benefit requirements and USEPA left the MCLs unchanged. The Senate report (Congress 1995) on the SDWA amendments of 1996 explains that [S. Rept. (Senate Report) 104-169] It is quite possible that a future Administrator will be required to issue or reconfirm an existing standard with costs that the Administrator does not believe are justified by the benefits. Because the valuation placed on the benefits achieved by a regulation is

320

REGULATING RADIONUCLIDES IN DRINKING WATER

necessarily shaped by the subjective judgment of the Administrator, it is to be expected that some future occupant of the position may find a standard issued by a predecessor too costly for the benefits obtained. Nevertheless, section 1412(b)(9) would require that the standard be reissued or retained.

An MCL is set as close to its MCLG as feasible. With an existing rule, because systems are meeting the MCL, its feasibility is demonstrated and the standard cannot be weakened in the future by a determination that attainment of the standard is not feasible for systems today. The issue of MCL relaxation is tied to changes in the MCLG. The Senate report (Congress 1995) added a footnote that an ‘‘existing standard may be relaxed, but not on the grounds of a cost-benefit analysis. If new science shows that a less stringent standard would provide the same level of health protection, the MCL may be revised upward’’ (S. Rept. 104-169). Earlier, the Senate report explained the circumstance in which an MCL may be relaxed: if new scientific information causes the MCLG to be revised to a higher number than the MCL, the MCL would change upward because it need not be more stringent than the MCLG and overshoot the goal. The MCLG is set at a level of no known or anticipated adverse effect on the health of persons with an adequate margin of safety. Calculations of it therefore contain safety factors. These may change as better data on health effects decrease the uncertainty, accounting for extrapolations from animal to human, between humans, for a short- versus a longterm study, for a lowest observable adverse effect level (LOAEL) instead of a no observable adverse effect level (NOAEL), or if the relative source contribution changes. With better data from a better study, the uncertainty becomes smaller, and the calculated MCLG can rise. For example, if an uncertainty factor were to change from 1000 to 100, the calculated no-effect level, the MCLG, would rise by a factor of 10, say, from 5 to 50 parts per billion (ppb). If the MCL had been at the MCLG, it would also rise to 50, ‘‘since it need be no more stringent than the MCLG.’’ The same is true for carcinogens. If one is reclassified or assigned a threshold, the MCL need only be at that threshold of no effect. However, both numbers do not necessarily rise. If the MCL has been set at 20 and the MCLG rises from 0 to 10, the MCL remains the same (or can drop if feasible). The aim is to get as close to the MCLG as possible and the MCL of 20 is now numerically closer to the new goal (MCLG) of 10 than it had been to the old goal of 0. Only if the MCLG were to overtake the MCL (rise above 20) would the latter also rise. Can an MCL be revised to maintain risk? In the case of radionuclide regulations the estimate of risk is constantly changing as new information from more sophisticated modeling methods becomes available. If the models change, presumably for the better, can MCLs be adjusted regularly to match the risk we envisioned as being protective? In other words, what if the risk of a radionuclide is less than originally thought? Could its concentration then be increased to the originally anticipated risk while still maintaining our same protection? The answer to the question is ‘‘No,’’ for two reasons: 1) The focus is wrong—are we getting closer to the goal, the MCLG, or

13.6 1996 SDWA AMENDMENTS AND RULE REVISIONS

321

not? The MCLG is the target and while the MCL selected represents some risk, the risk number is not the target. 2) As discussed below, any increase in concentration also increases risk relative to what it used to be. An illustration demonstrates the fallacy of equating ‘‘maintaining risk’’ with ‘‘maintaining protection.’’  Assume a risk of 1 in 10,000 of getting cancer was considered to be 7 pCi=L for a nuclide.  A new model now shows the 1 in 10,000 risk to actually be 14 pCi=L instead.  If the MCL was 7 pCi=L, raising it to 14 pCi=L would not ‘‘maintain’’ the risk of 1 in 10,000. Instead using the same new model to calculate the risk of 10 pCi=L will reveal 7 to be actually 1 in 20,000 risk.  Increasing the MCL from 7 to 14 actually increases the real risk in the water 2 fold (from odds of 1 chance in 20,000 up to one in 10,000) and thereby simultaneously decrease protection 2 fold.  Protection is maintained by retaining the MCL of 7, regardless of the risk it was perceived to be. The radionuclide did its damage irrespective of the correct or incorrect risk assigned to it. Maintaining the concentration maintains the protection by maintaining the associated actual risk at any concentration. A greater concentration will be a greater risk. Only by using the two different models, one old and one new, can the same risk be assigned to two different concentrations, leading to the false conclusion that risk was maintained. Only with an odd U-shaped dose–response curve could 7 and 14 have the same risk. Or, if a threshold is discovered at 30, both 7 and 14 are the same, no risk—although it could be argued that 10 is still more protective, allowing a greater cushion before an effect would be felt. Otherwise with a linear model of risk, a lower number is always a lower risk than a higher number regardless of the numerical risk assigned. An MCL’s actual risk at the numerical limit may be simply lower than was thought, thus making the protection greater than imagined. Maintaining risk by changing to 14 is an illusion, because the erroneous calculated level of risk is used, and protection is not provided to the same actual level of risk. Protection cannot be maintained while increasing the concentration of the substance regulated. Protection is inversely related to the amount of exposure to the harmful agent from which protection is desired. From another perspective, even if the calculation (modeling) of risk changes, the radioactive element itself does not. If today’s model shows that 7 pCi=L of an element is half the risk thought years ago, could someone drink twice as much radiation, 14 pCi=L, and have the same protection from radiation? The question is whether the effect on one’s body is still the same at 14 as it is at 7. In the absence of a threshold of effect, it certainly is not. Therefore, all things being equal, a higher concentration (picocuries, or mrem, or risk) is less protection. Only the misperception of equal risk would be maintained, but the actual risk to the public would rise.

322

REGULATING RADIONUCLIDES IN DRINKING WATER

In Waukesha et al. v. EPA (2003), the Court stated: Because the SDWA defines ‘‘maximum contaminant level as ‘the maximum permissible level of a contaminant in water which is delivered to any user of a public water system, (emphasis added), EPA is right to focus on the level of contaminant set by the original MCL rather than the degree of protection that such a level was anticipated to provide . . . Petitioners contend that this provision does not prohibit EPA from revising an MCL upward when (as here) scientific advances show that a contaminant poses less risk than previously believed . . . This argument requires inferring the following bracketed and italicized qualification to the actual language of x[1412](b)(9): ‘[E]ach revision shall maintain, or provide for greater, protection of the health of persons [than the agency initially thought it was providing].’ But there is nothing unreasonable about EPA’s decision to decline to read such a qualification into the section and instead to regard it as a straightforward instruction to maintain the level of protection that the initial MCL actually provides.

Some ask if a revised risk estimate puts the risk of a contaminant below the low end of the Agency’s risk range of 106, should the MCL be increased to that policy de minimus threshold? Again the Senate Report (Congress 1995) speaks to the issue: This amendment to the law does not provide the Administrator with authority to set MCLGs based on a finding that the cancer risk is negligible or so small as to be acceptable; the Administrator is not authorized to use the authority to set a ‘‘policy’’ threshold below which increased cancer risks are not considered in standard setting.

As a practical matter, if USEPA regulated simply based on risk estimates, the MCL would rise and fall with the fortunes of each risk model. For example, the original calculation for radium 226=228 was a risk between 5
105 and 2
104 for a combination of 226Ra and 228Ra at 5 pCi=L. In later years, estimates could be made separately, and by the 1991 proposed rule, 1
104 was thought to be 22 pCi=L for 226 Ra and 26 pCi=L for 228Ra using the RADRISK model. Today’s model shows Ra226 at 22 pCi=L to be 10 fold higher at 1
103 and Ra-228 at 26 pCi=L to be 3.3 times higher risk (USEPA 2000a). The proposed MCLs selected in the 1991 proposal were 20 pCi=L each. The 1991 standard would have allowed up to 40 pCi=L total radium or about 4–20 times higher than high=low combinations of radium (e.g., 5 pCi=L of 228Ra and 0 pCi=L of 226Ra in the 1976 combined standard). Had those limits been selected as MCLs in 1991, a new rule would again have to be promulgated to bring them back in line with the new model’s calculations, resulting in a moving target and extreme uncertainty for the drinking water industry and for the public.

13.7

2000 FINAL RADIONUCLIDES RULE

Briefly, the final 2000 rule (USEPA 2000b) maintained protection from radionuclides, but made certain changes to monitoring requirements. MCLs are summarized

13.7 2000 FINAL RADIONUCLIDES RULE

323

in Table 13.1. Monitoring frequencies for gross alpha, uranium, and radium 226=228 are summarized in Table 13.2. Figure 13.1 illustrates the basic components of isotopes. Entry points to the distribution systems are now monitored to ensure that some consumers were not on ‘‘hot’’ wells while others were on ‘‘cool’’ wells, as when monitoring was done at a representative blended point. Hence, greater protection should result. Monitoring for beta and photon emitters is only conducted by vulnerable systems. BAT and small system compliance technologies for radionuclides are summarized in Tables 13.3 and 13.4, respectively. Compliance deadlines are listed in Table 13.5. USEPA has developed several helpful guides to assist states and water utilities in complying with this rule (USEPA 2001, 2002a, 2002b). Nontransient, noncommunity systems (those serving less than 15 service connections for more than 6 months out of a year, e.g., schools, nursing homes, hospitals, churches, campgrounds, prisons) are not regulated because of doubts regarding the exposure that the typical consumption would represent, although these systems may be included in the rule in the future as they are for other drinking water rules. Approximately 60% of risk is given in the first 18 years of a lifetime of exposure to radium. Because schools can represent an important point of exposure to children, any doubts regarding regulation should be decided in favor of children’s health. 13.7.1

Alpha Emitters

In the 2000 rule, the gross alpha count continued to include 226Ra, an alpha emitter, and exclude uranium and radon as in the original 1976 rule. It is more helpful to think of the MCL as ‘‘net alpha’’ (the total alpha from the gross alpha analytical method, minus radon and uranium for a net result). The proposal in 1991 removed 226 Ra from the count of alpha emitters because it was proposed to be regulated separately, terming the level ‘‘adjusted gross alpha.’’ However, removing it would allow other alpha emitters to be present at higher levels replacing the 226Ra, which was no longer counted toward the 15 pCi=L limit. The new standard would therefore not maintain protection. The term ‘‘gross alpha’’ is better used for an analytical method, and the term ‘‘net alpha’’ is better used to refer to the MCL, because it excludes Rn and U. However, because the term has been in use for so long, USEPA retained the old terminology. 13.7.2

Radium 226=228

As discussed earlier, USEPA was following whenever the models led in the 1991 proposal. It led in a proverbial circle from a perceived risk of about 1
104 at 5 pCi=L in 1976 to above 20 in 1991 to 10 in 1994 and back to 5 in 2000. In the case of 228Ra, the calculated risk using the Federal Guidance Report 13 (USEPA 1999) is now greater per picocurie than previously thought, resulting in a 2
104 risk at 5 pCi=L, the upper risk range estimate in 1976. However, to maintain the same protection from radiation, the numerical limit could remain the same. After all, the results in the body of the same amount of radiation did not change even though our perception of its risk did. USEPA considered the feasibility of lowering

324

b

Source Naturally occurring Naturally occurring Naturally occurring Contamination

MCL 15 pCi=L (no radionuclide þ uranium) 5 pCi=L 30 mg=L 4 mremb=year

A total of 179 individual beta particle and photon emitters may be used to calculate compliance with the MCL. Milliroentgen equivalents per man.

Zero Zero Zero

Radium 226 þ 228 Uranium Beta particle and photon emittersa

a

Zero

MCLG

Radionuclide

Gross alpha emitters

Radionuclide MCLs

TABLE 13.1

Cancer risk Kidney toxicity; cancer risk Cancer risk

Cancer risk

Health Effect

13.7 2000 FINAL RADIONUCLIDES RULE

TABLE 13.2 226=228.

325

Monitoring Frequencies for Gross Alpha, Uranium, and Radium

Initial 40 CFR 141.26(a)(2)

Four consecutive quarters of monitoring at each entry pointa

Reduced 40 CFR 141.26(a)(3) Gross Alpha and Uranium One sample every

Systems may composite up to four consecutive quarterly samples from a single entry point if analysis is done within a year of the first sample

Nine years if average of the initial monitoring for each contaminant is below the detection limit listed in 40 CFR 141.25(c) Six years if average of the initial monitoring results for each contaminant is at or above the detection limit but at or below 1 2 MCL Three years if average of the initial monitoring results for each contaminant is above 1 2 MCL but at or below the MCL If the results from the composited sample are less than 12 MCL, reduce in accordance with the given schedule above

Combined Radium 226=228 Four consecutive quarters of One sample every monitoring at each entry point

Nine years if average of the initial monitoring for combined radium 226=228 is below the detection limit listed in 40 CFR 141.25(c) Six years if average of the combined initial monitoring results for combined radium 226=228 is at or above the detection limit but at or below 12 MCL Three years if average of the initial monitoring results for combined radium 226=228 is above 12 MCL but at or below the MCL (continued )

326

REGULATING RADIONUCLIDES IN DRINKING WATER

TABLE 13.2

(Continued )

Systems may composite up to four consecutive quarterly samples from a single entry point if analysis is done within a year of the first sample

If the results from the composited sample are less than 12 MCL, reduce in accordance with the schedule given above

Systems may substitute the gross alpha results that are 15 pCi=L for uranium to determine compliance and the reduced monitoring frequency. Systems with gross alpha result greater than 15 pCi=L must collect uranium sample(s) to determine compliance and reduced monitoring, per 40 CFR 141.26.(a)(5).

a

the absolute limit of 228Ra to 3 pCi=L (within a combined limit of 5) to reduce the overall risks of 228Ra individually or as a combination of 226Ra and 228Ra. After examining the costs and benefits of a lower limit of 3 pCi=L for 228Ra, the Agency concluded that the incremental risk reduction benefits was not justified by the costs. Radium is treatable to a lower level, closer to the MCLG, but the high cost is not worth the fraction of a life saved as a result.

Figure 13.1 Isotope nomenclature, illustrating two ways to describe radioisotopes.

TABLE 13.3 BAT for Radionuclides Contaminant Gross alpha emitters (less Rn and U) Radium 226 þ 228 Uranium Beta particles=photon emitters

BAT Reverse osmosis Ion exchange, reverse osmosis, lime softening Ion exchange, reverse osmosis, lime softening, coagulation=filtration Ion exchange, reverse osmosis

327

13.7 2000 FINAL RADIONUCLIDES RULE

TABLE 13.4

Small System Compliance Technologies for Radionuclides Compliance Technologies Appropriate for System Sizea

Technologies Ion exchange (IE) Point of use (POU) IE Reverse osmosis (RO) POU RO Lime softening Green sand filtration Coprecipitation with barium sulfate Electrodialysis=electrodialysis reversal Preformed hydrous manganese oxide filtration Activated alumina Enhanced coagulation=filtration

25–500

501–3300

3301–10,000

C,a B,b Uc C, B, U C, G,d B C, G, B, U C C C

C, B, U C, B, U C, G, B, U C, G, B, U C, U C C

C, B, U C, B, U C, G, B, U C, G, B, U C, U C C

C

C

C

C

C

C

U U

U U

U U

a

Combined radium 226=228. Beta particle activity and photon activity. c Uranium. d Gross alpha-particle activity. b

Previously, the gross alpha method was used to screen for alpha emitters and the result was to be presumptive for 228Ra, a beta emitter in the following manner. Alpha emitters were limited to 15 pCi=L. If the test result was at or below 15 pCi=L, the MCL was met for gross alpha. If results were above 5 but below 15, the result could represent what the 226Ra concentration in the water was, and 226Ra would have to

TABLE 13.5 Radionuclide Rule Compliance Deadlines Date

Requirement

June 2000

Dec. Dec. Dec. Dec.

8, 8, 8, 8,

2000 2002 2003 2003

Dec. 8, 2004 Dec. 31, 2007

Data collected between June 2000 and Dec. 8, 2003 may be eligible for use as grandfathered data (at state discretion) to satisfy the initial monitoring requirements for gross alpha, radium 226=228, and uranium Final radionuclides rule published in Federal Register State primacy revision application package due Effective date of the rule Systems must begin initial monitoring under their state-specified monitoring plan unless the State permits grandfathering of data collected between June 2000 and Dec. 8, 2003 State primacy revision application package due for those states requesting a 2-year extension All water systems must have completed initial monitoring

328

REGULATING RADIONUCLIDES IN DRINKING WATER

be measured. If results of gross alpha measurement were below 5 pCi=L, including the high-side measurement error at 1.65 sigma (standard deviation), 95% confidence, the result could be taken as the proxy amount of 226Ra. If the result were below 3, the 228Ra need not be measured even though it is a beta-particle emitter. The reason was a presumption that there was a relationship in occurrence between the two. If 226Ra were not above 3 pCi=L, then 228Ra would be unlikely to occur above 2 pCi=L, and the sum of the two would not exceed 5 pCi=L. However, even though they are both radium by virtue of their 88 protons, the two are independent of one another; 228Ra (a beta-particle emitter) is derived from the thorium decay chain and the 226Ra (an alpha-particle emitter), from the uranium decay chain. Minerals containing U and Th co-occur, but the thorium is relatively insoluble, and it, or its daughters, may not be found in the groundwater. A reconnaissance study by USEPA=USGS=AWWA of 100 wells showed a weaker co-occurrence relationship between the two isotopes of radium than between 228Ra and 224Ra, which are in the same decay chain (Focazio et al. 1998). Because 226Ra cannot be used as a predictor of 228Ra, the new rule requires the testing of 228Ra. The result is added to either the proxy alpha result for 226Ra or actual 226Ra result for comparison to the MCL. Instead of lowering the standard for 228 Ra, USEPA was able to maintain the stringency of the standard with the same allowance of radiation, yet increase overall public health protection by decreasing the amount of 228Ra in approximately 270–320 water supplies around the country affecting 380,000–460,000 people, via the change in the monitoring requirement, which will result in finding and treating the contaminant. At least one state knew the existence of 228Ra in their water, which, when added to the amount of 226Ra present, would exceed the standard, but the old rule did not require monitoring for 228Ra separately from 226Ra results. The SDWA does not hold a system in violation of the standard even if it is exceeded, if the monitoring to discover the contaminant was not required in the first place. Uncertainty over the future standard resulted in a holding pattern for radium fixes in some states even where monitoring was required. In this case, it resulted in a policy of emphasis on the letter rather than the spirit of the law to ignore the excess radium because there was no requirement to look for it. Thus, monitoring directly for the existence of 228 Ra, a beta emitter from a different decay chain was common sense and good science, and fixing this loophole a sensible thing to do while still retaining the MCL. 13.7.3

Radium 224

Tests of gross alpha done within approximately 48–72 hours may reveal alpha activity due to 224Ra or its daughters. Radium 224 is a short-lived (3.66-day halflife) great granddaughter of 228Ra on the thorium decay chain. Within a little over a month (10 half-lives), the 224Ra as well as its very short half-life progeny has decayed. The 1976 rule had not focused on this radium as a problem in drinking water, although testing for gross alpha would include it, and it could be present in drinking water with a short residence time in the distribution system. Radium 224 surfaced as an issue when New Jersey did some sleuthing for an environmental cause

13.7 2000 FINAL RADIONUCLIDES RULE

329

for some mysterious brain tumors in children in South Jersey. Comparison of gross alpha results from four composited quarterly samples taken during the year as allowed by rule, and quick turnaround samples, showed a large disparity. The culprit (of the gross alpha result, not the brain tumors) was deduced by some keen observations to be the short half-life radium isotope 224Ra, which had decayed to a stable 208 Pb lead isotope by the time of the analysis composited samples, but was present in the tapwater. USEPA considered issuing a rule on 224Ra and decided that it was not a priority at that time for several reasons. First, mandating it would be a hardship on states who must complete the analysis within 48–72 hours in order to backcalculate the original concentration of 224Ra. The additional expense for a fast turnaround analysis and the problem of laboratory capacity was a telling factor. In addition, the risk of 228 Ra is 8 times more radiotoxic than 224Ra, and already includes the effect of 224 Ra on the body, as it is in the decay chain. A quick turnaround gross alpha containing 224Ra would count the 224 and its progeny. The progeny would contribute approximately 3 times the 224Ra level present. But because the risk of 224Ra contains the effect of its daughters, the committed dose, the gross alpha count with daughters of 224Ra would overstate by threefold the 224Ra count and apparent health effects of the gross alpha count. Additionally, 224Ra would not likely be present if its greatgrandparent 228Ra were not present. Although there is a fairly insoluble 228Th in between, the USEPA=USGS reconnaissance survey demonstrated a close correlation between the presence of 228Ra and 224Ra, as might be expected. Therefore, systems who know that they have 228Ra and 224Ra, but do not want to monitor and analyze for 224Ra, might multiply their 228Ra result by 3 and add this to the gross alpha result to show what the gross alpha result might be if a fast turnaround sample were taken. The 224Ra level could roughly be taken as equal to 228Ra, and the 224Ra risk added to the 226Ra risk and any other gross alpha emitters present to determine the total risk from net alpha emitters. Given a 1 : 1 relationship between 228Ra and 224Ra, if 228Ra were 5 pCi=L, 224Ra would be at 5 pCi=L also and contribute only 13% of the total radium risk. Additionally, treatment to remove 228Ra alone or in combination with 226 Ra would also remove 224Ra to the same extent. Treatment is normally well below the MCL, leaving the risk from 224Ra small. 13.7.4

Uranium

In 2000, USEPA regulated for the first time the naturally occurring uranium isotopes in drinking water. Natural uranium (‘‘U-nat’’) is fairly ubiquitous, occurring in rocks rich in silica such as granites and the sediments and metamorphic rocks derived from them. U-nat consists of three isotopes: 238U, 235U, and 234U. The crustal abundance of 238U is by far the greatest by weight, constituting 99.275% of the uranium found. Uranium-235 is next being 0.72% by weight, followed by 234U, a greatgranddaughter in the decay chain of 238U, a distant third at 0.005%. One microgram of uranium with these relative isotopic proportions has an activity of 0.67 pCi=mg. The radioactivity of the three uranium isotopes, however, presents a different story. Uranium 238, the top of the uranium decay series, has a long half-life of

330

REGULATING RADIONUCLIDES IN DRINKING WATER

4.5 billion years, and while its mass is 99.275%, its radioactivity represents only 48.9% of the total activity from natural uranium. Uranium 235, the head of the actinium decay series, has a 710-million-year half-life, and is 2.25% of the total activity. Uranium-234, although tiny in mass, has a high radioactivity and contributes 48.7% of the total radioactivity from uranium, an amount of activity almost equivalent to that of 238U. Because of its (relatively) short half-life of 250,000 years, it has a high radioactivity per weight, called specific activity. So, while 234U is uncommon in terms of mass, it represents about half the activity because it is so ‘‘hot.’’ In rock, the activity : mass ratio is 0.67, while in water the ratio is more variable, ranging from 0.77 to over 2 and generally around 0.9, where the activity is over 3.5 pCi=L. The reason is that a greater amount of 234U winds up in groundwater in part because of faster decay and the ‘‘alpha recoil’’ allowing 234U atoms to be ejected from the crystal matrix of the mineral into the surrounding water. USEPA established an MCLG, the health-based goal, of zero by virtue of its ionizing radiation. This was proposed in 1991 and finalized in 2000. In setting the enforceable limit, the MCL for uranium, USEPA had to decide to regulate it on the basis of its mass, its activity, or both. USEPA chose mass, but this decision was a circuitous, iterative process. Deciding the MCL was complicated by the fact that there are two endpoints of concern for uranium: heavy-metal toxicity affecting the kidneys, and cancer from radiation. Cancer is a potential effect, and any regulation of uranium had to recognize the carcinogenic effect of the element. The Agency intended that any regulation must stay within the 104–106 risk range. Despite this, cancer in humans has not been observed resulting from exposure to natural uranium, although it has been seen in animal studies using enriched, more radioactive species of uranium. For natural uranium, with its abundance of 238U, the toxicity of the heavy metal outweighs the cancer concern. The proposed MCL in 1991 was selected in the following manner. The proposal utilized a 1949 study of rabbits by Maynard and Hodge (1949) that found a 2.8 m k1 day1 dose to be the LOAEL. Dividing by a 1000-fold uncertainty factor (10 for interspecies, 10 for human variability, 10 for a less than lifetime study, 1 for LOAEL to NOAEL) resulted in an RfD of 0.003 m k1 day1. Multiplying by 70 kg divided by 2 L=day resulted in a DWEL of 100. Multiplying by a default RSC of 20% from drinking water converted to micrograms yielded an MCLG of 20 mg=L. By the time of the 2000 rule, the data had changed. USEPA selected the MCL in the following way. A panel of scientists expert in uranium health effects from the United States and Canada met in Washington in June 1998 for two days to discuss the available studies. Data by Gilman et al. (1998) from male rats resulted in a LOAEL of 0.06 after histopathologic examination of changes to the proximal tubule of the nephridium. This study was selected as the best evidence of damage from uranium in drinking water. The panel utilized a 100-fold uncertainty factor [10 for human variability: 3 for interspecies (animal to human), 3 for LOAEL to NOAEL, and 1 for a less than lifetime study], resulting in an RfD of 0.2 m kg1 day1. Multiplying by the standard factors of 70 kg and dividing by

13.7 2000 FINAL RADIONUCLIDES RULE

331

2 L=day consumption (95 percentile) resulted in a DWEL of 20 mg=L. However, new data by Fisenne and Perry (1985) and Fisenne and Welford (1986) on uranium consumption resulted in a new RSC value of 77% rounded to 80%. The resulting 16.8 mg=L was rounded to 20 mg=L which was the same as the proposed 1991 level arrived at via newer data. USEPA then utilized occurrence data to predict the number of water systems that would be out of compliance with various concentrations. While 20 mg=L is achievable, it could not be justified in USEPA’s opinion. A best estimate [between the lognormal distribution for the data and utilizing the data as directly reported in the National Inorganics and Radionuclides Survey (NIRS)] indicates that 850 systems would be affected at an annual cost of over $90 million to treat to 20 mg=L, while treating to 30 mg=L would affect only 500 systems (mainly small groundwater) at a cost of about $50 million annually. USEPA used a 1 : 1 activity : mass ratio to predict the benefits from the number of cancer cases and deaths avoided by regulation at different levels. Calculations could be made on the basis of cancer, but kidney toxicity was not quantifiable. The data do not provide a dose–response continuum from no effect, to cellular abnormalities, to frank effects and actual disease. But by using cancer risk numbers, cancer deaths are predictable. The excess cancer cases nationally at 20 pCi=L (equivalent to 20 mg=L) were predicted to be 0.9 resulting in 0.6 deaths. Going from about 20 pCi=L of radioactivity to about 30 pCi=L or 30 mg=L increases cancer deaths from 0.7 to 0.9, an incremental increase in 0.2 statistical deaths. The statistical valuation of a life is approximately 6 million, and the incremental lost health benefits of regulating at 30 mg=L calculates to a cost of a million dollars. When compared to the additional costs of about $40 million annually for regulating at the lower concentration, justifying the costs was difficult. Approximately 42% of national compliance costs and 20% of cancer risk reduction occur between 20 and 30 pCi=L. At 20 pCi=L the benefits are $4 million in avoided cancer deaths; at 30 pCi=L it is $3 million. In scrutinizing the data and the rationale for selecting 20 as the noncancer no-effect level for uranium, it was apparent that there was considerable safety built into the number. Selecting 30 mg=L would not have a demonstrably greater adverse effect on human health. The safety factors coupled with an assumption of 70 years of drinking uranium tainted water at 2 L=day all lead to a very conservative no-effect level. However, it is not apparent at what point some clinical effect would occur. Within these parameters of data and uncertainty, safety factors and cost, USEPA selected a limit it deemed both protective and practical and within the SDWA legal framework. The 1986 amendments to SDWA mandated a list of 83 contaminants be regulated, which included radon and uranium. This regulation also fit the new criteria established in the 1996 amendments in Section 1412(b)(1): that the contaminant may have an adverse effect on the health of persons and it is known to occur or there is a substantial likelihood that it will occur with a frequency and level of public health concern and its regulation represents a meaningful opportunity for health risk reduction. Therefore, when reviewing the best approach for regulating the contaminant, we attempted to regulate to maximize risk reduction at a cost that could be justified

332

REGULATING RADIONUCLIDES IN DRINKING WATER

by the benefits recognizing that there are intangible and unquantifiable benefits from protection from kidney toxicity. The statute Section 1412(b)(6) gives the USEPA Administrator the authority to select a limit the cost of which is justified (not offset) by the benefits. When those benefits from toxicity and cancer were taken together, they did not, in the opinion of the Agency, justify the costs involved. A number that maximizes public health protection at a cost justified by the benefits was selected. For uranium, 30 mg=L did that. And, because the Agency was regulating it for the first time, not revising an MCL, the provision to maintain or provide greater protection than a previous MCL was not a factor. Initially, USEPA considered a dual standard for uranium, which recognized both kidney toxicity and cancer, specifically, a limit of both 30 mg=L and 30 pCi=L. However, discussions over the activity : mass ratio and conversions from one to another, the chemical methods approved for drinking water analysis, and the costs to community water systems, led to a decision to select one parameter to regulate. Because gross alpha was already an MCL and measured in pCi=L, our MCLG was based on cancer from radiation, and as this was a radionuclides rule, it seemed logical to select radioactivity as the units of the MCL. However, it was soon apparent that the approach had flaws. Regions where the activity : mass ratio was low, limiting activity to only 30 pCi=L could result in an equivalent mass of 45 mg=L. While toxicologists were willing to admit that all factors considered, their best scientific judgment would support 30 mg=L, an MCL that would allow up to 45 was not acceptable. We analyzed additional data demonstrating that rarely was an activity to mass ratio of 1 : 1 exceeded at levels we were regulating. By regulating in micrograms, it served as a limit on the picocurie concentration as well.

13.7.5

Beta and Photon Emitters

The SDWA authorizes the rules applicable to public water systems to be written according to specific criteria. The limits also affect other aspects of U.S. water policy. The Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) or ‘‘Superfund’’ as amended by the Superfund Amendments and Reauthorization Act (SARA) makes the drinking water limits relevant and appropriate for cleanup at Superfund sites. CERCLA specifically refers to maximum contaminant level goals as relevant and appropriate for cleanup, and the National Contingency Plan interprets the cleanup levels to be MCL’s for contaminants whose MCLGs are zero. Implementation of one law requires the use of the product (MCL) of the other law. Policymakers in other agencies argued for higher standards in the revised radionuclides regulations for beta and photon emitters because drinking water standards also used as cleanup criteria did not factor in the high costs of cleaning aquifers. However, the SDWA does not allow consideration of Superfund cleanup costs in setting drinking water standards, which are based on health effects and treatment feasibility at public water systems. The SDWA considers relatively clean source waters have been contaminated, and requires an MCLG and establishment of BAT

13.7 2000 FINAL RADIONUCLIDES RULE

333

for removing the contaminant to a feasible level, the MCL, in a water treatment plant context. Superfund cleanup is very different from municipal water treatment. Yet, if the water is consumed, the health effect is the same and the safe level should be the same. That is precisely why CERCLA refers to SDWA goals and standards for selecting cleanup levels. Undeniably, the costs can be high, but USEPA allows considerable and reasonable flexibility in establishing a cleanup goal based on commonsense factors such as natural attenuation, and technological infeasibility realities. There is no need to attempt to sidestep the intent of Congress in SDWA and CERCLA in a rulemaking by making allowances not intended. Section 1412(b)(3)(C)(i)(III) specifically provides that USEPA is to analyze [q]uantifiable and nonquantifiable costs for which there is a factual basis in the rulemaking record to conclude that such costs are likely to occur solely as a result of compliance with the maximum contaminant level, including monitoring, treatment and other costs and excluding costs resulting from compliance with other proposed or promulgated regulations (emphasis added).

The argument that the costs of clean-up of groundwater contamination to achieve drinking water standards ought to be factored into the development of the standards themselves was rejected by the Court. Drinking water standards are referenced in CERCLA but what defines a safe drinking water level of a contaminant and what is feasible for public water systems to achieve should exclude costs associated with compliance with regulatory regimes other than the SDWA itself. . . Furthermore, CERCLA itself imposes no requirement that EPA consider the costs and benefits of compliance with MCLs as an element of clean-up standard, and certainly no requirement that the agency do so as part of its obligations under a separate statute like the SDWA (Waukesha v. EPA 2003)

In addition, the standard selected for drinking water is also relevant for setting standards for allowable contamination from long-term disposal of radionuclides at such sites as Yucca Mountain, Nevada, and decommissioning of Nuclear Regulatory Commission (NRC)-regulated sites. The Agency has a decade old policy through several Administrations of protecting ground water to drinking water standards. Underground sources of drinking water, whether they are current or future sources, are an extremely valuable resource that should be protected from degradation. The fact that future underground sources of drinking water would arguably be protected to drinking water standards proved a source of debate during deliberations for Yucca Mountain geologic repository for spent nuclear fuel and high-level waste. The Energy Policy Act of 1992 gives USEPA responsibility to set site-specific health and safety standards for Yucca Mountain. Because the radionuclides rule sets the amounts of contaminants allowed in the drinking water, it potentially affects the licensing of the site by the NRC, and the design and operation of the site by the Department of Energy to limit the human-made (synthetic) radionuclides that will eventually leak from the repository, if and when it is ever built. Should a clean

334

REGULATING RADIONUCLIDES IN DRINKING WATER

source of future drinking water be allowed to be degraded by operation of a facility? Or should design considerations take into account preservation of the resource to avoid imposing unreasonable cleanup burdens on future generations? Rational public policy is embodied in the law. The lesson of the Superfund is that an ounce of prevention is worth a ton of cure. The extent of cleanup at facilities does consider the background levels of natural contaminants already at the site. In the case of synthetic nuclides, the background is clearly at zero. Thus a limit for them is already an allowance to increase contamination from zero up to a selected MCL, rather than the usual case of decreasing contamination down to an MCL. In this sense the standards selected for the beta and photon emitters were not based on a goal of zero and the feasibility of treatment, which could be much lower, but the original conceptual ceiling of excess deaths of one in a million per year. Revision of the standards according to the SDWA would result in lower limits than now. Treatment feasibility is not an issue and neither is cost to public water systems, there being few impacted at this time. USEPA engaged in discussions with the Office of Management and Budget (OMB), Department of Defense (DOD), NRC, DOE, and others relative to revising the part of its drinking water rule concerning the limit on the synthetic radionuclides in drinking water. A higher allowed limit would make the task of decommissioning power plants or licensing future facilities easier based on projections of ground water quality available to future consumers years from now. A stricter standard would make licensing more costly and unsure for Yucca Mountain. USEPA determined the risks of each of the synthetic contaminants regulated using the most recent model of Federal Guidance Report (FGR) 13. The original ceiling of risk in 1976 was to have been 5.6
105, but 25 contaminants exceeded this limit. However, rather than lowering those 25 to the intended risk ceiling, some were pressing for increasing the remaining contaminants up to that cap. Alternatively, some inside and outside the Agency urged changing from 4 mrem to 4 mrem EDE. Using the units of effective dose equivalent, most contaminants increased their concentrations substantially. The associated increase in risks convinced policymakers that this was not a good idea. In the end, the decision was to retain the existing limits, that is 4 mrem and the corresponding picocurie limits. However, USEPA did correct omissions and errors in calculations. A total of 179 synthetic beta and photon emitters are regulated. After the radionuclides rule was published, petitioners and USEPA discussed the elements of their petition before going to the Court to adjudicate. During this time, the agency was preparing the Yucca Mountain rule and discussions pertaining to the beta particle and photon emitters were raised, having relevance to both rules. The limit applying to drinking water systems would also serve as limiting factors for drinking water contaminated by the Yucca Mountain repository. Should the current units and limits be used or units of 4 mrem EDE with correspondingly higher limits be used? That weighting factors to account for differences in effects on the cells was good science and not in dispute though the EDE units had been superceded themselves by even newer science. But, the issue was that it would translate into higher individual drinking water limits for almost all nuclides regulated since 1976. In this case, scientific issues were clouded by how the issue was framed and presented.

13.7 2000 FINAL RADIONUCLIDES RULE

335

Pejorative phraseology of ‘‘new and better science’’ versus ‘‘old, bad science’’ became the currency of dialogue in discussions of what the numbers ought to be. Under the guise of the purity of better science, the cry for newer units was, in fact, all about raising the allowable limits of residual man-made radiation in the drinking water. In the final analysis, what really mattered was not the newness of the science, but its proper application and the results it achieved. The application of good science should be to increase, not decrease, protection. The technology for removing contaminants closer to the public health goal gets better over time, not worse. Could USEPA legally, in light of the SDWA mandates, abandon the protections embodied in the existing limits under the rubric of ‘‘good science’’? Using 4 mrem EDE would result in over 90% of the nuclides exceeding the 1
104 risk ceiling. In contrast, about the same percent of the nuclides at current limits fell below 104 and above 106 risk range as demonstrated by FGR-13. Use of equivalent units, like changing 72 F to 22 C or changing 4 mrem to 0.9 mrem EDE would not work equitably in this case because 0.9 was only the average of the regulated nuclides. Individual concentration limits would increase or decrease to line up with the new limit and nobody liked that result. In addition, those who would prefer uniform risks were not willing to set a sum of the fractions of risks at a 106 level because, as the Court pointed out, EPA could not achieve the uniformity for which petitioners argue without lowering most of the 1976 beta/photon MCLs until they yield a risk level actually provided by the most protective of those MCLs – a result petitioners do not seek and that would defeat their aim in bringing this petition. (Waukesha v. EPA 2003)

With respect to good science, USEPA defended the use of Federal Guidance Report No. 13 as reflecting ‘‘the best available, peer reviewed science and supporting studies conducted in accordance with sound and objective scientific practices’’ [SDWA Sec. 1412(b)(3)(A)]. The effective dose equivalent unit from 1991 is in reality based on 27-year-old organ weighting concepts embodied the International Commission on Radiological Protection’s effective dose equivalent (EDE) Publications 26 and 30. FGR-13 uses the concept of organ-specific risk, the basis of the organ weighting factors employed in calculating the effective dose equivalent. It also goes a step further, using the Blue Book for estimation of radiogenic risk; developing gender- and age-specific risk coefficients accounting for water consumption, radionuclide intake, and physiological and anatomic changes with age. It also uses the most recent biokinetic and dosimetric models recommended by the ICRP, 1990 vital statistics, best information radiogenic human health effects from the National Academy of Sciences and other groups, and updated life tables from the National Center for Health Statistics to adjust risk estimates for competing causes of death. The model is clearly superior to the 1991 model, but for all its sophistication it simply pinpoints the actual numerical risk. However, SDWA asks only about the movement of the MCL. If the number is retained or decreases, protection is maintained or improved. The previous model was inaccurate, but improving risk estimation is not central to the rulemaking. Keeping the MCL under the risk ceiling of 104

336

REGULATING RADIONUCLIDES IN DRINKING WATER

is important, but this is not the MCLG. The newer model reinforced the underlying truth that concentration matters. Besides issues of the MCL itself, there were larger issues about at what point in a contaminated plume must water actually meet the standard? Should it apply to public water systems or individual well owners? The longstanding policy of protecting the resource would argue for the single private well, but in the case of Yucca Mountain, application of the MCL was as far out as 10,000 years under the law but far beyond that in reality. The size of the population and the distance to that population and the presumed dilution en route were major topics of discussion. Since the SDWA concerns public water supplies, applying the MCL to a public-water-sized system is reasonable but here reasonable people differed. A compromise position was reached applying the MCL at a point where people could be reasonably expected to consume the water and at a quantity of use by a reasonable-sized population drawing a large enough portion of uncontaminated water along with contaminated portions of the aquifer that together would meet the MCL. On June 6, 2001, one day shy of 7 months after the final rule for radionuclides in drinking water was published, USEPA finalized a rule regarding environmental standards to be applied at the Yucca Mountain site. The tension between setting a drinking water standard according to the law, and application of that standard in protection of drinking water sources from synthetic contamination far into the future was resolved, but not without considerable discussion and compromise to retain the spirit of the law and apply it in commonsense ways. The final rule is being challenged in court, so the debate will continue on to a legal resolution.

13.8

FUTURE OUTLOOK

How does regulation development become as entangled as it did for radionuclides? One reason certainly is the complexity of the subject that crosses many disciplines and viewpoints. An attorney may understand some of the science, but be excused for being unaware of the finer points or unique aspects of health effects or physics. The health physicist glazes over at intricacies of the law. An agency with a keen interest in a certain outcome and its own authority can easily forget that other competing interests and complementary legislative authorities exist alongside their own. The public wants healthy water, and the industry wants indemnity from future liability. Cleanup for a decommissioned facility is certainly cheaper at 400 pCi=L than 20 pCi=L. It is a simple matter of economics as are so many things. Is that necessarily wrong? No, managing scarce resources wisely means that tradeoffs must be made. Why should a contaminant in drinking water be limited when risks may appear to be much greater elsewhere? First, the law requires it. With better knowledge of relative risks perhaps the focus will shift to other pressing risks as well. Second, however, is that something can be done about identified risks to health from contaminants in drinking water despite other risks that cannot be changed. Lawmakers must fit different pieces of legislation together as they are created or amended into the overall risk reduction puzzle so that the eventual set of require-

REFERENCES

337

ments is harmonious. Future laws should continue to insist that better science leads to better, not worse, public health protection. Regulators do the law’s bidding. The challenges, conflicts, and opportunities experienced in regulating radionuclides will continue to reflect the diversity of interpretations and viewpoints involved with science, technology, law and politics, government, industry, and the consumer in a democratic society.

ACKNOWLEDGMENTS The author wishes to thank the individuals on the USEPA radionuclides work group for their exceptional work and devotion to a sensible standard, to good science, to safe drinking water.

REFERENCES City of Waukesha v. EPA,___ F.3d___, 2003 WL 431867 (D.C.Cir. Feb 25, 2003). Congress. 1995. Senate Report 104-169 (Nov. 7, 1995). Washington, DC: U.S. Government Printing Office. Congress. 1996. House of Representatives Report 104-632 (June 24, 1996). Washington, DC: U.S. Government Printing Office. Covello, V. T. 1992. Risk communication: An emerging area of health communication research. Communication Yearbook, 15. edited by S. Deetz. Newbury Park and London: Sage Publications. Covello, V. T. 1983. The perception of technological risks: a literature review. Tech. Forecasting Social Change, 23:285–297. Covello, V. T. and M. W. Merkhofer. 1994. Risk Assessment Methods. Plenum Press: New York. Fisenne, I. M. and P. M. Perry. 1985. Isotopic U concentration in human blood from New York City donors. Health Physics 49:1272–1275. Fissenne, I. M. and G. A. Welford. 1986. Natural uranium concentrations in soft tissues and bone of New York City residents. Health Physics 50:739–746. Focazio, M. J., Z. Szabo, T. F. Kraemer, A. H. Mullin, T. H. Barringer, and V. T. DePaul. 1998. Occurrence of Selected Radionuclides in Ground Water Used for Drinking Water in the United States: A Targeted Reconnaissance Survey. Water Resources Investigations Report 00_4273. Denver, CO: U.S. Geological Survey. Gilman, A. P., D. C. Villeneuve, V. E. Secours, A. P. Yagminas, B. L. Tracey, J. M. Quinn, V. E. Valli, and M. A. Moss. 1998. Uranyl nitrate: 28-day and 91-day toxicity studies in the Sprague-Dawley Rat. Toxicol. Sci. 41:117–128. Maynard, E. A. and H. C. Hodge. 1949. Study of toxicity of various uranium compounds when fed to experimental animals. In Pharmacology and Toxicology of Uranium Compounds. C. Voegtlin and H. C. Hodge, eds. National Nuclear Energy Series, Div. VI, Vol. 1. New York: McGraw-Hill. NRC. 1983. Risk Assessment in the Federal Government: Managing the Process. National Academy Press: Washington, DC.

338

REGULATING RADIONUCLIDES IN DRINKING WATER

Powell, D. 1996. An Introduction to Risk Communication and the Perception of Risk. University of Guelph, Guelph, Ontario, Canada. http:==www.foodsafetynetwork.ca=risk= risk-review=risk-review.htm Sandman, P. M. 1987. Risk Communication: Facing Public Outrage. EPA Journal 13:21. Slovic, P. 1986. Informing and Educating the Public about Risks. Risk Analysis, 6:403–415. USEPA. 1991. National Primary Drinking Water Regulations; Radionuclides; Proposed Rule. Fed. Reg. 56:138:33050–33127. USEPA. 1999. Federal Guidance Report 13, Cancer Risk Coefficients for Environmental Exposure to Radionuclides. EPA 402-R-99-001. Washington, DC: Office of Radiation and Indoor Air. USEPA. 2000a. National Primary Drinking Water Regulations; Radionuclides; Notice of Data Availability; Proposed Rule. Fed. Reg. 65:78:21576–21628. USEPA. 2000b. National Primary Drinking Water Regulations; Radionuclides; Final Rule. Fed. Reg. 65:236:76708–76753. USEPA. 2001. Radionuclides Rule: A Quick Reference Guide. EPA 816-F-01-003. Washington, DC: Office of Water. USEPA. 2002a. Radionuclides in Drinking Water: A Small Entity Compliance Guide. EPA 815-R-02-001. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 2002b. Implementation Guide to Radionuclides. EPA 816-D-00-002. Washington, DC: Office of Water.

14 RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISION MAKING MARK GIBSON Program Officer, National Research Council, Water Science and Technology Board, Washington, DC

MICHAEL OSINSKI Drinking Water Utilities Team Leader, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

14.1

INTRODUCTION

The provision of safe drinking water throughout the United States has been a major triumph in U.S. public health practice since the beginning of the twentieth century. The quality and reliability of this service is vital because it simultaneously provides a life-giving substance to our communities while having the potential to deliver harmful substances and microorganisms if not properly maintained. Maintaining drinking water quality in the United States has been accomplished through several layers of continually evolving federal, state, and local government laws, regulations, and guidance designed to protect public water supplies from contamination. Despite overlapping regulatory protection, however, many water sources and treated public drinking water in the United States periodically contain chemical, microbiological, and other types of contaminants at detectable and sometimes harmful levels.

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

339

340

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Annual manufacture and use of thousands of new chemicals that can reach water supplies and the discovery of emerging microorganisms that potentially can resist traditional water treatment practices or grow in distribution systems pose a regulatory dilemma: Where and how should the U.S. government focus its attention and limited resources to ensure safe drinking water supplies for the future? The widespread availability of increasingly powerful analytical methods for their detection and identification often serves to complicate these decisions. Regardless, the continuing presence of contaminants in water supplies, as well as documented outbreaks of waterborne disease and the many other outbreaks thought to go undetected, send a clear message that continuing public health vigilance is necessary to ensure that drinking water contaminants, including newly identified ones, are appropriately addressed. Certainly the most important, comprehensive, and widely enforced law designed to protect the public from hazardous substances in drinking water is the Safe Drinking Water Act (SDWA). As discussed in earlier chapters, the SDWA was enacted in 1974, authorizing the U.S. Environmental Protection Agency (USEPA) to establish federal (enforceable) standards to protect the public from harmful contaminants in drinking water. Since its enactment, the SDWA has been amended several times but most significantly in 1986, 1996, and 2002. This chapter reviews the important scientific and policy aspects of implementing the requirements mandated by the 1996 SDWA Amendments with respect to how drinking water contaminants are currently regulated and will be regulated in the United States for the foreseeable future. Specifically, a detailed discussion is provided regarding the Drinking Water Contaminant Candidate List (CCL), an overview of the development of the first (1998) CCL, the implementation status of the 1998 CCL, and most importantly, how future CCLs should be created and used effectively to regulate drinking water contaminants in the future. Much of this chapter originates and builds on select findings and recommendations of three successive reports of the National Research Council (NRC) Committee on Drinking Water Contaminants reports that were largely commissioned to assist USEPA in this enormous, complex task.

14.2

SDWA AMENDMENTS OF 1996

In 1996, after much debate and discussion, Congress amended the SDWA to provide increased funding and emphasize public health protection. The full text of the SDWA as amended appears in Appendix D. Notably, the amended SDWA significantly restructured the existing framework for regulating new drinking water contaminants and reviewing existing drinking water standards established under the 1986 SDWA amendments. Under the 1986 SDWA amendments, USEPA was required to establish drinking water standards within 3 years for 83 priority contaminants that the Agency had identified for study, and develop standards for 25 new Drinking Water Priority List (DWPL) contaminants every 3 years. The 1996 amendments to SDWA improved the standard setting

14.2 SDWA AMENDMENTS OF 1996

341

process by establishing a new risk-based focus for setting contaminant regulation priorities based on the adverse health effects of the contaminant, the occurrence of the contaminant in public water systems, and the opportunities for tangible reduction in health risk that would result from regulation. Specifically, Section 1412 of the amended SDWA requires the USEPA to ‘‘publish a list of contaminants, which, at the time of publication, are not subject to any proposed or promulgated national primary drinking water regulation, which are known or anticipated to occur in public water systems, and which may require regulation under this title.’’ This list, known as the Drinking Water Contaminant Candidate List—more commonly referred to by the acronym CCL or DWCCL—provides the basis for a mandated USEPA decision to regulate (or not) at least five new contaminants every 5 years. Moreover, each CCL is also used to prioritize related drinking water contaminant research and occurrence monitoring activities within USEPA. The first CCL was to be published within 18 months of enactment, and updated every 5 years thereafter. In addition, the 1996 SDWA amendments require USEPA to give priority to selecting contaminants for inclusion on future CCLs that present the greatest public health concern, taking into consideration the effect of such contaminants upon subgroups that comprise a meaningful portion of the general population (such as infants, children, pregnant women, the elderly, individuals with a history of serious illness, or other subpopulations) that are identifiable as being at greater risk of adverse health effects due to exposure to contaminants in drinking water than the general population [SDWA Sec. 1412(b)(1)(c), as amended].

In essence, USEPA must use each CCL development process to help ensure that vulnerable subpopulations have access to safe drinking water. Thus, Congress, through the 1996 amendments, clearly mandated that the future of regulating drinking water contaminants in the United States will be through a riskbased approach by the regular creation and use of CCLs by USEPA to target and prioritize contaminants for possible regulatory action. SDWA Section 1401(6) states ‘‘The term ‘contaminant’ refers to any physical, chemical, biological, or radiological substance or matter in water.’’ This definition has not been revised since the inception of the SDWA in 1974, and is so broad as to include nontoxic and potentially beneficial ‘‘contaminants.’’ The 1996 SDWA amendments require USEPA to regulate such a contaminant when [SDWA Sec. 1412(b)(1)(A)] (i) the contaminant may have an adverse effect on the health of persons; (ii) the contaminant is known to occur or there is a substantial likelihood that the contaminant will occur in public water systems with a frequency and at levels of public health concern; and (iii) in the sole judgment of the Administrator, regulation of such contaminant presents a meaningful opportunity for health risk reduction for persons served by public water systems.

342

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Furthermore, the amended SDWA allows USEPA to issue interim regulations for any drinking water contaminant that is determined to pose an ‘‘urgent threat’’ to human health without adhering to the newly revised process for making regulatory decisions (i.e., the CCL process) or completing a cost=benefit analysis [SDWA Sec. 1412(b)(1)(d)]. Hence, to select a contaminant for regulation, USEPA must eventually determine whether the contaminant meets the preceding SDWA criteria for regulation, or if the contaminant represents an urgent threat to public health. Does the contaminant have or may have an adverse health effect? Does it occur, or is it likely to occur in drinking water? Is there a meaningful opportunity for risk reduction if a regulation is established? Answering these questions is difficult.

14.3 ROLE OF THIRD-PARTY CONSULTATION IN SDWA REGULATORY DEVELOPMENT USEPA relies on expert advice and analysis from organizations outside the agency, referred to as ‘‘third parties,’’ throughout the regulatory development process. Two such third-party groups are the NRC and the National Drinking Water Advisory Council (NDWAC). Both of these groups have played an important role in the development of the CCL. 14.3.1

The National Research Council (NRC)

The National Academy of Sciences (NAS) was established in 1863 under a charter granted by Congress. NAS organized the NRC in 1916 to serve as the principal operating agency of both the National Academy of Sciences (NAS) and the National Academy of Engineering (NAE). The NRC, NAS, and NAE further knowledge and provide services to the government, the public, and the scientific and engineering communities. Together with the Institute of Medicine (IOM), all four organizations are now collectively known as The National Academies. Section 1412(e) of the 1974 SDWA mandated that the NRC conduct studies to identify adverse health effects associated with contaminants in drinking water, to identify relevant research needs, and to make recommendations regarding such research (NRC 1977). Subsequent amendments to the SDWA requested revisions of the NRC study and have mandated additional reports. Between 1977 and 1989, the NRC published a total of nine reports in a series called Drinking Water and Health covering a wide variety of issues related to drinking water safety, including toxicology, dose–response relationships, and risk assessment. In addition to this seminal series of reports, the NRC has continued to periodically release important reports on drinking water issues and specific contaminants. A complete listing of NRC drinking water reports is provided in Chapter 4. In late 1997, USEPA requested NRC to provide advice regarding the setting of priorities among drinking water contaminants in order to identify those contaminants that pose the greatest threats to public health as required by the 1996 SDWA amend-

14.3 SDWA REGULATORY DEVELOPMENT

343

ments. In response, the NRC Committee on Drinking Water Contaminants was formed in early 1998 and consisted of volunteer experts from academia, state and federal government, and private industry with expertise in public water system operations, water treatment engineering, public health, epidemiology, toxicology, water and analytical chemistry, risk assessment, risk communication, and aquatic microbiology. The committee’s activities were conducted in two discrete phases over a 3-year study period (Feb. 1998 through July 1999 and Aug. 1999 through Feb. 2001) to address two distinct sets of related tasks and issues. Three reports were released during the two phases addressing the following aspects of the study:
Developing a scientifically sound approach for deciding whether to regulate contaminants on the current and future CCLs and how to prioritize CCL contaminants for further research or monitoring (Phase I)
Convening a workshop that focused on emerging drinking water contaminants and the database that should be created to support future decision making on such contaminants (Phase I)
Creating a scientifically sound approach for developing future CCLs (Phase II) Setting Priorities for Drinking Water Contaminants (NRC 1999a) addresses the first of these topics, Identifying Future Drinking Water Contaminants (NRC 1999b) addresses the second, and Classifying Drinking Water Contaminants for Regulatory Consideration (NRC 2001) addresses the last. Collectively these reports address issues directly and indirectly related to the three central aspects of the study; many are noted or discussed in some detail within this chapter.

14.3.2

The National Drinking Water Advisory Council (NDWAC)

The NDWAC was chartered under the 1974 SDWA to provide independent advice and recommendations to USEPA on drinking water issues and related SDWA policies and is an approved advisory committee that operates under the authority of the Federal Advisory Committee Act (USEPA 1999c). It advises the Administrator on key aspects of the USEPA’s drinking water program. The NDWAC is composed of members representing the general public, state and local agencies, and private groups concerned with safe drinking water. Several working groups were formed within NDWAC after the 1996 SDWA amendments to help USEPA implement many of its new and revised statutory requirements. During the development of the first CCL, USEPA relied heavily on the advice and recommendations of the now-disbanded NDWAC Working Group on Contaminant Occurrence and Contaminant Selection. This group consisted of engineers, microbiologists, toxicologists, and public health scientists from government agencies, water utilities, and other stakeholder groups. USEPA has also looked to NDWAC for recommendations for developing decisionmaking protocols for making regulatory decisions once contaminants are placed

344

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

on the CCL, and to its Working Group on Research for support in developing a Comprehensive Research Strategy.

14.4

ROLE OF USEPA PROGRAMS

In accordance with the 1996 amendments, developing future CCLs will be coordinated and closely linked with two other drinking water programs: (1) the National Drinking Water Contaminant Occurrence Database (NCOD); and (2) the unregulated contaminant monitoring (UCM) program. The relationship of both of these programs to the CCL in terms of their mandated timelines is illustrated in Figure 14.1 (USEPA 1998a, NRC 2001). Implementing all three programs is the responsibility of USEPA’s Office of Ground Water and Drinking Water (OGWDW). USEPA completed the first working release of the NCOD and the UCM Rule (UCMR) prior to its amended SDWA statutory deadline of August 1999. 14.4.1 National Drinking Water Contaminant Occurrence Database (NCOD) The NCOD stores occurrence data for both regulated and unregulated drinking water contaminants throughout the United States. It is designed to support agency efforts

Figure 14.1 Current and future timeline for selected major regulatory requirements of the SDWA amendments of 1996 [source: adapted from USEPA (1997b), NRC (1999a, 2001)].

14.4 ROLE OF USEPA PROGRAMS

345

to identify and select contaminants for placement on future CCLs; conduct related research, monitoring, and regulatory activities; and periodically reviewing existing drinking water regulations every six years for possible modification, as required under the amended SDWA (USEPA 2000b, NRC 2001). Two additional purposes are to inform the public about contaminants detected in drinking water and make available the data sets that are used by USEPA to help form the primary basis for their drinking-water-related regulatory and research actions. The design, structure (e.g., data elements), and use of the NCOD were developed with input from the public, states, and the scientific community (USEPA 1997b). The first release of the NCOD became operational in August 1999 (as mandated) and included occurrence data (both detections and nondetections) on various physical, chemical, microbial, and radiological contaminants found in public water systems (PWSs) and ambient (source) water (USEPA 2000b). More specifically, it contained some summary statistics of PWS data stored in USEPA’s Safe Drinking Water Information System (SDWIS) and ambient water stored in the National Water Information System (NWIS) of the U.S. Geological Survey (USGS). The second release of NCOD became operational in late August 2000 and included several changes intended to increase its functionality such as the availability of an NCOD User’s Guide (USEPA 2000c). Despite these ongoing improvements, USEPA has reported several self-assessed data limitations (e.g., does not contain occurrence data from every PWS or from every state) and cautions to NCOD users (e.g., although data sets are updated over time, they may still reflect a lag time of at least 6 months from data provided directly from a PWS) (USEPA 2000b). For more information about the NCOD and its development status, go to http:==www.epa.gov=ncod=. 14.4.2

Unregulated Contaminant Monitoring Program

Section 1445 of the 1996 SDWA amendments required USEPA to substantially revise its previous regulations for unregulated contaminant monitoring (see Title 40, Code of Federal Regulations, Part 141) (USEPA 1999a, 1999d, 1999e; NRC 2001). The major requirements of the program include: (1) developing of a new list of UCM contaminants every 5 years; (2) monitoring by all large PWSs (i.e., serving >10,000 persons) and a representative sample of ‘‘small’’ PWSs (i.e., serving 10,000 persons or less); (3) placing of the monitoring data in the NCOD, and (4) providing public notification to consumers that monitoring results are available. The 1996 amendments also limit the number of unregulated contaminants that a PWS must monitor in any given period to a maximum of 30 and require USEPA to pay the reasonable costs of analyzing the samples taken by small systems. USEPA will use data generated by the UCM regulation to: (1) evaluate and rank contaminants on the first (1998) CCL (see further discussion below) and help develop future CCLs; (2) support determinations of whether to regulate specific contaminants under the drinking water program; and (3) support the ongoing development of drinking water regulations (NRC 2001). The final UCM rule replaced almost all of the existing monitoring requirements of the existing UCM rule when it took effect on January 1, 2001.

346

14.4.3

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Drinking Water Research Plan

Since the publication of the first CCL in March 1998, USEPA has made significant progress in establishing an overall CCL research strategy and associated schedule. The overriding goal of USEPA’s drinking water research program is to provide sufficient information for the Administrator to make regulatory determinations for CCL contaminants as mandated by the amended SDWA. More specifically, this research is intended to identify the scientific and engineering data needed, and to characterize the risks posed by individual 1998 CCL contaminants. Several recommendations from Setting Priorities for Drinking Water Contaminants (NRC 1999a) were incorporated by USEPA in the Agency’s CCL Research Plan (USEPA 2000a, NRC 2001). USEPA ultimately decided on a two-phase approach to form the basis for the 1998 CCL Research Plan (USEPA 2000a). Phase I is a screening level process in which individual CCL contaminants are evaluated with regard to their available analytical methods, health risk, and treatment information to determine if a contaminant should be moved directly into the regulatory determination priorities category of the CCL or moved into Phase II of the Research Plan (NRC 2001). In Phase II, a more thorough examination of the available data is conducted to determine whether an individual contaminant should be recommended for regulation, guidance development, or whether a recommendation not to regulate should be made. In summary, Phase II research involves the creation of a comprehensive database for each CCL contaminant on its available health effects, analytical methods, occurrence, exposure, and treatment options. The CCL Research Plan was developed by USEPA in close consultation and with the extensive input of several outside stakeholders, including the American Water Works Association (AWWA), the AWWA Research Foundation (AWWARF), other government agencies (e.g., Centers for Disease Control and Prevention), universities, and other public and private sector groups (USEPA 2000a). Furthermore, several expert workshops were organized and conducted to not only help develop the 1998 CCL but also identify preliminary research needs for specific contaminants. In this regard, USEPA endeavors to make all aspects of 1998 CCL research planning, implementation, and communication a collaborative process through a series of public workshops and stakeholder meetings—including the activities of the NDWAC Working Group—that will be held periodically in the coming years. Two key challenges face USEPA to ensure that adequate research is conducted to support sound regulatory decisions: 1. Mechanisms must be found to leverage funding support across governmental and nongovernmental agencies. It is almost certain that USEPA’s current and future drinking water program budget cannot alone sustain the depth and breadth of research on CCL contaminants that is necessary to meet the SDWA mandated deadlines. 2. USEPA must make research planning an ongoing process that commits to the CCL, a coordinated effort between program offices within the agency and other federal or state agencies and industry stakeholders.

14.5 DEVELOPMENT OF THE FIRST CCL

14.5

347

DEVELOPMENT OF THE FIRST CCL

As noted previously, the SDWA amendments of 1996 require USEPA to publish the CCL, a list of unregulated contaminants and contaminant groups every 5 years that are known or anticipated to occur in public water systems and that may require regulation. This list, the CCL, will provide the basis for USEPA decisions to regulate (or not) at least five new contaminants every 5 years, as indicated in Figure 14.1. Furthermore, as additional research or monitoring was needed for most of the contaminants on the 1998 CCL, each successive CCL will also be used to help prioritize such activities (NRC 1999a). For these reasons, the CCL will play a central and recurring role in the foreseeable future of drinking water contaminant regulation in the United States, notwithstanding future Congressional action to amend SDWA’s standard setting provisions. USEPA published the first draft CCL on Oct. 6, 1997 (USEPA 1997a), and the first final CCL on March 2, 1998 (USEPA 1998a). The draft 1998 CCL included 58 unregulated chemical and 13 microbiological contaminants and related contaminant groups and was made publicly available for comments in the Federal Register (USEPA 1997a, 1998b). Notably, the chemical contaminants were further divided into ‘‘preliminary data need’’ categories such as those requiring additional health effects data but not occurrence data. The 1998 CCL (USEPA 1998a) contains 60 contaminants and contaminant classes, including 10 microbial contaminants and groups of related microorganisms and 50 chemicals and chemical groups, as alphabetically listed in Table 14.1. A total of four microorganisms and eight chemicals and chemical groups were removed from the draft 1998 CCL while one chemical and one broad group of related microbial contaminants were added. USEPA relied extensively on the recommendations and advice of the NDWAC Working Group on Occurrence and Contaminant Selection for developing the draft 1998 CCL. Modifications to the draft CCL were also reviewed and formally approved by the full NDWAC prior to publication of the final 1998 CCL. USEPA acknowledged that the ‘‘first CCL is largely based on knowledge acquired over the last few years and other readily available information, but an enhanced, more robust approach to data collection and evaluation will be developed for future CCLs’’ (USEPA 1997a). Several public commenters on the draft CCL also noted the need for a more systematic and scientifically defensible approach to selecting contaminants for future CCLs (USEPA 1998b). Development of the 1998 CCL and its limitations have been described in detail (NRC 1999a, 1999b, 2001), especially various sociopolitical issues surrounding the development of future CCLs. To a large extent, the widespread recognition of these limitations contributed to the formation of the NRC Committee on Drinking Water Contaminants to advise USEPA on setting regulatory and research priorities for the first (1998) and subsequent CCLs, and the creation of future CCLs.

348

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

TABLE 14.1 1998 Drinking Water Contaminant Candidate List (CCL) Microorganisms Acanthamoeba (guidance) Adenoviruses Aeromonas hydrophila Caliciviruses Coxsackieviruses Cyanobacteriaa (blue-green algae), other freshwater algae, and their toxins Echoviruses Helicobacter pylori Microsporidia (Enterocytozoon and Septata) Mycobacterium avium intracellulare Chemicals 1,1,2,2-Tetrachloroethane 1,2,4-Trimethylbenzene 1,1-Dichloroethane 1,1-Dichloropropene 1,2-Diphenylhydrazine 1,3-Dichloropropane 1,3-Dichloropropene 2,4,6-Trichlorophenol 2,2-Dichloropropane 2,4-Dichlorophenol 2,4-Dinitrophenol 2,4-Dinitrotoluene 2,6-Dinitrotoluene 2-Methyl-phenol (o-cresol) Acetochlor Alachlor ESA and other acetanilide pesticide degradation products Aldrin Aluminum Boron Bromobenzene DCPAc monoacid degradate DCPA diacid degradate DDEd Diazinon Dieldrin Disulfoton Diuron EPTCe Fonofos Hexachlorobutadiene

CASRNb 79-34-5 95-63-6 75-34-3 563-58-6 122-66-7 142-28-9 542-75-6 88-06-2 594-20-7 120-83-2 51-28-5 121-14-2 606-20-2 95-48-7 34256-82-1 N=A 309-00-2 7429-90-5 7440-42-8 108-86-1 887-54-7 2136-79-0 72-55-9 333-41-5 60-57-1 298-04-4 330-54-1 759-94-4 944-22-9 87-68-3

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

349

TABLE 14.1 (Continued ) p-Isopropyltoluene (p-cymene) Linuron Manganese Methyl bromide Metolachlor Metribuzin Molinate MTBEf Naphthalene Nitrobenzene Organotins Perchloratea Prometon RDXg (1,3,5-trinitrohexahydro-striazine) Sodium Sulfate Terbacil Terbufos Triazines and degradation product of triazines (including, but not limited to Cyanizine [21725-46-2], and atrazine-desethyl [6190-65-4]) Vanadium

99-87-6 330-55-2 7439-96-5 74-83-9 51218-45-2 21087-64-9 2212-67-1 1634-04-4 91-20-3 98-95-3 N=A N=A 1610-18-0 121-82-4 7440-23-5 14808-79-8 5902-51-2 13071-79-9 N=A

7440-62-2

a

Added after publication of draft CCL. Chemical Abstracts Service Registry Number. c (Dacthal)dimethyl-2,3,5,6-tetrachlorobenzone-1,4-dicarboxylate. d 1,1-Dichloro-2,2-bis(p-dichlorophenyl)ethylene. e S-Ethyl dipropylthiocarbamate. f Methyl-tert-butyl ether. g Royal Dutch explosive. b

Source: USEPA (1998a).

14.6

PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

USEPA recognized that sufficient data are necessary to analyze the extent of exposure and risk to populations ( particularly for vulnerable subpopulations) via drinking water for each CCL contaminant in order to determine appropriate regulatory action (USEPA 1998a, 2000a). If sufficient data are not readily available, additional data must be obtained before any meaningful assessment can be made for a specific contaminant or contaminant group. In this regard, a major intended function of a CCL is to help prioritize research and monitoring needs for drinking water contaminants of regulatory concern. Once a CCL is developed, all contaminants are divided into one or more future action (‘‘next step’’) categories that are used to

350

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

help set the research priorities for USEPA’s drinking water program (see Fig. 14.2). Notably, there has been periodic reassignment of contaminants between categories since publication of the first CCL as additional data have been obtained and evaluated. The ‘‘regulatory determination priorities’’ category includes those contaminants considered to have sufficient data to evaluate both exposure and risk to public health that will support a regulatory decision (USEPA 2000a). Contaminants in this category were used to select five or more contaminants for which USEPA was required to make a determination to regulate or not by August 2001. In June 2002, USEPA announced a preliminary decision to not regulate all nine CCL contaminants that were included in this category (USEPA 2002). Final determinations are still pending. Although the first CCL is essentially a list of research and monitoring needs for several dozen drinking water contaminants, to be included on the list, contaminants had to pass a relatively rigorous screening process involving assessment of existing contaminant occurrence and health effects data recommended by a group of experts (the NDWAC Working Group on Occurrence and Contaminant Selection). Furthermore, a variety of stakeholders, including representatives of the water utility industry and public interest groups, provided comments on the draft CCL, and the final list was revised accordingly. Thus, the contaminants on the first CCL and future CCLs are likely to have a much higher likelihood of posing public health risks in drinking water than would a randomly assembled list of unregulated substances and microorganisms. However, key questions remain for USEPA in how to best determine which contaminants should be moved off this research list and when sufficient health, occurrence, and treatment data warrant making a regulatory decision for both the first and future CCLs.

Figure 14.2

The 1998 CCL and next steps [source: adapted from USEPA (1999b, 2000a)].

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

351

The first CCL began as an essentially unranked list of research needs for drinking water contaminants; that is, additional research and monitoring is needed for most of the contaminants on the current CCL as indicated in Figure 14.2. USEPA faces a complex and ongoing task of (1) assessing the available scientific information on individual contaminant risks; (2) making risk management decisions (based on such assessments) regarding which contaminants should be removed from the CCL through regulation, guidance development, or no further action; and (3) how to prioritize remaining CCL contaminants for further research or monitoring (NRC 1999a). 14.6.1 Applicability of Prioritization Schemes for CCL Contaminants Many government agencies and private industries have developed a number of schemes since the early 1980s that rank chemicals according to their importance as environmental contaminants. However, there are no equivalent schemes for microbial contaminants existing at the time the report was written. Upon reviewing many of these schemes, NRC (1999a) concluded that a ranking process that attempts to sort or prioritize contaminants is not appropriate for the selection of drinking water contaminants from the CCL for regulation, monitoring, or research. In the absence of complete information, the output of such simple quantitative ranking processes is so uncertain (although this uncertainty is generally not stated) that they are of limited use in making more than preliminary risk management decisions about drinking water contaminants of concern. However, the report did conclude that several existing methods and approaches for ranking environmental contaminants could prove to be useful if modified to sort large numbers of potential drinking water contaminants to be considered for inclusion on future CCLs. 14.6.2

Generalized Decisionmaking Framework

As noted above, a ranking algorithm may to appropriate to help determine contaminants to list on the CCL, but such an approach was deemed unsuitable for determining the appropriate disposition of contaminants on the CCL (NRC 1999a). Rather, the decisionmaking process requires considerable expert judgment throughout to address (1) uncertainties from the inevitable gaps in information about exposure potential or health effects, (2) evaluate the many different health effects that contaminants can cause, and (3) interpret available data in terms of statutory requirements. Therefore, such decisions necessarily involve subjective judgments, and the amended SDWA designates USEPA to make them. For each contaminant on a CCL, there are three possible outcomes of USEPA’s decision process: 1. Consider for regulatory action, as required by the 1996 SDWA amendments, if information is sufficient to judge that a contaminant ‘‘may adversely affect public health’’ and ‘‘is known or is substantially likely to occur in public water systems with a frequency and at levels that pose a threat to public health.’’

352

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

2. Drop from the CCL if information is sufficient to determine that the contaminant does not pose a risk to public health in drinking water. 3. Conduct additional research on health effects or occurrence=exposure if information is insufficient to determine whether the contaminant should be regulated. These three outcomes are not necessarily exclusive. For example, based on available evidence, USEPA could decide to initiate regulatory action for a particular contaminant and issue a health advisory, while simultaneously pursuing research to fill information gaps that might result in subsequent further modifications of the regulatory level. Figure 14.3 shows a simplified illustration of the general decision process NRC (1999a) recommended that USEPA use in deciding which of the three outcomes (or combinations thereof) listed above is appropriate for each contaminant on a CCL. The right side of the figure provides a suggested timeline to progress through each step of the process in order to help the Agency allocate their limited time and resources to meet the statutory requirements of the SDWA for the development and use of the first and successive CCLs. Notably, the recommended framework applies to both chemical and microbiological contaminants; differences in either their characteristics or the information available about them do not justify separate decision processes (as used to identify potential contaminants for inclusion on the draft 1998 CCL) (USEPA 1997a).

Figure 14.3 Recommended phased process decisionmaking process for setting priorities among contaminants on a CCL [source: adapted from NRC (1999a)].

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

353

Steps that NRC (1999a) recommended in the decision process are summarized as follows:
After publication of a final CCL, conduct a three-part assessment for each CCL contaminant to include (1) health effects data (to include effects on vulnerable populations), (2) exposure data, and (3) existing data on analytical methods and treatment options. An important component of the three-part assessments will be policy judgments by USEPA about the significance of the available data.
After completing the three-part assessment, USEPA should conduct a preliminary risk assessment for each CCL contaminant on the basis of the available data identified in the three-part assessment. Each risk assessment should be conducted, even if there are data gaps, to provide a basis for an initial decision about the disposition of the contaminant under consideration and to guide research and monitoring efforts, where needed.
Issue a separate decision document for each for each contaminant describing the outcome of the preliminary risk assessment (i.e., whether the contaminant will be dropped from the CCL because it does not pose a risk, will be slated for additional research or monitoring, or will be considered for regulation).
Issue a health advisory for each contaminant not dropped from the CCL after the preliminary risk assessment. (Health advisories are discussed in Chapter 6. Although recommended by NRC, this step is not necessarily practical from USEPA’s perspective because of resource limitations.)
Compile a regulatory package for contaminants to be removed from the CCL or conduct research and=or monitoring for each contaminant remaining on the CCL after the preliminary risk assessment. For contaminants not selected for a regulatory decision, such research and monitoring results should be fed back into another preliminary risk assessment, and new decision documents should be issued on the basis of the results of these subsequent risk assessments. Decisions to drop a contaminant from the CCL, to issue a health advisory or to proceed toward regulation should be based on public health risk considerations only, also indicated on Figure 14.3. However, filling data gaps in treatment technologies and analytical methods is needed to avoid delaying regulatory action for contaminants posing a public health threat. Involvement of all interested parties is important, and should include regulated utilities, state and local regulators, public interest representatives, and consumers (NRC 1999a). For example, public comments on the preliminary regulatory determination will offer independent perspectives and can ensure that criteria developed after consideration of all the relevant issues have not been overlooked. In the long run, considering the views of stakeholders will likely lead to a more transparent and less contentious regulatory development process. NRC did not provide or recommend use of a mechanistic tool (i.e., one free of policy judgments) for assessing contaminants. Indeed, the need for policy judgments by USEPA cannot and should

354

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

not be removed from this process. Ultimately, USEPA is accountable to the public for the decisions it makes about regulating drinking water contaminants. In making these decisions, common sense should be the guide and decisions should err on the side of public health protection. 14.6.3

NDWAC Regulatory Decisionmaking Protocols

The NDWAC was also asked for advice on specific protocols to assist the Agency in making regulatory determinations for both chemical and biological contaminants on the current and future CCLs. Separate protocols were developed for chemicals and microbes, using the three statutory requirements of SDWA Section 1412(b)(1)(A)(i)–(iii) as the foundation for guiding USEPA in making regulatory decisions. These protocols identify specific factors for consideration and define the relative significance and weight that should be given to inform decision-making. Although the protocols do not explicitly score or rank contaminants, they provide a consistent approach to evaluate contaminants for regulatory determinations, and organize the data underlying regulatory determinations in a logical, rational, and transparent fashion for public review (USEPA 2002). To address the issue of whether a contaminant may have adverse effects on the health of persons, NDWAC recommended that USEPA characterize the health risk and estimate a health reference level (HRL) for evaluating the occurrence data for each CCL contaminant. To evaluate the ‘‘known or likely occurrence of a contaminant,’’ NDWAC recommended that USEPA consider (1) the known or estimated national percentage of PWSs with detections above half the HRL, (2) known or estimated national percentage of PWSs with detections above the HRL, and (3) the geographic distribution of the contaminant. To address whether regulation of a contaminant presents a ‘‘meaningful opportunity for health risk reduction,’’ NDWAC recommended that USEPA consider estimating the national population exposed above half the HRL and the national population exposed above the HRL. To determine whether a pathogen poses a human health risk, NDWAC recommended assessment of the known treatment effectiveness of current treatment practices. If a pathogen is controlled by drinking water treatment techniques currently in place to comply with existing regulatory requirements, then a decision not to regulate would be appropriate. If a pathogen is not controlled by such practices, the protocol considers the following factors in assessing the risk of adverse health effects: (1) whether the pathogen is frank or opportunistic; (2) the concentration of the infective dose; (3) the duration of illness, and extent of secondary spread; (4) the method of detection; (5) the immune status of the host (i.e., sensitivity of vulnerable subpopulations), and (6) the long-term immunity conferred by exposure. Finally, if it is unknown whether the pathogen can be controlled by existing technology, the pathogen should be integrated into research track for further study. To assess a pathogen’s occurrence and exposure, the decisionmaking protocol examines the natural history of the organism, to include the existence of resistant

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

355

forms of the pathogen, its survivability in water and the host, the geographic distribution of the organism, and the extent of human and=or animal reservoirs for the organism. The determination of whether regulating the pathogen presents an opportunity for meaningful health risk reduction depends on whether current or pending regulations fail to provide adequate risk reduction, and there are no recognized indicators or surrogates that can be used to demonstrate the effectiveness of treatment in controlling the organism. 14.6.4

Regulatory Decisions from the 1998 CCL

USEPA developed and applied its evaluation approach to be consistent with the recommendations from NRC and NDWAC (USEPA 2002). The Agency evaluated the adequacy of current analytical and treatment methods, the best available peer reviewed science on health effects, and approximately 7 million analytical data points on contaminant occurrence. For those contaminants with adequate monitoring methods, as well as health effects and occurrence data, USEPA employed an approach to assist in making preliminary regulatory determinations that follows the themes recommended by the NRC and NDWAC to satisfy the three SDWA requirements under Section 1412(b)(1)(A)(i)–(iii). USEPA characterized the human health effects that may result from exposure to the contaminant of concern and estimated either an HRL or a benchmark value for each contaminant (USEPA 2002). For contaminants considered to be human carcinogens or likely to be human carcinogens, data on the mode of action of the chemical were assessed to determine the method of low dose extrapolation. When this analysis indicates that a low dose extrapolation is needed and when data on the mode of action are lacking, USEPA used a default low dose linear extrapolation to estimated oral exposures associated with incremental risk levels that range from one excess cancer per 10,000 people (104) to one cancer in a million (106). A 106 risk-specific concentration was selected as the HRL for this assessment. For CCL chemicals not considered to be carcinogenic to humans, USEPA generally calculated a reference dose (Rf D), corresponding to a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of adverse health effects during a lifetime (Barnes and Dourson 1988). It can be derived from a ‘‘no-observed-adverse-effect level (NOAEL),’’ ‘‘lowest-observed-adverse-effect level (LOAEL),’’ or benchmark dose, with uncertainty factors generally applied to reflect limitations of the data used. A Benchmark Dose (BMD) is usually defined as the lower statistical confidence limit for the dose corresponding to a specified increase in the level of health effect of concern over the background level (Crump 1984). Please refer to Chapter 7 for a discussion of uncertainty factors that may be applied, and also to Gibson et al. (1997) for a review of noncancer risk assessment approaches for the use of deriving drinking water criteria. For each CCL contaminant, USEPA estimated the number of PWSs with detections greater than one-half the HRL and greater than the HRL, the population served

356

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

at these benchmark values, and the geographic distribution using a large number of state occurrence data (approximately 7 million analytical points) that broadly reflect national coverage. If a benchmark value was used instead of an HRL, the same process was carried out with one-half the benchmark value and the full benchmark value. Use and environmental release information, as well as ambient water quality data, were used to augment the state data and to evaluate the likelihood of contaminant occurrence. The findings from these evaluations were used to determine if there was adequate information to evaluate the three SDWA statutory requirements and to make a preliminary determination of whether to regulate a contaminant. Table 14.2 provides the preliminary regulatory decisions with supporting HRLs and occurrence and exposure information.

14.7

DEVELOPMENT OF FUTURE CCLs

As noted previously, the CCL will be reviewed and updated by USEPA at least every 5 years. Also, the Agency is required to make determinations to regulate or not regulate for at least five contaminants on each list. In preparing the 1998 CCL, the Agency essentially did the best it could given the short timeframe to meet the statutory deadlines, with the intention of developing a more robust process for developing future CCLs. 14.7.1

Identifying Future Drinking Water Contaminants

The NRC Committee on Drinking Water Contaminants held deliberations following a series of presentations given at a December 2–4, 1998, workshop on emerging drinking water contaminants in Washington, DC. A report was prepared (NRC 1999b) that included papers on individual and groups of related emerging chemical and microbiological drinking water contaminants, analytical and treatment methods, and existing and proposed environmental databases for their proactive identification, research, and potential regulation. Following the presentations, the committee developed a conceptual consensus-based approach for the creation of future CCLs. This conceptual approach for developing future CCLs involves a two-step process (NRC 1999b). First, the ‘‘universe’’ of potential drinking water contaminants, derived from a wide variety of sources, would be first combined and culled using simple criteria and expert judgment to prepare a ‘‘preliminary CCL’’ (PCCL). In this regard, NRC recommends evaluation of several types of related potential drinking water contaminants that were not considered for inclusion on the first CCL, such as pharmaceuticals, biological toxins, and fibers. Next, the PCCL would be processed, using more information in conjunction with a semiquantitative screening tool and expert judgment, to prioritize which contaminants should be listed on the CCL to drive research and regulatory efforts. The process would be repeated every five years for each CCL development cycle to account for new data and as emerging drinking water contaminants are identified. Finally, all contaminants that have not been regu-

357

0.02% (2 of 12,165) 0.09% (11 of 11,788) Round 1: 0.16% (20 of 12,284) Round 2: 0.08% (18 of 22,736) 6.1% (60 of 989) 0% (0 of 13,512) Round 1: 0.01% (2 of 13,452) Round 2: 0.01% (2 of 22,923) 22.6% (224 of 989) 4.97% (819 of 16,495) — Microbes

Chemicals

Systems >12HRL

0.02% (2 of 12,165) 0.09% (11 of 11,788) Round 1: 0.11% (14 of 12,284) Round 2: 0.02% (4 of 22,736) 3.2% (32 of 989) 0% (0 of 13,512) Round 1: 0.01% (2 of 13,452) Round 2: 0% (0 of 22,923) 13.2% (131 of 989) 1.8% (295 of 16,495) —

Systems >HRL

0.02% (8700 of 47.7 M) 0.07% (32,200 of 45.8 M) Round 1: 0.57% (407,600 of 71.6 M) Round 2: 2.3% (1.6 M of 67.1 M) 4.6% (68,100 of 1.5 M) 0% (0 of 50.6 M) Round 1: 0.007% (5600 of 77.2 M) Round 2: 0.002% (1700 of 67.5 M) 18.5% (274,300 of 1.5 M) 10.2% (5.2 M of 50.4 M) —

Population >12HRL

Source: Adapted from USEPA (2002).

Population >HRL

0.02% (8700 of 47.7 M) 0.07% (32,200 of 45.8 M) Round 1: 0.37% (262,500 of 71.6 M) Round 2: 0.005% (3100 of 67.1 M) 2.6% (39,000 of 1.5 M) 0% (0 of 50.6 M) Round 1: 0.007% (5600 of 77.2 M) Round 2: 0% (0 of 67.5 M) 8.3% (123,600 of 1.5 M) 0.9% (446,200 of 50.4 M) —

Diease incidence: Keratitis, 1.65–2.01 cases per year for contact lens wearers; GAE infection, 64 cases in the United States during 1957–1998.

Do not regulate, issue guidance to contact lens wearers

Acanthomoebaa

a

Do not regulate, update health advisory Do not regulate, issued health advisory

Do not regulate

Do not regulate

Do not regulate

Sodium (NIRS) Benchmark ¼ 120,000 mg=L Sulfate (R2) HRL ¼ 500,000 mg=L

Hexachlorobutadiene (R1, R2)s HRL ¼ 0.9 mg=L Manganese (NIRS) HRL ¼ 300 mg=L Metribuzin (R2) HRL ¼ 91 mg=L Naphthalene HRL ¼ 140 mg=L

Do not regulate

Do not regulate

Do not regulate

Preliminary Decision

Preliminary Regulatory Determinations from the First Contaminant Candidate List

Aldrin (R2) HRL ¼ 0.002 mg=L Dieldrin (R