Uncertainty in Medical Innovation: Experienced Pioneers in Neonatal Care (Health, Technolgy and Society)

Uncertainty in Medical Innovation Experienced Pioneers in Neonatal Care

Jessica Mesman

Uncertainty in Medical Innovation

Health, Technology and Society Series Editors: Andrew Webster, University of York, UK and Sally Wyatt, Royal Netherlands Academy of Arts and Sciences, The Netherlands

Titles include: Gerard de Vries and Klasien Horstman (editors) GENETICS FROM LABORATORY TO SOCIETY Societal Learning as an Alternative to Regulation Jessica Mesman UNCERTAINTY IN MEDICAL INNOVATION Experienced Pioneers in Neonatal Care Nadine Wathen, Sally Wyatt and Roma Harris (editors) MEDIATING HEALTH INFORMATION The Go-Betweens in a Changing Socio-Technical Landscape Andrew Webster (editor) NEW TECHNOLOGIES IN HEALTH CARE Challenge, Change and Innovation

Forthcoming titles include: John Abraham and Courtney Davis CHALLENGING PHARMACEUTICAL REGULATION Innovation and Public Health in Europe and the United States Alex Faulkner MEDICAL DEVICES AND HEALTHCARE INNOVATION Herbert Gottweis, Brian Salter and Catherine Waldby THE GLOBAL POLITICS OF HUMAN EMBRYONIC STEM CELL SCIENCE Maggie Mort, Tracy Finch, Carl May and Frances Mair MOBILISING MEDICINE Information Technology and the Modernisation of Health Care Steven P. Wainwright and Clare Williams THE BODY, BIO-MEDICINE AND SOCIETY Reflections on High-Tech Medicine

Health, Technology and Society Series Standing Order ISBN 978 1–4039–9131–7 Hardback (outside North America only) You can receive future titles in this series as they are published by placing a standing order. Please contact your bookseller or, in case of difficulty, write to us at the address below with your name and address, the title of the series and the ISBN quoted above. Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke, Hampshire RG21 6XS, England

Uncertainty in Medical Innovation Experienced Pioneers in Neonatal Care Jessica Mesman Maastricht University, The Netherlands

© Jessica Mesman 2008 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted her right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2008 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–21672–3 hardback ISBN-10: 0–230–21672–2 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Mesman, Jessica, 1962– Uncertainty in medical innovation : experienced pioneers in neonatal care / Jessica Mesman. p. ; cm. – (Health, technology, and society) Includes bibliographical references and index. ISBN-13: 978–0–230–21672–3 (hardback : alk. paper) ISBN-10: 0–230–21672–2 (hardback : alk. paper) 1. Neonatal intensive care. 2. Medical innovations. 3. Neonatology. I. Title. II. Series. [DNLM: 1. Intensive care, Neonatal. 2. Intensive Care Units, Neonatal. 3. Intensive Care, Neonatal–ethics. 4. Neonatology. WS 421 M579u 2008] RJ253.5.M47 2008 618.92’01–dc22 10 17

9 16

8 15

7 14

6 13

2008020792 5 12

4 11

3 10

2 09

1 08

Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

Contents Series Editors’ Preface

vii

Glossary

ix

1 Neonatology: A Permanent Dynamic of Change 1.1 Medical innovation, societal responses, and practical implications 1.2 Exnovating competence and structures 1.3 Significant fields of insight 1.4 Mortality, morbidity, and morality 1.5 A journey begins

1 2 4 7 12 17

2 Newly Born and Indeterminate 2.1 Defining the specific situation 2.2 Constructing child and NICU The child as construct The NICU as construct 2.3 Styles of reasoning The use of repertoires

19 20 26 28 33 35 41

3 Co-travellers 3.1 Mutual expectations Staff members’ expectations of parents Parents’ expectations of staff Vocabularies of expectations 3.2 Parents as object of care 3.3 Parents as bearers of parental authority It is the doctor who decides It is the parents who decide It is the law that decides 3.4 Reconciling competing vocabularies Who decides? When is a decision taken? Styles of ordering and social orders

46 49 49 52 55 57 61 65 66 67 71 71 75 77

4 Uncertain Trajectories 4.1 Determining a treatment trajectory Time as risk and ally The use of protocols 4.2 Technology’s role on the treatment trajectory Technology and care

80 82 82 86 93 93

v

vi Contents

Disciplining the treatment practice Technology as problem 4.3 Numbers as compass A reliable instrument The interpretation of numbers Number versus number Competing knowledge frames The NICU as a quantitative practice

97 101 104 105 107 109 111 113

5 Beacons on the Horizon 5.1 Position and prognosis The specificity of the perspective The specificity of acting 5.2 A topography of experience Distance versus proximity Observing versus comparing Hands-on versus hands-off 5.3 Prognostic markers Exemplary cases Technological markers Rhythm and pace Spatial markers 5.4 Experience and expectation

116 119 121 122 124 125 130 132 136 137 141 144 147 150

6 The Moral Load 6.1 The nature of the decision Evidence Asymmetry 6.2 The functionality of relocation Robustness Creating a basis for consensus 6.3 Another trajectory, another order

152 154 156 158 160 161 166 170

7 The End of the Journey 7.1 Moments of reflection Rewriting the past in reflection 7.2 Situating knowledge and morality in treatment processes Opening up the in-between zone Experienced pioneers

179 180 184 187 187 190

Notes

195

Bibliography

205

Index

215

Series Editors’ Preface Medicine, health care, and the wider social meaning and management of health are undergoing major changes. In part this reflects developments in science and technology, which enable new forms of diagnosis, treatment and the delivery of health care. It also reflects changes in the locus of care and burden of responsibility for health. Today, genetics, informatics, imaging and integrative technologies, such as nanotechnology, are redefining our understanding of the body, health and disease; at the same time, health is no longer simply the domain of conventional medicine, nor the clinic. More broadly, the social management of health itself is losing its anchorage in collective social relations and shared knowledge and practice, whether at the level of the local community or through statefunded socialised medicine. This individualisation of health is both culturally driven and state sponsored, as the promotion of ‘self-care’ demonstrates. The very technologies that redefine health are also the means through which this individualisation can occur – through ‘e-health’, diagnostic tests, and the commodification of restorative tissue, such as stem cells and cloned embryos. This Series explores these processes within and beyond the conventional domain of ‘the clinic’, and asks whether they amount to a qualitative shift in the social ordering and value of medicine and health. Locating technical developments in wider socio-economic and political processes, each text discusses and critiques recent developments within health technologies in specific areas, drawing on a range of analyses provided by the social sciences. Some will have a more theoretical, others a more applied focus, interrogating and contributing towards a health policy. All will draw on recent research conducted by the author(s). The Health, Technology and Society Series also looks toward the medium term in anticipating the likely configurations of health in advanced industrial societies and does so comparatively, through exploring the globalisation and the internationalisation of health, health inequalities and their expression through existing and new social divisions. The locus of this book is the neonatal intensive care unit – a place at the cutting edge of medical science where many activities give rise to ethical concerns for both medical professionals and parents. Mesman vii

viii Series Editors’ Preface

provides a rich account of how scientific and technological developments in diagnosis and treatment may not only improve medical processes but also complicate them by raising new uncertainties about symptoms and expectations about the future health of newly born babies. The book takes us on a journey, from the arrival of a baby in an intensive care unit, accompanied by anxious parents, through diagnosis, treatment, and departure, sometimes to a healthy future, but not always. The staff in these units are highly skilled but each patient is unique and fragile, and each journey is different. Thus, the staff who accompany babies and parents are ‘experienced pioneers’. Andrew Webster and Sally Wyatt

Glossary1 Acidosis: a state of low blood pH that is marked by sickly sweet breath, headache, nausea and vomiting, and visual disturbances and is usually a result of excessive acid production. Anencephaly: congenital absence of all or a major part of the brain. Asphyxia: a lack of oxygen or excess of carbon dioxide in the body that is usually caused by interruption of breathing and that causes unconsciousness. Atrium: a chamber of the heart that receives blood from the veins and forces it into a ventricle or ventricles. Babylance: an ambulance that is specially equipped for transportation of newborns. Bilirubin test: tests the level of bilirubin in the blood. Bilirubin is a reddish yellow pigment that occurs especially in bile and blood and causes jaundice if accumulated in excess. Bradycardia: an abnormally low rate of heart contractions and consequent slow pulse. Bronchopulmonary dysplasia (BPD): a chronic lung condition that is caused by tissue damage to the lungs, and usually occurs in immature infants who have received mechanical ventilation and supplemental oxygen as treatment for respiratory distress syndrome. Caesarean section: surgical incision of the walls of the abdomen and uterus for delivery of a child. CPAP: continuous positive airway pressure: This is a technique of assisting breathing by maintaining the air pressure in the lungs and air passages constant and above atmospheric pressure throughout the breathing cycle. CTG: cardiotocography: an instrument that registers the foetal heartbeat and the contractions of the uterus during childbirth. CTG can be used to identify signs of foetal distress. DNR-code: a Do-Not-Resuscitate code which indicates that a non-resuscitation policy has been agreed upon. Doppler machine: Doppler ultrasound used to measure cardiovascular blood flow velocity for diagnostic purposes (as for evaluating valve function).

1

The descriptions in this glossary are based on the Web-based Medline Medical Dictionary and Wikipedia. ix

x Glossary ECMO: extracorporeal-membrane oxygenation is an extracorporeal technique (medical procedure outside the body) of providing both the heart and lungs oxygen. An ECMO is similar to a heart-lung machine as during open-heart surgery. EEG: electroencephalography, a brain function test, is the neurophysiologic measurement of the electrical activity of the brain derived from electrodes attached to the scalp. Hypoglycemia: an abnormally low concentration of glucose in the circulating blood. Immaturity: lacking complete growth, differentiation, or development. Intra-uterine condition: condition of the foetus within the womb. Intubation: tracheal intubation is the positioning of a tube of a ventilator for artificial respiration into the windpipe (trachea). Lasix: a diuretic is any drug that elevates the rate of bodily urine excretion. Neonate: a baby of no more than 28 days. Nurse practitioner: a registered nurse with advanced education in the primary care of particular groups of patient, who provides services, within the scope of nursing practice, in the areas of health promotion, disease prevention, therapy, rehabilitation, and support at all levels of the health care system. Parenteral nutrition: food that is provided by intravenous, intramuscular, or subcutaneous injection to bypass the intestines. Perfusor: an infusion pump that infuses fluids, medication or nutrients into a patient’s circulatory system. It is generally used intravenously. Perinatology: sub-specialty of paediatrics that is concerned with the period around the time of birth. Resident: a physician in (post graduate) medical training to become a specialist (in this case a paediatrician). Surfactant: a surface-active mixture that contains both proteins and lipids which prevents the lung from collapsing at the end of expiration. Steroid therapy: therapy that aims at growth of the lungs of the foetus (prenatal therapy) or of the neonate (postnatal therapy). Tertiary care unit: highly specialized medical care unit that can provide advanced and complex procedures and treatments performed by medical specialists in state-of-the-art facilities. Toxaemia of pregnancy: disorder of unknown cause that is peculiar to pregnancy, is usually of sudden onset, is marked by hypertension, albumin in the urine, oedema, headache, and visual disturbances, and may or may not be accompanied by convulsions. Trisomy 13: an extra (third) chromosome of the 13th pair. It causes severe deformation of the head and face, heart, eyes and brain and mental retarda-

Glossary xi tion. Trisomy 13 is typically fatal especially within the first six months of life. Also known as ‘Patau Syndrome’. Trisomy 18: an extra (third) chromosome of the 18th pair. It causes severe deformation of the head and face, heart, digestive and uro-genital tract and mental retardation. Trisomy 18 is typically fatal especially within the first year of life. Also known as ‘Edwards Syndrome’. Turner Syndrome: instead of the normal XX sex chromosomes for a female there is only one X chromosome and fully functional. There is no Y chromosome present. It causes severe deformations.

This page intentionally left blank

1 Neonatology: A Permanent Dynamic of Change

Academic Medical Center: Bleeping monitors, toiling ventilators, and alarm signals all around. As the door swings open, an incubator rocks over the threshold, flanked by a nurse and a resident. In the incubator Tom is lying amidst lines and tubes, barely fifteen minutes old. He was born prematurely at twenty-five weeks’ gestation; his weight of 610 grams sharply contrasts with the average 3500 grams of full-term newborns. Skilled hands carefully take him out of the incubator and put him into his own bed: an incubator surrounded by an array of medical devices. His twin brother died right after the delivery. Fed, warmed, and on a respirator: thus Tom’s life begins. Tiny and vulnerable, he is immediately taken to the intensive care unit of the hospital’s neonatology ward.1 The neonatology ward specializes in the care and treatment of newborns, also called neonates*.2 As a rule, children born after a term of less than 32 weeks or with a birth weight under 1000 grams are always admitted to this ward.3 The Medium Care unit treats children that only need basic medical care; if their situation may have been bad, they are on the road to recovery and mainly sleep, eat and grow. The High Care unit treats children that are not or no longer in a life-threatening condition, yet who (still) need special care. The ideal trajectory of a child admitted to the Neonatal Intensive Care Unit (NICU) is to be moved to High Care and from High Care to Medium Care, after which, finally, s/he is discharged and allowed to go home. A NICU leaves first-time visitors with an impression of controlled chaos – an amalgam of busy staff, noisy machinery, and a strange sense of order. Newborn babies are here because their lives are seriously at 1

2 Uncertainty in Medical Innovation

risk on account of complications linked to their delivery, congenital diseases, infections, or premature birth. Lined up side-by-side, the incubators are connected to an array of state-of-the-art medical technologies with all sorts of tubes and cables: monitors that display flickering numbers and respirators that rhythmically do what they were designed to do until they suddenly generate a bubbling sound or piercing rattle. Most babies are sleeping all day and night, most with the help of sedatives. A single baby may be crying – without making a sound, on account of the respirator tube in the throat. The fragility of the children calls for their constant monitoring. After all, these are the children who do not always respond to medical interventions according to the descriptions found in handbooks, while factual data on their conditions are ambiguous or not available yet. This also explains the round-the-clock presence of nurses and physicians in a NICU. One sees them engrossed in their work of caring for their very young and fragile guests. Much of the medical work involved centres on generating and keeping track of patients’ vitals, recording new data, or checking and rechecking them. The calm, soft-spoken interactions of nurses and doctors rely on a vocabulary that seems specially invented for this space. Parents too have a prominent presence. Regardless of the time of day or night, some are found sitting for hours at the side of the incubator containing their newborn baby. Clearly, then, the lights in the NICU are never turned off.

1.1 Medical innovation, societal responses, and practical implications The NICU is a unique domain for two reasons in particular: this practice finds itself on the cutting edge of medical science and many of its interventions give rise to ethical concerns. As such the NICU constitutes a domain where the challenges and opportunities of medical innovations converge. The medical-technological innovations of the past decades have been such that entirely new and unprecedented opportunities for diagnosis and treatment have become available.4 In hindsight we clearly see what new medical technologies do and what they do not, but initially the intrinsic promise of technology is still fully at work. From this perspective medical innovation is conceived as an answer to each and every problem. It is commonly assumed that with better equipment the problems encountered by today’s physicians will be solved, implying that physicians have to broaden and refine their technological arsenal even further. However, a substantial number of qualitative studies demonstrate how new medical advances not only improve medical processes but

Neonatology: A Permanent Dynamic of Change 3

also complicate them by giving rise to new uncertainties about symptoms and expectations involving diseases and health disorders.5 Technological development creates not only new treatment options, but also new questions, new dilemmas and new tragedies. Moreover, these challenges with respect to medical innovations do not limit themselves to medical practices but also have ramifications far beyond biological laboratories and hospital wards. A number of studies about recent medical innovations show how their implications can be found in other realms of society as well, such as in regulatory bodies, insurance practices, and family life.6 Although there has been substantial research on how diagnostic and prognostic innovations change the ways in which we think about issues of life and death, health and disease, and quality of life, there is still little insight into the effects of medical innovations and their societal responses in particular, on actual treatment trajectories. In other words, the social responses to medical innovations have quite specific consequences for medical intervention processes. This raises questions about the implications of the various social responses to new medical technologies for the day-to-day experience of health workers. What are, for instance, the consequences of the medical domain’s public accountability for the actual intervention process? How do legal frames affect hospital practices? What are the consequences of changing perceptions about life and death in society for the decision-making processes in medical care practices? In what ways are particular social concerns reflected by medical practices inside the hospital walls and how are these effects reshaped in the swirl of activity? This feedback loop needs to be taken into account when we study the consequences of medical innovation. Given the increasing intricacy of the overall health care process, it is also relevant to go beyond the effects of societal reactions and focus on what actually takes place at the interface of diagnosis and prognosis, of actors and technology, of medical facts and moral concerns. In other words, it seems increasingly important to examine these interlinked yet discrete processes. With the introduction of new medical technologies, the diagnostic and prognostic processes have become much more complicated, involving more options, other risks, new decision moments, and more pronounced dilemmas for everyone involved. Therefore, the multiple lines of action and reasoning in the day-to-day experience of health care workers are in need of scrutiny, notably as to how these activities intersect. This book provides insights into both the reshaping of the societal responses to health innovations in the concrete work setting of medical

4 Uncertainty in Medical Innovation

care practice, and into what actually happens at the interface of processes of diagnosing and prognostication, and of establishing facts and values. There is a need for this type of study because it is precisely the established routines and practices of actual medical intervention that are constantly destabilized in a health care system that is primarily marked by rapid change. The ensuing uncertainties have contributed to the emergence of a ‘grey area’, one where the established protocols and the conventional answers found in medical handbooks no longer apply. Yet, as before, many clinical conditions still require immediate medical attention and intervention. A focus on the complexities of human decision-making in the face of uncertainty helps us to identify not only the implications of societal responses for concrete medical intervention and decision-making processes, but also what actually happens at the interface of the different constituents of medical practice.

1.2

Exnovating competence and structures

This book zooms in on the NICU as a contact zone of high-tech medicine and exceptionally vulnerable human beings. As an outpost of today’s health care system where the pioneering spirit of medicine reigns supreme, the neonatology ward can serve as an exemplary case for studying some of the health system’s vulnerabilities triggered by a permanent dynamic of innovative change. At all levels, new professional routines and sophisticated technologies have radically altered the nature of medical practice. On many occasions, the NICU staff finds itself in uncharted territory, because a specific technology is new or a certain intervention trajectory poses unknown risks, especially given the extreme fragility of the patients. The intervention process has become much more complicated. How, then, do all those involved make sure that the medical intervention process does not founder? Despite the countless and frequently interrelated dilemmas involved, it is a major challenge for doctors and nurses to prevent the intervention process from stalling. How do the staff members deal with the uncertainties that are intrinsic to being on a scientificmedical frontier? What exactly is the role of medical technology and when or for how long is its deployment warranted? Which strategies help in tackling the many problems that present themselves on a daily basis? What are the mechanisms that ensure the continuity of medical intervention in the NICU? Aside from the ongoing need to develop pragmatic guidelines, the complex problems of our high-tech health care have to be addressed in other ways. It is argued that although various aspects of the contempo-

Neonatology: A Permanent Dynamic of Change 5

rary neonatal practice have been studied thoroughly, there is still little insight into the hidden competence of the actors involved to cope with the complicated interplay of problems and dilemmas associated with the actual processes of managing day-to-day operations in the NICU. Mostly a NICU appears to be a smooth, well-ordered practice with professionals who know what they are doing and who treat and care for children in a routine manner. Yet appearances are deceptive: the order has to be made and adjusted constantly because in the treatment of gravely ill, unstable newborns the existing rules and routines will not always be effective for addressing new and unexpected situations. In a NICU, the normal everyday practice tends to be the outcome of a deliberate ordering process that leads a fragile existence.7 After all, there are the children who do not always respond to medical interventions as foreseen or for whom normal values are not even available yet. Furthermore, frequently decisions have to be taken within a very short time span and in the absence of – reliable – factual data. No wonder that occasionally intervention processes falter and, in the worst cases, end in a deadlock. To create, reproduce, and uphold the socialtechnical order in today’s medical practice calls for a substantial effort on the part of those involved. This effort is the object of analysis in this book on the interplay of control and contingency in critical care units like a NICU. Therefore, this book should not be viewed as strictly an inductive effort, or as just another case study. Instead, its analytical scope and effort can be understood as an act of exnovation. Exnovation refers to the attempt to foreground what is already present – though hidden – in specific practices, to render explicit what is implicit in them (de Wilde, 2000). Where innovation can be defined as ‘to make something new’, exnovation pays attention to what is already in place and challenges the dominant trend to discard existing practices (p. 13). Importantly, ‘things or practices are not less valuable simply because they already exist’ (p. 13). A focus on exnovation allows us to bring to light implicit matters of actual practice and to develop a fresh perspective on the ingenuity of the professionals and the specific structure of their practices. More than innovation, exnovation does justice to the creativity and competence of the participants in their effort to produce a professional performance in the particular dynamic of the practice they are involved in. It offers insights into their specific modes of ordering day-to-day practices (Law, 1994). Exnovation, in other words, elucidates competencies of coordination and alignment of these modes of ordering of which those involved are not always aware. Excavating these coordinative resources can prevent that these kinds of accomplishments

6 Uncertainty in Medical Innovation

will always remain a hidden feature of the everyday medical practice. Moreover, an exnovation of hidden competencies reveals not only the complexity of treatment trajectories and the resourcefulness of the actors involved, but also the limited power of medical technology and formal protocols and regulations to ensure the continuity of medical intervention in the NICU. The methodological consequence of a focus on hidden competencies is threefold. First, to gain insight into these matters we cannot rely on quantitative data collection and analysis. More than quantitative research, qualitative methods such as ethnography are pre-eminently suited for studying the dynamics of critical care practices in situ. As a standard definition puts it, ethnography is ‘a particular method or set of methods which in its most characteristic form … involves the researcher participating overtly or covertly in people’s daily lives for an extended period of time, watching what happens, listening to what is said, asking questions – in fact collecting whatever data are available to throw light on the issues that are the focus of research’ (Hammersley and Atkinson, 1983, p. 1).8 To gain access to the various implicit methods and strategies of those involved in the NICU, it proved productive to focus on complex cases: situations in which it is not exactly clear what is going on or what has to be done. The study of difficult situations may render visible the implicit patterns that are deployed to move a newborn’s treatment trajectory into a positive direction. By highlighting such situations, the boundaries and dilemmas of intervention come more sharply into view. The exploration of complex cases, to be sure, is not meant to give the impression that NICU life is totally caught up in doubts and uncertainties. Social conventions and everyday routines are forms of action coordination that in most cases are forceful and compelling enough to shape and spin out a normal course of affairs. In most cases, NICU treatment is successful simply because it is clear what has to be done and who has which tasks and responsibilities. Based on ethnographic research in the Netherlands and the United States, a multi-level analysis is provided, the objective of which is to exnovate how treatment trajectories are preserved in uncertain circumstances in ways that are frequently unrecognized or misunderstood without such analysis.9 Second, in order to identify the various routines and the often, implicit methods and strategies of intervention used by the NICU staff, this book pays special attention to the in-between zone and focuses on the relationships between the general and the particular; human and machine; formal protocols and the swirl of the treatment trajectory;

Neonatology: A Permanent Dynamic of Change 7

public and local accountability; facts and values; risk and responsibility; expectation and experience. This intermediate zone is fascinating precisely because it facilitates individual conduct; this zone makes it possible for individuals to act precisely by setting boundaries as to what we as individuals want to do and can do. However, this book is concerned not so much with either one or the other, for example the collective structures or the individual case, but with the in-between zone, the hinge between the collective and the individual. As such, processes, rather than products, will be the centre of attention. In other words, the focus will be on gradualization (alignment, fine-tuning) rather than on a comparative/contrastive approach (e.g. certainty versus doubt; familiar versus the unusual; control versus contingency; cohesion version conflict). Third, to gain access to the various implicit methods and strategies of a reliable performance in the day-to-day routines requires special attention to the analytical points of entry. Given the importance of acting, reasoning and interacting as coordinative resources in the constitution of the NICU practice, their styles of ordering deserve attention. Analysis of several treatment trajectories shows how those involved rely on different styles of ordering to deploy their knowledge, experience, and technology in the treatment trajectory. For describing these modes I rely on several analytic categories. If these analytic categories (secondorder) are not part of the discourse used by those involved in the NICU (first-order), I deploy them to define and interpret the different modes of ordering that are deployed in the NICU to identify, prevent or solve problems. Analytical categories like ‘repertoire’, ‘vocabulary’, and ‘script’ will be used to bring to light implicit matters of actual practice. These meta-level categories make it possible to analyse different positions, different practices, and different repertoires at once while avoiding any absolute categories or distinctions.

1.3

Significant fields of insight

With its focus on coordinative resources in medical situations that involve a high level of complexity and uncertainty, this book uses insights from a range of academic fields, notably medical sociology and anthropology, science and technology studies (STS), and bioethics. Being based on ethnographic research, the field of medical sociology and anthropology played a major role in the realization of this book.10 I am aware, though, that the notion of ‘ethnography’ has multiple meanings in the field of anthropology and that it can be viewed as more or

8 Uncertainty in Medical Innovation

less theoretical or political (e.g. the reflexive turn). The study involved in this book can be characterized as ethnography if this notion is viewed as synonymous with ‘fieldwork’.11 However, compared to other field studies in medical anthropology it should not be viewed as strictly an inductive effort, a mere description of ‘life in the NICU,’ as some may conceive of ‘ethnography’. Instead, this book has a specific focus: how actors succeed in acting promptly and adequately in situations of uncertainty and doubt – situations that the actors to some extent actively pursue in this particular setting. Moreover, this book may well be characterized as a theory-guided ethnography since the analytical meta-level concepts direct its focus. In another meaning, ‘ethnography’ suggests the researcher’s explicit textual presence as narrator and/or reflection on the effects of her presence on the practice studied. In the rest of this book I make an effort to minimize my presence and use a passive voice. By using ‘passive voice’ I do not want to give the impression that my account is the only story that can be told about the NICU. In the tradition of qualitative research, however, it is commonly understood that each text embodies the author’s subjectivity. Because of this acknowledging of the author’s signature, explicit suggestion of this subjectivity through the usage of active voice is no longer necessary.12 As a framework for reporting and discussing my findings, I will use four case studies as backbone of my argument: Tom, Esther, Maureen and Robert. These cases were selected for their level of complexity and illuminating strength. Unlike in some ethnographic accounts, my cases are not ‘fictitious’, meaning based on threads from different real cases that are woven into one ‘model’ case.13 In this book the illness trajectory of the cases describe their actual course. Ethnographic study of medicine-in-action is not only a well-tried method in medical anthropology and medical sociology. The preference for medicine-in-action fits in with the tradition of modern STS aimed at science-in-action, as it has existed since the 1970s.14 Specifically, my analytical focus is based on the theoretical frame developed in constructivist STS. For one thing, this angle allows me to criticize the assumption that the fundamental structure of (medical) practice is constituted by principles, deductive patterns of reasoning, and decision protocols. Empirical studies of specific practices found no evidence for these assumptions.15 Instead, detailed studies of practices show how day-to-day operations can never be reduced to a mere application of rules and theoretical principles. Knowledge is not simply waiting out there to be applied in practice; it is constituted in the very same practice as it is used. Thus, in the analysis of medicine-in-action, clinical

Neonatology: A Permanent Dynamic of Change 9

work does not have the sole function of being the context in which busy NICU doctors and nurses are moving around and make their decisions on treatment and prognoses. On the contrary, a turn to practice capitalizes on a careful analysis of the concrete work performed as part of a treatment trajectory. This approach provides insight into how medical knowledge, standards and regulations are reshaped in the swirl of concrete activities, skills, and behaviours that are involved in treatment trajectories. The work of Berg (1997a), Franklin and Roberts (2006), Lock et al. (2000), and Mol (2002), to name just a few out of many, are excellent examples of this approach. The studies in these volumes are based on a fine-grained moment-by-moment analysis which gives attention to details in the cases and the work practice that is analysed. These studies, then, do not so much emphasize medical and ethical handbooks, prescribed rules and professional codes, but the everyday practice of acting and intervening and the intersection of the constituents of practice. Similar to these studies, this book focuses on intersections, interrelations, junctions, interconnections in the practical order as well. Yet, the specific value of this book is tied to the focus on exnovation of the ‘hidden competencies’ required to establish such a practical order. The ‘hidden’ constituent of practices is a significant line of research within the social sciences including the STS field. Especially studies on ‘invisible work’ bear a resemblance to the idea of ‘hidden competence’ as it is used in this book. Nardi and Engeström (1999) distinguish four kinds of invisible work: (1) work done in invisible places; (2) work defined as routine or manual; (3) work done by invisible people; (4) informal work processes (p. 1). The analysis in this book is primarily directed to the second and fourth kind. Furthermore, also with its focus on styles of ordering this book joins a large collection of STS. Order and ordering processes are major themes in STS.16 Some, like Davidson (1999), Hacking (1982; 1992), Rheinberger (2000), and Young (2000), provide insights about different styles of reasoning on basis of a historical analysis. Latour and Woolgar (1979), Law (1994), and Heimer (2001), on the other hand, have chosen for an ethnographic approach. Others have opened up styles of ordering by using notions like ‘ecologies of knowledge’ (Star, 1995); ‘logics’ (Mol and Berg, 1994); modes of ordering (Law, 1994); and ‘repertoire’ (Gilbert and Mulkay, 1984). Opening up ordering processes of practices is a well-established research topic within STS. The analysis in this book is inspired by these studies. Besides looking ‘upstream’ into medical practices, the STS approach has also inspired work that prefers to look ‘downstream’ at the social implications of medical innovations. These kinds of studies show how

10 Uncertainty in Medical Innovation

innovations in the field of medical technology have ramifications far beyond the walls of the biological laboratories or hospital wards. For instance, Brown and Webster (2004), Franklin and Lock (2003), Rapp (2000), and Webster (2006) demonstrate how new medical technologies change our temporal and spatial perception of the body, and life and death. In answering the question to what extent new medical technologies produce distinctively new orderings of the body, these authors show how these technologies alters, develops or sustains certain orders of embodiment. Moreover, they demonstrate how socioeconomic status, ethnic and cultural diversity are intertwined in the decisions that people make. Although highly relevant, their focus is on the social implications of diagnostic and prognostic practices, whereas the topic of my book is on the effect of these societal responses to medical innovations on NICU activities and choices. Since societal responses to new medical technologies have contributed to the development of guidelines and procedures, rules and regulations, the analysis of ordering processes will also include the implications of societal responses for hospital practices. Exploring complex cases on intensive care units for newborns implies that one is confronted with the ethical dilemmas parents and staff members have to face. However, I am not concerned with whether the actors follow the proper ethical guidelines, nor is the focus on an evaluation of the existing guidelines and procedures. Instead of studying how one should act, the emphasis in this book is on how actors in fact act in moral complex situations. With this focus this book favours an empirical approach and as such takes a critical stance towards a rule-based approach. Rule-based ethics starts from the assumption that a physician can arrive at proper decisions by applying ethical principles. Over the years this approach, however, is subject to criticism, mainly directed to the assumed relationship between theory and practice: acting is applying general theoretical principles. According to this model, the underlying structure of medical action is constituted by medical principles, deductive reasoning patterns and decision protocols. Based on this, rule-based ethics would ignore the concrete practices in which the dilemmas occur. General guidelines and procedures would ignore the specific situation, the experience of those involved and the complex, unique character of medical practice. To do justice to the latter, so the critics argue, reflection from within the profession is necessary. This calls for an ethics that does not negate the situation’s special character, but one that precisely starts from it. A situated morality, rather than universal principles, should be the focal point. In this sense some have referred to an empirical turn in

Neonatology: A Permanent Dynamic of Change 11

ethics.17 Despite different approaches to study moral choices, studies in line with this perspective share the view that reflection should start in the actors’ specific experiences.18 In this book, I join them in their effort to explore the actual process of moral decision-making. An important feature of this book is its focus on the process of ‘morality-in-the-making’ and demonstrates how the testing of facts and the testing of values go hand in hand. Questions addressing the nature of a case and how to act properly cannot be dealt with in isolation: a specific answer to one question implies a particular answer to another, and vice versa. On the basis of detailed analysis this book reveals how the description of the baby’s condition exceeds the domain of diagnostic and prognostic facts. Case descriptions suggest directions for action and have consequences for how one deal with moral issues in the NICU. Careful analysis of a decision’s genesis reveals how diagnostic and prognostic facts, values, and medical actions are mutually interdependent. A detailed examination of the processes in which case descriptions emerge, then, offers insight into how moral choices are made in the NICU. It reveals the processes and competencies involved to make morality custom-made, tailored to size in the NICU practice. This book aims to explain how in the NICU one encounters a situated morality in which the distinction between a general moral category and the unique case (the child’s situation) is resolved. Morality, as I will argue, means demarcation and it therefore explores the range of divergent activities involved by several actors in various sites and at different moments. Rather than addressing the correctness of decisions and actions, I consider how the actors – nurses, physicians, and parents – actually arrive at decisions and interventions and how they subsequently handle them. This is why this book concentrates on the roles of actors, the meanings of data and the functions of devices. In sum, the larger concern of this book is with uncovering the coordinative resources in medical situations that involve a high level of complexity and medical and ethical uncertainty. A consideration of how medical processes are resumed, adjusted, and perhaps jumpstarted again also provides insight into the ways in which actors in the NICU practice make moral decisions and, potentially, stretch moral boundaries.19 Based on this specific concern, this book pursues to give another perspective on today’s NICU practice, and, as such, it may generate insights that, besides social sciences, can also be interesting for neonatology or, for that matter, the ethics of health care. In order to explore the NICU as a practice with a permanent dynamic of change, the book follows doctors and nurses in their day-to-day

12 Uncertainty in Medical Innovation

production of workable knowledge and reliable performance, while also the role of parents as knowledge producers is considered. However, I will first address two main characteristics of neonatology: its ongoing concern with the most advanced medical knowledge and technology and its recurring moral dimensions.

1.4

Mortality, morbidity, and morality

The introduction of the incubator at the end of the 19th century meant the beginning of special medical care for newborns, culminating in today’s cutting-edge neonatology practice. Especially after the Second World War, medical technology rapidly advanced. The opening of the first intensive care units for newborns in the 1960s meant a radical shift of the boundary beyond which newborns qualified as viable and hence eligible for treatment.20 The survival rate of infants rose sharply in this period. In the Netherlands, for instance, the lower limit for treatment in the 1950s was a birth weight of 1500 grams. At the start of the 20th century a child weighing less than 1000 grams at birth had no chance of survival whatsoever, while children between 1000 and 1500 grams had an 11 percent chance of survival (Dubbelman and Hermans, 1990). The intensive care provided caused these rates to go up to 33 percent and 68 percent, respectively, for the Dutch NICU population by the first half of the 1970s. A decade later they further increased, respectively, to 46 percent and 82 percent (Ibid). Starting in the 1980s, birth, pregnancy, and the overall neonatal phase were increasingly seen as interrelated. This led to more collaboration between gynaecology, midwifery and neonatology, which evolved into perinatology*. In the past two decades the therapeutic techniques have improved such that in western societies, like the United States, the survival rate of extremely low birth weight infants (below 1000 grams) currently stands at almost 70 percent (Hack and Fanaroff, 2000; Wilson-Costello et al., 2007).21 Major developments that lay the foundation for this accomplishment are that in assisted ventilation in the 1980s and the introduction of surfactant* and steroid therapy* in the 1990s. At first the rise of the survival rate was accompanied by an increasing number of children with handicaps, such as blindness, deafness, spasticity, and reduced mental capabilities.22 By fine-tuning the equipment and an increasing knowledge and skills, this morbidity rate began to drop in the 1970s. By the mid-1990s, however, the chance of neurodevelopmental impairment and other handicaps increased again, probably because of the rise of the total number of premature births and these

Neonatology: A Permanent Dynamic of Change 13

babies’ chances of survival. Nowadays the perinatal (increased use of Caesarean section delivery and antenatal steroid therapy) and neonatal (use of postnatal steroid therapy, surfactant therapy, and assisted ventilation) approach of the 1990s has changed in some respect. For instance, to reduce adverse effects the use of postnatal steroid is decreased while sepsis prevention measures have increased. Due to these adjustments survival without (neuro-developmental) impairment increased again in the 2000–2001 period (Wilson-Costello et al., 2007). The statistics of the Dutch NICU population show a similar trend.23 If in the 1980s the mortality rate among premature babies in the Netherlands was 76 percent, in the 1990s it dropped to 33 percent. The overall mortality rate of the Dutch neonatology population dropped from 30 percent in the 1980s to 11 percent in the 1990s (Stoelhorst et al., 2005). By the 1990s the success of the treatment of premature babies had led to a significant increase in survival rate of children with a birth weight of at least 1000 grams after a pregnancy of more than 26 weeks (Health Council of the Netherlands, 2000). However, by the late 1970s, when neonatology had achieved many medical successes, in several western societies, including the United States and the Netherlands, a public discussion erupted about this practice’s ethical aspects. Physicians developed various new treatment options, many of which were based on high-tech, but to outsiders the dominant role of technology also proved a source of concern. Some argued that the developments in the neonatology practice were dictated by a technological imperative: all that is possible in technical terms should be deployed. Concerns were voiced from within the profession as well. In 1987 Versluys, a Dutch professor of Neonatology, worried about the increasing ‘machine room’-like outlook of NICUs with ever more specially designed ‘mini versions’ of tools and equipment: ‘We treat ever smaller newborns; a term of 32 weeks, once the limit, became 30, 28, and 26. Will we perhaps one day, together with our colleague embryologist and the embryology-trained nurse, care for a 12-weeker in an artificial placenta, or even go further, and start feeding the sperm cell?’ (Versluys, 1987, p. 1083) This neonatologist was critical of the one-sided ‘triumphant belief in progress’. If physicians had a powerful ally in technology, he reasoned, it would also be accompanied by new moral dilemmas. How to resolve them and who should decide? Increasingly, ethicists also raised questions

14 Uncertainty in Medical Innovation

about whether all that was feasible on account of technology should in fact be pursued. Given the severity of the disabilities linked with premature birth, not every child seemed helped by life-saving interventions. The traditional equivalence between ‘doing all that is possible’ and ‘doing well’ some believed to be no longer tenable. In specific cases it would be preferable to ignore technological options and refrain from treatment, notably in light of the child’s projected poor quality of life. Increasingly, physicians began to opt for a selective treatment policy. This was not limited to the Netherlands. In the United States, for instance, selective treatment began to be practised as well. In 1973 the American paediatricians Duff and Campbell (1973) published the first data on refraining from life-extending intervention in the neonatology practice.24 Because babies who were not eligible for treatment would die immediately, the question arose in which cases abstention from intervention was to be allowed and who should decide. The discussion was no longer just about the boundaries of life-extending intervention, but also – and even more so – about those of life-ending intervention.25 A growing number of ethicists and legal experts articulated their views on the boundaries of medicine, life-ending action in particular. In the early 1980s the public debate in the United States was concentrated around a life-ending intervention on a baby boy, known as the ‘Baby Doe case’.26 The controversy resulted in the ‘Baby Doe Regulations’ which prohibited the act of refraining from medically necessary treatment of handicapped newborns. From now on American physicians must treat all infants with life-threatening disorders unless the child is irreversibly comatose; treatment would only prolong dying; treatment would be futile in terms of survival and as such would only add to more misery. In this way the Baby Doe rules restrict the abilities of the parents and doctors to select individual treatment plans for severely ill infants. For that reason, some claim that these legal requirements are inconsistent with guidelines from the American Academy of Pediatrics that stress individualized decision-making based on the best interest standard (Koppelman, 2005). A similar case happened in the Netherlands, known as the ‘Baby-Ross case’. Also this child was born with Down syndrome and an atresia and died due to non-intervention. Likewise, the Baby Ross case triggered a fierce public debate.27 However, unlike in the United States, this debate did not result in a specific government policy. In the Netherlands the initiative to develop guidelines was taken by the Dutch medical associations. They felt the urge to formulate their position and develop professional guidelines. In the early 1990s, the Royal Dutch Medical Asso-

Neonatology: A Permanent Dynamic of Change 15

ciation (KNMG) and the Dutch Paediatrics Society (NVK) issued reports articulating procedural guidelines. Both reports propose criteria for rules of medical intervention in cases of life-extending and life-ending action. This attention for guidelines was prompted by the desire to get a handle on the complex moral problems in actual practice. The guidelines were meant to function as a mould, aimed at reducing ethical issues to manageable proportions. Moreover, the guidelines answered the felt need for regulation. A preliminary study by the KNMG revealed that in taking decisions on life-ending intervention all sorts of criteria played a role, such as chance of survival or handicaps, the seriousness of the handicap and the personal views of physicians and parents involved. There was hardly consensus on the criteria. As a result, comparable cases were decided differently in various hospitals. The Commission on Acceptability Life-ending Intervention (CAL) was in favour of some uniformity in this respect and it listed rules for physicians to adhere to (KNMG report, 1990). Clearly delineated procedures would contribute to situations in which decisions on life-ending treatment were taken with great caution. Although paediatricians use the standards for exercising due caution from the 1992 NVK report ‘Acting or Deciding to Forgo: Limits of Medical Treatment in Neonatology’, they themselves underscore the need for external social evaluation in the case of life-ending treatment. Active lifeending action without the explicit request of the person involved legally counts as murder in the Netherlands. In such cases physicians are obliged to notify the district attorney. To date not a single physician has run into difficulties, as all cases have been dismissed based on proof of exercise of due caution in intervention; however, Dutch physicians continue to show reserve in notification. They certainly favour monitoring life-ending intervention, but not just by the legal authorities. Preference is given to a retrospective assessment by a multidisciplinary team, a view supported by the KNMG. A clear notification procedure and a multidisciplinary monitoring commission that includes legal experts, medical experts, nurses and representatives from patients’ organizations, is something paediatricians have been pushing the government for over the past ten years. In 2006 the multidisciplinary monitoring commission was installed. In the meantime, one of the hospitals had set up a protocol for active life-ending intervention in close collaboration with a district attorney (Verhagen and Sauer, 2005). This protocol applies to situations in which children suffer unbearably and cannot be cured but still do not die. For this fairly rare situation (occurring 15 to 20 times per year in the Netherlands), this protocol articulates the standards for exercising due caution that agree with those in the 1992 NVK report. This so-called ‘Groningen

16 Uncertainty in Medical Innovation

Protocol’, however, defines the procedure to be followed in great detail. If physicians act in line with it they run no risk of persecution. The aim of this protocol, then, apart from exercising due caution in intervention, is to bring such situations into the open and ensure that physicians, backed by the guidelines, will formally notify such intervention. This will allow society to monitor whether indeed the procedures are followed. Since July 2005 the Groningen Protocol is valid for hospitals throughout the country. This is not to deny that the Groningen Protocol caused great worries abroad. Especially in the American and Slovakian media there was an outcry, couched in terms of Dutch barbarianism, Nazis and murderers. The Vatican also spoke of Nazi practices. Strikingly, the data on lifeending intervention in the Netherlands have barely changed in the past five years, despite the fact that in 2001 the Euthanasia Act came into force. Many, notably abroad, feared that the Netherlands was thus setting down for a slippery slope. So far this has proved not to be the case (Vrakking et al., 2005). From the start, there was consensus about the Groningen Protocol on life-ending interventions in the Dutch medical profession. The same is not true, however, for the proposal to give up the 25 weeks minimum. As a rule, Dutch physicians do not treat children born at less than 25 weeks. Yet some argue that this uniform boundary should be replaced by a case-by-case approach, based on each child’s individual chances, irrespective of age or weight.28 So far this proposal has met with great resistance in Dutch paediatrics. Although all agree that premature babies’ chances of survival have strongly improved since the publication of the 1992 report, this would merely apply to those born after 27–32 weeks. Of the children who survive after a term of 22–24 weeks most still die within a month or they have to live with multiple handicaps. This is why paediatricians are still quite hesitant to treat very premature babies (22–24 weeks). In this, the Netherlands differs from nearby countries such as England, France, and Sweden, where seriously premature babies are treated. These countries justify this policy by claiming that they withhold treatment when serious complications occur. This is different in Germany and Italy, where in most cases treatment will be continued (den Ouden et al., 2000). Also in the United States children are treated after a pregnancy of 22 weeks. Obviously, neonatology is not the only medical practice struggling with complex issues. Nor do procedural guidelines function as a solution that is specific to the ethical concerns involved in neonatology. Descriptive and procedural guidelines also occur in other health care domains, like medical experiments, organ donation, and genetic screening.

Neonatology: A Permanent Dynamic of Change 17

If the discussion on health ethics – the procedural approach in particular – is mainly conducted within the academic arena, the debate on the deployment of medical innovations and the decision power of physicians also takes place in the public domain. Outsiders are equally concerned about the decisions that are taken on the work floor. Social concerns about both life-prolonging and life-ending interventions have instigated a debate that has resulted in public accountability of the medical domain. Nowadays, medical-technological developments are also on the agenda of public forums while commissions periodically assess medical work. Clearly, the era when physicians were the only ones to evaluate medical work formally is over. Yet, social concerns do not only result in rules and regulation, but also motivate the medical profession to become more open. In the Netherlands the trend towards transparency and controllability on this issue is still ongoing. The Baby Doe Regulations, the guidelines of the Dutch Paediatric Society and the ‘Groningen Protocol’, are clear examples of the practical implications of public debate over medical innovations. The social concerns about the decisive power of doctors over life and death have resulted in specific rules which have an effect on the actual decisionmaking process on the NICU. However, having said this, the question remains how these rules actually operate in the swirl of daily activity. This issue will be dealt with in the following chapters.

1.5

A journey begins

A neonate’s treatment is like a journey that from the very start is full of unexpected incidents and incalculable uncertainties: both the young patient’s destination and the unfolding of his/her immediate future are anything but predictable. Even if the staff has accompanied children on similar trajectories numerous times already, the routes and circumstances are never entirely the same. Minor deviations from the chosen trajectory may have large and sometimes fatal consequences, but in most cases hope will appear on the horizon. The fellow travellers start out as mutual strangers: staff and parents mostly do not know each other beforehand, yet the baby is of course the greatest unknown stranger of all. The starting point of the journey has to be clearly identified. After all, what, exactly, is this new child suffering from? The child’s birth immediately raises the issue of his/her identity. The newborn is still a blank slate, unmapped territory. The following chapter describes the ways in which the infant’s identity takes shape as the outcome of a host of care activities and medical decisions. This process

18 Uncertainty in Medical Innovation

simultaneously offers an image of the NICU as practical order. Children who need NICU treatment also bring along their parents, who become part and parcel of the dynamic of their child’s dire situation as co-travellers. Chapter 3 focuses on how the social relationships in the NICU are organized. This practice is determined by the ongoing flow of activities associated with highly specialized care, by the continuous admission of new patients (including their parents), as well as by the instability of the conditions of the patients. As such the NICU can be considered as a ‘high-3 work environment’: high-technology, highintensity and high-reliability. The changeability of the patients requires around-the-clock monitoring and frequent adjustments, while the recovery of an admitted child can never be taken for granted. Life in the NICU is characterized by continuous struggle and entirely unpredictable changes that result in erratic flows of action. The fourth chapter describes the taxing trajectory that the actors in a child’s treatment have to follow. Treatment is never merely a matter of solving an infant’s medical problem; instead, it always involves the opening up of a very specific, individualized trajectory. Over time the staff’s room for intervention diminishes and in order to keep the child on the right track, the staff will have to find its way among a growing number of interventions that mutually affect each other. Each actor’s specific involvement with the child gives rise to a specific understanding of its prognosis. In Chapter 5 I analyse how divergent reference points function as prognostic markers. Occasionally, physicians, nurses and parents have to face extreme moral dilemmas in the NICU. Each treatment trajectory comes with its own moral load, which in some situations calls for radical choices. Chapter 6 analyses how the actors define or arrive at a moral turning point in a treatment trajectory. At the journey’s end we usually take a backwards glance in order to learn and digest all the many experiences and adventures, and perhaps revisit some sights. For those in the NICU, however, each journey is a new one by definition, which is why this practice continues to be a pioneering effort. But first things first: for NICU staff members, the beginning of a new trajectory frequently precedes a child’s birth. How, at the start of the journey do they determine the specific options and chances for treatment?

2 Newly Born and Indeterminate

Academic Medical Center Sunday, in the delivery room, a woman who has been pregnant for six months of her first child suffers from a toxaemia of pregnancy*, a condition that is dangerous for both mother and child. The father, a gynaecologist, a neonatologist and a nurse, who are gathered around her bed, are all looking at the CTG* that displays the child’s heartbeat. Repeatedly the heart rate drops, which indicates that the child is in distress. A loud beep from the CTG is sounding. When the number of heart beat interruptions start to succeed each other ever more rapidly and it takes longer for the child’s heartbeat to recover, the moment may arrive that its heart will no longer be able to correct. This is why direct intervention is called for. The child needs to be delivered as soon as possible. The baby’s condition, however, is so feeble that the chances of a happy ending are slight at best. What to do? How to define the child’s situation? The various treatment options are highly dependent on the child’s condition, if treatment is an option at all. This dilemma presents itself to all present in the delivery room of the Academic Medical Center. This hospital is one of the Dutch eight teaching hospitals. For four months I could be found nearly every day in this NICU. It offers space to 24 children, nine of which may be in its NICU. Most of them come from the same local region; only when there is a capacity problem in one of the other ten NICUs in the country, children from other regions are admitted. Despite the many ethical discussions on life-ending action in intensive care units for newborns, there is still little insight in the specific 19

20 Uncertainty in Medical Innovation

processes and actual practices involved. In this chapter I will provide a description and analysis of the ways in which the involved staff members and parents address these moral dilemmas and arrive at their decisions. By closely following specific decision-making processes it becomes possible to shed more light on the specific styles of ordering of activities, motives and forms of reasoning employed in these difficult situations. The first section concentrates on the decision-making process in the delivery room. To come to a shared decision, I will argue, does not require a shared set of reasons and motives. While focusing on how the involved staff members and parents define the situation and come to a decision, I will demonstrate how the prognosis, diagnosis and intervention policy are not produced sequentially but in one and the same process. To open up this process of co-production I will provide a detailed analysis of what happens at the interface of the intertwinement between prognosis, diagnosis and intervention policy. In the second part of this chapter we move to the NICU ward and follow staff members while they collect and organize data to define the situation of the child that is admitted to their ward. On the basis of this analysis, I demonstrate how both the identity of the child and the NICU are also co-produced in the same ordering process of activities and decisions. However, notwithstanding all the efforts, in a NICU ward there is no guarantee for a happy ending. The last part of this chapter therefore will focus on a discussion about the question whether or not to continue the treatment and the styles of reasoning involved. On the basis of a detailed analysis I will argue that specific styles of ordering play a crucial role in the process of decision-making. Before we move to the NICU ward we first go back to the delivery room. What will be decided? Will they withhold any treatment and let the child die a natural death or do they opt for a Caesarean delivery regardless of the involved risks? What kind of motives, arguments and principles are involved in their decision?

2.1

Defining the specific situation

In the delivery room a choice has to be made between intervention or letting the child die. The child is in such bad shape, however, that the nature of the intervention needed is not evident right away. But does the child have no chances at all? How do the medical staff and the parents assess all the specifics of the situation and how does their individual comprehension determine their decision? In this case, the lives of both child and mother are at risk. Toxaemia of pregnancy can be treated in several ways: with the help of medication or

Newly Born and Indeterminate 21

by terminating the pregnancy. A medication-based approach implies that only the mother is treated, while the child will die on account of its feeble, premature condition. A discontinuation of the pregnancy implies that the child will be delivered through Caesarean section, followed by an immediate transfer to the NICU for treatment. The child’s chances of survival are slim if her foetal condition proves to be weak, which raises the issue of whether these chances offset the risks for the mother associated with the surgical delivery (as well as with potential future pregnancies). Precisely the fact that the choice of treating both mother and child involves a single decision seriously complicates the situation in the delivery room. In this case, all data point to the necessity of the child’s immediate delivery. The CTG reveals that without intervention the child will only live for may be less than an hour. This is confirmed by the neonatologist, but he also points to this decision’s consequences for both child and mother: ‘We shouldn’t forget that this child is only twenty-five weeks old and has a serious growth disadvantage. This baby is not just far too young, but also way too small for its gestational age.’ Because of the pregnancy’s short duration, in combination with the child’s growth disadvantage and the worrisome intra-uterine condition*, there is a good chance that a Caesarean section will merely transform ‘foetal distress’ into ‘neonatal distress’ or even ‘neonatal death.’ When the father inquires after his child’s chances, the neonatologist replies: ‘It is very hard to say anything based on what we now know about the child. But my best guess is that your child has a ten percent chance of getting through and also a ten percent chance that she will live but with serious handicaps, yet she has an 80 percent chance of dying right after birth. But again, these are rough estimates.’ These percentages are hardly precise of course. But even if they were, what extra knowledge would they represent to those gathered in the delivery room? At this early stage, statistics is one of the few means for determining a child’s survival chance. It is derived from tables based on weeks of pregnancy and weight. During pregnancy a child’s condition is strongly influenced by these two parameters. A term difference of two or three days may already substantially affect survival rates. These rates go

22 Uncertainty in Medical Innovation

up with 3 to 4 percent per day in case of a pregnancy term of 23 to 24 weeks. In a pregnancy of 24 to 26 weeks this is two to three percent per day. Moreover, a small variation in weight also leads to a substantial variation in percentages. When a child turns out to weigh slightly more or less than projected, this also changes the chances of survival. This is why under these circumstances the question from parents on their child’s chance of survival is hard to answer.1 As a neonatologist from the Academic Medical Center comments: ‘We are aware of this problem. This is why one physician will evade mentioning a specific percentage while another will say that the child has a fifty percent chance to make it, which is in fact like saying nothing. Again another will give an estimate. It is mainly the parents who want to hear about percentages.’ In the delivery room situation sketched above, the figures mentioned are hardly precise, but they still provide the father a basis for assessing the options and making a decision on what should be done: ‘So there is a chance the child will get through?’ he asks. The neonatologist nods and points to the fact that if the child’s condition severely worsens it is still possible to withhold treatment. The father suggests that it is ‘a choice between either taking a risk and giving a chance or taking no risk and depriving the child of a chance to survive,’ and he concludes: ‘I feel we must give our child this chance. But I do not want this child to have to live its life with serious disabilities.’ For the father the mere chance of survival, no matter how slight, is enough to opt for intervention. His child, he feels, is a child with a chance. This knowledge he interprets as a moral appeal to intervene. In this context Tijmstra (1987) speaks of the imperative character of technology as resulting from ‘the anticipated decision respite.’ This means that a rejection of the technological possibilities means that one gives up chances of a positive outcome and that later on one may regret this. One opts for deploying technology to avoid possible feelings of regret. The anticipated decision regret, according to Tijmstra, would explain why it is hardly possible to reject the service of medical technology, which gives it a peremptory character. This inclination towards action or active intervention is not just characteristic of hopeful parents. Neonatologists tend to act on the slightest

Newly Born and Indeterminate 23

chance of survival as well. In some hospitals like the Northeastern General Hospital in the United States, physicians have a moral duty to do so. This American NICU has the same level of technological equipment as a Dutch NICU, but is much bigger. It provides medical care for more than forty infants and when all incubators are in use, they are put a little closer to each other to make room for an extra admission. Because of its size, this NICU is a unit all by itself. The region it covers is three times the size of that covered by the Dutch NICU. This is why in Northeastern it is not uncommon for children to be flown in by plane for treatment. The infants in Northeastern are much smaller and more of them are more seriously ill than those in the Dutch nurseries. Premature babies born after 24 weeks are no exception, nor are babies with very severe malformations and disorders. In some cases prenatal care could have prevented this outcome, but not every pregnant woman makes use of antenatal care. On the NICU in this American hospital there are no different ways of acting in cases of premature birth. Here one has to do everything possible to save the child regardless its premature gestational age. The prevalence of this rule becomes clear in their treatment of Maureen. Although she was born after a pregnancy that lasted fewer weeks than the minimum of 24, meaning that formally she was not eligible for treatment, this did not keep physicians from actively intervening right after her birth.2 The attending neonatologist of Northeastern General offers the following comment: ‘Well, you shouldn’t interpret the boundary of 24 weeks too rigorously. It is 24 weeks and its range. This child was 570 grams, so you should not draw boundaries too strictly. Moreover, we had to treat this child because it came out alive and also responded to treatment. A 23-weeker’s survival chance is 20 percent, which is not much. But for a 24-weeker it is already 50 percent. So we always go for those. For a baby of 22 weeks we do nothing; we do not even consider it. But when a 23-weeker comes out well, we give it a try.’ In Northeastern General physicians will only withhold treatment when they have full certainty about the absence of a survival chance. One of the nurse practitioners* of the Northeastern NICU comments on the implications of this striving for certainty: ‘The decision not to intervene is only made in those cases of which everyone knows right away that there is nothing left to be done. But

24 Uncertainty in Medical Innovation

this sort of clarity is rare. In most cases it is absent and therefore you act. In such a situation I go all out. Sometimes you do not know what to act on. Once I treated a child that had very bad blood values. According to the books this child had to be doing very poorly. But the baby was doing quite well, while children with better blood values succumbed. It only shows to tell that even medical knowledge is not fixed. It is rather a matter of interplay between your actions and the baby’s reactions. … Still, the children here all respond differently. Whenever a child with a beating heart is born, you give it a shot, no matter how bad its condition. We once had a child here without kidneys. This diagnosis was already known prior to his birth, which is why the parents wanted to keep the child with them after his birth so he could die in their arms. But when the baby was born his heart was beating, which implied that we had to act. We owed it to the child. It was terrible. The mother kept shouting that she wanted her baby back. Regardless, we took her child to the NICU for treatment, where he died a few hours later. But, well, if the heart is beating, you cannot turn your back on a child. Even if there is no heart rate it is not yet certain that all chances are lost. Generally I go on reanimating for twenty minutes. If after twenty minutes there is no heart rate, at least you know that you have done all you could. … This also puts you in a stronger position in case of a lawsuit.’ In Northeastern a beating heart forces physicians to intervene promptly. This intervention is justified with reference to a child’s chances and thereby it is irrelevant whether or not these chances reach beyond the present. However, these comments also reflect the influence of the legal system on the actual interventions. In these cases fear to get involved in a lawsuit creeps into the treatment trajectory. The juridical system outside the hospital walls is something that is taken into account when one decides about the actions to be taken inside. Besides the legal system, other reasons are also mentioned to legitimize life-saving actions. A resident in Northeastern frames such cases as opportunities for learning: ‘Each intervention leads to more knowledge and skill. Treating a child of 23 weeks also increases my chances of being able to save a

Newly Born and Indeterminate 25

child of 24 weeks. The experience I gain in treating the 23-weeker thus benefits other children.’ In the Northeastern General Hospital every child with a chance for survival – no matter how small – will receive maximal support and intervention. By contrast, in the Academic Medical Center’s delivery room a baby’s chance of survival is no automatic trigger for intervention. That a child has almost no chance also leaves room for other options, such as letting the child die intra-uterine*. In the Netherlands in these situations the final decision is up to the parents. In the first case addressed in this chapter the father asks the physicians to let his child be born because it has a chance to survive. The mother equally opts for this trajectory, but on other grounds: ‘At one point the heartbeat kept dropping and occasionally it even stopped. You cannot imagine what it means to hear your child’s heartbeat suddenly stop. I thought each interruption to be the final one. Those moments occurred ever more frequently. It was discussed whether the child should be taken out. Something had to be done quickly, for I felt that the child would not be able to carry on much longer. … I absolutely did not want the child to die in my tummy. That I would end up with a dead child in my tummy was just a terrible idea. It just had to be taken out.’ The way the mother frames her understanding of the situation not only differs from how the neonatologist frames it, but also from the way in which her husband interprets the situation. She hears no data, no survival percentages: she only hears an alarm that indicates to her that any moment her child may die in her uterus. To her, the numbers on the monitor are not so much an indication of the child’s neonatal condition, but of the baby’s foetal condition. She primarily has attention for the figures on the screen, the alarm that sounds ever more often and the thought that there is a child in her womb. She defines the child’s situation in quite different terms: any moment this child may die in my tummy. Overcome by fear, she looks no further than her body and her being freed of her child. All that will follow afterward remains outside of her scope for the time being. To her what counts is the present and this means only one thing: the child must be taken out. Evidently, the description of the decision-making process in the delivery room, shows how none of the individuals present has a comprehensive overview of the situation. Trying to find the answer on the

26 Uncertainty in Medical Innovation

prognostic question ‘what can be expected in the (nearby) future?’, they need a more precise answer on the diagnostic question ‘what is the matter?’ to be able to decide what to do. What, for instance, is the actual condition of the foetus’ heart? Will the child be able to survive at all? How serious is the immaturity* of the child? Prognosis, diagnosis, and the treatment policy are constituted in one and the same process. In the end the staff members and parents define the specific situation in their own way and from their vantage point. To gain more insight into the various options they interact: they share their knowledge, experience, views and feelings with each other. The physicians notice both foetal and neonatal distress, which makes them hesitate on how to go about best treating the toxaemia. But in the Netherlands the parents have the final say in these cases. While the mother primarily sees a child that needs to be taken out of her womb as soon as possible, the father views his child as one that has a chance of survival and that therefore needs to be treated. Their divergent definitions of the situation ultimately lead to a single choice: the baby is born via Caesarean section. Her name is Esther and she is immediately put on a respirator and transferred to the intensive care unit of the Neonatology Department. While the doctors and parents struggled hard to define the foetal and expected neonatal condition of Esther, this problem of defining the specific situation of a child is not limited to the delivery room but exceeds its boundaries. It recurs also in the treatment of newborns all the time. Now the question arise as to how they deal with this problem on the NICU? Which instruments are used? Where do the data come from that allow staff to begin to define the newborn’s condition when it enters their intensive care unit? Let us take a closer look at what happens when a baby is admitted to the NICU.

2.2

Constructing child and NICU

A child in the NICU has a first and last name; the baby’s data are put into the record, entered on the flow sheet and inserted into the computer. After birth children are also given a nametag around their wrist and a footprint is made. This is meant to prevent potential mistaking of children and other errors. A child has young parents, worried parents, parents who may or may not come to visit, or who may or may not be around all the time. Or s/he has only a single parent. The child has little brothers and sisters or the baby represents the only chance parents

Newly Born and Indeterminate 27

have to fulfil their wish to have children. The ‘child with a chance’ becomes ‘Esther, our daughter.’3 ‘The child that needs to be taken out’ changes into a child with a face, in which parents may or may not discover a striking resemblance. In short, once the child is born s/he becomes part of a larger configuration – be it a family, a family history, a community, and, as in this case, a specific medical practice. For as long as a child resides in the NICU the baby is a patient, implying that the physical dimension of the identity is the crucial factor. Upon admission to the NICU the child is almost a blank slate. This undetermined status is characteristic of the work in this unit, and much of its dynamic is geared towards finding out more about a child’s condition. To get insights into the way staff members describe the child as a patient, we will follow the actions they perform. For describing this process I follow the course of events involving the admissions of Esther and Tom to the NICU of the Academic Medical Center. Like Esther, Tom was admitted to the Academic Medical Center’s NICU right after he was born: The moment he is brought into the NICU, various staff members take action immediately. A nurse places Tom in his incubator and connects the tube in his throat to the ventilator next to his bed. The resident* and the nurse discuss the ventilator settings: ‘Let’s put him on 20 to 4, with a frequency of 55 to 100 percent. If he’s doing well enough, we can wean him down.’ Other questions are asked: how did the delivery go, how high is his glucose, what about his blood gases, how do his lungs sound, and what was his weight at birth? To answer these questions Tom is extensively examined. The physician listens to his heart and lungs, and cautiously she presses his side to examine the liver. After fixing the ventilator, a nurse attaches electrodes to Tom’s chest, and a few seconds later a list of numbers appears on the screen: 155, 32, 71, 89, 36.4, providing information about heart rate, respiratory rate and temperature. His blood pressure and the blood’s oxygen levels appear on the monitor alongside the other data. To keep Tom’s body temperature stable, the incubator’s temperature is regulated. One of the nurses puts a cap on his head, after which she and the resident discuss the blood tests to be performed. ‘Let’s do an astrup, glucose, a full blood count and a bilirubin test.’ Again a list of quantities is summed up. While

28 Uncertainty in Medical Innovation

the nurse prepares for the blood tests, the resident inserts an intravenous line for fluids and medication. The attending neonatologist comes in, glances at Tom, and looks at the monitor and back at Tom again. ‘How’s he doing?’ The resident: ‘Considering his age, he is doing fine. His Apgarscore is hopeful: 3, 6 and 9. I ordered an x-ray, so we can have a good look at his lungs shortly and check if the tube is in the right spot.’ ‘Did he get surfactant*?’ ‘Not yet, first I have to fix the needle of the infusion.’ She struggles with the needle, which often happens because tiny children like Tom have fragile veins. After two other attempts the job is done. This infusion supplies him with medication. Another one is for feeding. With the help of a perfusor* the infusion’s pace can be regulated. This makes it possible to divide his nutrition over a 24-hour period with great precision. To prevent the infusion from coming off, Tom’s arm is splinted with the help of a spatula. In no time his incubator is a jumble of tubes and lines. The nurse glues a boy-blue card to his incubator with TOM on it. The neonatologist and the resident turn around and walk to the desk, while a nurse makes sure Tom is comfortable. He looks exhausted and the monitor displays a fickle picture. The child as construct A child’s admission to the NICU marks the beginning of a flow of activities and treatment sequences, aimed at answering the question: Who is this newly admitted child? All activities are geared toward defining the baby’s situation as swiftly as possible, as a way of inserting the new child into the unit’s everyday order. However, the effort to ‘define’ the newly admitted child and learn more about who he is hardly involves a transparent process. Tom’s identity is constructed on the basis of countless directly gathered data on family ties, test results, outer features, and so on. Yet the ensuing identity is all but stable. Some test values or vitals may suddenly change, thus also drastically altering the child’s identity from being ‘a stable baby’ into an ‘unstable baby’ again. This happened, for instance, in the case of Esther. Academic Medical Center Monday: Esther is one day old. The ultrasound of her head shows a fourth-degree brain haemorrhage on the left side. 10.30 hours: The parents of Esther come in to visit. Early this morning her father was here already, but now her mother

Newly Born and Indeterminate 29

joins him. Her bed is manoeuvred to the side of the incubator. She does not dare to touch Esther and stares absently into the incubator. After ten minutes she wants to go back to the maternity ward. 12.00 hours: A second ultrasound of the head shows a spreading of the bleeding to the right half of the brain. 16.00 hours: Esther is not doing well. Her lungs are so underdeveloped that even maximal assistance from the respirator does not raise the blood’s saturation level. The glucose and iron levels are too low as well. In addition, Esther’s temperature and blood pressure are too low. Although the staff tries hard to counter these problems, not all interventions catch on. Despite all their effort, the staff do not manage to stabilize Esther’s condition. The ultrasounds reveal that Esther’s situation has radically changed. In this respect, one can frequently overhear neonatologists warn residents that a stable baby can turn into an unstable child overnight. Yet it would be wrong to suggest that therefore the effort to define a child’s condition is marked by arbitrariness and contingency. In the NICU, too, the establishing of a child’s condition is based on guidelines and protocols, rather than merely involving a random accumulation of heterogeneous elements. But what role do they play? What are the constituents of the baby’s identity and how are staff members able to spin a web of meanings out of them? If the identity of a child in the NICU consists of an amalgamation of rapidly changing data and insights, it is still an ordered amalgamation. As such it does not appear out of the blue. It is up to the medical staff to structure and interpret the various data it collects. In this ordering process the physician’s daily examination is a major constitutive factor. Academic Medical Center 8.45 hours: The resident* enters the unit. She just received the latest information from her colleague in the night shift. Right now seven of the unit’s nine beds are occupied. The resident sits down at the table in the middle of the NICU, grabs a pen and calculator and starts with the children’s daily examination. She fetches Tom’s record that is on top of the stack. She copies data on temperature, weight, intake of fluid and nutrition and output of urine and faeces from the flow sheet,

30 Uncertainty in Medical Innovation

as recorded by the nurses who checked Tom during the past twenty-four hours. Because in his case many parameters were measured every three hours, the resident calculates averages or records the range. After reviewing the flow sheet she reviews all results that came in yesterday from the biochemical and haematological lab. Based on the blood test results in the computer she calculates the settings of the respirator. She also goes through the reports on X-rays and ultrasounds. Leafing through the record, she reviews the previous day once more and reads the report of her colleague from the night shift. Tom’s condition appears to be stabilizing. Having a basic grasp of his condition now, the resident walks to Tom’s incubator for the physical examination. Meanwhile more than fifteen minutes have passed. She cleans her hands and opens the incubator. She slightly adjusts the light, so she can better assess his skin colour. Next, she picks up the stethoscope that hangs to the side of his incubator to listen to Tom’s heart and lungs, as well as the movement in his tummy. The radio is on, and even if it is not loud it still prevents her from hearing the fine details. This is why she asks a nurse to turn off the radio after which she listens to Tom’s heart and lungs again. Next, she examines the spleen and liver. In the groins and at the heel she feels the pulsations of the blood circulation. By softly pressing his side she checks if Tom is not dehydrating. She is attentive to the presence of sudden peculiarities, but she finds nothing that is uncommon. Carefully she puts him again in the most comfortable position, closes the incubator, and walks back to her desk to write down her findings. She opens the computer program and calculates the composition of the parenteral nutrition*. Based on his weight Tom is given a specific amount of glucose. She calculates the quantity of vitamins, minerals and fats, and enters them in the record. Later on, after having checked all children, she will send these calculations via the computer to the pharmacy, where the parenteral feedings are prepared. The physician now has all the data needed for an assessment of Tom’s condition and she lists his problems, treatment and the related tests. She closes Tom’s patient record and takes the next one from the stack. During rounds she will discuss

Newly Born and Indeterminate 31

her findings and suggestions for treatment with her supervisor, the attending neonatologist. After all, the responsibility for treatment is in the hands of the neonatologist in charge, the ‘doctor in attendance.’ Thus, the resident assesses Tom’s situation mainly while sitting behind a desk, pen and calculator in hand. These hardly seem to be the proper tools for examining a premature baby, nor would one expect a physician spending much time behind a desk. But a physician who has to assess a baby’s condition does not only look inside the incubator; he or she also studies the patient record, the most recent data and the monitor’s current data. The child acquires a specific identity not only on the monitor screen and in staff conversations, but also in lab requests via the pneumatic dispatch, on the light boxes for X-rays and in the hospital’s central computer files. The staff is constantly busy to ensure a steady flow of new data on the child’s vitals – be it on paper, photos or screens. As such the NICU’s everyday practice is geared towards spinning an ever denser web of meanings around the child. Various concrete elements contribute to this web: numbers, words, records, devices, instruments, bodies, decisions and also the unit’s architecture and spatial layout. To work towards defining the child’s condition the physician is constantly busy analysing the stacks of paperwork on the table in the middle of the NICU. This table functions as ‘ordering center’, to use a notion introduced by John Law. According to Law (1994), ‘An ordering Center is (probably) constituted by gathering, simplifying, representing, making calculations about, and acting upon the flow of immutable mobiles coming in from and departing for the periphery’ (p. 104).4 The table is the place where the resident compares and combines various data, so as to turn them into a maximally coherent, transparent story. Given the heterogeneous nature of the elements involved in the ordering process, one should rather speak of a sociotechnical ordering process. The record functions as a central node. The flow sheet, the test results and the physical examination notes are added to the record. Writing is not so much geared towards communicating information, but towards ordering information (Berg, 1996). In the record all data converge and gradually this results in a narrative sequence of the baby’s changing situation. The record is like a travelogue that depicts the baby’s itinerary during his/her hospitalization. Apart from the notes entered by physicians, the unit’s secretary is the one to keep the record up to date all the time. She processes the flow sheet from the day before and lists the various

32 Uncertainty in Medical Innovation

data on nutrition, medication and interventions. The record’s coloured curves of the temperature (red), respiration (black), pulse (blue) and weight (green) cross each other all the time. Nutrition, medication, infusion settings and respirator settings are listed in boxes side by side. A single glance at this information informs the physician of a child’s current situation, including the most recent trials and adventures. In this way the record serves as a ‘center of representation’ (Law, 1994, p. 26). Reading a record properly requires much specific knowledge. A lot of the information is expressed in abbreviations and numbers. Combinations of letters refer to treatments, instruments, medications, assignments, discussion forms, diseases, symptoms and disciplines. Quantitative language helps to build up the child’s identity in particular. Some numbers function as basis for calculations and live on invisibly in averages, range, ratios or device settings. Other data end up in the record unchanged, but acquire new meanings in the context of other data. Blood test results are linked to the respirator settings or the feeding infusion’s cubic centimetres and percentages are related to the child’s weight. From the very start a child’s identity is established in a process of reading, measuring, calculating, comparing, combining and discussing data, but also through sensory techniques like observation, feeling, and auscultation. As such the definition of the neonate’s specific condition is not the outcome of a haphazard accumulation of data, but the result of a deliberate construction. Various data only acquire meaning when combined with others. A test result is assessed in light of the normal value listed in the medical handbooks, the earlier results in the child’s patient record and other test results. A child’s yellow skin colour, for instance, may confirm the result of the bilirubin test*, while a high temperature with a positive outcome of a bacterial culture may point to an infection. Only in relation to other data a specific result counts as normal or abnormal, too high or too low, worrisome or hopeful. Thus a child’s identity is established through a specific arrangement of turning this way and that, calculating, comparing, observing, feeling, listening, reading, combining and deliberating. The physician’s arranging and assembling of all the bits and pieces of information is much like solving a jigsaw puzzle. The child’s fluctuating condition causes the established order to be a momentary one. In this context, it would be misguided to believe in ‘pools of order’; rather, there is an ongoing process of ‘ordering.’ As John Law suggests, ‘orders are never complete. Instead they are more or less precarious and partial accomplishments that may be overturned.

Newly Born and Indeterminate 33

They are, in short, better seen as verbs rather than nouns.’ (Law 1994, p. 1) This is why study of the process of ordering is more appropriate than study of the order itself. The medical staff is aware that its tools are limited and that the data it works with tend to have a high turnover rate. Given the fact that new data may undermine the established clinical picture at any moment, physicians hardly aim to determine an infant’s true or ultimate identity. This would be useless, which is why they are satisfied if their grasp of the infant’s condition provides sufficient information to allow them to do the next step of the treatment trajectory. Aware of the high turnover rate of data, the staff aims to establish a practicable identity instead of a true identity. If in the NICU a child obtains his/her identity as a result of a pattern of actions that recur on a daily basis, this very effort simultaneously structures, as I will argue in the next section, the unit as well: the child’s identity and the NICU practice are constituted in a single dynamic. The NICU as construct The various activities involving the child thus provide a specific order to the neonatology practice. It is characteristic of the NICU, for example, to learn more about a specific child who is hooked up to an array of devices that produce a constant flow of numbers on a screen. One is not likely to find this same practice in a child’s room at home. In the NICU, however, it is normal that the child’s name becomes entwined with numbers with two decimals or that one listens to the heart, assesses the liver and takes pictures of the lungs, or that constantly blood samples are sent to the lab. All these are done to get to know the child. Through these local procedures, routines and vocabularies, the child’s individual identity is gradually constructed. The patient record plays a major role as a structuring device for collecting data about the child. Apart from being a ‘center of representation,’ the record is also an ‘ordering center’ (Berg, 1998). The record’s structure and make-up organize the activities of nurses and physicians; they have to check or examine certain aspects of the child in order to fill the record’s empty boxes. Thus the record’s design structures the staff’s daily routines and activities.5 The record is also an aid in structuring professional conduct. Its shared use is only possible on the basis of a ‘code of practice,’ which all physicians have to learn and sustain (Heath, 1982, p. 59). However, the use of the patient record implies more than just a structuring of the practice. The record leaves more room for specific information than for other information and grants physicians

34 Uncertainty in Medical Innovation

a central role in the information-gathering process. This creates a politicization of the unit’s bureaucracy that favours the position of physicians (Berg, 1996). The NICU actors all focus on a child’s condition, which is complex and instable and this in turn has specific implications for the unit’s organization. The children’s vulnerability requires the constant presence of medical staff. This round-the-clock monitoring is covered by various shifts (day, evening, night and weekend). To facilitate communication of the latest information the shifts briefly overlap, implying that there are fixed transfer moments. The NICU children’s vulnerability also entails another temporal structure: periodically nurses check various parameters. Depending on a child’s condition, this is done at two- or three-hour intervals. These rounds of control play a dominant role in the organization of the medical work in a NICU and determine the scheduling of activities. Other treatment-related matters contribute to defining the unit in specific ways as well. Given the infant’s instable and life-threatening condition, for instance, miscommunications between staff should be avoided at all costs. Apart from informal interactions, NICU life is dictated in part by formal moments of communication about the children’s condition, which function as fixed points in the unit’s daily dynamic. Shift changes, however, involve more than just a few shared moments of mutual communication between nurses or physicians from two shifts. The instable condition of the NICU children calls for these interactions to be based on clearly presented and well-organized relevant information. These transfer moments function as time limits because specific information must be available before the shift ends, which again implies careful planning and timely performance of certain activities. Moreover, interactions in the NICU strongly rely on quantitative terms and abbreviations. The daily medical round, too, is more than a moment of shared interaction, and also functions as a time limit: prior to the medical round, the resident must have checked all children, have all data in order and be able to present a treatment plan. Furthermore, the round’s timing is determined by the way the hospital pharmacy is organized. Most NICU children have infusions based on a 24-hour cycle. The start of this cycle is set at five in the afternoon. If the pharmacy is to have the feeding infusions available at that time, it needs to have the necessary information by noon. This means that the medical round, during which possible changes in feeding are discussed, has to be over by noon. Thus potential changes in feeding put time constraints on what takes place during the morning hours in the unit. Given these constraints, meetings with parents tend to be scheduled

Newly Born and Indeterminate 35

afterward, also because new data may be available by then. Talking with parents about their child’s condition is not only a matter of providing attention, but also of providing clear information. The NICU practice cannot be seen in isolation of its actors – physicians, nurses, children, parents, monitors, computers, ventilators and other things – and their activities and discourse. The unit’s order consists in part of doing medical rounds, the nurses’ monitoring, the staff’s mutual interactions, the processing of quantitative information and the talks with parents; as such these factors contribute to defining the NICU. Over time its practice gives rise to more pronounced patterns, routines, and roles in which child and unit are co-constructed. Thus both the child’s identity and the unit’s structure take shape in one and the same process. Notwithstanding all the hard work involved in taking care of the NICU baby, it does not guarantee a successful treatment. Despite all efforts Esther’s condition is worsening and forces both staff and parents to reconcile their decision to intervene and provide Esther with a maximum of support. A detailed analysis of the process in which these kinds of decisions are made reveals another form of ordering: the competence to order different styles of reasoning.

2.3

Styles of reasoning Academic Medical Center Monday, 18.00 hours: Esther’s condition is worsening. The blood circulation continues to be very poor. The monitor shows changing saturation values and a fluctuating blood pressure. Oxygen level and medication are adjusted once again. Her parents are informed of the results of the second ultrasound scan. 19.30 hours: New problems have surfaced for Esther. The oxygen is raised to 60 percent. Slowly the saturation moves up to 90 percent. In her lungs respiration murmur is detected. 21.45 hours: Esther’s lungs sound worse. The radiology department is called in to make an x-ray. It shows fluid in the lungs and an enlargement of the heart. Tuesday, 10.00 hours: Esther is lying in her incubator, pale and weakened. During the night the monitor displayed irregular values ever more frequently and the same is true of the values that came back from the lab. The physician once more looks at

36 Uncertainty in Medical Innovation

the results. The attending has drawn his conclusion: ‘This is not going anywhere, we’d better stop. I propose that this afternoon we have a meeting on this case in the staff room at half past one.’ Because of the limited success of the various interventions, more frequently one can hear solemn comments coming from the side of threeday-old Esther’s incubator: ‘Whatever we do, it fails to catch on and thus we will never succeed’; ‘if nothing changes, I am pessimistic about her’; ‘she is slipping through our hands.’ Such phrases are prompted by the flow of bad lab results, whimsical frequencies on the monitor, medication that barely works if at all, and ultrasounds and X-rays that show no improvement. The attending observes that it is meaningless to go on treating her, while stopping treatment will result in Esther’s death. By removing technology support, such as artificial ventilation, preterm infants die within minutes to hours. If needed, additional medication such as analgesia is provided to prevent suffering after discontinuation of treatment (Abe et al., 2001). However, not in all cases the child dies because of this intervention. In Dutch NICUs, mortality occurs in 74 percent to 82 percent after withdrawal of treatment (de Leeuw et al., 1996). Although each month one member of the team of neonatologists is formally in charge, in Dutch hospitals these kind of difficult decisions are taken collectively by the entire team. The medical staff of the Academic Medical Center NICU consists of neonatologists and if needed other medical specialists, such as a paediatric neurologist, a paediatric cardiologist, and a gynaecologist are involved as well. In the staff meeting all data is re-ordered again. Various methods are used to do so. This ordering process can be described on the basis of the notion of ‘repertoire.’6 A repertoire involves a particular style of reasoning; as such it functions as a guiding principle that orders our ideas about what the world is and how it works. This guiding aspect should not be viewed too strictly, though. Rather than exactly fixing what will be said or done, a repertoire determines what those involved view as relevant, which arguments or strategies, they feel, matter. It outlines what is central or peripheral in a particular situation of condition. It provides a frame for legitimizing decisions. In this NICU context, this notion allows me to demonstrate how the participants in the staff meeting select and arrange data to articulate and determine the fine details of Esther’s situation. Which repertoires structure their thinking?

Newly Born and Indeterminate 37

Academic Medical Center 13.30 hours: The staff meeting begins. Its subject is the proposal to refrain from further treatment in the case of Esther. Seven staff members are present: the attending neonatologist, the resident on duty, two nurses who provide care to Esther, two other neonatologists, and a social worker. The resident opens the meeting with a report on the course of Esther’s condition since her birth and her current condition. She concludes her introduction with the question: ‘In light of the problems and taking into account the child’s immediate future and her very bad prognosis, what is our policy going to be?’ ‘What is “very bad”?’ the social worker wants to know from the resident. ‘A fourth-degree bleeding at both sides of the head makes her chance of survival very small and if she survives at all, what kind of life are we then talking about?’ ‘What do the parents think?’ one of the nurses asks. The resident: ‘If their child is going to be severely handicapped, they do not want us to keep going on.’ The social worker, who also talked to the parents, reports: ‘They do not regret their decision to have Esther treated, but they do have second thoughts if their child will be seriously handicapped. They told me: “It is something with which we want to burden neither the child nor ourselves.” In fact, based on the information they so far received they have already concluded for themselves that they have to reckon with their child’s passing. This is why at any rate they would like to have their child baptized. It might still be done today.’ In the first ten minutes of the meeting all staff members who in the past few days could constantly be found near Esther’s incubator articulated their concerns and points of view. Their stories suggest that they all use a repertoire in which the child’s specific situation is centerstage. These speakers mainly rely on a situational repertoire. In this style of reasoning the uniqueness of a situation is foregrounded, situating the issue involved – the child in this case – in the here and now. Explicitly the child is associated with specific persons, places, and moments, and thus individualized. Esther is a child in the here and now about whom a decision should be taken in the here and now. This repertoire dominates the meeting’s first ten minutes. The speakers

38 Uncertainty in Medical Innovation

address Esther’s situation only and their interaction basically consists of an exchange of questions and replies that all seem to point toward a policy of refraining from further intervention. Next, the discussion takes a turn, when the two neonatologists who are not directly involved in Esther’s treatment join in and challenge all that has been said so far. One of the neonatologists: ‘Has an EEG* been made as well?’ The resident: ‘No EEG was made. If you think it is useful for determining our course of action, we should have one made. But what are the criteria you use when the results come back?’ The neonatologist: ‘I believe that with a normal EEG you cannot make a life-ending decision.’ ‘Do you really think,’ the resident asks, ‘that with this cerebral condition there is a chance of getting a normal EEG? Don’t get me wrong, I only try to explore the boundaries of the new parameters on which to base our policy.’ The other neonatologist supports his colleague’s proposal: ‘I think that having an EEG made is useful. What we are faced with here is the task of trying to determine this child’s prognosis together. That is difficult if you only look at the ultrasound scan. I even wonder whether we are allowed to take such a decision based on a scan. The literature shows that a fourth-degree bleeding itself is not conclusive. The exact location and spread also play a role. There is a study of twenty children with a fourth-degree bleeding of which one is only moderately handicapped. I doubt, therefore, whether the currently available data allow us to start withholding treatment.’ Unlike the actors directly involved, these two neonatologists are not yet fully convinced of the hopelessness of Esther’s situation. They do not only present another opinion, however, but also introduce another style of reasoning: a case repertoire. In this logic the child is specifically considered in relation to other, similar cases. Thus this repertoire reaches beyond the boundaries of the specific situation of the child at hand by inserting the outside world. The meeting no longer only addresses Esther, but also criteria for assessing an EEG result, parameters that help to determine the course of action, and anonymous children that are described as statistical units in the literature. The case repertoire allows one to highlight other aspects of a child’s identity,

Newly Born and Indeterminate 39

thus breaking the dichotomy between the general case and the specific child. Esther too is a case.7 Again, we can see a process of co-production: in the staff discussion not just the child is being defined, but also the local NICU practice. The staff’s own responsibilities are at issue, as well as the reach of the knowledge and expertise of its members. With respect to the forging of this NICU self-identity, the case repertoire and the situational repertoire put other emphases and relate to other matters. In the case repertoire the elements that go into the construction of this specific NICU identity reach beyond the concrete situation. Much more this effort is geared towards general procedures and standards. In the staff meeting this is visible, for instance, when one of the neonatologists proposes to consult a radiologist: ‘I think that, regardless of the case at hand, it is more careful in this kind of decision process to ask for a written report from the radiologist.’ ‘What I sense here is that you need a certified declaration,’ one of the nurses observes, which triggers a somewhat puzzled look from the physician. ‘Speaking for myself I see little need for that,’ the attending cuts in defensively. ‘In this case it is a matter of being legally safe rather than having medical certainty. As for me, I don’t need that certainty.’ This proposal does not focus on Esther’s treatment, but on the procedures regarding this kind of decision. In this situation, this neonatologist believes, the NICU runs up against its own boundaries. This is why he proposes, with reference to the significance of maximal caution, to look for a substantiation of the diagnosis outside of the unit. This links up the unit not just with another medical discipline, but with a very different domain. It is no longer only a matter of medical certainty, but also of legal certainty. Not all present are enthusiastic about this. Yet the question remains whether there is enough expertise in the unit to interpret the ultrasound. Not everyone present believes so: ‘It is safer to have an expert look at it,’ the neonatologist argues once again. The attending disagrees: ‘I feel that meanwhile we are also quite capable of interpreting ultrasound pictures. We have been doing this for years and have more experience than the average radiologist

40 Uncertainty in Medical Innovation

who does it for us. Suppose we hadn’t identified that bleeding on the ultrasound scan and only had to work with the clinical picture of the child being hypotensive. We would have been very worried right now, knowing that from a prognostic angle it is a bad sign. In prognostic terms we have various criteria that together with the fairly objective bleeding are unfavourable. There is the previous history and this child is currently in a stable bad condition. Also because from the very beginning she had a very bad prognosis, I’m willing to go along with the advice of refraining from further treatment – or, at least, to recommend as much to the parents.’ The other neonatologist, however, persists: ‘I believe that in this group there is no 100 percent certainty on the prognosis that would justify a definite advice.’ The attending, slightly irritated, retorts: ‘But our field is not mathematics. I feel you will never get that certainty. There is always someone who on whatever basis will raise doubts. We may do so here as well. But if you add up all aspects you can say with a very large degree of certainty, indeed nearly 100 percent, that the child’s prognosis is extremely bad, if she still has a chance of survival at all. I believe that chance to be minimal right now.’ The legal system outside the hospital affects the process of prognostication in the meeting room. In his proposal, the neonatologist is trying to establish and re-establish a link between the NICU and the outside world. It is a practice, he feels, that has to reckon with external influences such as legal rules and norms involving certainty. The available expertise in the NICU would not be enough to guarantee the required level of certainty for this kind of decision. According to him, the NICU is a unit that in decisions regarding life-ending action has to be linked to other specialties, such as radiology. Here legal rules, that have been developed as an answer to the societal fear that medical innovations will lead to a moral ‘slippery slope’, enter the NICU practice and direct in their own way the decision-making process in the meeting room about Esther. In expressing the need for additional expertise the neonatologist disconnects the policy proposal from the child that is central in the meeting’s discussion. His proposal is ‘independent of this case’ and is geared towards ‘this kind of discussion.’ The attending, by contrast, speaks about his unit in different terms, suggesting that they have become experts as well ‘after all those years’ and that their field is not mathematics. The attending is geared towards the actual NICU practice, all the things that happen within the walls of the unit.

Newly Born and Indeterminate 41

The use of repertoires In this situation the attending and his colleague have divergent views on the definition of child and NICU. The physicians oppose each other and legitimize their positions with the help of divergent repertoires. Given this difference it almost seems impossible to come to a shared decision on Esther’s treatment. The neonatologists, after all, do not speak about the same child, nor do they have the same unit in mind. However, the staff meeting’s continuation will reveal that the gap between the repertoires can be bridged, even if this calls for an active stance on the part of the actors. The attending neonatologist adopts such stance by slightly adapting the proposal of his colleague. When the neonatologist proposes the need for an elaborate written report from a radiologist, the attending reacts surprised: ‘But that would take at least two weeks and we don’t have that time. In this case, it seems to me, a corroboration of the X-ray by one of the staff members will equally suffice.’ In order to prevent the negotiation process from stalling the attending searches for a formulation that would meet the two positions. Thus he addresses his colleague’s request as well as reformulates it in terms of his own position. Such re-conceptualization is a major means to move a decision process forward. The fact is that something has to be done. It is not just the attending who makes this move however. The neonatologist is also looking to find a course in-between as a way of persuading his colleague, without discarding his own conditions. The neonatologist: ‘I could not be so resolute in saying that this child will die.’ The attending: ‘Well, okay, it is more like a clinical hunch that it is bound to happen. But even if the child might survive, it is basically certain that the quality of her life will be extremely poor.’ The neonatologist: ‘Only, that is, if we are not mistaken in our interpretation of the ultrasound scan. Would it be an idea to ask Rob [paediatric neurologist] once again, and find out whether he has some thoughts on the matter? The only thing is that he cannot examine the child because of its sedated condition. But he may have a look at the ultrasound scan and give advice about an EEG’s level of information and whether or not it is sufficient.’ The attending: ‘I wonder if an EEG would help in making this decision. Personally, I would not do it. But if it would give someone

42 Uncertainty in Medical Innovation

else in this group the feeling that we would be acting more carefully, I think we should go ahead.’ Although the neonatologist refuses to change his position on the need for additional certainty from external sources, this proposal is less radical and more tailored to this individual situation. Strikingly, in this reformulation the neonatologist switches from the case repertoire to the situational repertoire. He is not talking about this kind of situation in general, but about that of Esther and her sedated condition. No longer does he refer to the neurology department in general, but to one specific paediatric neurologist – one who apparently everyone is familiar with because mentioning his first name suffices. It suggests that specific repertoires are not tied to specific individuals. If the gap between two people who use different repertoires can be bridged, both repertoires can also be deployed in support of one and the same conclusion, as at one point is done by one of the neonatologists: ‘But I sense that Marije [the third neonatologist] is somewhat fearful of making a decision only on the basis of an ultrasound image. The literature indeed suggests that it is possible to misread such an image. I also remember a child whom everyone called a four whereas at the post-mortem it proved to be a three.8 That surely is a big difference.’ This neonatologist shifts from his colleague’s doubt to studies in the literature, subsequently adding his own personal experiences. Apparently, repertoires are not mutually exclusive, nor does one repertoire have to lead to a single conclusion. The neonatologist uses the case repertoire to conclude, with reference to statistics, that Esther has a 5 percent chance, thus arguing for further examination. He mentions a study of 20 children with a similar condition, 19 of who became seriously disabled and one moderately disabled. One of the nurses subsequently uses the same example to argue for refraining from further treatment: ‘The child is very hard to provide care to. If you touch her, you see the values on the screen drop instantly. At this point the child is steady. Steadily bad, mind you. But if I consider the long term, and the prognosis is as bad as just sketched by Ruud [one of the neonatologists], I have my doubts about continuing treatment.’ To this nurse, the significance of the research results from the literature is not in the single exception (which for the neonatologist is the reason to bring in this research), but in those nineteen other cases. In

Newly Born and Indeterminate 43

her reaction she disconnects the relationship established by the neonatologist between results found in the literature and doubts about ultrasound scans as prognostic tool. The deployment of a specific repertoire, then, merely affects the nature of the reasoning involved, but it does not say anything about the final conclusion. In addition, the meeting suggests that repertoires are not automatically tied to medical roles or disciplines. If during the discussion the physicians do not limit themselves to the case repertoire, nor do the nurses restrict themselves to the situational frame. Step by step the meeting’s participants work toward a common course of action. In this process the role of speaking can be considered as a process of thinking. Speaking is a process of ‘thinking through talk’ (Amann and KnorrCetina, 1989) or in the words of Mattingly (1998) a form of ‘narrative reasoning’. After about 50 minutes there is sufficient consensus on the need to move toward a policy of withholding treatment, and the meeting is wrapped up. 14.30 hours: The attending: ‘Let’s agree that one of the staff members from radiology looks at the ultrasound scan and that the neurologist is asked for his opinion and whether he has something to add to this case. Subsequently we can go to the parents.’ The resident summarizes the decision: ‘If the radiologist arrives at the same conclusion as we did and the neurologist describes the same picture as we did, our advice will be to refrain from further treatment?’ All nod in agreement. 15.00 hours: In the NICU a nurse sits at the table preparing the paperwork. Binders with patient records are opened and closed. Notes and keywords are written on small scraps of paper that subsequently disappear in her uniform’s pocket. Amidst this bustle Esther is baptized. Her mother sits in a wheelchair to the side of the incubator. The father is standing next to her. A priest is saying prayers. The parents are in tears. Esther’s grandparents are watching through a window from the corridor. If normally a baptism ceremony is a festivity, in this setting it marks the beginning of biding farewell. Snapshots are being taken. For later. 15.30 hours: The radiologist corroborates the seriousness of the brain haemorrhage.

44 Uncertainty in Medical Innovation

15.45 hours: The paediatric neurologist enters the unit. He bends over Esther’s incubator. ‘Poor thing! What can I say? … I can’t tell whether the child reacts as she does because she is still so young or because she suffers from a brain haemorrhage. What’s more, I cannot even examine her, for she would not be able to bear it. But just seeing her there I am willing to agree to a discontinuation policy, if only on humane grounds.’ He walks to the table and writes in Esther’s record that he agrees to a discontinuation policy on the basis of the problems following from premature birth, a degree IV brain haemorrhage, constant blood pressure problems, and the multi-organ problem. The list ends with ‘on humane grounds.’ As he explains: ‘They won’t do it only on the basis of humane grounds. Medical reasons have to underpin the decision.’ Although at first the neurologist approaches the situation from the angle of a case repertoire, soon he switches to a situational repertoire. Although I have shown that repertoires are not tied to one person or a single conclusion, this is not to say that the deployment of a specific repertoire is always as effective. Obviously, some reasons are more persuasive or fitting in specific situations than others. The absence of a comparative dimension in a situational repertoire is not necessarily a problem. In this case, a single look into the incubator suffices and humane considerations are decisive: ‘This is enough; this cannot go on any longer.’ While at the side of the incubator this suffices, it turns out that the situational repertoire in the official context of the neurological report does not provide enough support. The neurologist is now forced to switch to the case repertoire. Rather than changing his conclusion, this changes the nature of the argumentation. While ‘humane considerations’, which are decisive in the situational repertoire, are mentioned last, the neurologist’s report first lists medical data in order to justify his decision. However, they do not come from a separate examination because Esther’s extremely vulnerable condition does not permit a physical examination. What is lacking, moreover, is a frame of reference for interpreting the condition of such a premature infant. The neurologist therefore derives data from Esther’s record. This closes the circle: to corroborate the medical data gathered by the neonatologist a neurological consult is requested, which the neurologist also provides, but he does so based on the same data. Even though he thus offers no new information about the child’s condition, his agreeing to the final decision still plays an important role.

Newly Born and Indeterminate 45

As my argument in this chapter underscores, an answer to the question ‘what to do?’ is closely intertwined with the question ‘what is the matter here?’ A reply to this question requires more than just a simple establishing of the facts. In a NICU facts are not so much given, but they are moulded and fashioned on the basis of various materials, such as words, numbers, actions, and things. The staff is constantly geared towards collecting data to determine a child’s situation. Thus the child develops from a ‘case’ into a child with its own identity. Yet more is going on. In the effort needed to do away with the child’s unspecified identity, the NICU itself takes shape as a practice. In this process repertoires are deployed, brushed aside or combined. Their value or use is neither fixed, nor do they coincide with a specific discipline, a systematic treatment trajectory or a specific argument. Academic Medical Center 16.00 hours: The attending communicates the decision to Esther’s parents. The parents feel further treatment not to be useful and give their consent to a policy of withholding. 16.30 hours: Esther is disconnected from the respirator and dies. The nurse frees her of all tubes and lines. The monitor screen goes blank.

3 Co-travellers

The admission of a child to the NICU sets in motion an ordering process that is geared towards the child’s incorporation into the unit’s everyday practice. But rarely, if ever, do children come in alone. In their wake their parents cross the nursery’s threshold as well. This implies that from day one parents too have to be integrated into its everyday life.1 If newly admitted children are usually wheeled into the NICU, often in a hurried fashion, their parents confront a closed door because no one is allowed to enter the ward without first taking specific precautionary measures. The vulnerability of newborns calls for a strictly controlled environment, which basically means that the NICU itself is one giant incubator. The temperature is constantly held at 28 degrees Celsius, the air’s humidity has to be kept up, and hygienic rules need to be strictly adhered to. This is why parents have to pass through an airlock that marks the transition from the cold, contagious outside world to the warm, clean, carefully regulated world inside. Overpressure and hygienic rules should ensure that bacteria are kept out. But the airlock is also a site of social selection. That there are separate airlocks for parents and staff members is indicative and constitutive of the NICU’s order. Thus social distinctions are expressed in its architecture. Only authorized individuals are allowed to go in, notably hospital staff members and parents who have a child in the NICU. Close relatives and friends are allowed to enter the ward only in the parents’ company. As such the airlock for parents not just marks a spatial transition but also an emotional one: it is where parents anticipate seeing their child again. Moreover, it is the site where they first have to perform several mandatory acts: store their coat, bag and jewellery (favourite hiding places for bacteria) in lockers, put on a gown and wash hands 46

Co-travellers 47

and wrists vigorously. Only after attending to these ritual procedures are they allowed to pass through the second door that leads into the NICU. However, the airlock is not watertight. Some bacteria still know how to find their way in. This is why upon re-entering one has to scrub one’s hands again. Thus parents, many of whom have barely recovered from learning that their child is not the healthy baby they were hoping for, enter the world of the NICU. When some time before a child’s delivery it is clear that s/he will need to go to the NICU, parents are normally shown around the unit, so they experience the place as less daunting. Information about the care and the treatment also tends to give parents more hope about their child’s prognosis. Research has shown that such preparation significantly reduces parents’ anxiety (Griffin et al., 1997). In most cases, however, it is not known in advance that a pregnancy will end in the infant’s admission to the NICU. The NICU is first of all a professional domain where staff members routinely perform all sorts of tasks. The incubators are lined up side by side and there is hardly any privacy. Regardless of the number of beds, the NICU is a confined space where the actors are always in each other’s proximity, often spending many hours together. The atmosphere can be relaxed, with nurses who calmly perform tasks, parents who chat with other parents or who are sitting next to their child’s incubator, relieved that their newborn seems to be doing well. But things may also work out quite differently. After all, the context in which staff and parents meet can hardly be labelled everyday. The NICU is a world of state-of-the-art technology and cutting-edge medicine where very ill children regularly and suddenly collapse, which causes everyone to jump up. At once there is tension and a flurry of activity: Will the child make it? Will the baby recover? The limited space makes it impossible for the staff to shield off parents from unforeseen emergencies. The inescapable presence of medical equipment underlines their child’s critical condition and makes parents feel uneasy, certainly when the NICU reality is new to them. A monitor that suddenly starts beeping and blinking often provokes feelings of panic, in part because parents do not know how to interpret such signals correctly. Similarly, staff members are not always relaxed; they may feel irritable because they are fatigued or frustrated over internal affairs, or they feel powerless because a child does not progress despite all their care and effort. Because of the closed-off nature of the NICU space parents are likely to notice such tension right away. Rarely, however, are they familiar beforehand with the many local rules and practices, and this makes it

48 Uncertainty in Medical Innovation

hard for them to interpret such moments of tension. Are they doing anything wrong? Are they standing in the way, or did they perhaps say something stupid? By entering the NICU parents do not just cross the threshold into another climate or atmosphere; they gain access to another world – a social and cultural practice with idiosyncratic customs and traditions. Parents with their individual views and backgrounds and staff members with their various specific rules and rituals meet each other around the incubator. Through their child, parents and staff become part of a shared social order, and they remain co-travellers as long as their child has to stay in this special ward. Since children come and go, the staff has to be flexible and capable of handling a continuous flow of new parents who all require an individual approach. After all parents do not leave behind everything in the airlock; they come in with their individual hopes, histories, and expectations which are fed by stories in the newspaper about the latest medical innovations. Each case is different and this means that parents with all their personal and cultural baggage have to find or be given a role in the everyday course of affairs in the NICU. How do staff members handle this continuously changing social presence of parents? How do doctors and nurses look at them and work with them? The NICUs everyday organization benefits from having parents become familiar with the unit’s culture as soon as possible. How are parents integrated? How do they become part of the unit’s social order? Which image of parents emerges from the manner in which staff members talk about them and how does this affect the way in which the staff deals with them? Put differently: how do ways of speaking about parents contribute to structuring the NICU social order? To address this concern I will explore in the next section various moments in which staff and parents actually discuss their mutual expectations. These allow me to gain insight into how NICU staff members tend to view parents and vice versa. What specific individuals are expected to do in the eyes of the other actors involved is partly tied to assumptions as to who those individuals are. This can be traced more easily by zooming in on situations in which the staff/parents relationship is somehow challenged. I would like to emphasize once more that my analysis mainly addresses such conflict situations rather than standard situations, and that this focus serves a methodological goal. Precisely in conflicts or boundary situations mutual views and preconceptions are explicitly articulated. By describing these moments of tension it becomes possible to identify not only which social order between staff and parents exists in

Co-travellers 49

the unit, but also which opportunities and limitations are available for solving frictions that may emerge during a child’s hospitalization. Let us start with the expectations and those of the staff in particular.

3.1

Mutual expectations

The parent/staff relationship in the NICU is based in part on formal, established routines. For example, parents who first visit the NICU first talk to the nurse who at that moment is responsible for their child’s care and treatment. He or she shows them around the unit, discusses some of its basic rules such as the washing ritual before entering, and explains the tubes and devices to which their child is connected to. Next, the parents soon have a meeting with the attending doctor to discuss their child’s medical situation. Apart from explicit information and formal rules, however, there are many unwritten rules and traditions. These have to do with expectations among staff members on how parents behave and react, how they deal with their child, and how they adapt to the NICUs everyday practice. Of course, parents also have their own ideas about their role during their child’s admission. Although many parents soon relate well to the unit’s culture, it is not uncommon that one of the parties involved feels irritated about some issue, which may lead to lack of understanding, irritation or, in the worst cases, an irreparable breach of confidence. A variety of factors can play a role, including factors that follow directly from the NICU structure, such as the unit’s organization, the regional referral system, and the ongoing rotation of staff. My argument focuses on moments at which one party fails to meet specific expectations. Because expectations mostly remain implicit as long as things go well, my analysis is geared to situations in which specific expectations are not realized. Staff members’ expectations of parents Physicians and nurses have more or less outspoken ideas on how parents are expected to behave in the NICU. Given the instable situation of their patients, the staff expects, for instance, that parents can always be reached (by phone), something that is explicitly mentioned in the unit’s information flyers. They also expect parents to contact them at regular intervals to inform about their child’s condition. This is not to say, though, that there is always time for giving

50 Uncertainty in Medical Innovation

elaborate feedback. As a resident from the Academic Medical Center explains: ‘The parents are allowed to come in the entire day. This is fine. But they also need attention from us all the time, and this is difficult for us in the morning, when we are very busy and have to check all children. The rest of the day we basically work alone, perform medical acts such as puncturing, and have meetings and arrange all sorts of things. Parents may ask for your attention when you are in the middle of performing a medical intervention. … Take, for instance, Dan’s parents. They are so pushy and want you to be available for them instantly. As if we have no other children. They are also very priggish, thinking they know it all. They are just difficult to deal with. Now they want to take their child home. That this is impossible they do not seem to realize.’ This physician is quite willing to talk to parents, but only when he has time. Parents who expect him to interrupt his medical work annoy him. Parents who disturb the NICU practice are seen as troublesome. In particular this doctor feels bothered by parents who think they know better; he rather expects them to acknowledge his medical expertise. Doctors and nurses do not only have expectations about parents’ behaviour vis-à-vis staff. Parents are also expected to be involved in what happens to their child in the NICU, and this implies more than occasional phone contact. If staff expects parents to visit regularly, this does not always happens. As one nurse from the Academic Medical Center explains: ‘We once had a father here who made more pictures of the child in the incubator next to that of his own child. And his baby was hardly two days old. “Well, he is such a sweet baby,” he responded. This is just ridiculous …. Or take the parents of Benny. His mother doesn’t say anything and never touches her child, and his father only smiles at us. They are a puzzle to all of us. No one gets through to them and they did not want to have anything to do with our social worker. These parents do not seem capable at all. When you explain something, they hardly listen. The mother once held her baby because John [a nurse] put it into her arms. But she did not really know how to handle it. We do not have their phone number and therefore we cannot call them if something goes wrong. To gain

Co-travellers 51

more information about them we tried to contact their general practitioner, but this person is nonexistent. The insurance company they listed never heard of these people. We suspect the father to be an illegal refugee. I already worry about the time when this child is ready to go home.’ Although these parents visit their child each day, they fail to meet this nurse’s expectations. They show not enough interest, she feels, nor are they cooperative. Another example of expectations among staff involves a reaction by the parents of Anne. She proves to have a serious heart condition and right after her birth she is transferred to the Academic Medical Center for surgery. The surgery is only partly successful and after some weeks Anne is completely exhausted. One day she is in critical condition and her parents are called. They refuse to come, however, because they have to go to a party. The staff is shocked and angry. When a few days later Anne collapses again, her parents cannot be reached because they are attending a theatre show. The doctor who calls them the next day with a request for a meeting only manages to convince them of the seriousness of the situation after much effort. When several days later Anne’s heart frequency sharply drops it is clear to everyone that she will die that same night. Again her parents are contacted, but they again refuse to come because it is in the middle of the night. The doctor keeps calling them, however. Only after three phone conversations the parents decide to go and they arrive just before their child died. Staff members tend to be perplexed and outraged about such conduct, and parents who do not meet their expectations are soon labelled in various ways. For instance, parents who cry during each visit are incompetent; parents who feel good about their having found a way of dealing with their ill child do not understand the seriousness of the situation; parents who on religious grounds refuse treatment are unreasonable; parents who show not a single emotion have a psychological problem; parents who are mad, frustrated, or desperate have a mental problem; well-educated parents think they know it all and parents with hardly any education do not grasp what they are told (Guillemin and Holmstrom, 1986). Such quick conclusions and judgements may of course negatively affect the quality of the parent/staff relationship. Interestingly, behaviour of parents from lower classes is judged more mildly than the same behaviour of parents from society’s higher echelons (Heimer and Staffen, 1995).

52 Uncertainty in Medical Innovation

Parents’ expectations of staff Apart from staff members who have specific expectations about parents’ behaviour, parents also have or develop a sense of their own role in the unit and this affects the expectations they have of staff members’ behaviour. Most parents are fully aware of their baby’s critical condition and they expect the staff to keep them informed about it, whereby interactions with the attending physician are seen as relevant in particular by parents. Staff members underline the significance of good communications with parents. This is why parents are allowed to call anytime to ask how their child is doing, even in the middle of the night. In practice there are two peak hours for such phone calls, however: right before parents go to bed and right after they rise in the morning. Parents’ expectation that they will be informed can become a source of tension when they feel excluded from the decision process. Anspach (1987) refers to the limitations of phone contact, such as no privacy, the absence of non-verbal signals and that only two people actively join the conversation. Given the significance and emotional side of many decisions, a phone call may fall short in respect to providing proper care to the parents. It may also lead to misunderstandings that staff does not immediately notice. Although the staff usually keeps parents posted, this does not mean that the amount of information supplied is enough for everyone. One of the parents in the Academic Medical Center complains about a lack of information: ‘They hardly tell you anything here. That’s wrong, isn’t it? He is our child after all.’2 Parents do not judge staff members on the amount of information they get, but also on how they are informed. They may take issue, for instance, with the choice of words, a paternalist tone, or medical jargon. However, what one parent considers honest, another may view as rude. In the nursery of Northeastern a parental couple files a complaint about one of the residents. The mother: ‘He first gave us a lot of medical information and then he suddenly said: “But we are not sure whether your child will make it.” This phrase was the only thing that stuck in my mind afterward. It was so rude of him to say this. As if we didn’t know. Of course, you reckon with something happening. It is floating around somewhere in the back of your mind all the time. But there is no need for such a snotty boy to remind you of this. We immediately went to the attending to complain. Not just for

Co-travellers 53

us, but also for all parents in our wake. You simply don’t say such things.’ Aside from expectations about how staff members deal with them, parents may also have preconceptions on how staff ought to deal with their child. Many parents claim to hope that their child is treated by an experienced professional who knows the child well and with whom their baby will feel at ease. Occasionally parents worry that their child does not receive enough attention. The parents of Lisa, for instance, were dissatisfied with the amount of attention she received. Lisa was admitted to the Academic Medical Center because of a cardiac abnormality. As there was no bed available they first put her in the cardiac critical care unit, where she received full attention from one nurse. Next, she was transferred to the NICU, where one nurse is in charge of two patients. The parents were bothered by this course of affairs and accused the staff of irresponsible behaviour. These parents, according to the NICU nurses, are too demanding and overreact. They also feel that their professional honour is violated. This mutual lack of respect seriously damaged the relationship between the staff and Lisa’s parents. Parents’ preconceptions about the expertise of the medical profession at large may also give rise to tensions. A nurse in the Academic Medical Center: ‘You have parents who would rather have a premature baby because they look so sweet and receive so much attention from us. They have no notion whatsoever what is really going on. Nor do they worry. They just say: “They can do anything here. You just put a baby into an incubator and he will start growing.” These people do not know what they are saying.’ A physician from Northeastern General Hospital comments in a similar fashion: ‘The problem with many parents is that they do not comprehend how difficult it is to know what is going on or how the child’s condition will develop. They come here and are so impressed by all the hi-tech that they think we can do everything. They think that we detect more and hence can also do more. The influence of television is enormous. They watch series in which in a one-hour episode everything is solved. That the real world is different they find hard

54 Uncertainty in Medical Innovation

to believe. They cannot fathom that we have test results come in only after days and that in some cases we may still not know what exactly is going on. They see you as a bad doctor and ask for another one. They are deceived by the hi-tech surroundings in which they find their child. As if we can accomplish all we want. Sometimes you need to go on because parents want you to, regardless of the child’s hopeless situation. The law prohibits us to refuse it. But when parents want us to stop merely because their child is less than perfect, we go on and they simply have to accept it.’ The current pace of medical innovations and the way it is presented in newspapers and on television creates expectations to which medical reality cannot live up to. Today’s medical soaps, this physician argues, create an image of doctors who always know how to make a proper diagnosis and what to do to save a child from certain death. The inevitable happy ending underlines the notion that these days the medical profession is capable of solving everything. In this image the physician is seen as all-powerful and this of course is bound to be uncovered as an illusion. When a doctor cannot say what a child is suffering from, some parents insist that another doctor takes over. In a rare case they want to take their child home with them, for if no one can tell them what is wrong with their child they assume that nothing is wrong. Various factors thus provide a breeding ground for unjustified expectations and a perplexed, annoyed, or outraged response. All sorts of tensions may drive a wedge between staff and parents. In the worst case there is a fundamental breach of confidence. A breach can occur in various ways. For instance, parents may view their child’s treatment as undesirable and refuse to give permission. If the staff deems this decision unwarranted and decides to treat the child after all, parents can (temporarily) be divested of parental authority. In another scenario the staff/parents relationship is broken off when parents take their child to another hospital. It also happens that parents leave their baby behind in hospital. This happened in the Northeastern General to a child born after a term of 27 weeks with various disorders. One of the nurses subsequently adopted this child. This rarely occurs, but when it does it is for both parties a dramatic event that often lingers on in the ward’s collective memory. The above-described examples clarify how mutual expectations are entwined with the child’s specific situation. The special nature of the social interactions in the NICU colours not only the mutual expectations of those involved, but also gives meaning to who they are. If the

Co-travellers 55

professional identity of doctors and nurses is constructed in part through their relation to the child, the parents’ role is determined fully by their relation to their child. Still, there are different ways how in the NICU context the relationship between parent and child is defined, and thereby various vocabularies are at stake. What vocabularies are involved when staff and parents refer to each other? Vocabularies of expectations As an analytical category, the notion of ‘vocabulary’ makes it possible to identify and label various styles of social ordering.3 The use of divergent vocabularies implies more than just a difference in phrasing. The same words may have other meanings in various vocabularies. Vocabularies have a collective nature. The meaning of words is not determined by the objects to which they refer but depends on the specific ‘language game’ in which they are used (Wittgenstein, 1958). Put differently: the rules of a particular language game fix the meaning of words. By applying a specific vocabulary a reality is created in which people, things, facts, values, and actions obtain a role and meaning of their own. Thus language does not so much refer to an already existing reality; rather, a reality takes shape as language is being used. In this way I study vocabularies from a pragmatic angle. From such angle, language is viewed as an element of the action context (Goodman, 1978). In other words, my focus on vocabularies is concerned not so much with whether articulated words correctly represent reality or not, but with how they construct a specific reality. We can define the vocabularies in a NICU in more detail by analysing examples of expectations of doctors and nurses vis-à-vis parents. I discuss two examples: the expectation that parents visit their child in the NICU and that they can always be reached. In the first case the relationship between parent and child is defined in terms of emotional bonding. Or, as one nurse from the Northeastern puts it: ‘We, nurses, apply a holistic perspective to baby and parents. They belong together. They are a single whole.’ Together with their child, parents constitute a single unit of care and attention. Parents too are seen as object of care and this implies that the medical staff invokes a psychosocial vocabulary. The expectation that parents can always be reached mainly applies to information exchange and decisions. It suggests some sort of contract: we (staff) will keep you (parents) posted if you make sure that we can also reach you. Put differently, the right to information comes with the duty to be reachable. The notion that parents are kept

56 Uncertainty in Medical Innovation

informed and need to be involved in the decision-making process in part defines the parent/child relationship. The medical staff may provide care to the child, yet its parents are still responsible. In this vocabulary parents are couched as bearers of parental authority, involving particular rights and duties. This way of speaking about parents draws on a juridical vocabulary. There are, it appears, various ways in which staff members talk about parents. Ways of speaking about parents are not static because they are linked up with changes in the child’s condition and other changes. The juridical vocabulary tends to surface more when actors believe (further) treatment to be either undesirable or without prospect of recovery. If the reality construed on the basis of the psychosocial vocabulary highlights the wellness of both child and parents, the juridical vocabulary establishes a NICU order that is marked by formal decisions and negotiations, the various actors’ rights and duties, and the filling out of official documents. The two vocabularies not only outline the limitations and opportunities for parents and staff in their role as co-travellers; they also give rise to quite different conceptualizations of the social role of parents in the NICU order. It should be pointed out that parents are also spoken about in other vocabularies. Consider, for example, an economic vocabulary, which highlights parents as health care system consumer, or a vocabulary which prioritizes the masculinity or femininity of parents. Therefore Michael (1996) argues for studying the origin of vocabularies: ‘Often the impression one is left with is that these constitutive resources float around in some linguistic or discursive or cultural ether. The point is that these are specifically embodied in particular textual and representational forms, realized in specific and particular social contexts or patterns of social contexts’ (p. 12). Vocabularies are coined in specific disciplinary domains. In the use of different vocabularies similar repertoires can be deployed. In a psychosocial vocabulary one may rely on a case repertoire as well as a situational repertoire. In a juridical vocabulary one may also rely on both repertoires. Vocabularies and repertoires, then, have a matrix relationship: in each vocabulary (psychosocial or juridical) both repertoires (situational and case) may be deployed as a way to organize the practice. This is not to suggest that only repertoires actively contribute to ordering the actual practice. Vocabularies give shape to the actual practice in which they are used as well. In the remainder of this chapter I will limit myself to the psychosocial and juridical vocabulary and further explore their role in ordering the NICU practice. Let me start

Co-travellers 57

with the psychosocial vocabulary which conceptualizes parents as objects of care.

3.2

Parents as object of care

A reality built on a psychosocial vocabulary prioritizes the wellness of parent and child. A major element thereby is of course the child’s hospitalization. The ensuing separation of parent and child disturbs their normal relationship.4 To optimize the child’s chances of survival s/he is usually transferred to the neonatology department right after the delivery, if at least the hospital involved has such a tertiary care unit*. As a result of lack of beds or understaffing it can happen that mother and child do not stay in the same hospital or that twins cannot be admitted to the same NICU. Although the hospitals involved will try to reunite mother and child as soon as possible, this may still take a few days, for instance because their physical condition does not allow them to be transported. An illustrative case is Robert, with whom right after birth in a Dutch hospital a serious heart condition is detected. Instant intervention is called for, but the hospital has no NICU. As his mother recounts: ‘I was so happy when Robert was born. The doctor said: “Congratulations, you have a healthy son”, but then she cut the umbilical cord and we saw Robert fade just like that. He turned blue all over. They immediately took him away from me. I hardly understood what was happening. I was wheeled to the maternity ward and assumed that Robert would be joining me soon. I was in fact already eating Rusk with aniseed comfits when they told me that he was no longer in the same hospital and already on his way to the Academic Medical Center. My husband and father-in-law followed the ambulance, but it was driving so fast that they couldn’t keep up with it. And I knew nothing. I didn’t know at all that Robert was in such bad shape. It frightened me enormously when they told me he had a severe cardiac abnormality. I just couldn’t believe it. Earlier they had said his heart was fine. They saw it on the ultrasound scan during pregnancy; fortunately I didn’t know it, for it would have been stressful. We’ve been lucky they responded so promptly.’ An unforeseen admission to the NICU of another hospital has a major emotional drawback because usually parents arrive later and then they do not know who their child is. A nurse or physician will have to tell

58 Uncertainty in Medical Innovation

them this. The specific NICU circumstances complicate their bonding with their child. Frequently they will feel as if their child is assigned to them. This happened to Esther’s parents, as her mother explains: ‘After we had decided to have Esther born via a Caesarean section I was administered anaesthetics right away. Twenty minutes later Esther was born. She was brought to the NICU immediately. I did not see her until hours later. They took me there, with bed and all. But I must say that seeing her that first time did little. You have to imagine that they bring you to an incubator where they tell you: “This is your baby.” Sure, you have to start somewhere, but I felt no bonding whatsoever. She was so tiny and I could not see her very well because my bed could not be put near enough to the incubator. You must not forget that she was only a twenty-seven week term. But the second time I saw her something broke inside of me. I was just able to put my hand into the incubator and carefully moved my finger toward her. Esther instantly tightened her tiny hand around. And later she very briefly opened her eyes. It really does something to you.’ It is hard for the mother to imagine that the foetus in her womb and the infant in the incubator are one and the same child. Because she bore her baby for months, she has to get used to the new situation. In this case the reaction of the child clearly contributes to the mother’s sense of contact. Because the infant’s frail condition renders the treatment’s course highly unpredictable, parents often fear a bad outcome. This fear may prevent them from bonding with their child. Feelings of disappointment about the pregnancy’s conclusion or grief and feelings of guilt about their not having a healthy child complicate first contacts with their child (de Leeuw and Bakker, 1995). Not each child, moreover, is capable of responding to the presence and affection of the parent; the baby is perhaps too ill or sedated to react. Rituals such as a baptism may contribute to breaking the detachment. Esther’s father: ‘Esther’s condition grew worse all the time. You saw her fighting but then she got that bleeding in her brain. We had her baptized that same afternoon. It was very difficult because at once she became much more “Esther”, more an individual.’ If this ritual gives her more of a place of her own within the family (history), the child’s inevitable hospitalization caused a disturbance of

Co-travellers 59

the natural bonding between parent and child. Because parents were separated from their child and did not get the healthy baby they hoped for, they suffer in fact a double loss. This triggers a mourning process that may include a variety of reactions including anger, disbelief, bewilderment, emotional withdrawal, and resignation. A parent with a child in the NICU is a parent who needs care and support. Normally the staff is very aware of the difficult position of parents, but staff members are not always able to provide sufficient attention to them. An intern’s frustration, when he claims that you do not ‘just go talk to the parents when you are in the middle of something,’ can also be seen as an example of the fact that mounting job pressure may easily go at the expense of the quality of the contact with parents. Nurses do not differ in this from physicians of course. As one nurse from Northeastern explains: ‘We have to perform an increasing number of medical tasks, such as putting in an IV-line or taking blood arterially. We do it all today. A doctor will only stare at you when you ask him to do it. Only residents are occasionally willing to do it, for thus they can practice. But all this causes us to be busier all the time, and parents are the victim. It is very frustrating when you notice that you have less and less time for parents and you are increasingly curt with them. It is not a good thing.’ This nurse refers to parents and children as constituting a single unit of care. Unlike the child, parents, in the psychosocial vocabulary, have no medical problem but a psychosocial problem that calls for a psychosocial solution. They are considered ‘second-order patients,’ so to speak (Guillemin and Holmstrom, 1986, p. 175). Apart from treating newborns, the support of parents plays a major role in a NICU. Nurses in particular are in charge of providing this support, and their specific discourse on parents contributes to structuring the unit’s social order. They instruct them about their child’s treatment, show them around the NICU, inform them in everyday language about the child’s condition, and function as their help and stay. If physicians have authority over a child’s condition, nurses deal with the parents. In this respect, nurses act as ‘emotional managers’, a work force that functions to evaluate and manage parents’ attachment to their babies’ (Anspach, 1993, p. 144). The psychosocial vocabulary may contribute to parents’ medicalization. In this vocabulary, their conduct will be understood as confused

60 Uncertainty in Medical Innovation

rather than as misbehaving; parents are in a stage of denial rather than recalcitrant; they are not aggressive but afraid. By speaking of parents in this way, guilt and responsibility is situated outside of the parent. Conversely, the staff is more tolerant of abnormal parental conduct. Even parents who do not visit their child are not immediately rejected by those who embrace this vocabulary. Within a psychosocial vocabulary this is understood as ‘coping through distance’ (Anspach, 1993, p. 112). Parents stay home or seem detached, not because of a lack of interest but out of a fear for loss. A psychosocial vocabulary tends to view parents’ social background as much a factor as their emotions. In this respect, there is substantial difference between the parent population in the NICU of the Northeastern General and that of the NICU in the Academic Medical Center. In this American hospital, many parents proved to be either single mothers, very young (teenagers), hardly educated, addicted, or living below the poverty line. Carson (1996) points to a rise of ‘socially complex families’ in American NICUs on account of drug use, teenage parenthood, illiteracy, poverty, AIDS, and aggression. However, in most NICUs the organization of the interaction with and the care for parents is based on ‘the competent parent’, who understands information, takes part in the decision process, and is capable of bearing the responsibility for the child after its discharge. The psychosocial problems that are part of complex family situations may confront staff with new dilemmas and problems. Therefore, Carson argues for a larger role of the social worker in the care and aftercare aimed at parents.5 If the medical staff is confronted with a large diversity in social cultural backgrounds and a multitude of specific views and customs among parents, this can easily give rise to misunderstandings. To prevent these, the staff tries to assess the parents’ social background and select an appropriate interaction style, taking into account issues such as class, ethnicity, and level of education. The influence of social background, however, involves more than just communication. The social stability of parents sometimes also influence the timing of a child’s discharge or a decision on withholding treatment. The deployment of a psychosocial vocabulary, however, is not always fitting and may lead to frictions. Speaking of parents in psychosocial language may enlarge the tolerance of staff vis-à-vis parents when indeed such vocabulary agrees well with the individuals involved. But there are also parents who would rather not view themselves as object of care; they experience well-intended care as belittling or discriminatory. Staff members who approach parents psychosocially assume a certain level of

Co-travellers 61

inequality in their mutual relationship. As a consequence, parents may experience this vocabulary as paternalist. The reverse situation can lead to frictions as well. When parents talk about themselves in a psychosocial vocabulary, to some degree the staff members need to go along with them. But this does not always happen. The mother in Northeastern, for example, does not feel her mental needs recognized when the doctor claims to be unsure about whether her child will make it. With reference to the overall mental burden involved, the mother qualifies the physician’s statement as unsuitable and files a complaint. This also shows that a psychosocial vocabulary does not always suggest concern and understanding. It is also deployed to express anger and frustration. In some cases staff members are irritated about parents who hardly visit their child or because they are excessively focused on their own traumatic experience. Such attitude can cause the staff to view them as egotistic or cold-hearted. A psychosocial terminology, it appears, does not automatically go together with compassion and tolerance. Another ambiguous aspect of this vocabulary is that staff may also deploy it to avert criticism. Negative comments from parents on the staff’s functioning can be neutralized and dismissed as expressions of emotional imbalance. Displeased parents, then, are not seen as angry but as confused. Yet this may give parents the feeling that they are not taken seriously, which in turn may cause them to voice their concerns even louder. Similarly, incorrect assessment of the parents’ social background can create problems. There is always the risk that parents feel left out or that staff does not try hard enough to engage them in their child’s treatment and the decision processes involved.6 Aside from situations in which the psychosocial vocabulary itself generates tensions, this vocabulary contributes to a social order in which providing care is center-stage. The positive influence of this vocabulary on the quality of the parents/staff interactions generally rests on the actors’ inclination to anticipate and feel for the other party’s behavior. At the same time, such attitude makes it easier to solve problems in their relationship. The juridical vocabulary, as will be explained in the next section, goes one step further: it is primarily geared to a maximally clear demarcation of the roles, so as to prevent problems from occurring.

3.3

Parents as bearers of parental authority

A juridical vocabulary assumes quite another NICU order, geared to maximal clarity in the demarcation of roles so as to prevent problems

62 Uncertainty in Medical Innovation

from occurring in the first place. It pertains to actors’ rights and duties, making decisions, deliberations and negotiations, and filling out official documents. In this vocabulary the active, intellectual input of parents is given much weight. They are first and foremost bearers of parental authority and responsibility. Instead of being dependent or in need of help, they are autonomous, informed partners in the decision process. The juridical vocabulary is geared to organizing the social interactions between staff and parents transparently. A major notion in this respect is ‘informed consent,’ implying that parents need to give their consent for specific interventions, such as those with higher risk levels. To make those decisions, parents have to be informed by the medical staff. The ‘informed consent’ rule forces staff members to explain clearly to parents the aim, nature, and consequences of a planned intervention. Based on this information they can then make a decision. This principle ensures that parents have a say in the decision process. Various studies have shown that this regulation works differently in actual practice (Pols and Mol, 1994; Schlomann, 1994). Although at first sight ‘informed consent’ appears to grant parents a clear role in the decision process, this does not yet mean that each parent indeed considers himself a party in the decision process. A study of parents of children in the NICU in the US indicated that although parents feel basically informed about decisions, they do not feel that they are being involved in the decision process (Raines, 1996). This does not yet mean, though, that each parent regrets this situation. Some indicate to be glad that others perform this task. If in the psychosocial vocabulary nurses play a mediating role in staff/parents interactions, within the order of the juridical vocabulary physicians take on this task. This becomes obvious, for example, when doubts arise about the continuation of a treatment. Before a decision is made to stop treatment, there is interaction between the attending physician and the parents. Although such a decision is the doctor’s responsibility, the parents must be involved in it. Without their consent doctors are formally not allowed to stop treatment. A study of the input of parents in the decision process on withholding treatment in the Netherlands shows that Dutch parents are highly involved in that process. In most cases by far the decision to discontinue treatment was based on consensus between parents and doctors. Dissension played a role in a single case. In situations in which parents favoured continued treatment this preference was respected. In a few exceptional cases dissension arose after doctors refused to honour the

Co-travellers 63

request of parents to stop treating their child. In these cases the medical staff’s view prevailed and treatment went on. The authors conclude that in disputes the position of the parents proved decisive, if at least they favoured to go on treating (van der Heide et al., 1998). In the United States conflicts between patients or their relatives and hospital staff frequently end up in the legal arena. There are several reasons for this. For instance, patients, in this case the parents of children in the NICU, are approached by lawyers – so-called ‘ambulancechasers’ – with the offer to help them in a legal case against a physician or hospital. Because in the US lawyers work on a ‘no cure, no pay’ basis, patients’ financial risks are limited. This is why in the Northeastern General a juridical vocabulary occupies a more prominent place than in the Academic Medical Center. Even if this vocabulary encourages formal staff/parents interactions, it does not guarantee consensus about decisions. For example, parents who disagree with a doctor’s proposal to stop treatment may ask the staff to go on treating their child. Because parents are assigned a prominent role in the decision process, the staff will normally do so. An experienced nurse from Northeastern dwells on the dilemmas involved: ‘Today we no longer look at weight and weeks. Each child is simply treated, regardless of its size and number of weeks. There are parents who want us to go on while it is clear to all of us that the child will die. The law prohibits us to let a baby die. A child may be turning blue already, but parents deny this and refuse to sign a DNR-form*. In those situations, I basically pull out emotionally and merely do what they want me to do. For at that point I am not really helping the child any longer. Thus I switch over and direct all my attention to the parents. I do it for them and try to help them. This is the only way to handle it. Much has changed. Earlier we did unhook children from the respirator. ‘Let it go,’ the motto was. But this is no longer possible on account of the legislation. It is meant as protection, but it is rather a weapon, a danger to the children. You have to go on. Nowadays these children hardly get a chance to die.’ The juridical vocabulary in the nursery of the Northeastern General Hospital is much more tied to formal regulations than that in the Academic Medical Center. The juridical concern in Northeastern is in part the result of a number of cases in the US in which it was decided to discontinue treatment. As have been described in the first chapter of this

64 Uncertainty in Medical Innovation

book the famous Baby Doe rules were intended to protect the medical decision process and keep doctors on ‘the proper track’. The wider effect was that medical action became more solidly anchored in juridical provisions. This potential assessment of specific treatments in legal contexts has increasingly turned medical intervention into a form of strategic action, with fewer options in terms of non-intervention or withdrawal of treatment in particular. In this context, the notion of ‘defensive medicine’ is used. In such cases, medical intervention is primarily geared towards solidifying one’s juridical position rather than being motivated by the child’s medical condition. This does not automatically mean that such acting is not in line with medical standards. It is not uncommon, however, that juridical rules lay down the various NICU decision practices and procedures in ambiguous terms. The juridical vocabulary in which these rules are couched is far from watertight. What happens, for instance, when parents ask to stop their child’s treatment on grounds considered insufficient by the medical staff? A nurse from Northeastern explains: ‘In practice very few children get a Do-Not-Resuscitate code*. Treatment will be stopped only when the child is not viable. The rights of parents come first here, but when they want doctors to stop because their child is less than perfect and the doctor sees no ground, we go on treating and the parents simply have to accept it. Fortunately this happens only sporadically.’ This nurse refers to the possibility of taking away the decision authority from parents. This is legally possible in cases in which parents are no longer seen as party in the discussion, but as a party against whom the child has to be protected. Specific circumstances can thus put the power to decide into other hands, but it is not always clear in whose hands this power ends up. It is not uncommon for those involved to have very different ideas about the locus of this decisional authority. Even if in normal circumstances all rely on the same vocabulary, this not yet guarantees that there is agreement on who has which rights and duties or on where exactly the limits of decision authority and responsibility are. The consternation surrounding the treatment of Maureen in the Northeastern General offers a case in point. Maureen is born after a term of 23 weeks. Tests show that she suffers from a fourth-degree brain haemorrhage on both sides of her head. But who, according to those involved,

Co-travellers 65

decides what to do? Is it the law who decides, or the parents? Or is it the doctor? According to the nurse who takes care of Maureen right after her admission on the NICU, it is the doctor. It is the doctor who decides Northeastern General Hospital 11.00 hours: Maureen is born some hours ago. Now her heart rate is repeatedly dropping below the critical level. The nurse who is constantly stimulating the infant’s heart is hesitant about her intervention: ‘This sort of children makes me go home frustrated tonight. It has no use treating this child. Still, the doctors keep going on. They feel safer after they have done everything. Moreover, it is rather heroic to be able to say that you are treating 23 weekers here. I hope it will be over soon. You put a lot of energy into this child while she has a head full of blood. Emotionally I do not support this, so you understand how I feel when I get home tonight. This is such a sad child. Wherever I touch her, the skin becomes rash. She only spent 23 weeks in her mother’s womb. She should not have been here yet. Do you find it strange that she constantly tries to sneak away? That seems to be the only healthy reaction the child still manages to have. Children feel it, you know. They know when they should withdraw. Each time she tries hard to die. But I constantly have to intervene, for she is not yet given a DNR-code. But I suspect that later on Frank [the neonatologist in attendance] will talk to the parents and that then it will soon be over. I know Frank. He does not negotiate with parents. He tells them that it is over and that they still have an opportunity to bid their child farewell. I think this is right, because some parents only want you to keep going.’ 11.30 hours: The attending neonatologist on his medical round. ‘How are things going?’ The nurse reports on Maureen her ongoing bradycardic* situation. The attending checks the blood tests in detail as well as the data from the monitor. His face clears up: ‘The saturations are not all that bad, given her condition. I feel this may not be a matter of regular cardiac arrests at all because her saturations are too high. Something else must be going on here. It looks more like a heart problem than a poor condition. Hook her to another monitor, so we may get other information. We have to pursue this.’ The nurse

66 Uncertainty in Medical Innovation

is perplexed, but does not say anything. The attending turns around and walks over to the next incubator. The nurses involved feel that this child has no chance of survival and that treatment must be stopped. The attending, however, indicates that the brain haemorrhage as such is not a lethal affliction. Moreover, he claims, it is not possible to be sure about whether this child will be handicapped because in a number of cases the outcome proved less serious than anticipated. This means that in practice one cannot say anything about the outcome of a fourth-degree haemorrhage. This is why there is neither a conversation with the parents, nor a discontinuation of the treatment. On the contrary, the treatment is intensified. As the day progresses, the condition of Maureen seriously weakens and the attending is constantly at her incubator. He balloons, has blood tests performed, and keeps his eyes on the monitor. Maureen has one cardiac arrest after the other, but each time the attending manages to get her back. The nurses, who observe the attending in his interventions, believe this child is trying hard to die, but that he does not let her go. After all, it is the doctor who decides. But is this the case? For some it is not the doctor but Maureen’s parents who decide. It is the parents who decide Other physicians and nurses talk about Maureen as well. But they locate the authority to decide elsewhere: with the parents. The parents, they feel, determine the course of their daughter’s treatment. This is why the resident, who shares in the responsibility for Maureen’s treatment, defends the attending: ‘Several times already he tried to talk with the parents about stopping the treatment, but they refuse to discuss the issue. They left the NICU and never returned. They do not come back because they know that they will be asked to sign a DNR form. As long as the child has no official DNR-code, we just have to go on. Yet these people have such unrealistic image of what the situation is. They think they will always have such a lovely small child. But it is clear that they don’t understand it at all. They talked about a small bleeding in the head and that their child does not have to be a whiz kid. If later on she will hop around cheerfully they are satisfied as well. As if there will be any hopping around in this case.’ For this resident the authority to decide does not lie with the attending, but with the parents. She feels that the attending does not have

Co-travellers 67

the ultimate say, and therefore he is not the one to blame. It is the parents of Maureen who are responsible for this situation, she believes, but they are unrealistic and uncooperative. They are not prepared to talk, let alone sign, and thus it is impossible for the staff, including Frank, the attending neonatologist, to stop Maureen’s treatment. However, another resident sides with the parents: ‘I understand these parents very well. I also believe that everything should be done to save this life. You are not allowed to break off a life. It’s murder. This is why I cannot imagine that someone even considers breaking off a pregnancy. You cannot do that.’ Although the two residents judge differently about what is to be the proper decision, they both locate the authority to decide with Maureen’s parents. Some of the nurses question this. As one of them puts it: ‘I heard that Frank [the attending] would have told the parents that their child has a minor bleeding and that there is still a 50 percent chance that it will grow up as a normal child. It is understandable that, given the positive image painted by Frank, they insist that everything be done to save the child. From that corner, then, we have little to expect as well.’ For this nurse Maureen’s parents are not unrealistic or in denial or out to protect newborn lives, but they are merely wrongly informed if not misled. These parents can therefore not be held responsible for the treatment’s continuation. In fact, this responsibility, according to this nurse, does lie with the attending. He is the one who goes on treating the child and tells the parents that things are not that bad. However, the resident also referred to the DNR-code as a reason why they were not able to stop. So, is it eventually the law that decides what will happen on a NICU practice? It is the law that decides The first resident also tones down the physician’s omnipotence by pointing to the consequences of the DNR regulation: ‘As long as the child has no official DNR-code, we just have to go on.’ She feels it is the juridical structures that make it impossible to withhold treatment, thus situating the decision authority in the law. In the Northeastern General Hospital the juridical vocabulary’s structure forces physicians to perform specific actions. For example, the

68 Uncertainty in Medical Innovation

attending has to ask parents of a child whose treatment will be stopped to sign a DNR form. The attending who at the beginning of the new month will take over her colleague’s shift feels that in fact you should not pose this question to parents: ‘I want to save parents the misery of having to sign a DNR. It is as if you ask them to sign their own child’s death verdict. … I feel you cannot do that to a parent. For this reason I hate the DNR regulation. It is a good regulation for physicians and I am glad with it. But I find it inhumane to ask parents to put down their signature. It seems so simple, but it is so symbolic.’ It is the law, rather than doctors or parents, which is made responsible for the child’s situation. Juridical rules make up a solid network here, with a bias in favour of treatment because signatures are needed for ending a treatment. The DNR form proves to be an actor with a logic of its own that in this case complicates withholding treatment. The DNR form and its practice also complicate the contact with parents. At first sight a DNR form seems merely a tool that regulates how to act when it comes to discontinuing the treatment of newborns. In this respect staff and parents have to follow a specific procedure in which the parental signature on the DNR form counts as a crucial step. Thus they very concretely formalize the decision not to continue their child’s treatment, while also knowing that they are guaranteed a formal say in the decision process. But for Maureen’s parents this parental authority, rendered concrete in the signature on the DNR form, seems an impediment for visiting their child rather than a validation of their having a say. When the first attending indicates he wants to see them, they leave the nursery in order not to return. After all, the doctor could confront them with the DNR form. In this case the DNR regulation, a means to protect the exercise of parental authority, contributes to a lack of parental autonomy and increases the level of mistrust between parents and medical staff. Although from the angle of the doctors Maureen’s parents’ oral consent suffices to abstain from resuscitation, the nurses continue to be formally required to intervene as long as there is no parental signature on the DNR form. This is why the nursing staff is concerned if not outraged when at the start of the new month the new attending does decide not to resuscitate Maureen anymore,

Co-travellers 69

even if only on the basis of her parents’ consent by phone. This implies that from a juridical perspective, and despite the fact that all parties agree that there will be no resuscitation, the nurses are still obliged to intervene. They still find themselves in a juridical danger zone and will have to act when Maureen’s condition deteriorates. One of the nurses is unable to do it; she feels that intervention only contributes to further suffering and in order to do justice to her own moral views she decides to hand over the care for Maureen to a colleague.7 Such a decision is exceptional. Research in the US among NICU nurses indicates that generally they follow up on the doctor’s orders, regardless of their own standpoint in the treatment policy (Raines, 1996). The policy involving Maureen’s treatment seems entirely determined by juridical structures, but is it really? This case illustrates that juridical structures locate the decision authority in a single point, but that this does not yet totally fix this power at the same location. In the Northeastern those involved situate the authority to decide in various locations. For the nurse who provides care to Maureen on the first day this power is entirely in the hands of the physician. Unlike the resident, this nurse does not conceive of the doctor as someone who is caught in a web of juridical rules, but as someone who paddles his own canoe. When asked, this doctor in fact confirms this: ‘If I want to stop treating, I stop, regardless of what the parents think. If I consider further treatment to be senseless, I withhold it. No one wants to treat children whose only prospect is a vegetative life. Therefore, when I think it is best to discontinue treatment I will do so. Even if the parents want me to go on. After all, I am responsible for the patients as well as for my own actions. If parents disagree with me, I gather the opinions of four other neonatologists and have them recorded in a form. If the parents will decide to sue me, let them come. After all, I already have four medical experts who back me up, so the parents won’t stand a chance. But in most cases all this is not necessary. I generally prepare parents and tell them that there is really nothing we can do anymore.’ This neonatologist is averse to any juridical pressure. One of the nurses attributes the fact that the second attending in the case of Maureen does not instantly embark on a withholding policy to other factors than the absence of parental consent or the pressure

70 Uncertainty in Medical Innovation

of juridical rules. The nurse believes psychological aspects to be a factor: ‘I think it is difficult for her. But we have made these children the way they are. And since they are not like we had wanted them to be, no one wants to be involved anymore.’ This nurse legitimizes this attitude in a psychosocial vocabulary. Similarly, the second attending chooses a psychosocial vocabulary to explain her decision: ‘Last night I had a nightmare about this child. So I decided to wait a day before unhooking her from the respirator. I have to be able to live with it as well, isn’t it? Time and again it is a burden. You feel it deep within you. It is hard, which is why I waited a day.’ The physicians deal with the legal frames involved in their own ways. This is not to say that they ignore the law. They certainly take the juridical rules into account, but they decide for themselves how exactly to interpret them. Put differently: the physicians also determine the functioning of the rules rather than that the rules completely determine the functioning of the physicians. For example, the second attending tries to bypass specific rules by postponing her request for the parents’ signature as long as possible, thus hoping that at one point it is no longer needed. After all, once a child is terminal, verbal parental consent suffices. Thus she tries to disburden the parents. Similarly, the first attending is looking for a solution. If parents will resist his decision, he will start collecting expert reports to back up his juridical position. If the exclusive reliance on one vocabulary does not seem to guarantee a workable social order in the NICU, a combination of vocabularies is even more likely to cause tensions and undermine the normal sense of order. After all, speaking about parents as object of care is at odds with a discourse in which it is assumed that parents take decisions on whether or not their child’s treatment should be ended (Guillemin and Holmtrom, 1986). This conflict is caused by the contradictory priorities of both vocabularies. In the juridical vocabulary the authority of parents comes first, but in the world of the psychosocial vocabulary parents are viewed as in need of support and even protection. If it is assumed that they have decision authority at all, the kind of decisions they have to take are seen as impossible decisions. Is it nevertheless possible to reconcile the two vocabularies? Let us return to the

Co-travellers 71

Academic Medical Center, where, unlike in the Northeastern, staff meetings are organized about such decisions. It underscores that two vocabularies indeed can go together, even without undermining the practical order.

3.4

Reconciling competing vocabularies

An analysis of the staff meeting’s further discussion on Esther shows how the staff tries to guarantee parental authority without leaving parents alone in their decision. This is why the meeting does not end once a decision about the child is made; also the position of the parents needs to be addressed. The central concern is how to approach Esther’s parents and to what extent they should be given responsibility for the decision at hand. In the meeting the staff tries to steer a middle course between guaranteeing parental authority and minimizing the parents’ burden. In the staff meeting on Esther the presence of both vocabularies first causes a dilemma, but subsequently a solution comes into view. The staff indeed manages to find a position in which parental authority is ensured while the burden for the parents is minimized. Which manoeuvres are needed to arrive at this situation? Who decides? The staff meeting on Esther concentrates on the question whether or not it is better to refrain from further treating Esther in light of her bad prognosis. As a consensus about stopping treatment is gradually emerging, increasingly the parents become the object of the discussion. Academic Medical Center The social worker points to the situation of Esther’s parents. ‘I believe it is very important to those parents that our proposal contains no more doubts. That they won’t feel they have to take a decision once again, as they had to in the delivery room.’ The resident responds by saying that they offer ‘advice’ to the parents: ‘They will have to take the ultimate decision on what is to happen now anyway. We cannot decide for them. They will have to do it for themselves. We can tell them how we view the situation, but they still have to decide what they are going to do with it.’

72 Uncertainty in Medical Innovation

When the neonatologist claims the issue to be ‘a little more subtle’ than that, the resident agrees: ‘Yes, the actual practice is always a little more subtle.’ In this situation the psychosocial and juridical vocabulary are pitted against each other. The social worker, out of a sense of solidarity with the parents, invokes the psychosocial vocabulary in which the parent functions as object of care. This vocabulary, however, leaves no space for the parents’ decision. Disburdening parents means not letting them take the decision, which also implies that the staff has to do it. In response, the resident puts the entire decision burden back onto the shoulders of the parents, claiming that the staff cannot take this decision. It is something the parents have to do themselves. The social worker: ‘And who will talk to the parents?’ The attending nods in the negative. ‘I don’t know them. You know them, do you Margaret?’ The resident nods. ‘Yes, I talked with them a few times. Okay, so if the results of the radiologist correspond to ours and the neurologist paints the same picture as ours, our advice is to withhold further treatment?’ The attending: ‘Right, if at least we can all agree?’ The neonatologist: ‘I’m not too comfortable with the term “advice.” I think you have to be clear to the parents about the very bad prognosis. As I just said, nineteen out of twenty children will be seriously handicapped, both mentally and motorially. I feel that these are the hard facts.’ The attending: ‘Yes, and then the parents will also ask you, “well, doctor and what now?” You can’t say, “sir, you have so much chance that it will not work out well and now it is your problem.” The parents next will ask you: “Yes, doctor, what is your opinion about it?” I feel it is not enough to tell the parents how much chance they have to get a handicapped child. Parents expect you to give them advice. So this is what we will have to formulate.’ The social worker tries to understand the neonatologist’s aversion. ‘But Rudy, you claim to be not very happy with the term “advice.” Do you mean that you want to see the parents decide after all?’ ‘Yes, in part I do, even if we will not put it that way. Of course we will discuss it with them in more subtle terms than

Co-travellers 73

we are doing right now. But I feel that at least to some extent they should decide.’ The attending: ‘I fundamentally disagree with you. If you give them the impression that it is their decision, we put an enormous weight onto their shoulders. There will be the risk that whatever choice they make they will always feel guilty. I believe we should give them compelling advice, that is, basically make the decision for them by telling them that there is really nothing we can do anymore. Thus the parents will not have to carry that burden.’ At this point in the meeting the two vocabularies nowhere intersect. To none of the parties this situation is an appealing one, however. It breaks the collective of staff and Esther’s parents as it existed around the incubator and it pits both parties against each other at a time when harmony is called for. As we have seen before in the case of Maureen’s parents, the parents of Esther now emerge in the juridical vocabulary in various guises. During the staff meeting the discussion is also about parents who decide and can have a more or less active position in the decisionmaking process. In the first position parents actively participate in the decision process, while in the second position they merely have to agree with the choice already made by the staff. Anspach (1993) describes this position as the assent model, in which parents assent to a choice already made by the staff, and the first position as the informed consent model, in which parents actively participate in the decision process (p. 87). Despite this difference, these positions share the assumption that parents join in decisions because it is their right after all. If the right to decision power may have a prominent place in the juridical vocabulary, in the psychosocial vocabulary other values apply. Based on the psychosocial value system such decisions are qualified as burdensome. It would leave parents alone in their decision. If one speaks and thinks more in juridical terms about parents, such remark is disqualified as condescending. With the values of the juridical vocabulary in mind, the hidden paternalism of the psychosocial vocabulary is criticized. The switch to another vocabulary thus implies a switch to another value system. The psychosocial vocabulary foregrounds the contact between nurses and parents. In this vocabulary parents are dependent and emotional, and all decision power is with the physician, as a way to protect parents. The juridical vocabulary, however, puts

74 Uncertainty in Medical Innovation

more decision power into the hands of the parents and the law. In this vocabulary parents are autonomous, informed and, at some moments, equal to doctors. In this way each vocabulary has its own hierarchy and value system and each vocabulary creates its own (social) order in the practice. It defines ways of interaction, creates more or less room for solidarity or respect, and distributes decision power. Speaking another language also involves measuring other measures. Is a compromise possible at all? Through the distribution of authority (juridical vocabulary) and solidarity (psychosocial vocabulary) the staff tries to realize its objective: being of help to parents and continue to respect their parental authority. However, having two clashing value systems seems to render unachievable the objective of being supportive of the parents and go on respecting their parental authority. In the staff meeting about Esther, the participants go on searching for those parameters based on which it will be possible to do justice to the parents from the angle of both vocabularies. As the staff meeting is evolving, the participants increasingly manage to find this common ground by combining vocabularies. In the proposal of the attending to give the parents ‘compelling advice’, the situation is still presented in juridical terms but its concrete implementation implies that the final decision is a shared one, made by parents and staff. Furthermore, the neonatologist who mainly relies on the juridical vocabulary takes over increasingly more elements from the psychosocial vocabulary. ‘They have to make a choice’ is reformulated as ‘at least to some extent they should decide.’ Gradually those present manage to find a compromise between the two vocabularies. Academic Medical Center The neonatologist: ‘I agree that we should not leave that decision up to the parents alone, and that they should feel supported by us. But somehow they must also understand that they should go along with a decision that we took.’ The attending comments that this is ‘different.’ ‘It is, mind you, all very subtle,’ the neonatologist goes on. ‘They have to agree with something, for we cannot stop the respirator without them giving their consent.’ The attending: ‘You are right, but I feel we should act in such way that we do not burden the parents with feelings of guilt, irrespective of the decision that comes out of it. This is how you should frame it.’

Co-travellers 75

When is a decision taken? In their effort to respect the principle of parental authority as much as possible and be of support to the parents, the discussants do not only look for a solution in language but also in time. The social worker is the first to deploy this strategy: ‘I believe it is very important to those parents that our proposal contains no more doubts. That they won’t feel they have to take a decision once again, as they had to in the delivery room.’ At that point the parents opted for maximal treatment on the condition that later on treatment might still be withheld if the child proved to be seriously handicapped. By suggesting to them in the delivery room that the final decision might still be taken at a later stage, when more information would be available about the child’s condition, at that moment the neonatologist tried to relieve the parents from having to decide already. His strategy was to move the decision moment forward in time. As a consequence, the actual decision is in part bypassed while parental authority is ensured. In the staff meeting the social worker applies the same strategy, but instead moves the decision back in time by suggesting that the parents already made a decision in the delivery room. In the staff meeting another neonatologist also refers to the parents’ decision moment in the past: Academic Medical Center The third neonatologist hasn’t said anything for some time, but now he rejoins the discussion: ‘I believe that indirectly we know what the parents think, if I understood Ria [social worker] correctly. So we do not really have to leave it up to them. We already know how they feel about it. If we are convinced that the chance of a healthy child is zero while the chance of a seriously retarded child is very large, we think we already heard the parents’ opinion and we do not need to have them face making the decision.’ This is taking it too far, the resident feels: ‘But at that point you do want to be informed of the position of the parents at that very moment. I think you should simply inform them of the view of the medical team involved and still pay attention to what the parents think about it.’ The attending backs the neonatologist: ‘But, as I understand it, that position has long been taken and articulated in the presence of various people. So you know it.’ ‘Yes, but they did so under difficult circumstances,’ the resident maintains. ‘And it is evidently not so sure that they

76 Uncertainty in Medical Innovation

hold on to it if push comes to shove. I feel you always have to verify it.’ ‘Sure, of course.’ By shifting the parents’ claims up and down along the temporal axis one tries to disburden the parents in the staff meeting. This temporal manipulation of decision moments is a major tool in the NICUs social order for disburdening the parents. This manipulation, however, does not yet remove all friction between the two vocabularies. In the ensuing discussion the attention shifts from the parents to the prognosis. Those present at the staff meeting agree that in the upcoming meeting with the parents uncertainty about the bad prognosis should be excluded as much as possible, not merely as a way to justify the decision, but also to be of help to the parents. They should not be burdened with the few percentages of uncertainty with which the staff is grappling. Most feel that they should act as one in front of the parents. To realize this, the commitment of all involved is crucial. Academic Medical Center The resident: ‘I can imagine that you do not want a conflict with the parents afterwards, with them saying “No, this is not what we wanted after all.” You want to tell them this, but more in terms of the choice that was made and of which you inform them and to which they can respond. And so they won’t say afterward: “Well, we had wanted to give our reaction, which was that we suddenly had second-thoughts.” This, of course, you want to avoid.’ The attending: ‘It means that we have to make substantive predictive assertions about a prognosis and that we cannot say: “Yes, it would be possible that…” and “if, unless…”; this will only make them feel insecure. … … I think it is not enough to tell the parents how much chance there is that they will get a handicapped child. Parents expect that you give them advice, and this is what we should formulate. And then you may say: “Yes, the literature describes a case in which …” But it is not something you can work with!’ The resident: ‘If you are convinced of it, you can put it that way. But if you have doubts about it, this is more difficult.’ In the staff meeting about Esther various vocabularies are invoked. Yet the use of multiple vocabularies does not lead to a deadlock, but in fact

Co-travellers 77

offers the staff the possibility to support the parents while respecting their views. Ultimately there is no dominant vocabulary. What is more, both vocabularies need each other. Leaving the decision entirely up to the parents is impossible in light of the fact that they lack the knowledge to interpret the information correctly. But the staff’s taking full responsibility would have implied unacceptable paternalism. In short, for the overall order to function properly, one vocabulary cannot do without the other. In this respect the discussion about Esther strongly deviates from that about Maureen. In Northeastern the psychosocial vocabulary is suppressed, so to speak, by the juridical vocabulary. Although here too the staff occasionally falls back on a psychosocial vocabulary, the legal rules linked up with the juridical vocabulary basically deprive the staff the latitude to neutralize the conflict between them and Maureen’s parents. In addition, the institutional order of the NICU plays a major role. Styles of ordering and social orders In the Northeastern nursery it is based on individuality: the medical responsibility lies with the attending. Just like in the Academic Medical Center NICU, in Northeastern each month another neonatologist takes on the role of attending. This physician, however, bears full responsibility for the treatment of NICU patients, which is a major difference with the Dutch situation. The NICU of Northeastern can be characterized by its individual treatment policy, meaning that formal consultation with all members of the medical team is an exception rather than the rule, which implies that medical policies tend to be applied more personally or in varied ways. Also the responsibility for the nursing care lies with the individual nurse involved. The juridical vocabulary seamlessly fits in with this social order: it locates the decision authority in one point; a single person has final responsibility. In the Academic Medical Center, by contrast, the institutional order does not assume a one-to-one relationship; instead, a shared responsibility for care and treatment is in the hands of a team of doctors and nurses. This collaborative social order counterbalances the individuality implied in the juridical vocabulary. A combining of the two vocabularies offers the manoeuvring room needed for creating a new way of speaking which ties in with the institutional order of communality. This allows them to share the responsibility for the decision. The juridical vocabulary, however, creates an order that strongly relies on formal deliberations with parents and the filling out of forms. This does not mean that only the juridical vocabulary is about power,

78 Uncertainty in Medical Innovation

for a certain measure of inequality between staff and parents is also part of the psychosocial vocabulary. Nor does it mean that only the psychosocial vocabulary is about protection, as this equally applies to the juridical vocabulary. Yet it protects rights rather than feelings. It is important to underline, however, that the two vocabularies structure the NICUs social reality in divergent ways. If one vocabulary distributes roles on the basis of emotions, the other does so based on rights and duties. If in one vocabulary parents and child are pitted against the staff, in the other staff and parents have a common interest vis-à-vis the child. This analysis of two major vocabularies that structure the NICU social order also underscores how facts, values, and actions interact. The specific discourse on parents (who they are, how they should be defined) cannot be seen in isolation from the issue of how to deal with them. The parents as object of care call for another approach than the parents as bearers of parental authority. Vocabularies are more than just means of description. Vocabularies imply scenarios for action and have consequences for the ways in which interactions take place in a NICU. Analysis of the structure created by those vocabularies provides insight into how social structures are organized. It shows that those involved need to reach consensus about their interpretation of the situation (the facts) as well as the locus of responsibility. Among other things this calls for a coordination effort between the various vocabularies. In the discussion about the treatment of Esther the participants only managed to take a decision by bringing the juridical and psychosocial vocabularies into line. Such alignment must be established each time anew. In the two vocabularies the equality of both parties is understood differently and thus they also indicate who has the right to speak and who is responsible. The various ways of speaking intervene in their own way in the constitution of facts, values, and actions.8 An analysis of these ways of intervention demonstrates their ‘endogenous normativity’ (ter Meulen and Berg, 1999). The performative character of the various ways of speaking offers direction to the ways in which the intervention process takes shape. This can lead to friction in the decision process, for instance, regarding responsibility. To resolve it one searches for a way out by combining vocabularies and speaking in terms of ‘compelling advice.’ Vocabularies, then, are rarely found as pure forms. In the NICU, understood pragmatically, multiple social orders exist side by side and are mixed up with each other, rather than there being one overriding order. The discussions around the incubator of Maureen, however, did show that these orders are not equal. Some social orders, like the juridical, are more powerful than others due to larger societal structures.

Co-travellers 79

The discussions about the treatment of Maureen and Esther made visible how they offer staff members and parents more or less possibility of negotiation. In a constant process of interaction they can adjust and refine their mutual expectations. The co-existence of these various social orders that are brought in with the deployment of specific vocabularies can give rise to misunderstandings and even tensions. However, in the dynamic and organic interconnection of these vocabularies – historically they are not fixed – there is also room for flexibility and debate. After all, staff and parents are travelling companions: they need each other in their care and concern for the child so it will reach the most appealing destination. It is in everyone’s best interest, in particular that of the child, that on this journey staff and parents do not turn themselves against each other. Because they know they are in a relationship of interdependency, they deploy strategies to discourage the other party from prematurely pulling out. Their shared objective, in other words, is the child’s recovery. This involves all but an automatic process. Life in the NICU – even more than life in the everyday world – is marked by constant struggle and unpredictable changes. From its inception, life in the NICU is uncertain in extreme ways. It is the staff’s task to steer its erratic dynamic in the right direction. How they do so, including the problems that thereby present themselves, is the subject of the next chapter.

4 Uncertain Trajectories

The basic task of the NICU staff is to respond adequately to the frequently changing conditions of its patients. These changes are not always predictable and seldom unambiguous. They are indicative of the simple fact that the children in the NICU are alive. Like all infants they go on growing after birth, and this is of course an essential development. If, however, an infant is too small or too ill to realize this natural growth on his/her own, the baby’s condition will worsen without outside intervention. In such cases change implies no longer growth or development but a deadly threat, and direct intervention is essential. Furthermore, changes may be the effect of treatment and often this will mean recovery. By the same token, not all children in this unit respond to treatment according to the handbooks. This is why NICU efforts also have a more experimental basis: they rather involve attempts to influence the course of events so that a child’s condition will improve. Even if much of this work is routine and experienced staff will not be caught by surprises easily, it is also common that situations recur in which indecision prevails. In this sense, change in a NICU context always comes with an element of uncertainty. Uncertainty is a characteristic element of medical practice as a whole.1 The everyday practice of treatment is messier than the medical textbooks tend to suggest. Specific clinical traits prove absent in patients or several disorders influence each other. Given this specificity of individual cases, patients count as starting point of action for physicians. This principle turns medicine into a ‘science of individuals’ (Hunter, 1991, p. 28). An experienced physician can still have great doubts about how to proceed in a given situation, for instance, because there is no time to wait for a test result, because no protocol is available, or because a particular tool produces imprecise or even unreliable results. Active 80

Uncertain Trajectories 81

intervention may still be called for, however, despite the absence of hard facts. Precisely in such situations the best physicians will rise to the occasion. Uncertainty comes in many guises in the NICU. Will a treatment catch on or not? Will a child collapse because of heart problems or a lung infection? Should a protocol be brushed aside because the child responds very differently? A series of interrelated problems may completely destabilize a child’s condition. It is not always clear how a child is doing, where it is on the treatment trajectory, and how its condition will evolve. Doubts about the correctness of test data, the reliability of equipment or the ability of the parents to move along with the changing situations are bound to crop up at some point. Although uncertainty is an integral component of medical knowledge in a NICU, staff members do not constantly feel it as a burden when performing their daily activities. Doctors have an array of strategies at their disposal for dealing with uncertainty: they keep their options open, they think in terms of probability, they consult others, they decide not to take a decision yet, they identify it as a subject for scholarly study, or they rely on humour or irony to put their lack of insight into perspective (Bosk, 1980). In actual practice, the uncertainties that are an essential part of medical knowledge are compensated for by the suggestion of control that tends to characterize medical action. Routines, clinical experience, skills, knowledge – all contribute to a position from which physicians can begin to act in ways they are used to. Based on professional expertise, commonsense reasoning, and a pragmatic attitude they make decisions and perform actions, thus finding their way, generally fully aware that complete control of a particular treatment’s unfolding is a fiction. The initial challenge is not simply to find a solution to a specific problem but how to begin treatment. After all, the staff is unable to oversee the entire trajectory, and even if the goal of its effort – a healthy baby – is clear from the start, it is unsure how this objective can be realized. Because treatment in the NICU often involves new or boundary-transgressing situations, it is not uncommon for doctors to enter uncharted territories where routines are absent and unanticipated challenges present themselves. This chapter focuses on the issue of how the staff deals with intrinsic tensions involved in the need for intervention and making adjustments on the one hand and the associated risks and uncertainties on the other. How do doctors keep their young patient on track and replace doubts with firm knowledge and control? As a framework for reporting

82 Uncertainty in Medical Innovation

and discussing my findings, I will rely on two case studies. My main case involves Robert, a full-term baby with a severe malformation of the heart; the complex trajectory of his care and treatment constitutes the core of my argument in this chapter. The other case involves the trajectory of Tom again. My use of the term ‘trajectory’ is the same as the one formulated by Anselm Strauss: ‘Trajectory’ refers to the development of an illness, the work that has to be performed in relation to this development, and the influence of this work on those who perform these activities (Strauss et al., 1985, p. 8). In the case of Tom and Robert the treatment involved is all but routine-like, which is why they are useful for demonstrating the various guises of uncertainty in the neonatology practice – how they complicate the treatment process, and which strategies are deployed to overcome hesitation and even gridlock. Major factors are time, guidelines, equipment, and numbers. It will become clear that strategies for dealing with uncertainty also give rise to new problems and concerns. However, before they are in a position to deal with these concerns neonatologists first have to decide which trajectory to follow. How do they select which path to follow? Therefore I will first take a look at the moment neonatologists have to decide how to intervene, and the role of time and protocols in particular, before I discuss the strategies staff members use to keep the child on the right track.

4.1

Determining a treatment trajectory

The choice of a specific treatment trajectory highly influences the success of the treatment of NICU children. Because of their great vulnerability, promptness of action can be life-saving.2 In the staff’s effort to control a child’s condition two factors are essential: time and guidelines. Time as risk and ally The decision on how to start a specific treatment trajectory is obviously of great significance as to how a newborn’s life will evolve, yet the need for such decision mostly presents itself at a time when the physician involved has the least insight in the child’s situation: he sees the child for the first time or only few data are available. Basically, at this point, the staff only has information on the course of the pregnancy and the delivery. Still, specific decisions have to be made and doctors

Uncertain Trajectories 83

thereby fall back on the narrative structure of medical knowledge (Hunter, 1991). They read the human body as if it is a story. They sort, combine, and interpret what they hear and see and try to detect a medical story in it. To find their way through the many potentially relevant details they not only rely on scientific medical knowledge, but, as experienced readers of the body, they also use their knowledge of plots and similar medical stories. Their familiarity with specific plots raises their chances of recognizing particular disorders. The significance of recognizing the source of a baby’s medical problem is not so much in defining its clinical picture, but in the implied perspective on possibilities for treatment (Strauss et al., 1985). Based on its diagnosis, the staff takes into account the implications for its course of action, including the interventions needed, their proper order, the tools and the expertise that have to be deployed, and the potential (negative) effects. From the start, the staff performs diagnostic and therapeutic acts concurrently, rather than successively. Diagnostic tests may confirm suspicions and provide the key for sound decisions on the various treatment options available. Performing such tests, however, takes time, while in acute cases immediate action is called for. Time is a scarcity in the NICU. In some cases there is no time for extensive diagnostic testing or carefully preparing interventions, meaning that one has to choose a trajectory on the basis of incomplete knowledge. Obviously, this is not without risks, and therefore one does not only select a treatment trajectory, but also – and perhaps more so – the kind of risk one is willing to take. If physicians decide to refrain from intervention, they run the risk of letting a child die that still has a chance of survival. Yet intervention may also come with the risk that the child will have a life entirely devoid of quality.3 The physician present at Robert’s birth did not first try to establish a diagnosis, but immediately opted for admission to a more specialized unit of another hospital. Although most Dutch hospitals have a medium and high care unit for neonates, only academic hospitals have intensive care units for them as well. Thus he sought to limit the risk of Robert’s premature death. When he was born his health seemed to be perfect, but after the phyician cut the umbilical chord the newborn turned blue all over. This physician had no information on what caused it, but he had to act immediately in order to keep the child alive. What to do? How can he select a treatment trajectory in the absence of a diagnosis? Fully aware of the seriousness of the situation, the physician contacted the Academic Medical Center’s NICU right away. As will become clear, this decision saved Robert’s life.

84 Uncertainty in Medical Innovation

Robert is taken by babylance* from the peripheral hospital to the NICU. The staff has an incubator ready with all the necessary equipment, while the paediatric cardiologist is informed of the infant’s impending arrival. This cardiologist examines the baby immediately. The Doppler machine’s* monitor clearly shows how his oxygen-rich blood directly flows back into the lungs, which means there is no doubt about the transposition of the blood circulation system. The cardiologist, neonatologist, and surgeon discuss how to proceed, and they decide in favour of corrective surgery. Without it, Robert will die, yet the procedure is very risky: it either succeeds or fails; there is no in-between. If it succeeds, there is a maximal return: the child will grow up like any other healthy child. Given the complexity of the trajectory that is called for in Robert’s case, detailed and time-consuming preparations are absolutely necessary, which means that somehow extra time needs to be created. This is done through medication that slows down the closure of the ductus arteriosus Botalli. This is an opening in a foetus’s heart. Normally after birth this opening – the ductus – closes within twenty-four hours, so as to allow the oxygen-rich blood to enter the circulatory system. If the ductus fails to shut, the oxygen-rich blood flows back into the lungs. This will fill up the lungs, while the blood in the body’s circulatory system does not contain enough oxygen. By keeping Robert’s ductus open, the oxygen-rich blood may still enter the circulatory system. The decision to apply this medication provides the staff additional time – a temporal niche that allows it to do preparations for the next steps of the complex trajectory involved. Occasionally the staff has more time available for diagnosis. Its effort, then, is likely to concentrate on excluding specific uncertainties. For instance, one changes the medication or a device setting to see how the patient responds. In such situation, treatment is a matter of ‘try, wait and see’ instead of assured success. This method is a muchused strategy in medical practice for coping with uncertainty. Although one has no certainty on whether the treatment selected will catch on, the situation forces one to intervene. An experimental space is created in which it is not the staff’s move but the child’s. The staff has acted and subsequently it has to become clear how the child responds, if s/he will. Another way of handling doubt or uncertainty is by taking recourse to the future: a decision does not have to be taken or cannot be taken because all the information needed is not available yet. This

Uncertain Trajectories 85

is a frequent reason for refraining from medical action. Consider the following scene that occurred during a medical round: Academic Medical Center The attending neonatologist, three residents and a nurse are grouped around one of the incubators. The neonatologist challenges the resident by asking ever more questions. In one case, the resident does not quite know what to do. It is insufficiently clear to him what the child is suffering from. He therefore proposes to wait until they have more data, suggesting that the blood test results are due to come in while other tests have already been requested. During medical rounds residents frequently hide their uncertainties by using phrases like ‘it seems to be,’ ‘it might be the case that,’ ‘my best guess is,’ ‘it reminds me of’ and so on. Or they explicitly claim to be unsure of the diagnosis or therapy they propose. Their presentation relies on phrases like ‘as far as I can tell now I think that’ and ‘if we assume that X is the case we perhaps have to do with Y.’ The resident refers to the future as the site of more information or more certainty – and hence the site of decision. Instead of admitting his ignorance about what is going on or what has to be done, he refers to the blood test results that are due. This does not mean he is a bad doctor. After all, decisions have to be based on sufficient information, and thus he merely follows the established rules of the game. In some cases the staff decides to do nothing at all, hoping that the situation will improve by itself and that after some time the problem will have gone. In some cases this happens indeed, meaning that time is on their side. This temporalizing of doubt, however, is not without risks, because meanwhile the problem – whatever it is – may also develop into a disorder that proves even harder to address. In such cases time is an unreliable ally. As a factor, time is made and deployed. Having time is not alldetermining, though. In some cases the staff will never find out what exactly is wrong with a child. This can be frustrating, as one of the nurse practitioners* from Northeastern suggests: ‘My worst case was a baby that was here for months and I did not know what was wrong. For all those months you do not have a basis to work from. It is a matter of constantly trying and seeing what works. All those months I could not explain to these parents what

86 Uncertainty in Medical Innovation

was going on. You have no label for it. Nor could we find out if it was hereditary. After six months the child went home and after one year the parents got another child. The same story; it was hereditary after all.’ It is important to be able to define the disorder, even if it is one for which there is no treatment, such as trisomy 13* and 18*. In these cases the physician involved at least knows that his failure to help is beyond his responsibility. Such situations, though, are exceptions; in most situations there are various specific treatment options available. The use of protocols The vulnerable condition of NICU children makes it crucial to plan their trajectory carefully, based on established routes that have proved successful: guidelines or treatment protocols. As a set of instructions, a protocol tells staff what to do in specific situations. The protocol’s conditional ‘if-then’ structure guides staff members along complicated treatment trajectories. This reliance on protocols is a major strategy for coping with risks. When planning a treatment trajectory, the staff can make use of these itineraries that guide them step by step through specific medical situations. Protocols prescribe which diagnostic tests have to be performed, which treatment criteria are valid, which direction the treatment should go after a certain outcome, when informed consent is needed, or which lab tests no longer are relevant. As such protocols structure the flow of activities in the NICU. In performing their tasks doctors and nurses delegate part of their coordinating activities to protocols, which indicate who or what has which role. A protocol functions as a ‘focal point of reference to which different staff members refer, can orient themselves, and can find instructions on what to do next’ (Berg, 1998, p. 232). The route specified by a protocol is the most ideal trajectory – one with a progressive line. This also requires an ideal patient – one that responds to the treatment swiftly (Guillemin and Holmstrom, 1986). In a NICU, however, one will rarely find such ideal patients because those who are well enough tend to be transferred to high care. The use of protocols assumes not only a specific patient, but also a specific practice. Treatment protocols, after all, are geared toward standardizing medical work. This objective is at odds with the specificity of the situation in which medical acting occurs. Yet the NICU caters to children whose clinical picture is erratic and complex. Their development is marked by ups and downs rather than straightforward improvement.

Uncertain Trajectories 87

This is why guidelines, which presuppose a more or less linear development, seem at odds with the erratic disease patterns found among NICU patients. Physicians may be put in a position of having to make a choice between sticking to a protocol and ignoring the child’s idiosyncratic development or, conversely, ignoring the protocol and foregrounding the child’s changing clinical picture. This inflexibility limits the effectiveness of protocols as a tool in the decision process. Other elements of criticism are that this instrument would reduce medical intervention to ‘cookbook-medicine’ and result in ‘deskilling’; that the nature of protocols (generic) is different from that of practical medical intervention (individual) and that therefore they need to be translated; and that the normative aspects of protocols are disguised, as if they were neutral instruments. Protocols suggest that medical data consist of transparent bits of information; that medical criteria are clearly demarcated rules that merely need to be applied; and that medical intervention consists of a sequence of successive steps (Berg, 1997b; 1998). It can be argued that because of their relative inflexibility they do no justice to the complexity of medical practice. They assume clinical decisions to be purely rational, unrelated to their social and material context. The need for making specific choices is most urgent right after a baby’s birth, when generally there is minimal insight in his/her condition. Yet such medical knowledge is essential for choosing the most suitable protocol. Protocols start from a specific practical order, including the availability of diagnostic information or the time to collect this information. But what should be done if both are absent? How should one proceed? Which strategies do the staff members deploy to come to reliable performance despite the odds? Based on the diagnosis of transposition, a custom-made trajectory is prepared involving surgery, medication, pre-care measures, aftercare, a timeframe, and possible complications. The specific protocol, also formulated by the cardiologist, guides the staff members, but they are all fully aware of the many pitfalls, of the risks tied to each phase and transitional moment.4 Academic Medical Center Monday, 14:00 hours: Nurses receive instructions for tomorrow’s transposition. The meeting is well-attended. Using slides the cardiologist explains the problem and tells them what will happen during tomorrow’s operation. Next, he discusses everything that

88 Uncertainty in Medical Innovation

has to be done and what deserves special attention before the child leaves for the OR and after its return in the NICU. The cardiologist emphasizes that a ‘heart baby’ such as this demands another attitude: ‘Normally we have to do with insufficient lung function. In those cases there is still some time to discuss what you are doing. You change the settings of the respirator or things like that, yet basically every move can first be discussed. But a child with this kind of heart problem requires immediate action. There is no time to discuss options in some other room. Therefore it is crucial that in advance you closely study the protocol I have written. Although the protocol is essential, it is the actual situation that at all times will determine what should be done.’ Tuesday, 10:00 hours: This morning at eight o’clock Robert entered the OR. There is a tense atmosphere in the NICU. Repeatedly one can hear someone inform how the operation is going and how much longer it will take for Robert to return. The attending neonatologist is somewhat irritated by all this attention: ‘So much concern for this one child. All the other children are important as well.’ Robert’s condition is so fragile that the surgery is extremely risky. As a step-by-step planner, a protocol always contains several transitional moments that are frequently also the most risky ones. During the operation such moments occur from the very start. The cardiologist is very aware of this risk: ‘These cases give us a lot of stress. When all goes well, I feel like conquering the world tomorrow. I really get a kick out of it. But if things go wrong with such a child, I feel terrible for a few days. I am extremely exhausted and unable to handle anything. You give all you have in such a case. There are always a few anxious moments during which you fear that things will go wrong. One such moment is when the child is unhooked from the ECMO* and he has to perform those functions on his own. If he doesn’t do so, you’re stuck. In the previous case it did not work out and the child died.’ In Robert’s case it worked out well. With his successful disconnection from the cardio-pulmonary bypass a major hurdle in the trajectory is taken. Although the surgery is successful, in the first 48 hours there is

Uncertain Trajectories 89

still a substantial risk that he will die. The next transitional moment presents itself right after his return in the NICU: Tuesday, 17:00 hours: Robert returns accompanied by a ‘green train’ of cardiologists, anaesthesiologists, neonatologists, and nurses. It is hard to see a baby at all, enwrapped as he is in lines and tubes. Taking Robert out of the reanimation cart and putting him into his incubator involves a tricky procedure. Much has been said about this moment during the preparatory session. As foreseen, there is spaghetti of lines and tubes that can hardly be unravelled. Extreme caution is called for, but it also has to be done fast, for the child is extremely vulnerable. The sooner it is comfortably back in its incubator, the better. After prudently lifting him out of the cart with all his technological baggage, they carefully put him into his incubator. Everyone is looking tensely at the monitor that briefly shows his levels jump up and down, after which they become stable again. It worked out well and the staff members are all relieved. The probability of specific risks differs: it is always uncertain whether complications will indeed occur. In the NICU, however, risks are considered an inevitable element of the everyday practice. One way to deal with risks is to ignore their probability and consider them as established facts. The staff acts as if a specific risk will in fact become a reality, until the evidence shows otherwise. Unforeseen and nontransparent situations, in other words, are taken into account in the planning of the treatment trajectory. In the guidelines for Robert’s treatment all the possible complications are calculated in advance. A bleeding after surgery is not viewed as a possibility but as a certainty. This ‘certainification’ in turn generates another choice: between one risk (a bleeding that cannot be reached in time due to the closed chest) and another risk (an infection due to the open chest).5 In Robert’s case the first option is chosen and his chest is only covered with a patch – one that only a few hours after surgery proved to be justified. Tuesday, 19:00 hours: Robert’s chest is covered with a large patch. During the last half hour it slightly bulged. All his parameters are okay except his venous saturation level. The staff decides not to wait for the other parameters to go down as well, because the message of the venous saturation is clear: the child

90 Uncertainty in Medical Innovation

suffers from a bleeding. Direct action is called for, leaving no time to transfer him to the OR, which is too risky and not absolutely necessary. While the surgeon is putting on his sterile garb, two nurses do the required preparations. They create a sterile field on a cart, and put the surgical tools on its lower part. One staff member goes to the OR to pick up a suction pipe. The anaesthesiologist is ready. The plastic patch is cut open and the old blood that accumulated is drained off. The larger clots are removed with a pair of tweezers and weighed. Probably these blood clots clogged the drain, which no longer allowed the fluid from the wound to be discharged. This caused higher pressure around the heart, which reduced its pumping power. The draining diminishes the pressure and the parameters on the monitor slowly improve. Initially things do not seem as bad, but quite soon fresh red blood appears. The pressure has indeed caused a bleeding. The surgeon puts on his surgical glasses and starts looking for the leak. Meanwhile Robert receives a blood transfusion. Soon the leak is found and the hole closed. As the bleeding stops, the surgeon begins to curse so as to express his relief. He remained calm throughout, but now the danger is over his apprehension comes to the fore. He managed to stop the bleeding in time. Everyone responds with relief. The monitor clearly indicates how Robert is coming out of his crisis. Deviations from standard procedures are sometimes necessary to prevent dramatic outcomes. In cases like these, medical staff looks for a compromise between assessing the child’s condition completely and saving time to act. They try to find a sufficient assessment of the situation for action under uncertain conditions. By deviating from the protocol rule that operations have to be performed in the operating room, the child was saved in fact. This was also possible because after the operation the access to the heart was not closed but merely covered with a patch. By taking into account the possibility of a bleeding in the protocol for Robert, the staff was prepared for it. Once the bleeding indeed occurred, immediate action could be taken: after all, there was still open access to his heart. It was merely a matter of removing the patch that covered the wound. This ‘emergency entrance’ ultimately turned out to be life-saving. But it cannot be kept open for long because the closure of the wound on his chest equally constitutes a risky transitional moment.

Uncertain Trajectories 91

Transitional moments in specific medical interventions carry extra risks, but the same applies to the inevitable moments of transition in the organization of the daily activities in the NICU. A daily moment of transition, for example, is the staff’s rotation, which involves the transfer of information. One of the specific risks is that data are not communicated or understood incorrectly. A complex trajectory like that of Robert requires the input of a variety of medical specialists, including cardiologists and neonatologists. In such intricate cases it has to be clear which person has final responsibility, as misunderstandings have to be avoided at all cost. This is why the collaboration between the various medical disciplines and the mutual gearing of the various sorts of medical knowledge are of great importance. In this respect one of the neonatologists of the Academic Medical Center observed the following: ‘The child is in our ward, so ultimately we are responsible. With this kind of surgery the cardiologists of course have more experience. But we have more experience with infants. Much deliberation, therefore, is needed, which is not always easy. One interprets children’s saturation levels differently than those of adults, while a cardiac patient has yet again different saturation levels. What we consider too high, a cardiologist may view as too low or the other way around. But it would be wrong if a nurse were told by one specialist to do this while another tells her to do the opposite. There is an agreement that as neonatologists we are in charge here and that therefore we are consulted on everything that takes place. The nurses only do what we tell them to do, so as to avoid conflicts. This does not mean that we tell the cardiologist: You have to do this and that. No, it is a form of deliberation in which you propose solutions and subsequently decide things together.’ Thursday: Robert’s condition noticeably improved. He is given a bath and his skin has more colour. His drain still barely generates fluid. This is a good sign. Now that his condition is stabilizing, the monitoring team is lowered from three physicians and two nurses to two physicians and one nurse. The number of machines to which he is hooked up, however, is not reduced yet. Friday: Today Robert’s chest will finally be closed. Around his incubator a sterile field is created once again and the surgeon prepares for surgery. First the band-aid is removed, followed

92 Uncertainty in Medical Innovation

by the patch that covers his heart. Each stitch is cut carefully. In the chest’s cavity the heart, beating calmly, becomes visible. The drain is put to the side, so that the catheters can be removed that measured the pressure in the heart: first the catheter from the left atrium*, then the one from the right atrium*. At this point something goes wrong. When removing the second catheter something else is pulled loose. The effect is a bleeding. The surgeon acts instantaneously. ‘Where is the extra blood?’ the neonatologist asks the nurse at his side. ‘It has been ordered, but it is still in the blood transfusion department.’ The neonatologist appears shocked: ‘Go get it right away.’ The nurse leaves running, while the attending pulls the other nurse aside: ‘The blood should have been here. Here in the ward. Not at the blood transfusion department. The bleeding is not so bad yet, but had it been worse, we would have been in serious trouble.’ The nurse fully agrees, and next time it will be done differently. To everyone’s relief, the surgeon manages to halt the bleeding. Now the chest can be closed. Layer after layer, everything is sewn up. First the breastbone that still only consists of cartilage; next, the dermis that mainly consists of fat; and finally the outer skin. Each layer has its own knots and stitches. The inner layers are done with dissolvable stitches, unlike the ones of the outer layer that will be removed later on. After everything is sewn up, the nurse washes Robert and this makes him look much better already. Despite the surgeon’s great care, he could not prevent a bleeding from occurring. This is a familiar risk, which is why the staff ordered extra blood, but it had not yet been delivered to the NICU. In this case a cool-headed doctor, a nurse in good shape, and some luck contribute to averting a bad outcome. In actual practice errors function not only as cues indicating limits but also as occasions for learning. The unfortunate experience with Robert in this respect will challenge the staff to ensure that in the next child’s treatment plan the availability of extra blood in the NICU will be listed as a prerequisite. As such near-misses in treatment trajectories also provide specific learning moments for staff. Guidelines give shape to actual interventions and these interventions in turn lead to more knowledge and new guidelines. However treatment should never be just a matter of following guidelines. If guidelines offer support and show the way, they offer no guarantee whatsoever that a

Uncertain Trajectories 93

child’s treatment will follow the projected itinerary. Sometimes treatment requires one to sidestep a guideline or even to do nothing. Treatment consists of more than just direct intervention or taking decisions under time constraints. It may be better in some cases to refer a child to a more specialized unit of another hospital. Moreover, time, as we have seen, is not always an opponent. NICU staff also may use and construct time to prepare a specific treatment or have it catch on – situations, of course, in which patience needs to prevail. Uncertainty is not just an element of specific diagnostic doubts or the risks of particular interventions, but also a more general feature of the NICU practice. It is quite normal that a NICU child suddenly collapses, that its heart action sharply curves downward without clear reason, or that during a routine act a child suffers a bradycardia*. In these cases the staff has to intervene promptly to correct the child’s condition. To intervene quickly the staff has to be on guard all the time. Thereby technology is a major aid, but, as we will see in the next section, as such it does not automatically guarantee smooth or unproblematic treatment trajectories.

4.2

Technology’s role on the treatment trajectory

Staff members are very aware of the infant’s vulnerability and the potential unintended consequences of their own interventions. At any time, without warning, newborns may suddenly collapse. This means that staff must keep close tabs on their vitals in order to be in the best possible position to prevent disastrous outcomes. The most accurate, up-to-the-minute information is required, often on an ongoing basis. Via information or results from the monitor, blood tests, X-rays, ultrasound scans, daily examinations and data entered on the flow sheet, the condition of children is monitored and, if needed, their treatment adjusted. The more control, the better the chances that the infants will survive the incubation period without serious complications. Each irregularity, no matter how small, activates an alarm attached to the monitor. Considering the enormous amount of technological devices, the role of technology in patient care deserves further attention. What, actually, is the position of high-tech equipment in the care of newborn babies? Technology and care The various intensive monitoring practices and the detailed recording of the child’s condition are mainly the responsibility of nurses, who strongly

94 Uncertainty in Medical Innovation

rely on all sorts of devices that measure various vital functions of the child. The treatment of children like Tom, Esther, and Robert is unthinkable without the technology that surrounds them. The electrodes on the infant’s skin register all sorts of bodily functions. The monitor screen above their incubator displays the frequency of the heart rate, pulse, and respiration, the saturation value of the oxygen in the blood and blood pressure. The question of how the child is doing is mostly followed by a look on the screen of the monitor. In a single glance the staff thus learns much about the child’s condition. As such, this device mediates between staff and child. Staff members use the monitor as a lens through which they look at the child up close. They watch the monitor, but see the child – as if monitor and child are one. Similarly, the terminology used for discussing the child is rife with references to the machines and their output. The articulation of the child’s condition is closely bound up with technology. The post-phenomenologist Don Ihde (1990) identifies two ways in which artefacts can play a mediating role in the relation between humans and world. The first involves an embodiment relation, whereby the technology is incorporated into the bodily dimension of people. A clear example of this form of mediation is a pair of glasses. We look through them to see the world around while forgetting that we are wearing them. In this sense glasses are transparent. The second mediation relationship Ihde identifies is a hermeneutical relation, whereby the artefact provides a representation of the world. An example is a monitor, which is not transparent. The staff members have to look at the monitor, and subsequently they have to read and interpret what the monitor tells them. A nurse in the Academic Medical Center checks and registers Tom’s condition every three hours:

10:00 hours: The nurse takes the flow sheet that is on top of Tom’s incubator and writes various readings from the screen next to the previous data she entered three hours ago. With a stethoscope she listens to his heartbeat and counts his breathing. She judges his colour and enters her finding next to the other data. The intravenous systems are checked. One has nutrition and two others have medication. The perfusor* regulates their flow and dosage. The readings are entered in the flow chart, as is the case with data

Uncertain Trajectories 95

on the ventilator’s pressure and rates. She also takes a blood sample for analysis. In three hours the nurse will again check, measure, and record these same aspects and data. Depending on the instability of the child’s condition, the frequency of the checks and rechecks is raised. The extreme instability of Robert’s condition requires the utmost concentration of all staff members involved. His record consists of so many parameters that it is no longer possible to identify intervals between the various checks; those in charge of him hardly pause for a moment. Robert is not just a neonate but also a cardiac patient. For such patients other rules apply and other forms of reporting data are used. So far this particular NICU has seldom admitted an infant with transposition, which means that the nursing staff has had little opportunity to gain experience with this type of case. Therefore, double-checking is called for when new data are added to the flow sheet. The complexity of Robert’s monitoring proves to require a kind of specialized knowledge and experience that is still hardly there. Although this NICU has the expertise needed for his treatment, it turns out to be hard to distribute this expertise in a balanced way among the staff that is present at any given moment:

Thursday, 15:00 hours: When the evening shift enters, a light panic erupts. Of the certified nurses no one feels experienced enough to take over responsibility for Robert’s care. One nurse explains: ‘The same staff members usually provide care to these children. This is why others, including me, do not gain experience in treating them. So you can’t expect us to do this straight away. Just look at how this child is hooked up to everything. I can’t do it.’ To solve the problem an experienced nurse scheduled to work in High Care changes place with a less experienced NICU nurse. 19:30 hours: All is calm in the unit. The attending is sitting behind the table and writes his report in the record. The nurse enters data on Robert in the flow sheet. She grabbed a high stool to sit on and uses a lectern as desk. To everyone’s relief, the other children remain quiet tonight. The staff has no use now for an emergency. From the bordering high care unit

96 Uncertainty in Medical Innovation

sounds from a music box can be heard, a surprisingly playful and innocent sound in this clinical setting. The best staff members cannot be present around-the-clock, nor do they all have as much experience. Because of staff’s rotation and specific disciplinary differences between neonatology and cardiology intensive monitoring, which is aimed at ruling out uncertainties, it becomes also a source of uncertainty. This applies to both the nursing staff and the medical staff. In a complex case such as that of Robert, experienced physicians are present at night and during weekends. And for a good reason: Friday, 22:00 hours: Unexpectedly, the saturation level of Robert’s blood drops sharply. The attending just left, and the resident is hesitant about what to do. Everyone is anxiously looking at the monitor. At that moment the anaesthesiologist and the cardiologist enter the NICU. They intervene instantly, and after some minutes the screen indicates that the saturation level improves again. The anaesthesiologist is grumbling at everyone because upon entering she noticed that instead of doing something they were all just watching the monitor. Monday: Robert is doing well. His condition has improved noticeably. His technological pack is substantially smaller. If last week there were eight lines and tubes – supplying him with medication, extra blood, oxygen, and nutrition and draining his urine and the fluid from the surgical wound – now only two or three are left. The trajectory of Robert is strongly dominated by technology. The staff is constantly busy watching the monitor, writing down numbers, hooking up tubes and removing them. Do they only focus on the input and output of all the various devices or do they still have attention for the child itself? This question assumes a contradiction between care and technology, whereby attention for technology is conceived to be at the expense of the care for the infant. Such view, however, is based on a definition of technology that altogether ignores its social dimension and a definition of care that ignores its material and technological aspects (Widdershoven and Verkerk, 1995). By conceiving these two dimensions as interrelated, however, the oppositional character of care and technology disappears. Technology does not so much conflict with providing good care, but precisely contributes to it.

Uncertain Trajectories 97

The NICU is a prime example of a social practice that aims at providing care to children with the help of technology. The close interconnection of child and technology in this practice implies that machine-related activities are also care- and child-related activities. Right after Robert’s surgery his instability is such that in the first hours his care is entirely a technological affair: he is only touched, so to speak, by the tubes of the many machines to which he is connected. This does not imply that the staff has nothing to do. On the contrary, Robert requires all the attention they can give, but most interventions take place outside the incubator. Caring for him means checking devices, listing readings, comparing results, and adjusting settings. Rather than being an obstacle, technology functions as a material passageway for the staff to the child as long as his fragile condition requires a ‘hands-off’ policy. In the NICU the child and his/her technological pack cannot be separated. In the NICU attention for the technology implies care for the child and vice versa. Disciplining the treatment practice The medical-technological machinery provides the staff with a major basis for treating and providing care to the children in the NICU. The treatment trajectory is shaped by different kinds of technologies, like a respirator and a patient record. Given the prominent presence of technology in the treatment trajectory of children like Robert, Tom, Esther, and Maureen, it is relevant to analyse its role more closely. For instance, what is the exact function of the respirator in how a specific child’s treatment trajectory takes shape? The practice of having a child on artificial respiration implies that one constantly has to look for the right balance. This calls for precise registering and constant correcting of the direction in which the child is moving. To determine the optimal setting of the respirator, the saturation of the oxygen is determined. This is possible by means of an infra-red light that is attached to the child’s hand or foot. The monitor above the incubator indicates the saturation level measured in percentages. This level can also be determined by taking a blood sample via a line in one of the (main) arteries and subsequent lab analysis. Anti-coagulation drugs have to prevent this line from becoming clogged. Based on the saturation the physician determines the respirator’s setting. If a child’s respiratory rhythm does not run synchronously with that of the respirator, however, the oxygen supply is disturbed. Medication thus serves to suppress the child’s spontaneous breathing, so that he cannot start struggling with the respirator. Moreover, a premature child is also administered surfactant* to improve the ventilation function of the undeveloped lungs. With the

98 Uncertainty in Medical Innovation

help of X-rays and regular blood tests the staff monitors lung development. The single decision of putting a child on the respirator thus implies a number of as yet undefined future decisions. The respirator plays a central role in the treatment of NICU children, the premature ones in particular, which is why its correct usage matters a lot. In the case of the respirator this involves more than just inserting the tube in the trachea and setting the device. It also requires a specific organization of the practice and the availability of specific knowledge and experience. These kinds of facts are not listed in the manual. Nevertheless they are crucial for the working of the device. In other words, in the device a so-called ‘program of actions’ is inscribed (Verbeek, 2005). This implicit user’s manual I refer to as the script of the device. Script as second-order notion allows me to investigate how the usage of medical technologies such as the respirator structures the NICU practice. ‘Thus, like a film script, technical objects define a framework of action together with the actors and the space in which they are supposed to act’ (Akrich, 1992, p. 208). In designing an artefact, such as a respirator, one has a specific use in mind. This script is inscribed into the instrument’s design (inscription) and also prescribes a certain use (prescription).6 This script defines not so much the device’s actual usage but the various roles and tasks of the nurses, physicians and means that are involved in its deployment. This does not imply, however, that staff members always subscribe to this specific usage. They may decide to ignore a device, use it in part, or use it in idiosyncratic ways. Not just technological apparatuses imply a script. The patient record, for instance, also comes with a specific script that prescribes which aspects of the child the physician has to examine and what counts as professional conduct. Other examples include protocols, the airlock system at the entrance of the unit and the monitors above the incubators. They all come with a specific program of actions that disciplines the NICU practice in such a way that their usage or application becomes possible at all. The script that, for example, is inscribed in a protocol disciplines the practice so that despite its heterogeneous, dynamic character it becomes a stable, uniform and predictable whole.7 Let us take a closer look at the script of the respirator. What is implicitly presumed to make it function properly? By attaching Tom to the respirator the staff conforms to a specific ordering of activities that the use of this appliance prescribes. The respirator’s deployment implies a whole range of additional activities. For instance, a careful setting of the respirator requires blood testing on a daily basis in Tom’s case. During each check the nurse writes down

Uncertain Trajectories 99

both the device’s settings and the values on the screen. What is more, the respirator tube may slip out. To prevent Tom from accidentally removing the oxygen tube, it is fixed onto his upper lip with tape that has to be replaced a few times each week. The respirator tube causes irritation in Tom’s throat, which means that a few times a day a nurse has to clear his throat. For optimal respiration Tom has to breathe along with the pace of the device. To prevent him struggling with the tube of the respirator he is administered sedatives. The decision to put him on respiration, then, requires that a series of other activities be added to the treatment trajectory. To make devices function as designed their scripts have to be followed systematically. As such devices have a disciplining effect: the alarm of the monitor above the incubator makes a nurse jump up and check and if needed, intervene. Intravenous infusion of nutrition feeds a child a whole day, but after exactly 24 hours a nurse has to replace it. If this is done too late, the infusion will be empty and this is risky. To avoid this, a perfusor warns the nurse with a loud alarm right before it is empty. But if an infusion is changed too early, the child will not get its full quantity of feeding. The nurse who takes care of the child has no choice: when the alarm sounds he or she has to act. Devices provide care only when used properly or when certain conditions are met. The monitor will guard the child if the electrodes are attached correctly, the lines to the monitor lie freely, and the alarm settings are used. The incubator will keep the humidity at the right level if it has enough water. An infusion system will feed the child if it is properly connected. In a way the NICU children can be seen as users as well. Like all other users of medical devices, they have to follow the inscribed script. The infusion will feed the child if it stays calm, but wild moves may cause the line to slip out of the vein. The respirator will provide the child with oxygen, unless the baby accidentally removes the tube or take breaths out of tune with the rhythm of the respirator. Through their script, devices impose specific norms on the practice.8 The precision of the devices thus compels specific conduct. This is why many interventions, such as taking blood samples and administering sedatives, have become part of the unit’s daily routine. Thus the regular deployment of specific devices allows specific practices to harden into normal practices (Harbers and Koenis, 1997, p. 117). Technology equally plays a role in the structuring of the social order in the unit.9 The usage of devices assumes specific competence and thus distributes tasks and affiliated risks among various persons. It indicates

100 Uncertainty in Medical Innovation

who is qualified to use a device and who is not (yet). The use of complex technology gives rise to a distinction between more or less competent staff members, who is skilful enough to intubate* a child and who has the courage to care for a child like Robert. Thus the script plays a role in how the social order takes shape. This order may cut across formal hierarchical or disciplinary lines. In the Northeastern General Hospital, for instance, one of the neonatologists has great trouble setting infusions. To burden the child no longer, he asks a resident who is known for his unusual skill with ‘difficult’ arteries to take over. He indeed manages to insert the infusion in one try, thus again commanding respect. In other ways as well devices contribute to structuring the NICU’s social order. The treatment team determines which device is deployed, but the technology determines the make-up of the treatment team. The technology that facilitates Robert’s surgery brings other disciplines onto the trajectory: paediatric cardiology, surgery, and anaesthetics. As soon as he needs fewer lines and tubes, these specialists will recede to the background again. Technology binds and unbinds (Harbers and Koenis, 1997). Technology thus structures the NICU social order, but also has different ramifications for various NICU children. Technology provides possibilities, but not to all children in the same way. To a premature child born after a term of 23 weeks the respirator offers fewer chances and more risks than to a child born after 28 weeks. Thus time and again technology constructs new categories of children, while a particular device’s options are not fixed. In the past a child like Robert would have had no chance of survival. Technology enabled his survival. Similarly, until recently children like Esther and Tom had no chance of survival. The script that comes with a device thus draws boundaries and breaks them down. Devices, then, actively contribute to how treatment trajectories evolve. However, their deployment does not turn the NICU into a domain where technology rules supreme – as if the options of human actors are fully dictated by the available technology and its social and moral consequences are given. In this way a critique of instrumentalism would be taken too far and lead to a technological determinism in which technology is the all-decisive factor. A technological-determinist standpoint, in turn, totally ignores the contribution of the users in the structuring of the trajectory and the specific ways in which technology itself also changes in its use. To study to what extent and how technology and other things coshape the treatment process it is necessary to make the inscribed script (inscription) explicit (description). This rendering explicit is a basic step for being able to discuss and, possibly, change the script (Akrich, 1992).

Uncertain Trajectories 101

Technology does not just shape treatment trajectories, but is itself shaped in those trajectories as well. How and in which measure a device’s formal requirements are met depends on how the device is deployed. Thereby technology is just one of the means deployed by the staff in providing care and treatment. Devices, including the specific knowledge and actions they entail, become part of an array of activities and interactions within the NICU. The performance of activities and interventions is determined by countless factors, such as the child’s condition, organizational aspects, time constraints and test results. There is a mutual coordination of technology and user, and only in this dynamic devices take on their specific shape, function and significance. However, each deployment of technology has a temporary status. Initially a treatment trajectory is highly characterized by the presence of medical technology, which supports or takes over a child’s vital functions. As the time passes, however, the child will gradually have to regain control of its bodily functions and get by without technology. This is not to deny that the lives of patients such as Tom and Robert would never have been possible without hi-tech equipment. The deployment of machines, though, is not without risks. This is obvious in the artificial respiration of severely premature children. Following the trajectory of Tom it becomes clear how a life-saving machine like a ventilator can become the reason for concern. Technology as problem Tom’s vital functions, for instance, are quite unstable and insufficiently developed. He lacks the capability to self-regulate his body systems properly. Being born prematurely his lungs are too immature to provide enough oxygen for survival. In these cases, withholding treatment means instantaneous death. If the doctors and parents opt for intervention, the child will be immediately connected to the respirator for oxygen supply. But prolonged mechanical breathing with high oxygen concentrations or high air pressure can damage the fragile lungs of premature babies to such an extent that it reduces their intake capacity to an unacceptable level. This phenomenon, called bronchopulmonary dysplasia (BPD)*, is a ‘calculated risk’ when there is no other option but artificial respiration. In these cases it may prove to be hard – and sometimes even impossible – to get the infant off the machine again. Thus the deployment of the respirator, brought in as a life-sustaining apparatus, also marks the beginning of a balancing act for the staff. Its complexities are visible in how Tom’s treatment evolves.

102 Uncertainty in Medical Innovation

In the first weeks already there are signs that Tom does not respond well to the treatment; his tendency is to get off course. His lungs do not begin to function properly on their own. And worse, Tom is developing BPD. The staff is very aware that if the damage to his lungs will be too large, he will depend on artificial respiration for the remainder of his life, which is why he needs to get off the device as soon as possible. The physicians try to limit the damage by lowering the machine’s activity. This fails however. Bad blood gas values and complications force them to raise the pressure again. Each attempt to get Tom off the respirator seems bound to fail. He responds directly with bad saturations that hardly improve again. Consequently, after four months Tom is still in the NICU. While other infants are dismissed after a few days or weeks, his recovery seems a distant possibility at best. The respirator does not turn out to be the reliable ally the staff hoped for when Tom was hooked up to it right after his birth. The device has gradually transformed from a life-saving tool into a life-threatening one. As such this machine proves to be ambiguous: it has both a constructive and destructive effect on the treatment. The option of taking an alternative route, one in which no respirator is needed, is no longer available. Because of his advanced BPD, Tom lacks the lung capacity and the power to regulate his own breathing. Essentially, the technology itself has cut down on the number of options. This means that the staff has no alternative but to pursue the road already taken – carefully balancing how far it can go and trying to find the point at which the device is most effective and least harmful. Tom’s BPD* is an example of a calculated risk of which the staff was aware when it decided to hook up the child to the respirator. If it involves a decision that is inevitably accompanied by potential risks, in Tom’s case one such risk has become a reality, meaning that the device has a hold over both child and staff. The sheer need for oxygen goes together with damaging effects and this seriously restricts the staff’s intervention options. The same technology that right after birth created room to move, thus opening up a specific trajectory toward recovery, some weeks later blocks all other potential treatment options. As such a respirator’s script pre-structures the available treatment options and it may render situations more complex, force one to change course, or block any alternative options, thus in part defining the actual treatment trajectory. In the months that follow additional problems crop up. As Tom grows larger, his agility increases. To prevent him from accidentally removing his oxygen tube, he is given sedatives. There are other complications as well. The fragile veins in his head failed to withstand the blood pressure

Uncertain Trajectories 103

and have burst. This brain haemorrhage causes an increase of the pressure in his head, which requires a drain to discharge fluids. Apart from these various complications, Tom contracts several infections. Given this long list of problems it is no wonder that as many as four times he entirely collapsed, his heartbeat going down without automatic correction. At the explicit request of his parents, the doctors do everything to save him and each time successfully so. What they do not succeed in, however, is to get him off the respirator. Other challenges present themselves as well: Academic Medical Center Thursday: Tom has had trouble eating for several days. This morning’s test results establish that he has an ulcer. During their round the doctors discuss his situation. They suspect it is caused by the corticosteroids. Monday: Tom appears to have become addicted to sedatives. Each effort to cut back on sedatives he responds to with bad blood results. Tuesday: Tom has been having coughing fits for days. The tube in his throat irritates him so much that it causes mucus build-up. He tries to get rid of the mucus crying and coughing. This in turn makes him feel oppressive, what makes him cry even louder. Silently, for the tube in his throat makes it impossible for him to produce sounds. The nurses try to break this vicious circle by regularly cleaning his throat. But this treatment is torture to him so that he starts crying and becomes short of breath again. It’s enough to make the nurses despair. In the afternoon Tom’s mother and grandma come to visit. They do not know either how to calm him down. Whatever they try, nothing helps. He continues to cry. Ultimately, grandma is no longer able to handle it. Crying she waits for her daughter in the airlock. While Tom’s mother is standing next to the incubator a nurse approaches her, asking her to stay around for a while because the doctor wants to talk to her. It gives her a fright. ‘What would be the matter this time? If doctors want to talk to you it is never innocent.’ Her suspicion confirmed, the resident tells her that Tom probably has spasms. It also turns out

104 Uncertainty in Medical Innovation

that the drain in his head is leaking a bit. It is not alarming, the physician adds, but they have to keep an eye on it. The treatment of Tom shows how the side-effects of interventions can develop into a substantial problem. As problems are accumulating, it becomes ever harder to keep the child on track and the staff constantly has to map out new routes. When complications occur, new treatment protocols are added to the trajectory. This causes a condensation of the number of treatment lines that converge in the child. Each new intervention needs to be given a place in a growing system of interactions, but this is not always easy, nor does it keep other treatment options open. What to do with respiration pressure or medication that cannot be gradually reduced? One gridlock after the other may present itself. Unintended effects on other medication or physical processes may give rise to new problems that call for new interventions. A ‘locomotive phenomenon’ comes into being, whereby one intervention elicits another (Guillemin and Holmstrom, 1986, p. 131). The pattern of interventions loses its coherence while the initial goal – the child’s full recovery – has to be adjusted to ‘second best’. In situations like these, the staff runs up against the boundaries of the NICU practice and it is reminded that eventually its mother’s womb is the best place for a child to spend the first nine months. The malleability of the practice is limited by natural and technological boundaries. A respirator with high pressure, side-effects of medication, bacteria, fragile tissue – all these may be encountered along the way. A difficult case like Tom demands constant adjustments and involves much more than just guiding a child along a well-paved path. Specific treatment trajectories in the NICU are the outcome of a delicate interplay of multiple actors and factors, including a child’s responses to them. Complex treatment trajectories make it ever harder to maintain a sense of overview and predict the result of interventions. But numbers commonly provide a basis. The medical technology in the NICU produces each day a flow of quantitative information and activities. The babies are connected to machines with numerical input and output; lab tests are executed, yielding fractional data; nutrition and medication are carefully calculated, calibrated to a specific pace. Staff members collect numbers, calculate ratio and proportions interact in numerical discourse, and list numerical outcomes on special forms suited to that purpose. In what specific way do these ongoing calculations constitute the NICU’s daily routine activities? In other words, what is the position of numbers on a NICU?

Uncertain Trajectories 105

4.3

Numbers as compass

If the staff members want to know how the child is doing, they check the numbers in the patient record and the flow sheet or throw a quick glance at the monitor. They read HB values, saturation values and input and output quantities of the fluid balance, instead of ‘paleness’ or ‘fluid loss.’ Likewise, the transition between shifts and rounds are marked by the exchange of a great number of all sorts of numbers: hours, weeks, days, frequencies, percentages, pressure, sizes, and assessments. As a result, the condition of the neonate is increasingly represented by a variety of numbers: numbers related to technological devices as input or output measures; numbers generated by the staff, such as pulse rate or calculated ratios; numbers that are visualized and displayed on a monitor or verbalized in discussions during rounds. In the NICU numbers are not always considered hard facts. Often numerical data count as an objective representation of the child’s physical condition, but they can also be the focus of a negotiation process. How is that possible? This section’s concern is not so much with the intrinsic value of numbers, but with their usage and meaning in the construction of knowledge.10 A reliable instrument Through exquisite technological instruments, all kinds of body processes are measured and visualized in a numerical language. Because of its factual precision and comparability, a numerical transcription of the body is considered more reliable than its qualitative counterpart. The deployment of numbers is reflective of ‘mathematics of certainty’ (Gigerenzer, 1990). This quality is thought to eradicate the subjective dimension of knowledge, which is seen as an obvious obstruction in the mutual exchange of data. The unambiguous nature of numbers allows one to discuss them side by side. Numbers are considered hard facts, objective measurements of reality. They suggest a standardized method of reasoning that can reveal the nature of the problem objectively. In the ceaseless collection and comparison of numbers, medical personnel find a compass that guides them. Academic Medical Center The resident is studying the lab results that came in a minute ago: glucoses – 0.7 mmol/l. A quick comparison with the normal reading tells that Ronald is hypoglycemic*. The resident decides to give him once a bolus of 2 ml glucose 20 percent. Half an

106 Uncertainty in Medical Innovation

hour later this intervention proves to fall short. Though less serious, the lab results still reveal a hypoglycemia of 0.9 mmol/l. Therefore he decides to an intravenous infusion of 15 percent glucoses. But this is not the only problem with which the child is struggling. The monitor reveals a dropping blood pressure: 47/33, 45/29, 38/25. The resident decides to raise the cardiotonica dose so as to make the blood pressure go up again. In the NICU physicians are guided by numbers rather than patients’ complaints. The effects of their intervention are reflected in new results that tell if things are moving into the right direction. Through their relative exact and unequivocal nature, numbers are a major component of the NICU’s practical order. By translating physiological processes into a numerical language, detailed measurements become possible. Comparison of past, present, and future medical data reveals the course of internal physical processes through time. Is the infant more or less stable? Is he recovering or is his condition worsening? By evaluating factual representations of bodily processes, the medical staff attempts to detect the presence and extent of pathological processes. This too is ideally expressed accurately, with the help of detailed numbers. Because also the functioning of medical devices is represented in numerical units, doctors can set their devices with great precision and raise or lower the medication with minimal quantities. The quantitative description of both the devices’ settings and the child’s condition provides the staff with a shared language. This contributes to avoiding lack of clarity and misunderstandings. No vague description in terms of ‘add a little more oxygen’ or ‘he has trouble breathing somewhat,’ but numbers that allow one interpretation only. In a quantitative language they tell each other how the child is doing and to what extent the treatment needs to be adjusted. The unambiguousness, comparability, and accuracy of numbers has led to the situation that a quantitative picture of the body counts as more reliable than a qualitative picture. Furthermore, in other medical and social domains numbers count as objective data as well. This explains why carefully collecting and recording data reaches beyond medical practice. Notably in the United States juridical interests easily come into play in all sorts of practices, including medicine. This high degree of juridicalization forces doctors to found their decisions more solidly on quantitative data. To avoid court cases there has been a tendency to deploy all sorts of means and technological tools to generate a large body of test results. Because also in court numbers function as objective and

Uncertain Trajectories 107

trustworthy data. As a neonatologist from the Northeastern General Hospital tells about this aspect: We run less risk to be persecuted because we do all we can. We have it all here and we also use it. Thus they can never charge us with negligence. What we don’t do does not exist. So no one has ground for lodging complaints. If there is still a malpractice suit, the numbers will establish that the right decisions were taken. In the decision-making process, doctors do not only have to act in accordance with the standard level of care, but they also have to reckon with juridical rules. In the uncertain world of neonatology numbers provide the only basis for legal certainty. To have evidence in advance, much blood-testing is requested and each medical step is recorded. Thus the legal system in the world outside the hospital further pushes up the production of numbers and their recording in medical practice. The interpretation of numbers At first sight, medical intervention in the NICU appears mainly a matter of gathering and comparing numbers. Based on quantitative facts doctors opt for a particular treatment. If they want to know whether it is successful, they gather new numbers and compare them to the old ones. However, quantitative data seldom speak for themselves; often they have to be made informative, implying the generation of other activities yet again. Academic Medical Center The neonatologist comes in and they can start with their round. The resident offers his diagnosis based on the extra test results he now can present. Given the fluctuating blood pressure he suspects that Ronald has had a brain haemorrhage. He already put in a request for an ultrasound scan. The attending, though, is not so easily satisfied: ‘Of course, this conclusion is quite possible and the ultrasound scan will have to show if you’re right. But more roads lead to Rome, so we should not exclude other potential causes for this erratic picture.’ They talk about the issue in terms of ‘it could be this, but I do not exclude that.., and did you think about this.’ Various possibilities pass in review. The result of this

108 Uncertainty in Medical Innovation

discussion is a new series of blood tests and hence a new series of numbers. Significantly, numbers do not tell the staff members what to do: they have to decide for themselves. In so doing, the numbers may serve as a basis. For instance, what does raised or lowered blood values tell about the cause? The test results can point to several possible causes that require divergent interventions. In order to become effective, quantitative data must be embedded within a frame of reference. Outside an interpretive framework, numbers lose their meaning. Numbers are not isolated entities; they are related to other numbers within particular contexts of practice and meaning. Where numbers correspond to other numbers, each set may be made more meaningful. In this way making a diagnosis is a matter of excluding possibilities. However, the competence to interpret numbers correctly largely depends on knowledge and experience. Moreover, numbers are not always unambiguous and they may even be a source of confusion: Academic Medical Center The resident is standing near the incubator of Ronald. With concern he is looking at the numbers on the screen and doesn’t quite know what to do. The numbers display an erratic picture of the blood pressure: 55, 31, 38, 50; heart action: 181, 167, 158, 199. He checks the flow sheet. Its picture, too, is not very helpful: sodium – 124; calcium – 3.2; potassium – 4.3. Again he switches to the monitor. ‘What is this? Am I missing something?’ He decides to have additional blood tests. Maybe those numbers will provide a better picture of the situation. At once the unproblematic trajectory of gathering and comparing is no longer self-evident. Do the numbers point to a problem? Is there a serious situation, or is the child likely to come round soon? A young doctor with little experience in the NICU may be hesitant in certain situations. The hierarchy in a hospital leaves more or less room for uncertainty. A paediatrician who is still in training is inclined to blame his lack of insight on his lack of experience.11 There is, in other words, a process of distribution, whereby certainty is attributed to the numbers while conceiving his own expertise as marked by uncertainties. The question ‘What is going on?’ is replaced with ‘What am I doing wrong?’ In his indecision, the resident also loses sight of the child. Although usually numbers are helpful, this time he does not know what to do

Uncertain Trajectories 109

with them. Through their inconsistencies the numbers on the monitor have lost their power to resolve a situation and have become a problem in their own right. The numbers create confusion and raise doubts, thus breaking the once so natural unity between the child and its representations. To remove this uncertainty the resident decides to gather still more information. If this renders the already existing dilemma even more complicated, numbers are often deployed as a crowbar to move beyond an impasse. If the treatment trajectory has faltered in a pulp of data, new data may give a new impulse. Fresh blood samples disappear in the pneumatic dispatch. The new results that come back via the computer have to break the stalemate. Number versus number Doctors’ numerical tools work, but only under certain conditions. First, the right interpretation of numbers depends on specific knowledge and experience. Second, numbers have to fit in a pattern. But what if numbers contradict each other? Academic Medical Center The resident looks at Ronald’s blood results. They look better than on the previous days. The child seems stable. Or does he? The latest blood results already seem to announce the following wave of distress. The resident consults the attending: ‘I fear this child will start to haunt us again. His latest bilirubin* test results fluctuate between 50 and 200.’ The attending looks at the other results. ‘This is not possible. This has to be a lab mistake. Just look, the other values do not point into that direction at all. What’s more, such fluctuation is impossible. Do it once again and you will find it to be normal.’ Relieved the resident walks away to the table to write a new request, hoping to get reliable data in return and write off the previous result as erroneous. The physicians do not simply accept the result of the bilirubin test, but assess its value in light of the other data about the child. Even if the results are crystal-clear, the hardness of their factuality proves not yet guaranteed. Whether the result is valuable or useless is certainly reason for doubts and is here also dependent on other results. The established data on the child constitute a whole in which one data is confirmed or undermined by another one. But in this dynamic the bilirubin result is all alone, lacking all corroboration by some other data. The bilirubin

110 Uncertainty in Medical Innovation

result does not fit into the image that the other results provide. Each effort to link up the result with other data leads to frictions, rendering a solid embedding in the existing configuration of information impossible. The second test should produce a result that does fit the picture. The measure in which numbers are considered to be guiding proves to depend on knowledge and experience, as well as on other numbers. In contrast to the erratic values on the monitor, the bilirubin result is understood but not corroborated. If a result is to serve as basis for action, it has to be anchored in a network of other data. Physicians collect data and challenge them to be able to get a logical clinical picture. But what to do when the same incomprehensible test results persistently come back? Northeastern General Hospital Christine is lying in an open incubator. A heater above her keeps her temperature steady. For two weeks she has had high blood pressure in the right half of her heart and no one is able to explain it. At first the staff members hoped that it would simply go away by itself. They thought they were perhaps dealing with a persisting foetal circulation. This means that the blood circulation still follows the same direction as during the intra-uterine phase. As always in such cases, medication is administered, but this time without result. This is why more data are gathered. Although the stack of numbers continues to grow, the riddle remains unsolved. Other specialists are called in, but there is no indication on how Christine’s predicament can be ended. One ultrasound scan after the other is done, new medications are tried, and nothing works. Each measurement reveals that the pressure is too high: 90, 88, 96. All other values, however, remain normal. It boggles everyone’s mind. If one staff member advances a possible cause, another one dismisses it with reference to the other values that are normal. Exasperated by the contradictory values, each day the staff runs up against its lack of power to solve the riddle called Christine. But isn’t there an obvious solution? The numbers on the right ventricle simply do not fit. However, in this case the result’s persistence urges the staff to take the numbers seriously, while the contradiction cannot be marginalized as insignificant. The doctors first try to solve the problem by

Uncertain Trajectories 111

gathering more data. Medication is deployed and the pressure is remeasured each time with various devices, other data are collected and combined with the earlier data. But the contradiction between the numbers continues to exist. Next, they try to refer the child to a cardiologist. Although he refuses to take the responsibility for this child’s treatment, he briefly looks at the problem, but is unable to solve it. Meanwhile, they are silently hoping that as time passes, the problem will still go away by itself. But this doesn’t happen. Ultimately the staff decides to refer the child to another hospital that has other equipment. Perhaps they will manage to solve the riddle. It becomes clear here that not each doubt can be resolved through collecting numbers. Not a single effort helps to change the ambiguous picture: raised blood pressure in the right ventricle in combination with normal values. In this case the blood pressure result is too much anchored in a heterogeneous network to dismiss it as an error. Competing knowledge frames In the NICU relevant knowledge and experience are spread unevenly among those involved. Of course, for parents it is often hard to interpret all they see. Although they do not have to take medical decisions, this is not to say they ignore all the numbers. Academic Medical Center In the course of the afternoon Ronald’s parents again drop in. His mother appears exhausted. When she is watching the monitor, suddenly the alarm goes off. Both parents are looking anxiously to their child. The nurse responds instantly and checks the screen. It shows strongly fluctuating breathing. She turns off the alarm and looks into the incubator. The parents watch her fearfully. ‘Is there something wrong?’ Ronald in fact is sound asleep, his chest going up and down rhythmically. No, he is doing fine. It is the electrode that registers the respiration that caused all the panic. The nurse shows the loose tape to the parents and attaches the electrode again to the child. Instantly a stable breathing frequency appears on the monitor. The parents are confronted here with a high breathing frequency accompanied by a loud alarm, which is ample reason to get worried. The nurse, however, is not automatically convinced by the numbers. The child that is soundly asleep in the incubator renders the numbers

112 Uncertainty in Medical Innovation

on the monitor unreliable. The loose tape on Ronald’s body subsequently makes it plausible that his chest that is regularly moving up and down offers better information than the fluctuating numbers on the screen. If the parents, understandably, are put in a flurry by the alarm, the experienced nurse, who immediately notices the detached electrode and the child’s clinical picture, has no trouble distributing doubt and certainty. Conversely, clinical symptoms do not always determine things. In some cases, like in Peter’s, knowledge and experience or other numbers are not so much decisive, but other sources of knowledge. Northeastern General Hospital In the corner of the NICU, Peter, a premature baby, unexpectedly enters in a serious crisis. The nurse-practitioner is called in right away. Peter is terribly pale. The monitor indicates very high PCO2, meaning that the child’s blood is highly acidotic*. Until now he was doing so well that this morning his oxygen tube was removed and he switched to a non-invasive form of respiration. Although all day Peter continued to have good blood gas values, now he no longer seems able to cope. The nurses urge for re-intubation. But the nurse-practitioner decides to wait for the blood results that should be back anytime. After several minutes of anxiously watching Peter, the result comes in: the child has excellent values. Meanwhile the PCO2 went down as well. The nurse-practitioner is glad she did not intubate. The nurse-practitioner: ‘I have to acknowledge that briefly I got the wind up as well. They all suggested intubation, but I have seen this happen to Peter before and each time he manages to climb out of it on his own. This is why I wanted to wait as long as possible with invasive treatment. I decided to wait for the result and not immediately intubate him again because I saw that the PCO2 started to go down. But, as you know, if his blood results would have been bad, I would have intubated him after all. But I don’t like to intubate this child. It always makes him feel so uneasy that he gets into a crisis. And that doesn’t help either. Often, you really have to search for a solution. Sometimes the numbers don’t even match. You saw how high his PCO2 values were. And you also saw how intensely pale he was. Yet his

Uncertain Trajectories 113

saturation was perfect and his other results as well. Sometimes you don’t know what to trust. It seems to be rather a matter of the interaction between your actions and the child’s reactions, whereby your knowledge is a major aid. But it is certainly not the only thing to depend on, because at times the children here respond very differently than expected.’ The nurse practitioner ultimately chooses the least aggressive intervention. This decision is not so much courageous as it is based on experience. For experienced doctors the meaning of quantitative data exceeds labels like ‘normal’ or ‘abnormal’. Rather, their interpretation is a practical issue of what is acceptable or unacceptable in that specific case. This offers more space for action and the possibility, in the case of gridlocks, to move the treatment process forward again. The NICU as a quantitative practice As we have seen, numbers function much like a compass that guides NICU staff members in their effort to pilot the infant safely along the planned treatment trajectory. By constantly gathering numbers, calculating ratios and quantities, entering quantitative data on special forms and classifying them, and by interacting with each other in numerical language, the staff is able to monitor the condition of the child. Numbers, however, do not speak for themselves, nor can a large flow of quantitative data do away with each and every uncertainty: medical intervention trajectories in the NICU are always rife with risks and uncertainties. As tools, quantitative input and output are constructed in their actual application in practice. This contributes to a tension between contradictory numerical frameworks. On the one hand, numbers are considered as objective reflections of reality; on the other hand, they are constitutive elements of this same reality. In other words, numbers do not possess a solidity of their own, but become facts through processes of negotiation. Although claiming and seeming to rely on the exactness and objectivity of digital data, the staff engages in continuous negotiations about their meaning. A closer look at their daily activities reveals the contextual reframing of numbers. Knowledge derived from hard numbers cannot be applied irrespective of their context. There are always situation-specific activities and decisions. They are related to other numbers within particular contexts of practice and meaning. However, the existence and knowledge of such a framework is no guarantee that medical decisions and interventions will proceed smoothly. The distribution of numerical certainty and

114 Uncertainty in Medical Innovation

credibility appears not to be arbitrary, but is rather affected by contextual constraints and possibilities. Again, there can be a mutual confirmation of numbers or continuing struggles for credibility. Analysis of this contingent interaction shows how the meaning of numbers in terms of true or false or certain or meaningless, is the outcome of a distribution of credibility within a network of quantitative data and other frames of knowledge. As such, the meaning of numbers is shaped by specific contexts. Incessantly, so it seems, NICU staff members gather numbers and calculate ratios and quantities. They interact in a numerical language and classify all sorts of quantitative data on special forms. The examples discussed illustrate, however, that the ongoing flow of data cannot avoid that uncertainty continues to be a major element of the treatment trajectories. Working in the NICU implies quantitative dexterity and insight in numbers for doctors and nurses alike. The NICU is a quantitative practice in which both man and machine contribute to the numerical edifice erected around the child.12 Rather than the stethoscope, pen and calculator are the tools most used here. The cases of Tom, Ronald and Robert reveal how the NICU staff deploys preventive measures, constant monitoring, and an ongoing flow of quantitative data to pilot a child along a complex trajectory. However, in the decision-making process, doctors do not only have to act in accordance with the standard level of care, but they also have to reckon with juridical rules. Here we see the consequences of the medical domain’s public accountability for the actual intervention process. It shows how legal frames affect hospital practices. A trajectory marked by erratic and unexpected turns does not automatically exclude a happy ending. Although in the case of Robert not all works out with complications, in some ways he is the ideal patient. Robert’s treatment demonstrates the potential of neonatology at its extreme. Born as a child with little or no chance of survival, he was turned into a child with normal life expectancy. His condition stabilized and gradually he was rid of his technological pack, which allowed for the re-emergence of the baby in him. After two weeks he proved to do so well that minimal monitoring sufficed. He was returned to the hospital where he was born, from which he was dismissed after just one week: sent home, healthy and well. NICU staff members gain experience in dealing with the risks that turn up at specific moments and they learn to take them into account in a trajectory’s planning. My account of Robert’s treatment suggests how the staff manages to pilot an infant along an extremely risky and

Uncertain Trajectories 115

complex trajectory. Precautionary measures and ongoing monitoring may avoid much suffering, but unforeseen and undesirable developments are always possible. Calculated risks or unanticipated complications may still cause a treatment trajectory to founder. Their hands tied, staff members can only watch how the infant increasingly strays from the established course. Tom’s case shows how a sense of control has to make room for contingency and powerlessness if a trajectory becomes more complicated as it unfolds. His treatment trajectory reveals how he – despite all knowledge, numbers and medical technology – moves away ever further from his planned destination rather than coming nearer to it. The result is that after four months he is still in the NICU. At this point, the optimism that to some degree is always part of treatment efforts has nearly evaporated. In his case the medical staff has failed to realize its objective, and this forces it to reconsider the objectives of treatment and reformulate the child’s prognosis. But how, exactly, does one arrive at a prognosis? This question is the subject of the next chapter.

5 Beacons on the Horizon

Academic Medical Center The parents of Tom are standing at the side of his incubator. His father is checking the flow sheet. His mother winds up a toy bear and carefully puts it next to her son’s head. He is not really looking at her. One can hardly hear the toy bear’s music amidst all the other sounds – the hum of the respirator, the beeping of monitors, a ringing phone – and this painfully suggests that the bear was designed for use in the silent seclusion of the baby room at home. Tom has now been in the NICU for over four months. His condition has fluctuated a lot during the last few days. If one moment he appears to look well, an hour or so later his condition may have seriously deteriorated. Tuesday, 9:30 hours: Two nurses are bent over his incubator. One holds his head while the other fixes the oxygen tube with tape. Tom tries hard to resist the procedure. With soothing words the nurses try to calm him. But to no avail. The monitor above the incubator is blinking rapidly. ‘Wow, he grows stronger all the time.’ The nurse again tightens her grip. ‘It is already the second time he pushed away the tube. And yesterday his respirator pressure went up again to 35. This doesn’t get us very far. I wonder where this is leading.’ ‘I’m afraid we’re cultivating a second Mick here; in the end we did not manage to get him off respiration either.’ 11:00 hours: The medical staff is doing its round. Today the social worker joins in as well. She knows Tom and asks the neonatologist how far they will go in treating him. She is clearly not the only one interested, as the nurses eagerly listen as well. 116

Beacons on the Horizon 117

The neonatologist: ‘Given that the child is still doing well from a neurological angle, there is no reason to change course. Theoretically, moreover, he still has a chance of getting off the respirator. But we are aware that this chance is diminishing over time. This is precisely the problem with BPD children. The longer they’re on artificial respiration, the harder it is for us to get them off the respirator again. Therefore we try to reduce the pressure. But so far our efforts fail because of the bad test results and we have to go back to the old level. If there is really no chance of improvement, we’ll have a problem and need to discuss what to do. But as long as there is a chance we will go on.’ No one responds, so they move on to the next child. Over lunch some of the nurses are discussing Tom’s predicament. One nurse, who is still in training, tells that his case was addressed in an ethics class: ‘Eighty percent of the nurses do not want to go on with this child because he will never get off the respirator. We continue to provide care, but at night I sometimes can’t sleep because of his sorry case. The doctors, with the exception of Rijnders, want to go on, notably Geert Hogebeemt, probably because a few weeks ago he managed to get Tom out of a crisis. After all this effort he now wants to go all out. Rijnders and Margaret have their doubts. But, well, he is the youngest neonatologist and she is only a resident. I do not understand why they keep going on any longer because, as you can see, this child’s chances are slight and if he will live, what kind of life are we talking about?’ Tom’s parents see his condition get worse as well. But in contrast to the staff they concentrate on his recurring ability to cope with it and improve. His father: ‘We feel the physicians to be very pessimistic. Four times already they told us that Tom was not going to make it anymore. But so far each time he proved them to be mistaken. Perhaps he is stronger than everyone believes. This is why we feel they should not give up. After all, the next time things go wrong he may again recover.’ Tom’s condition is serious indeed and as his problems have increased, it has even become an issue whether or not they should go on treating

118 Uncertainty in Medical Innovation

him. The various people involved, however, assess his condition quite differently, and this is also reflected in their prognosis. Some nurses view Tom as a child whose suffering is no longer acceptable. They consider this suffering to be caused by the physicians involved, who do not know ‘when to stop’ and who go on treating the infant largely because the various technologies at their disposal allow them to do so. The doctor in attendance also considers him a child with a bad prognosis, but he refers to research of children in a similar situation that clearly suggests there is still a chance of his pulling through. This is why this doctor feels that stopping treatment is no real option yet. After four difficult months, Tom’s parents have an entirely different take on the situation. They see their son fighting for his life and each time recovering, which is why they continue to be hopeful. Why do individuals and groups involved – doctors, nurses, and parents – arrive at quite different prognoses of a specific case? Where do these prognostic differences come from? The aim of this chapter is to explore the origins of prognostic differences and the role played by these various aspects in shaping them. Such differences are not just relevant among the professionals involved. In their own ways, the parents of NICU children also actively pursue and produce prognostic knowledge. Despite the daunting circumstances and the always very specific NICU context, they too develop their modes and mechanisms of prognosis. A study by Anspach (1993) has revealed how the organization of the NICU as a work environment provides divergent frames of knowledge for ordering the NICU reality. The specific position of doctors and nurses, including their respective tasks and expertise, influences their perception of a child’s treatment and prognosis. The work of Anspach provides an understanding of the various positions that allows me to examine the emergence of various prognoses. This will be my focus in the first part of this chapter. In the second part of this chapter I address the multi-layered dynamics at work in the process of prognostication by identifying additional features of prognostication. In particular I concentrate on the specific elements that define a child’s condition and the effect of who happens to be in charge of the treatment trajectory. In the last part wider concerns, such as the input of technology, the styles of reasoning actors rely on, the local history of the nursery, the spatial and organizational structure of the NICU involved, and the personal knowledge and experience of the staff and parents will be taken into

Beacons on the Horizon 119

account as well. Finally, my analysis shows that a number of factors that are more or less tied to competence and technology, as well as the specific spatial and temporal order of the concrete setting, contribute to actors’ articulation of different prognoses. Before developing these concerns, I comment on ‘position’ as an analytical concept for explaining the relation between formal roles and the production of prognostic knowledge.

5.1

Position and prognosis

In the struggle for wanting to be put in the right, the views of others can be easily interpreted psychologically. Tom’s parents, for instance, blame the doctors for their pessimistic attitude. The staff members in their turn interpret the interpretation of Tom’s parents as a case of denial. They would not yet be able to face their child’s reality and let go of him. Also the view of Maureen’s parents is thus explained. A neonatologist from the Northeastern General Hospital: ‘So far her parents have denied her situation. I want to see them at any rate, so as to explain to them again what the current situation is. That we are ready to end this situation does not say that this also applies to the parents.’ A nurse who cares for Tom also has an explanation for the staff’s dividedness: ‘As soon as something happens with Tom, they [the physicians] go all out to get him back on track. It seems as if they do not know how to stop. And then they point to his parents. They want us to do everything we can. But, well, their decision also depends on how the doctors explain the situation to them. When they include a positive note in their story, parents immediately jump on it of course.’ Another nurse adds: ‘I have the feeling that physicians increasingly hide behind theoretical chances and use all hi-tech they have at their disposal, so they don’t have to take responsibility for stopping a treatment. It seems as if they no longer dare to take such decision

120 Uncertainty in Medical Innovation

and therefore keep postponing it. Thus a child like Tom gets no chance to die.’ One of the neonatologists in the Academic Medical Center: ‘When the unit is full of critically ill children this has repercussions on the nursing staff. As if this burden exceeds their bearing power. Then a chronic child such as Tom is like the last drop that makes the cup run over.’ The situation around Tom’s incubator is similar to that of Maureen. In the Northeastern General Hospital the prognosis of Maureen is also subject to discussion; moreover, there is a similar distribution of views on whether or not to pursue further treatment, one that seems largely based on the role or position of those involved. As in Tom’s case, the parents indicate that Maureen’s treatment has to be continued, the nurses at one point prove to be reluctant about further treatment, while the neonatologist is the one to argue that the child still has a chance to survive and that withholding treatment is therefore no serious option. In general, nurses are quicker to question the continuation of a treatment than physicians (Zussman, 1992, p. 69, n. 7). This is not to say, however, that always the same positions are taken. It is possible that physicians favour a DNR policy while nurses object to withholding treatment (Eliasson et al., 1997). The underlying concerns related to these perhaps stereotypical roles – denying parents, unstoppable physicians, and emotional nurses – only come to the fore in the NICU when serious differences of opinion emerge that also linger. As such these concerns become part of a rhetorical struggle, causing parties to dismiss the views of the other parties as irrational or as tied to fear or denial, which at the same serves as justification for one’s own opinion. Who is right? Is the response of the nurses involved too emotional? Are Tom’s parents denying the reality of their child’s condition? Are doctors too much focused on technological possibilities and treatment opportunities? A simple yes or no does not just reinforce stereotypes; it also fails to do justice to the motives of those involved and to the complexity of the situation. Therefore it is more pertinent to find out what accounts for these prognostic differences. Where do they come from and how are they constructed? To answer this question I first take a look at the role of the social or professional position of actors in determining a prognosis.

Beacons on the Horizon 121

Therefore, I will study the role of both their perspectives and their actions in this subsequent order. The specificity of the perspective The notion of ‘position’ offers a good starting-point for studying the emergence of various prognoses in divergent social groups, because it assumes a relationship between the interpretation of a situation and the perspective from which it is perceived.1 From a particular position one may observe something that cannot be seen from another position, or only in part. As such ‘position’ may also refer to someone’s role in a social order. What one specifically notices or perceives in a NICU setting depends on one’s social or professional role in this practice. Parents look at their child’s situation with other eyes than staff members. Parents primarily see a child that is their own. The staff’s more professional gaze perceives a severely ill child admitted to their unit. Although physicians and nurses consider the child’s recovery as the collective aim of their effort, they each have their own focus, respectively, treatment and care. The various positions provide divergent frames for ordering their practice’s reality. Each frame holds assumptions and views that influence perceptions.2 Accordingly, specific aspects of the situation are foregrounded while others are ignored. The position in part determines what is seen, what counts as foreground or background and how this is interpreted. Another position will reveal another world, another child. How do the NICU actors deal with these differences? In the everyday NICU practice different perspectives on a child basically exist side by side. Such differences are even seen as functional. Each position provides a unique perspective on a child’s condition and the combination of these various perspectives allows for a careful and more complete mapping of his situation. But differences may also be counterproductive rather than complementary. The different positions may give rise to contrasting interpretations. Strikingly, all parties claim to have a more realistic picture of a child’s situation. For instance, physicians claim to rely on their technical competence – read: scientific objectivity – to give a more realistic picture than the ‘more emotional’ gaze of nurses, while nurses claim to do more than just consider the numbers. How differences in perspective are no longer seen as functional can be illustrated by the case of Tom. In their own way all actors initially consider him a child with a chance that needs maximal intervention. They do so despite their different positions or roles. But gradually, as

122 Uncertainty in Medical Innovation

suggested by the nurses quoted above, this consensus began to erode, and the functional, complementary nature of having divergent perspectives starts disappearing. The various views become linked up with competing treatment trajectories and hence different prognoses. At first, after Tom’s admission, his future is open-ended and all actors accept the uncertainties involved in his treatment trajectory. When after some weeks problems begin accumulating, the staff starts losing its hold on Tom’s situation. His treatment proves all but a routine matter and its outcome becomes less and less predictable. This doubt allows room for different interpretations of what is ahead. The initial consensus on his prognosis is gradually undermined. Tom’s parents see him as a fighter who will also take the next hurdle. This is why they do not really consider another treatment trajectory as an option. Although the attending does not share their optimistic view, he also feels that a radical change of course is not yet necessary. There is still a theoretical chance that the treatment will be successful. Although the parents and the physician each have different prognoses, they prefer the same treatment trajectory. However, some of the nurses not only have another perspective on Tom’s future, but their prognosis implies a radical change of direction: abstaining from further treatment. Thus the complementary nature of different positions disappears once different prognoses give rise to the formulation of competing treatment trajectories. However, the specificity of the perspectives is closely related to their functions and as such to their set of activities. What, then, is the role of acting in the process of prognostication? The specificity of acting Anspach (1993), who analysed cases in which doctors and nurses disagree about a baby’s prognosis as well, has convincingly demonstrated how the position of doctors and nurses involves particular tasks that give access to specific sources of information, which in turn results in various kinds of knowledge. For example, the output of diagnostic equipment, such as blood test results, monitor data, ultrasound scans and X-rays, contributes to what can be labelled ‘technical knowledge,’ while direct observation – in the sense of seeing, listening and touching – leads to ‘observational knowledge.’ A third category Anspach identifies, ‘interactive knowledge,’ derives from social contact with the child. The extent to which those involved (are able to) make use of these various sorts of knowledge appears closely tied to their set of tasks. Consequently all participants have ‘partial’ and selective views. It turns out that neonatologists, for instance, have more technical knowledge

Beacons on the Horizon 123

than interactive knowledge. By contrast, nurses, given the nature of their tasks, have more or less access to all three sorts of knowledge. Various social positions, then, lead to various activities and these generate divergent sorts of prognostic knowledge. The concept of ‘position’ allows us to examine the emergence of various prognoses among divergent social groups. Such knowledge cannot be understood without taking into account the organizational structure that generates it. In this context Anspach (1993) refers to the NICU as ‘ecology of knowledge’ (p. 60). In this approach, the production of (prognostic) knowledge is explicitly situated in the specific circumstance in which it is produced. Insight into the effects of the various professional positions in this ecology largely accounts for prognostic differences and their emergence. By paying attention to the link between professional position and prognostic knowledge there is the risk that parents as producers of such knowledge are ignored. Parents, however, also develop their own mode of prognostication. It is important, therefore, to study how parents try to find their way within the complexities of NICU life and how they arrive at specific expectations about their child and his future.3 The existence of various positions renders the emergence of divergent prognoses insightful. The specific position of the actors determines what they see or can see, but also what they can do. By paying attention to the pragmatic level of acting in the analysis of the prognoses articulated, the notion of ‘position’ acquires a broader theoretical reach that also strengthens its explanatory force.4 As such, the articulation of specific prognoses proves to be closely related to the various social positions and their associated set of tasks. Although a focus on ‘position’ can account for several essential aspects of the prognostic process, there are also circumstances in which the specificity of roles hardly provides more insight into prognostic differences. By assuming a close link between professional position and prognostic knowledge there is the danger that prognoses are merely seen as a group-bound issue, leaving no room for conflicting prognoses within a single group. Furthermore, several factors will remain underexposed, including the specificity of the situation, the history of the unit and the significance of personal knowledge and experience. This is immediately clear when, for instance, we want to understand the occurrence of prognostic differences among nurses. Where do these within-group differences come from if they cannot be explained on the basis of different functions or social roles? To study these differences in expectations of Tom’s future requires a closer look at the experiences of

124 Uncertainty in Medical Innovation

those involved. Before I do so, I will first discuss the notion of prognosis as being defined by Koselleck (1985). His framework gives me the opportunity to avoid a static framing of prognosis and take into account the dynamics involved over time.

5.2

A topography of experience

In the NICU context a prognosis is a claim about the expected conclusion of the treatment. It is a formulation of what one thinks is likely to happen. This expectation is usually based on experience and as such these two are closely intertwined. Koselleck (1985) argues that this does not yet explain that experience and expectation are complementary notions. They in fact apply to quite different levels. Expectations, for one, still may change, which is not true of experiences. They are in the past and have become integrated in ‘the space of experience’ (p. 270). Expectations are distributed over unlimited temporal branches that reach beyond the horizon. Behind this horizon of expectation a new space of experience opens that is not yet visible (p. 273). Here is also a problem: also the unexpected may happen and we have to reckon with it. In the prognostication process expectations are disclosed that are not deduced from experience. The space of experience falls short to determine the horizon of expectation (p. 275). Where experience does incorporate the spectrum of expectations, the space of experience is too limited to survey the entire gamut of future possibilities. All actors in the NICU find themselves in a dynamic of experience and expectation that allows them to form and articulate a prognosis about a child. This not only applies to knowledge that is tied to education and position, but also to knowledge gained through all sorts of personal experiences involving the child and other NICU actors. In the specific ways in which nurses relate to a child – how they approach him, what they know about him, what they feel for him, how they speak for him – the subjective knowledge and experience emerge that are also used in prognoses. In other words, study of how those involved relate to Tom may concentrate on issues of role distribution and position, including the associated knowledge and prognosis, and on issues tied to the individual dimension of dealings and interactions with the child. Aside from professional knowledge, which generally emerges based on training, professional literature and deliberation, there is also knowledge and/or experience based on contact with the child and others involved. By focusing on this interactive dimension, room emerges in

Beacons on the Horizon 125

the analysis for the individuality of those involved as well as for other factors that play a role in the formulation of the prognosis, such as temporal aspects of specific situations and emotions like hope and frustration. The ways in which actors deal with children in the NICU can be conceptualized in terms of distance versus proximity, observing versus comparing, and hands-on versus hands-off. In the following sections I will discuss these conceptualizations in subsequent order. Distance versus proximity The spatial and temporal pattern of the movements of actors involved in the treatment of NICU children varies substantially. Parents, for instance, can almost always be found very near to their child’s incubator, while lab technicians who analyse his blood samples never see the child. What is the exact role of these factors in the formulation of a child’s prognosis? We can study and map this particular dimension in both literal (radius of action) and figurative (emotional bonding) speech. Let us start with the spatial movement and compare the radius of action of those involved. In the NICU the movements of parents concentrate around their child’s incubator. When they come in they immediately walk to their child’s incubator and in most cases they stay there during their entire visit. In this way many parents spend countless hours each day close to their child. The nurses also spend long hours near the incubator. For parents, then, the contact with nurses is the main source of information and this also influences their expectations of what will happen. Physicians, however, cover a much wider range. They are, after all, not just responsible for one or two children but for all children in the NICU. As such, physicians can be seen in the entire unit but also outside of it: in the parents’ room, the hallway, the secretary’s office, the conference room where lectures are held, and the library. Considering specific cases can refine this general spatial pattern involving the main NICU actors. Tom’s parents visit their baby every day, generally around two in the afternoon. The first months they come together, but later on his father mostly comes in only at night, after work. Tom’s mother, who is on sick leave from work during his hospitalization, always comes during the day, occasionally in the company of his grandma. After they come in they walk to his incubator right away and stay there, so they can observe him and his condition as closely as possible. They check the monitor, consult the flow sheet and compare his condition with that of the day before. Is he improving or not? Are things better or worse? For months they spend a number of hours

126 Uncertainty in Medical Innovation

near his incubator each and every day. There they meet the nurses who happen to provide care to Tom and who are also a very important source of information for them. The physicians do not spend much time at Tom’s incubator. But in Robert’s case, at least in the first days after his surgery, they could be found near his incubator all the time. Robert’s parents visited regularly as well, but unlike those of Tom they do not stick around for hours because that would be too much of an emotional burden to them. Moreover, because of the equipment and the staff’s activities there is hardly any room at the side of Robert’s incubator, which also explains why his parents never sit there for long. Still, they are around often and at quite different moments of the day. The first days they slept in the hospital and could be found near their son’s incubator at odd hours, such as at three in the night. The complexity of the surgery and their constant presence led to intensive contact with the staff. Robert’s treatment forced nurses and doctors to spend many hours together around his incubator, and this turned them into a close-knit team. In the Northeastern General Hospital, however, the nurse and neonatologist who work together much at the side of Maureen’s incubator hardly seems to have developed a shared sense of mission. Moreover, the child’s parents never visit and prefer keeping a distance. Rather than seeing their daughter’s condition with their own eyes, they listen to what the various doctors and nurses tell them on the phone about her. Tom’s parents, by contrast, rely on a variety of sources; their daily presence, their observations, and their interactions give them a much more versatile perception of their child’s condition. Likewise, organizational structures influence the spatial patterns of nurses in the NICU. In this respect there are substantial differences between the Northeastern General Hospital and the Academic Medical Center. In the former, for example, the organization of the work is based on primary nursing and twelve-hour shifts, meaning that one nurse is primarily responsible for the care of a child during his admission. The nursing staff has fixed twelve-hour shifts: either day shifts or night shifts. Some of them had been doing the night shift for more than twenty years. The advantage of primary nursing is that it brings continuity to the care provided. Nurses spend much time with a child and thus they gain much specific knowledge about him. They tend to pick up signs on what the child wants or how he feels more easily. As none of the other actors, nurses gain knowledge of a child and his specific trajectory. This kind of organization also tends to foster a more personal contact with the parents involved and offers doctors a useful

Beacons on the Horizon 127

source of information about the child. As one of the nurses from the Northeastern points out: ‘You try hard to provide the best possible care to the child. Because you are constantly caring for the same child you know well how he is doing. No one else follows his medical course as closely. But you also learn to know the child very well in terms of what he likes or dislikes. Of course you grow attached. It’s almost like he is one of your own children. And you also become involved with the parents, for you see them almost every day as well. There is a constant balancing between emotional bonding and knowing that one day he will go home, at least, if all goes well.’ In the Academic Medical Center the nursing work is organized in eight-hour shifts and in teams of nurses instead of a one-to-one relation between nurse and child. This means that no nurse bears exclusive responsibility for a child’s care. One tries to strike a balance between continuity and rotation. Nurses may provide care to the same child for three or four days in a row, but not weeks in a row. Moreover, the nurses also rotate in the high care and medium care units. This policy is meant to prevent them from becoming emotionally overburdened. This is a contrasting picture with the position of the nurses in the Northeastern General Hospital, who are regularly coping with stress. The downside of this hospital’s mode of organization is that nurses may become too emotionally involved with a child, especially when their stay in the NICU lasts several months. In such cases the emotional bonding process may cause nurses to identify with the parents.5 For example, Beth, a nurse from the Northeastern General, has been in charge of caring for Dave, born prematurely, who has been in the NICU for six months already and fully depends on artificial respiration. Like his parents, Beth believes that one day Dave will be able to go home. Although her colleagues do not share this prognosis, they understand her position, for they know that caring for a chronic child frequently leads to a high level of attachment. The members of the nursing staff are aware of how such processes work and they know how to support Beth. As one of them says: ‘It is obvious that Beth is denying Dave’s hopeless situation. But we have all been in charge of caring for a chronic child and we know what it means. You do not see how attached you have grown to a child when you are still fully engaged in providing care. Only when

128 Uncertainty in Medical Innovation

it is too late you realize how much you bonded. It is as if the child is your own baby. As if your own baby is slowly dying there. I couldn’t accept it either when someone else cared for him because I was off duty or working in another unit. That caused enormous strain. Now Beth finds herself in such a situation. Her life is now turning around Dave. We have tried to make her keep more of a distance by having her care for another child. But this is precisely what she does not want to do, for only she can care well for Dave, she feels. This is why we let her go ahead and support her as much as possible. But later on, when eventually the child will die, she’ll need us the most. For then she is done up with emotion. Some of us try to help her by telling her to anticipate that things may go wrong. But she does not feel helped by this at all. She knows the risk is there, no one has to tell her. These days she is extremely agitated. Yesterday Dave was sitting on her lap when we were doing rounds. When one of the residents said that the child looked so blue, Beth exploded. Dave did not look blue at all, she felt; it was just the fluorescent lighting. Well, what can you do?’ Although the system of primary nursing assumes the nursing staff to have good insight into a child’s condition, it turns out that keeping some distance is important as well. According to her colleagues Beth bonded too much with the child and this clouded her professional gaze. This said, it is important to note that a strong emotional engagement does not always distort prognostic judgement, for it may cause staff to understand families’ suffering (Benner, Tanner and Chesla, 1996). Nor is emotional detachment always a virtue, given that it can blind people’s awareness of suffering and lead to needlessly aggressive medical interventions. In the example above, it is assumed that more contact and more proximity inevitably lead to a greater degree of concern and participation. More distance, however, does not automatically imply disengagement. For instance, parents may not always be with their child in the unit, yet many still tend to be there all day in their mind. Regardless of where they are, they travel along with their child. A neonatologist from the Academic Medical Center explains: ‘Parents develop a tunnel vision. Everything centers on their child, regardless of whether they are at home or in the unit, or whether it is day or night. … The world of the parents narrows down to life in the NICU, irrespective of where they find themselves. The fear for

Beacons on the Horizon 129

all that can go wrong has a permanent hold on them, not just when they are with their child, but also, and perhaps even more so, when they are at home and do not see what is going on in the unit. They do not just become part of the hospital world – the hospital is their world. They are either here or they are at home, but that means they are here in their mind. As their world narrows, everyday things may take on grotesque proportions: the phone can become the messenger of the most horrible messages and it is used as little as possible so they can be reached. Their daily trip to the hospital puts unbearable pressure on them. Each day the same question: How is he doing, better or worse? … There is no escaping. Parents are not on a schedule with days off that they can look forward to.’ A child’s condition usually dominates the everyday life of the parents to such extent, that there is no more dividing line between the hospital world and their private life. Likewise, staff members may be bothered by work-related worries at home. As one nurse from the Academic Medical Center explains: ‘Some of us don’t think about work anymore the minute they cross the threshold of the unit. I don’t have that at all. If I had it my way I would call at night to ask how things are going. But I hold back, otherwise my colleagues start saying: she is deranged, calling the unit during her time off to enquire after the children. Tomorrow I go on vacation for four weeks. And I have to admit that I can’t stand it that someone else will handle my child. For this is how I feel it now: she is my child. I provide care to this child the entire day.’ The degree in which staff members worry about their work at home differs. Some do not sleep at night while others leave the unit and forget that it exists. These differences imply that distance versus proximity is not just a matter of the professional role. Among physicians there are differences in this respect as well. By contrast, parents are more or less expected to be concerned about their child around-theclock; it is a sign of healthy parental involvement. Physical distance and emotional proximity, then, colour the expectations of actors about a treatment’s course. Some are quite aware of this, such as Tom’s father: ‘As soon as we came home and added it all up, we thought it best to ask the physicians to withhold any further treatment. When you are

130 Uncertainty in Medical Innovation

sitting at home it is a theoretical decision. But once you are back in the NICU and you see your child, you don’t know it anymore. At such moment you can’t imagine that hardly one hour ago you decided that it was better to let your child die.’ The earlier decision of Tom’s parents is undermined by them being in their child’s immediate presence. At home other laws and choices are valid than in the NICU, at the side of their child’s incubator. Prognoses about a child’s physical future are not only based on levels of distance and proximity, but also based on direct observation. Also what one sees of a child influences the process of prognostication. How are the NICU children observed and what is seen while looking at them? Observing versus comparing Not everyone ‘observes’ a child as long or in the same way. It is mainly the nurses who spend much time observing a child and looking at the data on the monitor; it is perhaps even their main task. Tellingly they consider themselves ‘the doctors’ eyes rather than their hands’ (Schlomann, 1994). After some time nurses develop their own clinical gaze. They start noticing specific details and register changes in a child’s posture or how he moves. They interpret the colour of the child’s face, the size of his head, and how he digests the food. Does the child have a swollen tummy? Are the fontanels round or hollow? The nurses observe how the child reacts to daily medical actions, and whether he is restless or apathetic. Staff members also rely on these physiological aspects when determining a child’s prognosis, even if they are commonly hard to quantify. Having years of experience results in instinctive insight: intuition. This form of knowledge also plays a major role in the articulation of a prognosis. An experienced nurse or physician ‘senses that something is wrong with the child’ or ‘discerns that the child is not in good shape’. At one point the nurse who cares for Tom ‘knows’ that ‘he will never get off respiration.’ A judgement that is in part based on intuitive knowledge may be hard to back up. Although this form of knowledge does not fit the rational image we have of the professional medical gaze, the importance of this form of knowledge is all but underestimated.6 In monitoring a child’s condition, the nurse’s experienced gaze is as important as the numbers on the screen. But what do nurses actually see in the many hours they spend near Tom’s incubator? They see a child that is asleep under the influence of

Beacons on the Horizon 131

sedatives; that struggles with his respirator tube; that feels short of breath as a result of mucus in his throat; that coughs and cries without making sounds. They also see a child that is pale from being tired or red from exertion and that within minutes can change from a calm child into a troubled one. They see numbers on the monitor and lab results, which in Tom’s case never move long enough in a positive direction to be seen as signs of improvement. This accumulation of daily impressions and observations becomes a reason for the nurses involved in Tom’s care to consider him a child that too often is not ‘well’ enough, in the broadest sense of the word, to recover. The many hours that nurses and physicians actually deal with ill children also normalizes such cases. Staff members observe a child’s appearance with other eyes than parents. Robert’s parents were shocked when they saw him again in the NICU after his surgery. When a nurse welcomed his father saying ‘He is looking well, isn’t it?’ the man was stunned. What he saw was a tiny being covered with blood and iodine and hooked up to countless lines and tubes. As the cardiologist comments: ‘We’re already so conditioned that we don’t notice it anymore. Of course, the child really looks like he has been in the wars.’ The way Robert looked at first provided his parents little hope for the future. Amidst all the hi-tech around his incubator, all the tubes and lines, the dressing, the band-aids and red traces on his body, they no longer saw their child. To the mother of the child next to Robert, he even looked frightening. After she no longer dared to visit her own child anymore, it was decided to put a screen between the two incubators when she visited. In some cases all care and support from staff members cannot prevent that parents cannot deal with the way their child looks. This may cause them to leave the unit in despair and, in a rare case, never return. Staff members may also be bothered by how a child looks. In unit C of the NICU of Northeastern a child has severe malformation as a result of Turner Syndrome*. If her mother is bold and commits herself completely to the child, the malformation of the child is so hard to deal with for the staff that the primary nursing model is left behind and the child receives care from a different nurse each day. The nurses consider it a burden to provide care to this child for twelve hours in a row. Extreme malformation may even put pressure on a child’s ‘human’ aspect.7 Notably facial malformation makes it hard if not impossible to develop an emotional bond with the child. Without the familiar human look the affinity is absent that serves as basis for such bonding. These and other impressions resulting from direct contact also influence the prognostication process.

132 Uncertainty in Medical Innovation

What is observed while looking at a child does play a role in the process of prognostication. Likewise, the experiences one acquires while touching the child plays a role in the prognostication process as well. In the next section I will study the way the children on the NICU are touched to gain insight in the way sensory experiences contributes to specific prognostic knowledge. Hands-on versus hands-off Staff members acquire knowledge on the basis of the many children they see as well as on the basis of concrete medical acts. Nurses and physicians develop a tactile form of knowledge in their fingertips that helps them to assess a child’s situation. They perceive minimal changes in muscle tonus; the size of organs, or the child’s physical reactions, and this also contributes to their perspective on the child’s physical condition in the near future. There are all sorts of medical acts whereby the staff needs to touch the child, such as listening to his heart, measuring his temperature, clearing his mouth, fixing the respirator tube, percussion of his chest, washing and weighing the child, assessing the liver or blood circulation, attaching electrodes, and inserting an infusion. These are all hands-on experiences. In some cases the child’s condition is so bad that each touching poses a risk. This calls for minimal handling and a hands-off policy.8 In the case of Robert this applies in the first days after his heart surgery. Literally a finger should not touch him, which is why not even the iodine spots can be removed. Although less sick children may be caressed, 90 percent of the touching has a functional character (Zahr and Balian, 1995). Many of the regular procedures are a burden to children. The daily physical examination and the splinted arm to keep the infusion in place can lead to much discomfort. The intubation of the respirator tube, the clearing of the mouth cavity, and injections and punctures may even destabilize a child. All routine acts may cause stress in children.9 Painkillers and sedatives can only control this in part.10 After some time a child will get sleeping problems, become restless, and cry a lot. Environment factors, such as light, noise, cold and heat, can also be bothersome for a child. Especially glaring light and noise are a source of stress. The light, noise and daily series of interventions can lead to an increase of the frequency of the heart rate and respiration and a lowering of the saturation levels (Allen, 1995; Zahr & Balian, 1995). In particular chronic children have so many negative experiences that

Beacons on the Horizon 133

they respond in a stressed way to each stimulus, no matter its intention, which makes it hard to provide care. In the case of Tom each physical touch is one too many at one point. Sitting on his mother’s lap is not possible at all anymore, because it makes him panic. What is meant to be a moment of bonding proves to be a risky situation. For Tom’s parents this means that their role is reduced to that of spectator. Tom’s panic is a result of too many painful experiences. As one nurse explains: ‘Despite our maximal care we hardly manage to alleviate his suffering. Tom has been injected so often that with each touching – no matter how well intentioned – he feels stress. We cannot do more for him than try to alleviate his suffering as much as possible. But slowly, it seems, we have exhausted our options.’ Still, it is very hard to provide good care without touching Tom. Nurses have to hold him to be washed and cleaned, and they hold him more tightly when they fix his oxygen tube with a band-aid on his upper lip. They also have to take blood samples for lab tests. When nurses try to calm Tom, they unwillingly contribute to his anxiety. If they try to create a situation in which he feels as comfortably as possible given the circumstances, each effort on their part contributes to his stress. This of course is frustrating for the nurses. Time and again, they notice how their maximal effort has a minimal or even contradictory effect, and this calls for exceptional stamina. After four months most nurses who regularly care for Tom have reached the limit. Their daily struggle with him of course affects their view of his prospects. This case underscores that nurses not only have their responsibilities, but also their problems, risks, frustrations and dilemmas. In this respect, Tom has increasingly become more of a burden to them. His stressful behaviour is characteristic of children who have to stay long in the NICU. Another such case is Dave, who has been for months in the NICU of the Northeastern. A nurse tells about what it means to provide care to him: ‘I rather provide care to other children. Children such as Dave are so easily agitated. Little is needed to disturb their balance. It makes him turn blue all over and he needs extra oxygen. We raise it until his saturations are fine again, after which we return the setting. When he feels stressed and does not seem to get out of it we also give tranquilizers. We will look and see if we can put him in another

134 Uncertainty in Medical Innovation

spot. It is too noisy here. He is so close to the phone and the door where everyone enters and leaves. This makes him feel uneasy, so he becomes agitated and needs more oxygen. It is a lot of extra work to put him in another place with all these devices and such. But it seems worth it.’ Given the effort it takes nurses to calm down a child again, they try to avoid stressful situations as much as possible. They resent the long examinations performed by residents or their bad timing. Dave’s incubator even has a listing of times when he can be woken up to be examined, but not all residents comply with it. As an irritated nurse from the Northeastern General Hospital explains: ‘What drives me so mad is that this announcement at his bedside lists the times in big letters. But no one reads it. They all come in and go as they please. In this way children never rest. And every time Dave is woken up it takes hours before he sleeps again. This makes him feel stressed, so that the oxygen has to be increased again because the saturations go down. Children become totally deranged when they are woken up every five minutes.’ Like nurses, residents have physical contact with the child. Given that undergoing many interventions has a negative effect on a child, acting speed and skill are of prime significance. The extent to which staff members are skilled with their hands also affects the prognostication process. When physicians can still ‘handle’ a child, this gives them the feeling that they have a hold on the situation of the child and that thus they can ‘manipulate’ his future. Skilled hands, then, are a major asset in the NICU. In this respect a resident from the Northeastern comments as follows: ‘A NICU is a place where top specialists work and also the nurses are very skilled and experienced. This may cause them to turn the life of a resident into a hell. Fortunately, I am quite good at performing manual skills, such as inserting tubes and infusions. This impresses the nursing staff, for their own work largely consists of it: working with your hands and heart rather than just with your head. If you manage to insert a tube within a few minutes or set an infusion within two tries you’re in clover here. If not, you’re lost. You see it immediately with John [another resident]. John lacks the manual skill. Now he is counting the days and hours before he can leave

Beacons on the Horizon 135

here. They don’t trust him as physician and therefore they rather not entrust their children to him. This makes John so unsure of himself that he feels blocked and is no longer capable of doing anything.’ Touching children in the NICU requires experience and skill. Residents who lack this skill start viewing the NICU as an intensive scare unit (Frader, 1979). Because they are still in training, their medical actions are also exercises, ways to gain experience. However, it is not always inexperience that bothers them. The veins of some children simply make it hard to take a blood sample, while others may have had so many intravenous infusions that it has become difficult to find a good spot. This is a source of concern because in acute situations it is important to have direct access to the child’s bloodstream as it provides the option to administer medication instantly. Apart from skill, the actual touching of small infants also requires courage. Parents are often afraid to hurt their child. Most are reluctant to stick their hands into the incubator. When children are somewhat larger and may be held on your lap, this fear does not simply go away. Robert’s parents, for instance, did not dare to touch him for days after seeing him after surgery, let alone take him on their lap. It took some effort on the part of staff to convince them that Robert was stable enough to be handled and touched. When in the subsequent days Robert sits on his mother’s lap a few times, she grows convinced of his improved condition and also begins to be more hopeful about his prospects. The level of direct contact with NICU children results in specific knowledge that may play a more or less influential role in how actors develop prognostic insights. The same applies to how they use or experience the NICU as a spatial setting. They develop their knowledge of a specific child’s prognosis individually, based on sensory experiences such as observation and touching, while also emotional aspects – the tension between involvement and detachment – may have much effect on how staff and parents articulate prognoses. While the notion of position has helped me to explain divergent views on a child’s prognosis, the specificity of roles proved to be too limited to clarify for instance within group differences. By focusing on features like proximity, observing and touching I was able to understand the particularity of experiences of those involved. However, the situation of a child in the NICU is never static and always subject to change. The things that happen as a result of the treatment trajectory itself obviously

136 Uncertainty in Medical Innovation

affect the various prognostic processes. The actors that travel along on the same trajectory will gain new knowledge and experience all the time while they observe and touch the child. When one studies processes of prognostication it is important to take these changes over time into account because it will affect the interpretation of the situation. In other words, the experiences that the actors gain over time will influence their expectations of what will happen next. This calls for an exploration of the boundaries of their space of experience and how these boundaries may shift. To do so, I focus on various reference points that act as prognostic marker and are used in the NICU to determine a child’s prognosis. Through a description of these prognostic markers and how they function in the prognostication process I aim to capture the influence of the situation’s dynamic.

5.3

Prognostic markers

The experiences of those involved in the NICU practice elicit specific expectations. Regardless of whether they are realized, parents and staff members are constantly involved, both directly and indirectly, in learning processes, whereby the extent to which expectations are realized plays a major role. Prognoses, then, are the outcome of multifaceted prognostic processes that reflect a certain amount of conflict and disparity between the experience and expectation of the actors involved. They all deploy experiences from the (recent) past to interpret both the developments that occur along the trajectory and – by extrapolation – the ones that are likely to occur in the near future. This gives rise to specific ‘horizons of expectation’, whereby the measure in which expectations are realized plays a palpable role as well (Koselleck, 1985). In their formulation of new expectations, actors rely on subsequent knowledge. This may cause them to adjust their prognosis in a specific direction. It is in the reflection on the past that the future takes on a specific significance. Thus prognostic processes involve a linking of past and future in the present, resulting in a ‘current past’ and a ‘present future.’ In other words, a multi-layered understanding of such prognostic processes requires an exploration of the boundaries of the ‘space of experience’ of those involved and the way in which these boundaries shift. This I will do by studying the kind of markers they use and exploring the way their relevancy and meaning changes over time. What kinds of reference points are used on a NICU ward? In a NICU various prognostic markers are used to determine a child’s prognosis. Their analysis may reveal the ways in which individual

Beacons on the Horizon 137

experience and the dynamic of the treatment trajectory (including the learning processes among those involved) influence the various prognostic processes. A number of reference points play a major role: other children that serve as an exemplary case, the level of technological support on which the child depends, the use of the NICU space or the specific pace and rhythm of the care and the treatment provided. They can be said to function as prognostic markers, on the basis of which those involved conceptualize the child’s position along the treatment trajectory. In order to do justice to the input of the individual experience and expertise of all involved in the prognostic process, including the parents, the analytical category of ‘prognostic marker’ seems useful. By grounding my analysis in the concrete sequence of actions, the insight that specific positions matter in the production of prognostic knowledge is retained, but professional identities no longer play a restrictive role when analysing within-group differences. I want to discuss four types of prognostic markers: exemplary cases, technological markers, rhythm and tempo, and spatial markers. Let us take a look at the first category: exemplary cases. Exemplary cases In the production of prognostic knowledge, experiences with children in similar conditions function as a crucial marker. Physicians compare the situation of a child with relevant outcome studies that are described in the literature. These theoretical children allow doctors to form categories of more or less chances of recovery, expressed in percentages. Tom’s situation is compared to that of other young patients in a similar predicament and this plays a major part in determining his chances of survival. Thanks to these others, Tom, theoretically at least, ‘still has a chance of getting off the respirator.’ Statistical chances play a crucial role in prognostic processes. But chances of course are never identical. Some are calculated based on years of research and experience, while others are merely based on studies of small populations. And in some cases the results of large-scale studies are not yet available.11 If in such a situation less ambitious case studies have to provide at least some intimation of what the options and chances are, such results may not belong to the standard practice of care of professional knowledge. In practice this means that neonatologists found their prognoses on different experiences, which may also lead to different prognoses. What is more, the availability of outcome studies does not automatically prevent disagreement about prognosis. Divergent opinions can also result from conflicting interpretations of the same data, or

138 Uncertainty in Medical Innovation

from variation in the interpretation of follow-up data. How one interprets existing data also depends on one’s beliefs about whether it is morally acceptable to resuscitate babies with a limited chance for normal survival, the local culture of the particular nursery and its level of technological sophistication, and the limits of viability as legally defined. An extreme case of prognostic difference involves Maureen, the baby girl I presented in the third chapter, who was born in Northeastern General Hospital after a term of 23 weeks. One neonatologist estimated there was a 50 percent chance of survival for Maureen, while another saw no chance whatsoever, meaning he favoured discontinuation of her treatment. In a case like this the physician’s individual experience and prognostic competence play a decisive role. Having professional experience with children in similar conditions performs the same function. Prognosis is not so much based on a ‘paper’ or theoretical frame, but on the NICU’s local history. The nurses in the Academic Medical Center, for instance, compare Tom with other children in their NICU. In this respect one of the nurses has the following comment: ‘We have had children like Tom before. At one point you know how things stand. The respirator will destroy his lungs. Yesterday his “sets” went up once again to 35. This of course doesn’t help either. I wonder where this is going. I am afraid we are generating a second Mick. Ultimately we also failed to get him off the respirator.’ This nurse’s future scenario is not so much based on his experience with Tom but on his experience with the clinical history of other children: ‘I am afraid we are generating a second Mick. Ultimately we also failed to get him off the respirator.’ As this example suggests, ‘ordinary’ children are transformed into exemplary children that subsequently begin to function for those involved as a prognostic marker. This does not only apply to regular trajectories that have become part of the NICU’s or the profession’s collective memory; the exceptional case may become an exemplary one as well. Staff members sometimes try to qualify a colleague’s not so hopeful prognosis by referring to that one child that made it against all odds. If one person refers to the usual outcome, another will point to that one exception to the rule.12 The actors do not necessarily rely on the same prognostic markers. Not everyone has the same knowledge and experience. For most parents the markers that are based on theory and casuistry are outside their scope. Nor are they commonly familiar with the NICU’s local history.

Beacons on the Horizon 139

Their prognostic competence is rarely based on knowledge derived from the library or the NICU’s collective memory; rather, it is primarily based on their daily visits and the things that happen when they are there. The various developments involving the other children in the NICU also play a major role in the prognostic understanding of parents. When their own child is not doing well, other children may serve as a reference point. This comment by Tom’s father is telling in this respect: ‘They can do so much these days. Take Robert, the baby that some weeks ago had heart surgery here. We heard he was doomed to die. Now he is back home again. And healthy! They’ll take anything on these days. So why wouldn’t they find something for Tom. You never know.’ If Tom’s father refers to another successful case as basis for his hopeful attitude regarding his son’s treatment, other parents, by contrast, use Tom’s case as evidence of their sense of despair: ‘When I consider his case I sometimes wonder: Are we also headed in that direction? I hope not.’ To this mother, Tom’s situation represents a future perspective she does not want to think about. Again other parents will refer to his condition in terms of ‘our child is not doing all that bad.’ Rather than the healthy children outside the NICU, the other neonates inside are used as rationale for parental hopes or fears. The reliance on other children as points of reference in prognostic efforts comes in various guises. They may be based on the international medical literature, be tied to the NICU’s local history, be grounded in personal memories of those involved or in the experiences of others as made public in research reports or popular media. Some situations may have occurred years ago, while others may have happened only a week ago or less. These various kinds of references also reflect a specific kind of temporal understanding, a way of relating to both past and future that can also be characterized with the help of the notion ‘repertoire.’13 Whereas the situational and case repertoire can be defined as ‘spatial related’ forms of reasoning, we can also find temporal repertoires at work in the NICU context. There are two dominant temporal repertoires at work in prognostic processes: the cyclical repertoire and the linear repertoire. In a cyclical repertoire the course of events is conceived as following a steady pattern. Present, past and future are part of one and the same temporal system. They each function as a fixed moment in a natural cycle in which the past, for instance, is also the ‘future present.’ This cyclical structure

140 Uncertainty in Medical Innovation

renders the future both predictable and inevitable. For example, in statements about Tom as ‘a second Mick’ the past functions as a mirror that reflects lessons for the future. In this repertoire, a child’s prognosis can be logically deduced from past situations: ‘we have had children like Tom before. At one point you know how things stand.’ Also the reference to Mick implies the message ‘this is how this kind of situation ended before and how it will happen again. Let this be a warning!’ In a linear repertoire another conceptualization of time prevails. Arguments that exclusively point to the past are rare. A linear repertoire is more open, meaning that what lies ahead is not dictated by some natural order. This kind of repertoire basically comes in two versions: the future is seen as entirely open-ended, or, in a more moderate version, the future is seen as tied to probability (Koselleck, 1985). For example, a posteriori the attending physician articulates Tom’s prognosis in terms of chances: ‘Theoretically, Tom still had a chance to get off the respirator. But as the days went by this chance would grow smaller.’ In this moderate version the various options that are looming behind the horizon of expectation can be arranged by probability.14 The future is neither given nor entirely open. The number of possibilities, however, is finite and the likelihood of their occurrence is still tied to past experience. The reference to ‘chances’ also implies that the past can repeat itself and that thus it renders the future predictable. But in a linear repertoire the future is never entirely dictated by the past. Surely, this may give rise to uncertainty, but it also leaves room for the view that the future can be construed – that history is a product of human intervention. The temporal structure of this historical awareness implies progression. When reasoning from a linear repertoire, one is not concerned with individual moments in a repetitive whole, but with conceptualizing past, present, and future as subsequent stages in a development marked by progression. In this repertoire, for instance, statistical data always have a provisional character rather than a lasting validity. New knowledge and experience will increase children’s chances of recovery in the future. Therefore, in the linear repertoire the future holds the promise of recovery. It is this promise to which Tom’s father is hanging on: ‘they’ll take anything on these days. So why wouldn’t they find something for Tom?’ He assesses his own child’s future in light of the things that happen to other NICU children, Robert’s successful heart surgery in particular. In the present he already perceives the promise of the future. This future is one in which everything is possible and doable. This openness is carried further than the probability principle allows. In the repertoire used by Tom’s father the future transcends any sense

Beacons on the Horizon 141

of natural order: it is ‘transnatural.’ Characteristic of this historical awareness is the acceleration of time and the ignorance about what lies ahead (Koselleck, 1985). This future has no restrictions; it comes with a sheer infinite potential. This total openness knows no repetition anymore. It is irreversible. In the dynamic between experience and prognosis, then, the relationship between present, past and future can be conceptualized in various ways. The nature of such conceptualization leaves more or less space for human intervention, more or less space for certainty or change. Prognostic markers, then, are partially determined by the temporal character of the repertoire (cyclical or linear) in which they are articulated. This implies that reliance on the same markers can still lead to different prognoses. When reasoning on the basis of the cyclical repertoire, one will view Mick as a pre-eminent example of a BPD child* and as such his case is seen to reveal the senseless character of Tom’s continued treatment. Conversely, in the linear repertoire, one will emphasize the provisional value of Mick’s case and as such it is perceived to contribute to a hopeful expectation of Tom’s future. Thus the temporal nature of the repertoire (cyclical or linear) provides the dynamic in which the prognosis is articulated a more or less open character. In the NICU as social-medical setting, the thinking in terms of progression – including the related linear repertoire – clearly prevails. Staff members teach parents to view their child’s admission in linear terms and to recognize the signals of his improved or worsened condition in that repertoire. Besides the patient population nearby or far away and anonymous also the child’s technological embeddedness is indicative of his/her prognosis. What kind of technology do we have to think of? In what way do these technologies act as prognostic marker and how stable are their prognostic meanings over time? Technological markers The form of air supply, for instance, frequently tells a lot about the child’s condition. It is a positive sign when a child needs only minor support in this respect. In such case extra oxygen is inserted into the incubator via the circulation system. If this is not enough, the oxygen is inserted into the nose directly via a tube. When a child needs more support in breathing, the oxygen is brought into the lungs with the help of overpressure. This CPAP* respiration still lets the child breathe on his own. But in extreme cases, when children are too small or ill to breathe on their own, they fully depend on the respirator. This device,

142 Uncertainty in Medical Innovation

which inserts a mixture of air and oxygen into the windpipe, allows one to regulate the insufflation pressure, its frequency, quantity, and oxygen concentration. Both the form of respiration and the device’s settings are seen as yardsticks for a child’s condition. Other technological markers or reference points are the kind of incubator (open or closed) the infant is in. The incubator is important for both the warmth supply and the prevention of the danger of infections. Very small children are therefore lying in a closed incubator. As they grow they are better able to regulate their own body warmth and an open incubator will do. In many of these cases they are allowed to move to the high care unit. The mode of nutrition also functions as a major indicator for a child’s condition. When a baby is too small or too ill to suck and feeding by mouth or using the gut is not possible he receives parental feeding. In these cases a baby is fed intravenously. Otherwise the baby will receive enteral feeding, which is fed through mouth by a feeding tube or bottle. Thus the switch to enteral feeding means an improvement of the child’s situation. Such reference points specify the course of a treatment trajectory in terms of progression or deterioration. This gives the child’s development a graduation-like character, including the festivities that belong to it (Layne, 1996, p. 635). In many NICUs one will find certificates on the incubators with phrases like: ‘Hurray, I weigh 1000 gram, I am a champion.’ As Layne (1996) argues, the significance of such ward rituals reaches beyond merely being a nice gesture toward parents. Making such major transitional moments explicit is important to parents and staff alike. The graduation metaphor, however, comes with expectations of hope and success, and these are never automatically warranted in a NICU context. Parents may even be confused by this metaphor because it does not quite prepare them for the erratic treatment pattern of NICU children, which is more like a roller coaster ride. This metaphor more aptly suggests the unpredictable nature of NICU trajectories, including the ups and downs of collapsing and recovery and all the anxieties involved. Layne (1996) points to the cyclic nature of a roller coaster ride: you always come back again where you started and there is thus no progression. The roller coaster metaphor ties in with a cyclic historical awareness, while the graduation metaphor fits the linear form. An assumed linearity of treatment trajectories in the NICU is also undermined by the fact that technological indicators have no unequivocal meaning. Some devices and tools are deployed in various situations. For example, that a child is lying in an open incubator is not

Beacons on the Horizon 143

automatically a good sign; an unstable child may be put in an open incubator so that it is easier for the staff to keep an eye on the child and intervene promptly in case of a crisis. Nor is the number of technologies deployed necessarily a meaningful factor in a child’s condition. Children who suffer from trisomy 13* or 18* hardly depend on technologies during their time in the NICU, yet their minimal technological kit by no means justifies a positive prognosis, as most who suffer from this disease will die within 18 months. Similarly, the meaning of technological markers for the prognostic process during treatment tends to vary. Their prognostic value is all but fixed. On the first day of Esther’s treatment, for instance, the respirator is central. But on the second day doctors detect a fourth-degree brain haemorrhage and this means that her prognosis largely depends on the results of the ultrasound scans of her head. Accordingly, the respirator loses its central role in the staff’s prognostic process. This is also true for Esther’s parents, who attach more significance to the result of the ultrasound. Once those involved begin to attach different meanings to the same indicator, the prognostic views about the child start to diverge strongly as well. For instance, for the parents of Iris lung function remains the primary beacon for assessing their child’s condition. Like Esther, Iris is a premie who after a few days suffers a severe brain haemorrhage. Initially, however, the parents were told that her chances of survival would mainly be related to the respiration. As a nurse of the Academic Medical Center explains: ‘You see it happen more often that parents focus all their attention on the lungs. That, to them, is the only thing that counts when it comes to whether their child will survive or not. It is understandable of course, given that we ourselves stressed the importance of lung function in the child’s chances of survival.’ For the staff the severe brain haemorrhage that is discovered a few days after Iris’s birth is an important point of reference in the prognostic process. Her parents, however, invest all their hopes in the improved functioning of the lungs and view their baby’s future with confidence. In their conception of the situation the brain haemorrhage plays a much lesser part in their child’s prognosis. The doctor’s insistence on the seriousness of the brain damage hardly changes their view. Prognostic differences between those involved also emerge as a result of the accumulation of new experiences. If physicians and nurses can rely on their already acquired theoretical and practical knowledge,

144 Uncertainty in Medical Innovation

parents frequently have to start from scratch. At first they have no choice but to rely on what the staff tells them about their child. Staff members who inform parents of their child’s condition generally do more than just communicating data. In intelligible language they explain to parents what is going on, what the therapy means or how a device works. By processing the information provided to them, parents obtain more knowledge. They also undergo experiences of their own during their child’s admission and find ways of making sense of his condition. Especially after several months, parents gain a certain level of expertise that allows them to draw conclusions on their own. As, for instance, Tom’s father argues: ‘We no longer take for granted what the doctors say. At first you do of course, for you do not have a clue. But now I have been here and looking at that monitor for months as well. You become experienced to some extent and you no longer accept everything unquestioningly. We also find it more pleasant to talk to an experienced nurse than to the nurse who happens to take care of Tom on a given day. We want someone who knows what they are talking about.’ Thus there is a continuous dynamic between the various signals and their prognostic contents or meanings. Occasionally, technological or other markers will lose their central position and shift to the margin. Prognostic factors, in other words, have no fixed value. What one views as a decisive indicator, another hardly considers meaningful anymore. Of course, parents may experience such shifts as a source of confusion. If, after some time, they feel to have gained a basic idea of what to focus on, this knowledge may later turn out to be no longer valid anymore. To conclude, technological markers do not speak for themselves, but their prognostic meaning is embedded in a web of markers. The changes over time do not only influence the meaning of markers. They themselves act as a marker as well. The child’s prognosis is also assessed on the basis of the extent to which his/her development agrees with specific temporal patterns. By focusing on the extent children align with the expected rhythm and pace the role the temporal order plays in the prognostication process comes into view. Rhythm and pace In the interpretation of temporal markers, such as rhythm and pace, actors also make use of various repertoires. In the cyclical repertoire the temporal pattern is the fixed point from which a child’s development

Beacons on the Horizon 145

is measured, while a linear repertoire focuses on the time pattern itself and the extent to which it coincides with the experience gained in the everyday practice. In the neonatology practice, many sorts of time occur. For example, there is clock time as reflected by the unit’s central clock, marked by continuity, linearity, regularity and measurability. In the NICU we also find task-oriented time that is linked to a specific child’s care and treatment. This applies, for instance, to medication plans, infusion systems and monitoring rounds that take place every other hour or less frequently. In addition, there are schedules for day and night shifts, weekend shifts, breaks and visiting hours and deadlines, such as for informing the pharmacy of new nutrition and medication data.15 Acts, devices, and organizational principles all come with their own temporal frame, which often is of a cyclical nature. That many interventions recur a lot gives rise to a certain rhythm of activities. The temporal order of the treatment process is mainly based on this regular pattern of everyday acts. As such life in the NICU is cocooned in a web of temporal relations in which different notions of time coincide.16 These temporal structures closely interrelate. Specific acts tie in with each other, follow each other or take place simultaneously, such as the medication schedule and that of monitoring rounds. They may also exclude each other, as in the case of the different shifts. Medication time and sleeping time may be at odds with each other, or they may perfectly go together as in the case of intravenous feeding. Aside from aspects that have to do with rhythm and timing, life in the NICU also has a pace of its own, which is high, in terms of the schedule of activities as well as the circulation of patients. In Northeastern General Hospital the slightest uncertainty about possible complications after birth is already enough to have a child admitted to the NICU. In many cases it proves to be matter of undue precaution and after a few hours the child is discharged from the unit again. But precisely these pit stops come with bustling activity. It is exceptional for children to stay on for more than two or three weeks in the NICU. This means that there is a constant dynamic of children being admitted and being discharged again, which call for a strict temporal regime of care and treatment. This temporal regime also influences the prognostication process. If new admissions have to be integrated into the NICU’s temporal structure, this does not always involve a smooth or automatic process. An individual child’s situation can clash with this regime in various ways. First, children bring in their own temporality. They claim, as it were,

146 Uncertainty in Medical Innovation

their own process-oriented time: the time nurses need in order to provide care to them, which is primarily determined by the specific needs of the child involved. For instance, one child may need more time for bottle-feeding than another. Together with other temporal structures, process-oriented time may become a normative factor, in the sense that it divides the NICU population in ‘easy’ and ‘difficult’ children. Children from the second category require more than the usual amount of care time, thus posing a threat to the prevailing order and the everyday pace of activities. Process-oriented time may also be linked to a child’s prognosis. For example, a child that had no trouble with finishing the bottle but starts drinking much worse calls for more time, and this becomes a cause for concern. Deviations from the pattern lead to new assessments of his/her situation. However, process-oriented time is hard to determine as such, be it in advance or in retrospect. Activities may be performed simultaneously or be interrupted; implying that time is part of many different specific ongoing processes. This makes it difficult to indicate when something begins or ends. Process-oriented time cannot be measured or scheduled, mainly because its boundaries are too vague (Davies, 1994). Second, specific disorders and conditions come with their own temporal frame. For example, for a premature child the first two days after birth are full of risks. High blood pressure may lead to severe brain damage, but there are many more potential problems: a delicate fluid balance due to a thin skin can bring a premie into a state of super cooling and dehydration; kidneys often have trouble functioning; and there are moments in which the premature baby ‘forgets’ to breathe because of the immaturity of the respiratory mechanism in the brain. In response to this, the heart rate can drop rapidly. This is why physicians hardly dare to look beyond these first 48 hours. Time functions as a prognostic beacon. When a child survives the first two days, s/he proves to have vitality and this comes with expectations about his/her future. However, these expectations do not always have a positive connotation. The fact that Maureen survived the first 48 hours some nurses precisely viewed as a disadvantage. It gave the child less of a chance to die on her own. Third, specific kinds of treatment also have their timeframe. The child is expected to finish his/her trajectory within a specific period, but some need more or less time. A nurse from the Academic Medical Center emphasizes the basic NICU pattern in this respect: ‘we are not used to dealing with chronic babies. In IC everything is short-term. Children that are here for months we are not accustomed to.’ A lengthy

Beacons on the Horizon 147

admission puts children in the category ‘chronic cases,’ or ‘problematic treatments.’ For example, a child that has to be on artificial respiration for a long time may develop BPD* and as the time passes his prognosis becomes less positive. Tom is such a case. His development conflicts with the usual timeframe of treatment and recovery. Apart from temporal elements, spatial elements may equally influence how those involved arrive at their prognosis of a child’s chances at recovery. In the next section I will describe some spatial markers to understand the role of change the dynamics of the situation in the process of prognostication. Spatial markers One of the spatial elements involves the way in which the actors move around in the NICU setting. The various activities of staff and parents at or near an infant’s incubator tend to affect their prognostic process. All actors in a NICU develop a sense of its spatial dimension and what, for example, a nurse’s busy doings reveal about a child’s condition. Doctors or parents who see a nurse constantly walking up and down the NICU in response to alarms from one of the incubators will interpret this activity as a sign of the child’s instability. In such a case, constant presence or proximity is not a good sign. How parents interpret such spatial elements can be illustrated by a comment by Robert’s father, made in the wake of his child’s heart surgery: ‘Last night I managed to sleep a little, but the first night was awful. I couldn’t sleep. When I was up and just taking a glance into the ward and seeing everyone gathered around Robert, I immediately thought something serious had happened. But seeing only a single nurse at his side, I feared that something might happen and that no one would be around. Thus I was constantly on tenterhooks.’ In many cases the immediate presence of staff members around the incubator is an indication of a bad prognosis. But it does not have to be that way. Although the condition of Robert requires a lot of attention from the staff, his prognosis is favourable. Tom has to get by without all this attention, even though his prospects are significantly less favourable. Yet Tom’s condition, in contrast to that of Robert, is stable. Despite his bad prognosis, his condition is not in a critical stage. This stability means that he receives less direct attention from doctors. The time spent on a child by a doctor is a matter of time distribution. Doctors only pay attention to a child if their expertise is necessary at

148 Uncertainty in Medical Innovation

that point. Nurses in both the Academic Medical Center and Northeastern tell that sometimes they have the sense that doctors tend to take the condition of chronic patients for granted. A nurse from Northeastern General Hospital explains: ‘They [the doctors] think they know everything already. This you notice during rounds. It always begins with “This is Dave, wellknown to all of us,” followed by a quick reiteration of his story, after which they go on to the next. Chronic children like Dave receive comparatively little attention. They know them too well, so they do not look carefully anymore.’ In this same vein, special locations are also interpreted in prognostic terms. Many parents consider a doctor’s invitation for a meeting with them in the parents’ room as a bad omen. Good news is generally told to them in the presence of their child in the incubator, but for bad news doctors select the more private space of the parents’ room. When parents see other parents leaving this room with a subdued expression on their face, they see an image that sticks. Aside from these concrete movements of actors in the NICU, the space in which they define a child’s prognosis also has a more general geographic dimension. Parents of NICU children tend to conceptualize locations in terms of distance. At home means ‘being far away’ or ‘not being with their child,’ while the hospital parking means ‘almost being there’ or ‘leaving again.’ The geographic awareness of physicians is marked by their task or responsibility regarding various concurrent treatment trajectories. As such they have a wider geographic sense that includes other NICUs. For a neonatologist ‘far away’ refers to what cannot be reached by babylance or helicopter within several hours. ‘Close by’ is the radiologist on the fourth floor or a NICU in another hospital in a nearby city. The way in which this geography affects the prognostication process becomes visible in the discussion on Anne. Academic Medical Center In the NICU meeting room two neonatologists, a paediatric cardiologist, and three nurses have gathered to talk about the bad condition of Anne as a result of cardiac problems. The attending describes Anne’s condition. Heart surgery appears to be the only option, but the child is still too small for it. The cardiologist: ‘It can be done only after one year and a half, and this is quite long for a child that has to fight each

Beacons on the Horizon 149

day to reach the next one. In addition, there is the option of a heart transplant. Age is no issue here, but this hospital does not offer it. They do it in Paris.’ By suggesting a heart transplant the cardiologist introduces a new trajectory. It causes a shift in the discussion, and suddenly Anne is a transplant child. The others respond enthusiastically. Paris is only a few hours by plane. The cardiologist puts his proposal in perspective by referring to the potentially long waiting list and the high cost that is not covered by insurance. It is agreed that he will enquire first before discussing this option with Anne’s parents. In this case a situational repertoire prevails: Anne’s current condition is center-stage. What matters in this discussion, though, is that the boundaries of her situation are redrawn as widely as possible by inserting the outside world. Because the cardiologist introduces the French transplant unit, Anne’s chance of survival increases, as if a new door to the future opens. Anne’s prognosis is much better, if only briefly. She changes from a hopeless case into a potential transplantation patient and the prognostic attention shifts to the question whether the organization of a transplant abroad is possible at all. But in this case a mere abstract displacement is insufficient: the child itself has to be moved as well, to another facility – even if the discussants have strong doubts about its feasibility. The outside world is an integral component of the geographical reality of neonatologists. This not only pertains to exclusive treatment options elsewhere. The outside world is where patients come from and to which they will go back if all works out well. This permanent dynamic underscores the social nature of the NICU practice. Neonatologists are responsible for admission and discharge policies, which implies that they keep themselves informed about the goings-on in the delivery room, the admission capacity of other NICUs and the available options in peripheral hospitals. Ideally there is always an open space for a new admission because this option will greatly enhance a child’s prognosis. In Northeastern a new admission is sometimes a matter of creating a little extra room, suggesting a more flexible notion of space. All equipment is movable and in the worst scenario incubators are temporarily put in the hallway. Such local flexibility is absent in the Academic Medical Center yet here the number of available beds is linked to a nationwide system that provides detailed information about available beds in the country’s other NICUs. As such these units are part of the

150 Uncertainty in Medical Innovation

geographic reality of the neonatologist in the Academic Medical Center as well. The spatial reality of nurses is more limited to the unit itself, which follows from the specificity of their task definition. Nurses provide care to the child, monitor his or her condition, and perform most of the treatment and diagnostic tests. For these responsibilities the outside world is less directly relevant. The height of the monitor, the work space between two incubators, the distance from the desk in the middle to the alarm on the monitor – all these matters are part of the spatial awareness of nurses in the NICU. The physical set-up of the unit thus takes on significance in light of its everyday use. In this context ‘far away’ refers to the pneumatic dispatch in the hallway or the lab on another floor where the last blood sample has to be delivered quickly. More than any of the other actors, nurses have a very concrete sense of the NICU as a spatial reality that in various ways also provides markers for prognostic processes.

5.4

Experience and expectation

If the formulation of a NICU child’s prognosis is based on several closely intertwined factors, the various social roles and positions of the actors involved give rise to divergent forms of prognostic knowledge. Multiple factors are tied to the way in which those involved actually engage with the child, their mutual interactions, their prior knowledge, and their subjective experiences during the child’s treatment trajectory. My analysis of prognostic processes, while starting from the social position of the actors involved, shifts towards prognostic markers, because it allows for a more detailed understanding of the specific role of circumstantial elements, individual knowledge, and the subjective experience of those involved. What matters, moreover, is how the various markers are deployed. By capitalizing on prognostic markers as an analytical category, it becomes possible to widen our understanding of how actors arrive at their prognosis of a child’s future. Some markers are more likely to be used by certain individuals and/or groups than others. For example, the number of staff gathered around an infant’s incubator is much more likely to be used as a marker by parents than by nurses or doctors. Although staff members might also notice this same fact, they will not use it as ground for prognostication. Furthermore, the use of the same markers hardly guarantees prognostic consensus. The contrast between parents’ and professionals’ use of the same technological markers illustrates how the interpretation of indicators determines their significance in the production of

Beacons on the Horizon 151

prognostic knowledge (Heimer, 2001). Even when staff and parents rely on the same style of reasoning, they may still arrive at conflicting conclusions, as my argument about the different uses of exemplary cases has demonstrated. Apart from elucidating the divergent opinions on prognosis between and within groups, a careful analysis of prognostic markers offers the opportunity to take ‘time’ into consideration as a factor. The various actors formulate prognoses as the child’s treatment trajectory is unfolding. During this process their experiences are constantly subjected to change. The diversity of the knowledge and experience gained causes them to use different prognostic markers and/or the same prognostic markers differently. Likewise, study of the various contextual elements – such as subjective emotions, spatial and temporal aspects, and technological devices – shows how a specific kind of understanding influences the interpretation of these prognostic markers. Those who rely on a frame of knowledge provided by a cyclical repertoire yield different meanings and attach other significance to prognostic markers than those who rely on a linear repertoire. The various markers have no fixed prognostic value. What presents itself as prognostic marker, and which meaning one attaches to it, is closely related to the topography of experience of those involved and the repertoire applied. Even if prognoses are rooted in and construed on the basis of frames of knowledge – including their repertoires of (temporal) reasoning – as well as on skills and experiences that always have an individual dimension, prognostic processes are essential for a proper understanding of the social dynamic in a NICU. Each new admission sets in motion a new series of prognostic processes, as each child embodies a new horizon of expectation. My argument underscores that only a multi-layered analysis of these processes accounts for actors’ divergent expectations and formulations of what is looming behind the horizon. The treatment trajectories that various actors conceive may be at odds with each other, while the ensuing frictions may stifle the treatment process. This calls for specific strategies to get this process moving forward again. After all, something has to be done. In the previous chapters several corrective forms were addressed already, such as diplomacy and an institutional order of collectiveness in the form of teams. In the next chapter I explore in more detail a specific sort of tension, namely the moral dilemmas tied to medical action in the NICU and how they are resolved.

6 The Moral Load

Academic Medical Center Friday: An EEG is made of Tom’s head. His ventricles prove strongly enlarged, which suggests raised pressure on account of fluid accumulation. He needs an internal drain to remove the fluid. Wednesday: After implanting the drain on Saturday Tom’s condition clearly deteriorated. The respirator setting is still at thirtyeight. The blood test results in no way allow the respiration to be reduced. Tom’s father has not visited for several days. He can no longer handle his son’s condition and stays home. He is already dissociating because he expects he will soon lose his son. In his conversations with his wife this sense keeps recurring. Thursday: Tom’s internal drain leads to complications. The fluid he loses causes a diastolic pressure that is so substantial that his cranial sutures fall over each other. As the day progresses, his condition worsens. During the medical round the diastolic pressure in his head is the central issue. Yet the doctors decide to hold on to their policy. They feel the risk of brain damage to be too large to remove the drain. The problem of the diastolic pressure can be solved through medication. Tom’s incubator has to be tilted upward as well. Two nurses change the position of his incubator, but they are not convinced. ‘What a dismal case,’ one nurse mumbles. At noon he looks a little better. His fontanels have their normal shape again. The diastolic pressure in his head appears to be gone. Friday: Tom is doing badly. He turns acidotic all over. Treatment needs to be adjusted again. 152

The Moral Load 153

Monday: The drain in his head appears to leak somewhat which seems not to be serious. But since the raised pressure in his head he has not been doing well. Over the weekend, the respiration pressure was again put back to the previous setting. This suggests that its gradual reduction again failed to succeed. Wednesday: Tom has kidney stones. Diuretics each time ‘empty’ his kidneys to such extent that their calcium level goes up steadily and stones may form. In addition, there proves to be an increase of pressure in his heart. On account of the corticosteroids the ventricle wall has grown thicker. With each beat this has decreased the flow towards the aorta. When Tom is five months old his treatment is marked by a number of setbacks that succeed each other rapidly. His parents have basically lost all hope of his recovery. The physicians also are sombre now about his chances of ever getting off artificial respiration. If in the preceding period each problem was dealt with separately, now the full scope of his situation presents itself. Is it still useful in this case to go on with treatment? Or is it better to stop? Both options are highly problematic. By stopping they risk missing out on his last chance of recovery, but by further pursuing treatment there is the risk that their efforts only contribute to more suffering. When the two available options are both unattractive this puts pressure on the treatment’s progression. This chapter concentrates on the question of how one arrives at decisions in such situations. Such choices are not limited to either stopping or going on. In some cases the decision whether or not to start treatment is a dilemma that already comes up before a child’s birth, as happened, for instance, in Esther’s case. The staff members involved were grappling with a moral decision: a decision of which it is assumed that it influences the quality of life of the child. To analyse how those involved deal with moral issues, I focus in this chapter on the decision-making process. Where are the moral boundaries of the practice in which these children are treated? How do actors deal with decisions? What are the consequences of a decision to stop treatment? Who bear responsibility and how is it distributed? Before addressing these concerns, I discuss the problematic nature of decisions on life-ending action based on the decision process involving Esther in both the delivery room and the staff meeting. Although both cases were addressed in Chapters 2 and 3, here I explicitly address the moral dimension of decisions on

154 Uncertainty in Medical Innovation

life-ending action. At the end I return to Tom’s treatment and its related moral aspects.

6.1

The nature of the decision

It is almost impossible for neonatologists to arrive at an accurate prognosis about a child’s condition before birth (Polin, 2005). The fact that Esther is both premature and immature* and suffering from failing heart activity provides those in the delivery room insufficient basis for adequately assessing what her condition will be right after birth. She may live, but serious handicaps cannot be precluded. There is also a good chance she will die right away. Should one still intervene? The option of withholding treatment would imply that the child dies before birth. From this perspective, the choice in the delivery room is about whether or not one should opt for life-ending action.1 The complexity of decisions on life-ending action led to an extensive debate in the 1980s.2 It resulted in several guidelines. First, the need for guidelines was motivated by the desire to get a hold on the complex moral problems that occur in medical practice. The rules could function as a template in which moral problems and conflicts take on manageable proportion. Second, the call for guidelines was also triggered by the societal need to regulate decisions on life-ending action. In practice all sorts of motivations proved to play a role and there was hardly any uniformity. Guidelines in the form of evaluation schemes and decision-making procedures would support and steer physicians in their actual practice. In the early 1990s in the Netherlands two reports in particular provided guidelines for decisions on life-ending action. The Cal-report (1990) issued by the Royal Dutch Association of Medicine (KNMG) and the report of the Dutch Association of Paediatrics (NVK) entitled Doen of laten? (1992). According to the latter, treatment is senseless in children that medically have no chance at all or if treatment contributes to ‘unliveable life.’3 This report formulates criteria based on which physicians opt for one of three proposed treatment trajectories. The first option is the choice to start up a treatment by using all available means (trajectory A). This option is to be preferred if the child falls in the first category. This is ‘…in all cases where the child’s chance of survival, respectively its chance of a liveable life, is considered so high for the time being, that this interest offsets the damage and the suffering that can be the effect of medical treatment. This is reason to give the patient the

The Moral Load 155

benefit of the doubt…. But if one begins treatment, one should make it clear from the start that withholding life-extending actions will continue to be a valid option…. What one should weigh is: how much suffering, or how much uncertainty, is counterbalanced by which chance that the prognosis is still acceptable and the treatment therefore will benefit both child and family.’ (NVK, 1992, p. 43) A second option involves a trial phase with limited intervention consisting of, for instance, basic care, warmth, oxygen and intravenous nutrition (trajectory B). This applies to children who fall in the second category with no clear indication that intervention will be useful. In cases of doubt a decision for this trajectory may contribute to the emergence of a selection ‘…between vital children with a reasonable good prognosis who are served by intensive care versus children who are not served by it and who are better off with care that is geared toward a humane death.’ (Ibid, p. 44) A third option is called ‘primary abstention’ (trajectory C). This can be selected if children fall into the third category. This is the case ‘…when there is reasonable certainty about the diagnosis that gives rise to a very bad prognosis (for instance, anencephaly*, other already diagnosed serious prenatal deviations or extreme prematuria, especially in combination with serious perinatal asphyxia*), there will be at the moment of birth, notably in vitally threatened children, only one option available: terminal care. There is a communis opinion about this.’ (Ibid, p. 45) These guidelines are still part of today’s Dutch paediatric practice. In the shady area of decisions on life-ending action clearly formulated criteria can serve as a tool for making such decisions. Specific guidelines allow those in the neonatology practice to reduce and regulate the complexities involved in life-ending action. Given the uncertainties with which those in the delivery room have to struggle, the guidelines from the NVK report seem to be a perfect aid. It is, after all, designed for such situations. These guidelines have an open and procedural character and leave room for situational considerations, such as pursuing further treatment because the prognosis offers not enough certainty and the (subjective) concern of the parents’

156 Uncertainty in Medical Innovation

ability to care for a seriously disabled child. Still, in the decision process on Esther’s treatment these guidelines are not used because they assume a different problem and practice than those in the delivery room are facing. When formulating rules and procedures, after all, one does not only have a specific image in mind of how (moral) issues have to be tackled, but also of the problem and the practice in which they occur. The structure assumed by the guidelines is not only seen as a ‘desirable order’, but also as a ‘realizable order’. In other words, the guidelines present a ‘realizability claim’: that what is desired is also possible. With this claim they assume a specific practice, one in which these rules work. This implies that the guidelines not only imply moral judgments, but also preconceptions about reality (Swierstra, 1994). Which preconceptions on the issues of life-ending action are at the basis of these guidelines? How do they relate to the challenges that those present in the delivery room when Esther is born have to face? Evidence The report tries to provide an answer to the question: should we intervene or not? The physician opts for trajectory A when the child falls into category 1, for trajectory B for children in category 2 and for C with children in category 3. The choice between intervening or not intervening is thus reduced to a matter of categorizing the child by interpreting facts and chances. As such, this script ties in with the rationality of evidence-based medicine. The problem involved, however, precisely pertains to the child’s categorization. No matter how clearly the guidelines indicate in which situation life-extending or life-ending treatment is justified, it is still up to those involved to determine whether the situation indeed warrants it. What is the condition of Esther? Are her chances ‘so high for the time being’ that they offset the negative effects of medical intervention? Or does this child belong to the group of ‘doubtful cases’ and does she fall in the second category? How large should their certainty be if they are to be ‘reasonable’ enough to justify a policy of ‘primary abstention’? The boundaries between the categories are vague. Esther can be subsumed under each of the three prognostic categories. The reports also refer to criteria such as professional experience and statistic data on survival rates, as well as to the assessment of the child’s future suffering, ability to cope, dependency and possibilities for communication (NVK, 1992, p. 32). But also with this further reference the physicians in the delivery room are not helped much. The instrument provided to them precisely calls on what they hardly have insight in:

The Moral Load 157

the prognosis of this as of yet unborn child. Thus the guidelines seem rather a reformulation of the problem instead of an aid for coming to a decision. The guidelines provided for establishing a moral order assume a transparent, homogenous practice, one in which the facts are unambiguous and within reach. In the guidelines the facts constitute the starting point for the decision process. For those in the delivery room, however, this starting point lies in a domain that is outside their reach. The problematic nature of the choice between intervening and not intervening is precisely lying in the impossibility to categorize Esther. It is the lack of clarity about the situation that complicates the decision process in the delivery room. What prevails is not so much what has to be done but what is the matter. The uncertainty in the delivery room regarding the treatment policy, then, cannot be resolved with these guidelines. They can only guide intervention as soon as it is clear what is actually wrong. At the very moment that those in the delivery room hardly have a basis for grounding their decision, they have to make one. This is why they look for an alternative that allows them to move the process forward, and at this point the parents take over. Parents, like staff members, rely on repertoires. Esther’s mother, in her decision on Esther’s birth, relies on a situational repertoire. She is not so much geared to the child’s situation after its birth, but to the child’s situation at that moment, namely her presence in her womb. The heart frequency on the monitor of the CTG makes clear that the child can die any moment. The possibility that her child will die in her womb is unbearable for the mother. Caught by fear she looks for a way out and finds it in the Caesarean sectio*. The mother’s choice is not geared to the child’s treatment. She opts for the Caesarean section* because it offers her a way out as to the risk of her child’s intra-uterine death. The mother’s panic leaves no room for thinking beyond the delivery itself. All that comes after the delivery, such as the possibility of treating Esther, falls outside her scope. Yet the decision on her own treatment trajectory also determines the child’s trajectory: maximal intervention after birth. Esther’s father feels the pressure as well. In his fear for what may happen and can go wrong, he looks for a hold in numbers: which chance of survival does the child have? The child proves to be not entirely doomed, and this reality he embraces as a way out. Relying on the case repertoire, he converts this hope into a decision. The statistics given to him by the physician do not only introduce a specific distribution of survival rates. It also provides Esther’s father with a language in which he can articulate the situation (‘it is a child with a chance’) and justify

158 Uncertainty in Medical Innovation

his choice (‘the child has a right to that chance’). Based on the same rates, Esther’s father could have chosen a trajectory of withholding treatment. Then they would have represented not so much a chance of survival, but the risk they articulated: the slight chance of survival and the large chance of handicaps. At first sight the randomness with which statistics can be deployed seems to undermine its constructive contribution to the decision process. But it is precisely this flexibility that turns this case repertoire into a major component of the decision process. Percentages do not so much urge one to specific action, but rather make such action possible. Statistics props up the decision and functions as support at a moment when each footing is welcome. With reference to concrete chances Esther’s father justifies his decision to take a risk. To him Esther is not a child with a chance by definition, however. She becomes one by relating her situation to the trajectory of maximal intervention. Without intervention she would not have had a chance at all. In the delivery room the questions ‘what is the matter’ and ‘what should we do’ are bound up with each other. The eventual choice – after Esther’s birth – proves to be the outcome of a process in which simultaneously the situation is defined and the proper course of action is determined. In the NVK report the facts function as starting point for the decision on how to act. In the delivery room, however, one does not know the facts, so a solution is sought in the available options for action. Not the facts but the treatment options function as starting point. In contrast to what the guidelines assume, the availability of various treatment trajectories is not so much a dilemma but allows one to jumpstart the process again. The report presents the various treatment options as equal choices. But is this correct? The NICU practice of both the Academic Medical Center and the Northeastern General Hospital show a clear preference for the trajectory of maximal intervention after birth. Where does this preference for maximal intervention come from? Asymmetry Of the three trajectories listed in the NVK report – withholding treatment, limited deployment of means, or maximal intervention – the NICU of the Academic Medical Center and that of the Northeastern General both seem to have a clear preference for the trajectory of maximal intervention. This asymmetry between the three options has various backgrounds. First, this asymmetry emanates from the nature of the consequences. If it is decided to refrain from all treatment, the implication is that the

The Moral Load 159

child will die after being born. The irreversible character of withholding treatment often makes one reluctant to choose this option. Second, this asymmetry can also be tied to the treatment imperative that is bound up with chances. In the Northeastern General Hospital non-intervention means that the child has no chance whatsoever. This calls for a clarity that in most cases is absent. Basic clarity is only provided by diagnoses such as trisomy 13*, trisomy 18* and anencephaly*. The second trajectory, which favours a policy of exercising restraint, in fact reduces the chance of a positive ending, according to some physicians. A limited deployment of means would leave a potentially present vitality untapped. For physicians, however, vitality means the child’s power to remain alive, even if with the help of maximal technological support, for also the acceptation of this artificial support would be evidence of the child’s vitality. By refraining from deploying technological support this ability may disappear. As one of the neonatologists in the Academic Medical Center comments: ‘We do not engage here in a “survival test” by having a child first “show” us for ten hours how strong it is. If you would still decide to act afterward, you already lost out on several specific opportunities. The sooner you address a problem the better, or it will eventually have damaging effects for the child. Still it is impossible to say much about it. In the past you would let a child die that suffered from a specific disorder. But after you read a study that shows that children sometimes have a chance after all, the next thing you do is try to treat these children.’ The asymmetric nature of the consequences is precisely exploited here. The trajectory of maximal intervention allows the risks that are part of the other two treatment options to be avoided. Third, there is an asymmetry in the option to displace the decision in time. Only the trajectory of maximal intervention provides the possibility to refrain from treatment at a later stage, when one has more insight into a child’s situation. The decision process thus offers an escape route from the dilemma with which those in the delivery room are struggling: they have the option to postpone the decision. The neonatologist explicitly indicates this option to Esther’s father. The possibility to come back to the decision has in part contributed to the fact that Esther’s father preferred the trajectory of maximal intervention over withholding treatment. In the delivery room the decision is bracketed off, so to speak: as if it does not yet involve the final decision. He

160 Uncertainty in Medical Innovation

postpones this decision to a point when there will be more information on the child’s condition and thus a decision on a treatment policy can be taken with more certainty. Finally, there is an asymmetry in the room for action. The trajectory of maximal intervention follows naturally from the trajectory of a surgical delivery. One action (Caesarean section) elicits another one (the child on the respirator). The choice of Esther’s mother for surgical delivery puts the staff onto the trajectory of maximal intervention. Accordingly, Esther is admitted to the NICU. The guidelines of the NVK report appear to assume that the basic question in the delivery room is how one should act, but, instead, the focus is on the child’s undefined condition and his/her prognosis. In this discussion the various treatment trajectories are not considered as equal options. On the contrary, this difference offers those involved the possibility to make a consideration and thus the chance to postpone their decision. The moment that a decision is called for is not far off anyway. Three days later one has to decide whether to go on treating Esther or refrain from further treatment. We know the outcome of this decision. But where, in what place, and when, at what moment, did it come into being? In the next section I will follow the decision to answer this question.

6.2

The functionality of relocation

The staff meeting is commonly regarded as the context for making decisions on life-ending action. In this case a meeting was called because at Esther’s bedside the physician arrived at the conclusion that her further treatment seemed meaningless – a conclusion that functions as the occasion for the meeting rather than its outcome. Still, the staff team’s deliberation seems more like a legitimization of what ‘informally’ was already decided at Esther’s bedside. This is not problematic in and of itself. On the contrary, it shows that reflection or empathy is not limited to moments of formal deliberation. Typically, a decision on life-ending action is not articulated in a single moment or place, but gradually takes shape in a process spread out over time and space. This relocation of the decision also serves a purpose, for it allows the decision to be tested for its robustness. In view of the consequences of decisions on life-ending action, this robustness is necessary. The decision’s relocation, aside from ensuring its carefulness, offers the opportunity to distribute the responsibility for the decision among the various parties involved, thus creating a basis for consensus.

The Moral Load 161

The relocation of a decision on life-ending action also has several practical advantages. It introduces moments of pause into the overall process, so that it becomes impossible to rush a decision. As such, relocation involves a reconsideration of the decision. By specifically addressing it in various locations, different evaluation opportunities emerge. This contextual diversity allows those involved to judge the problem all over again, analyse it, and look at it from another angle. What in context A is a given, may be subjected to debate in context B. The location in which the problem is discussed influences its definition and may therefore emphasize other aspects of it. Depending on the context, some facts may be irrelevant, specific remarks are improper, certain arguments do not apply, or there is more room for emotions. Let us follow the decision throughout the decision-making process. For the analysis of both the testing of the robustness as well as the distribution of responsibility I return to the decision process on withholding treatment in the case of Esther. Robustness It is at the side of Esther’s incubator that physicians assess her situation as hopeless and decide that it is best to discontinue her treatment. They interpret a fourth-degree haemorrhage at both sides as a reason enough not to go on. As such a bleeding of this size thus constitutes a moral turning point, convincing them to shift from treatment to withholding treatment. It involves a medical fact that establishes the difference between liveable and unliveable life, demarcating the boundary between meaningful and meaningless treatment. Because such facts and decisions are tested much more seriously than others, a staff meeting is called which displaces the decision from the NICU ward to the NICU staff room. The question posed by the resident – ‘what is our policy given the current problems and taking into account the child’s immediate future and her very bad prognosis?’ – suggests openness that is not really there. From the start of the meeting, this question is linked to a specific answer, namely, a life-ending policy. Precisely this answer, rather than the question, is the reason for them to be sitting around the table. The staff members of this particular NICU share the view that, at bottom, a case in which treatment does not lead to improvement and only contributes to an unliveable situation justifies a policy of withholding treatment. But is Esther’s such a case? Yes, the doctor’s answer was when standing at the side of her incubator. Yet do the staff members involved arrive at this same conclusion during their meeting in the staff room?

162 Uncertainty in Medical Innovation

The staff room discussion first focuses on Esther’s situation only. As expected, the dialogue that goes on appears to move all present into the direction of a discontinuation policy. But the discussion takes a turn when the fourth-degree haemorrhage as a turning point is subjected to a trial of strength, and this puts pressure on all that has been said up to that point. One of the neonatologists who are not involved directly in Esther’s treatment takes on the role of devil’s advocate. He does so not because he mistrusts his colleagues’ intentions. Rather, by embracing this role he does exactly what he is expected to do in this setting, namely to put pressure on the proposal so that its potentially weak spots can be traced. In this way, a certain level of detachment is created that offers room for addressing other facets of the problem. Some enquire after the EEG result or refer to the literature. This latter challenges the fact that the haemorrhage implies a hopeless prognosis. In their decision-making they raise the issue of both the hopelessness of the prognosis of the brain haemorrhage and the value of ultrasound scan as prognostic instrument. Clearly, the brain haemorrhage proves to be less robust as a moral turning point in the staff meeting than at the side of the incubator. It should also be pointed out that it is not fixed in advance which fact serves as the moral turning point. In the Northeastern General Hospital Maureen also has a fourth-degree brain haemorrhage. The attending argues that a fourth-degree haemorrhage on both sides is even an incentive for treatment. A policy of withholding treatment is not at issue because the haemorrhage as such, as he puts it, ‘is no lethal affliction’. Moreover, he argues, ‘you cannot even say with certainty that this child will be disabled. After all, some cases have been reported in which the outcome was less serious than anticipated. This means that in practice you cannot say anything about the outcome of degree four bleeding.’ In this case the brain haemorrhage does not become a moral turning point. One person views a fourth-degree haemorrhage on both sides as a non-lethal affliction and even a reason for treatment, while another interprets this affliction as inextricably bound up with an unliveable future and will opt for breaking off treatment. Furthermore, the staff meeting challenges not only the fact that Esther is a child that falls in the category that justifies refraining from further treatment, but also the category itself. The ‘oppositional’ neonatologist questions the boundaries that determine what does or does not belong to the category ‘unliveable’ or ‘hopeless’. For him, mere certainty about the presence of a fourth-degree brain haemorrhage is not enough, as its location in the brain is equally relevant. The child, he

The Moral Load 163

feels, is in the category that will legitimize to stop treatment if it has a brain haemorrhage of a certain size and in a specific part of the brain. Correct information on these two aspects may cause this neonatologist to believe that a moral turning point has been reached. These categorical boundaries also require other prognostic evidence, such as an EEG, as well as other procedures and standards. The ultrasound result has to meet other conditions. As the neonatologist claims: ‘I think that, regardless of the case at hand, it is more careful in this kind of decision process to ask for a written report from the radiologist.’ In the case repertoire deployed here the boundaries of the category are cut back and outside help is necessary to guard these boundaries. Unlike his colleague, the attending neonatologist does not raise the issue of ‘such cases in general.’ He assesses Esther’s prognosis based on a situational repertoire rather than a case repertoire. This means he has to deal with other factors, such as lack of time. This is why merely ‘a corroboration from one of the staff members of the X-ray department already suffices.’ In this situational repertoire also another list of facts applies on the basis of which the child falls into the category of withholding treatment, such as the fact that ‘for a longer time the child also has been seriously decompensated’ and that ‘the hypotension picture rests almost certainly on a still active bleeding.’ But this is not the end yet. Aside from agreement on the nature of the fact and the content of the category, agreement has to be reached as well at the level of repertoires. Both styles of reasoning, however, put forward another standard regarding the moral turning point based on which a decision has to be made. In the normative order of the situational repertoire used by the attending an ultrasound scan will provide sufficient evidence, an EEG plays no role, and a brain haemorrhage degree four on both sides of the head is enough to justify a discontinuation policy. In his reasoning another list of facts applies on the basis of which the child would fall into the category of withholding treatment, such as the child’s ‘seriously decompensated’ state and his conjecture of a still active bleeding. The case repertoire used by his colleague works with other norms to justify such decision. In this style of reasoning an ultrasound scan alone is seen as providing insufficient basis, which is why an additional formal report from a radiologist is needed. In this neonatologist’s case repertoire an EEG is a prerequisite, as is a specification of the exact location of the brain haemorrhage. In short, the two styles of reasoning each incorporate another normative order, which influences the standards used to define a moral turning point. This means that in addition to agreement on the nature

164 Uncertainty in Medical Innovation

of the facts and the content of the category they still need to reach agreement at the level of their respective styles of reasoning. How to proceed now? As have been described before we know that both the attending and the neonatologist actively undertake an effort to reconcile the two proposals to move beyond the impasse. The neonatologist modifies the general standard of having an EEG examination by agreeing to only a consult by a neurologist. The standard that an extensive report from the radiologist is necessary is also adapted to the circumstances of the situation. It is agreed that consenting to the existing report suffices. Thus the neonatologist accommodates the specificity that Esther’s situation calls for. The attending, in turn, reformulates his proposal in such a way that it incorporates his colleague’s conditions sufficiently: ‘Let’s agree that one of the staff members from radiology looks at the ultrasound scan and that the neurologist is asked for his opinion and whether he has something to add to this case. Subsequently we can go to the parents.’ The staff meeting’s final outcome is entered in the record. In line with its standards, the discussion was reduced to merely summing up those present, a brief description of the prognosis and treatment policy. In this case the policy is tied to the condition that the tests by the radiologist and the neurologist have to corroborate the prognosis. Next, the request for advice is mailed to the paediatric neurologist and the radiologist. Thus the number of experts involved is broadened. The radiologist looks at the ultrasound scan of the head and confirms the interpretation of the attending. So does the neurologist. If at the side of the incubator science and clinical experience have a much more pragmatic value, this context also offers more space for expressing emotions (‘What sadness’), once the tragedy of the situation becomes real. One glance into the incubator is enough for him to believe that in this case humane grounds alone would warrant abstaining from further intervention: ‘This is enough; it is unacceptable to go on.’ This should not lead one to conclude, however, that at the side of the incubator one will always employ a situational repertoire. Repertoires are not automatically tied to specific locations. For example, specialists who examine children in the NICU and discuss their findings with the attending, frequently tend to refer to the medical literature, thus relying on a case repertoire.

The Moral Load 165

In the formal context of having to write a neurological report, however, a strictly medical vocabulary once again prevails. Even if this does not change his conclusion, it does change the nature of the argument. ‘Humane grounds,’ a decisive argument at the side of the incubator, is now just one of several reasons and as such it is also the last one mentioned. The report couches the tragedy of the inevitable decision in formal medical language. To describe the medical nature of the tragedy involved the neurologist refers to cerebral pathology (ultrasound), hypotension problems, and experience with other cases involving potential multi-organ problems. These facts together contribute to a morality that situates a continuation of treatment in this case beyond the boundary of the acceptable. The decision once again proved its robustness. One more time the decision is relocated in order for it to be strengthened. The attending communicates the decision to Esther’s parents. They have the last say. And they too agree to it. In the decision process on stopping Esther’s treatment the participants mainly speak in terms of an active bleeding, a child that is seriously decompensated, hypotension as clinical picture, a prognosis, and an ultrasound image. Obviously, a medical terminology prevails.4 At first sight, only the factual aspects seem to count, with no role for moral views and positions. Yet, such a picture does justice neither to the integrity of those involved nor to the complexity of the decision. What is more, such a conclusion would confirm the idea that there is a strict distinction between facts and norms. Although in the staff meeting – the formal context for taking moral decisions – medical data are discussed, this does not mean that the course of action is exclusively determined by the facts. On the contrary, morality proves to play a definite role. If we consider how the facts were established when they were not recognized as such, it shows that simultaneously moral choices were made. The testing of prognostic facts and the testing of values go hand in hand. To categorize a child in the group for which withholding treatment is deemed acceptable also involves assessment of what is factually hopeless and normatively unliveable. But when are a child’s prospects hopeless or unliveable? In the case of a two-sided brain haemorrhage degree four, says the one. Not true, says the other, in the case of a two-sided brain haemorrhage degree four plus a specific location. To establish this, one needs an EEG result and a radiological report on the ultrasound scan of the head. In Esther’s case it is impossible to meet the two standards. Situating morality calls for more than the application of general moral rules to a specific case. In the context of the staff room, the boundaries of the category are shifted back and

166 Uncertainty in Medical Innovation

forth. The standards for falling within the category, as well as the categories themselves, have to be made to fit. If the child’s situation is assessed in general terms of the category, the category itself is also modelled after the specificity of the child’s situation. The degree to which the standards are allowed to be flexible and the boundaries of categories are allowed to be stretched so as to cover the child’s situation is the outcome of a negotiation process in which those present manage to combine the case repertoire and the situational repertoire with each other. In one and the same movement the child’s situation (the facts) and the norm (the category) are defined.5 This leads to a situated morality in which the contrast between ‘general categories’ and ‘the child’s specific situation’ has been resolved.6 However, morality does not only involve the testing of a prognosis, the assessing of standards, and the bringing into line of repertoires. It also, and perhaps even more so, involves the consequences of acting.7 The tragedy that is implicated in life-ending medical action comes with a moral cost.8 How is this moral burden shared? Let us take a closer look at the way responsibilities are distributed and the effect of relocation on this process. Creating a basis for consensus Before a child is born the decision power is in the hands of the parents, the mother in particular. The child after all is still part of the mother’s body. At this point physicians merely have an advisory role.9 Esther’s mother could have chosen, for instance, to refrain from medical treatment. She would have simply walked out the hospital. But this does not happen. She stays on and has to decide on what should happen next. The juridical principle of parental authority divides the group in the delivery room in two: the medical staff and the parents. Because of this principle all responsibility is in the hands of Esther’s parents: they have to make a decision, whether they want to or not. It is up to the doctors to help them in doing so, whereby these doctors themselves bear no responsibility for the final decision. Several nurses from the Academic Medical Center and the Northeastern General Hospital question the practice on parental decision right. They feel that parents are only given a say after physicians no longer know what to do or further treatment seems not useful anymore. This would happen only if all medical options are exhausted. ‘It seems as if parents are allowed to join the conversation only after the doctors no longer know what to do,’ a nurse from the Academic Medical Center claims.

The Moral Load 167

In case of hospital admission after a child’s birth, the attending neonatologist will engage in medical intervention and this means that his or her merely advisory role is left behind. By connecting the child to the respirator the physician in charge becomes responsible for the treatment. In the Academic Medical Center this responsibility is not only in the hands of the attending, as it is shared with other medical staff members. After the attending decides at the side of the incubator that further treatment is meaningless or undesirable, this view is discussed in a staff meeting. Yet by exercising their right to join the discussion the participants also join in the responsibility over the final decision. This no longer rests on the shoulders of the attending physician, but is shared by the entire medical staff and the others involved. In Esther’s case this distribution of responsibility even exceeds the boundaries of the NICU: a radiologist and a neurologist are also asked to share in this burden. Their advice, unlike the advisory role of the neonatologist in the delivery room, does come with an element of co-responsibility for implementing the policy of withholding treatment in Esther’s case. Similarly, Esther’s parents have to make an explicit statement about their decision to refrain from further treatment of their child. This is their right as bearers of parental power. The staff is quite aware of the moral burden implied in this decision power. At the staff meeting, the staff tries to remove part of the moral burden from the parents without tampering with their decision power.10 In relocating the decision more and more individuals become involved in the decision while simultaneously distributing the moral burden among the various actors. With this increased involvement the individual doctor’s decision on the ward at the side of the incubator gradually evolves into a shared advice of the staff to the parents. Much of the discussion in the staff room is geared towards an outcome that all participants can support. In this search, ‘the decision’ changes into ‘an advice’, a formulation that all can agree to. From this angle, the function of the relocation is also to render the tragedy that is implied in a decision on life-ending action malleable. In the process of the decision’s hardening and unhardening the actors get the opportunity to appropriate the decision, make it their own. The development of ‘who we are’ and ‘the decision’ evolve jointly. The form of the final decision and its outcome mirrors the moral identity of the unit. This kind of decisions and the ways in which they are dealt with allow the NICU staff to show its moral face. An analysis of the genesis of a decision also renders visible the normative character of locations. It is perhaps possible to make a convincing

168 Uncertainty in Medical Innovation

case for discontinuing treatment at the side of the incubator, but the same may not be true in the staff meeting where that decision has to be substantiated. Emotions that may emerge at the incubator’s side can be misplaced or play no significant role in the meeting room. Analysis of the process in which a decision was articulated, however, does not only reveal which norms and values are used in the NICU and when or where they are deployed. It also shows how staff members deal with each other. In other words, an analysis of the decision’s design also reveals the internal power structures involved. In decisions on life-ending action the actors do not have an equal say. For example, doctors can call a staff meeting, but nurses cannot. In the case of Esther, the nurse involved does not see this as a problem. But in the treatment of chronic infants like Tom this is another matter. Nurses who care for individual children like Tom, for several weeks or months in a row may earlier reach the point at which they feel continuation of treatment to be meaningless than the doctors involved. Even if they can inform residents or more experienced nurses of their view, only when the attending has arrived at the same conclusion a staff meeting is called and there is a formal context for discussing major treatment decisions. Nurses do not carry responsibility for the treatment and are in no position to determine whether a treatment will be stopped or when this ought to be done. Yet nurses are directly involved in the consequences of specific decisions on treatment. At first sight, then, it seems only reasonable that nurses should also be in the position to call a staff meeting. In practice, however, the issue is more complex. First, my analysis of the decision-making process has shown that the staff meeting is not so much the ‘site’ where the decision is made, but where an already taken decision is hardened by subjecting it to a trial of strength. After all, the decision to stop Esther’s treatment provided the occasion for the staff meeting. Second, my analysis also underscores that the setting in which the decision is discussed certainly matters. The kind of data to be discussed during the decision-making is, for instance, affected by the power hierarchy of the NICU. Although doctors and nurses may share repertoires, the specific setting of a staff meeting will offer more or less space for various types of arguments. If statistical data from the medical literature play a role in the outcome of a decision process, it is mainly the doctors who are familiar with this case repertoire. If doctors do not share the nurses’ view, in part because of the role of this type of argument, their claim is soon settled to the advantage of the physicians. Other contextual constraints may be

The Moral Load 169

related to the issue of a practical culture in which there is a preference for informal consultations between physicians and the dominance of evidence-based information. These differences in power, access to preferred data and moral perspectives complicate the position of nurses in participating in end-of-life decision-making processes.11 An official staff meeting cannot remove the unequal doctor/nurse relationship.12 This does not yet mean, however, that mutual deliberation is useless. What happens when there is no deliberation at all in a unit becomes clear when we call to mind the process in which the decision on Maureen’s treatment in the Northeastern General came into being.13 There is no collective deliberation in this hospital, neither in staff meetings nor at the side of the incubator. In the decision process on Maureen’s treatment each form of communality is absent. Those involved either agree or feel excluded, given that there is no formal basis for sharing views on a child’s situation. Similarly, there is no formal context for articulating positions or distributing responsibility. In the Northeastern General Hospital everyone but the attending on duty feels both powerless and responsible. The neonatologist who was the attending in the second month of Maureen’s admission didn’t sleep at night because of this situation. The parents feared the signature and stayed at home while some of the nurses refused to care for Maureen any longer. Maureen eventually died on the twenty-third day of her admission, several hours after being disconnected from the respirator. In the NICU of Northeastern the decision on whether or not treatment should be refrained is neither discussed collectively nor born collectively. Those involved in the treatment of Maureen are not united in a ‘we’. Instead, there is grumbling, uncertainty and indignation. In this situation power and powerlessness are magnified. In Maureen’s treatment from the start contention prevails between nurses and physicians, between the attending and the assistant physicians as well as between the staff and the parents. There is a dividing line between the powerless without decision power and the powerful who have to decide, whether they like to or not. In the daily course of affairs this redistribution of authority and its related responsibility does not have to lead to problems because it creates a clarity that is of fundamental importance for the construction of a practical order. The mutual relationships are evident and all actors do what they have to do. But a conflict may turn this transparency around and lead to a hazy situation in which it is not so clear anymore where decisions are made. Who, then, is authorized to decide? Is it the physician? Is it the parents? Or is it ultimately the law that regulates the course of affairs? Evidently,

170 Uncertainty in Medical Innovation

there is not a single centre of power, but a jumble of opinions, problems, heterogeneous values and varying facts. No matter how responsibilities are distributed, decisions about life and death are taken and in most cases it implies the end of the trajectory. When in Esther’s case the treatment was actually stopped, she died within minutes, but the decision to turn to a Do-Not-Resuscitate code does not always swiftly end the trajectory. However, this last part of the trajectory deserves attention as well. Therefore I return to the NICU ward in the Academic Medical Center to follow the trajectory of Tom.

6.3

Another trajectory, another order

Like the trajectory of Esther and Maureen, that of Tom is marked by maximal intervention from the start. But unlike Esther and Maureen, Tom is still in the NICU after five months. An accumulation of problems has made his situation hopeless. Academic Medical Center Thursday: One by one the neonatologists enter the NICU. This morning a round will take place at which all children are discussed extensively by the entire medical team. Awaiting the round’s start, two neonatologists together review the last data on Tom. The round begins. In each case the resident reports the course of affairs, the problems involved and the suggested policy. The group moves from one incubator to the next. One child may take up more time than another. Some children are stable and doing well while others involve a more complex picture. Tom belongs in the latter category. The group gathers around his incubator. Although known to all, the resident sums up the long list of problems: a high oxygen need, kidney stones resulting from the lasix*, a stomach ulcer on account of the corticosteroids which makes it hard for Tom to digest food, spasms caused by the increased pressure in the head, the drain that had to solve this appears to leak somewhat. The neonatologist on duty adds further data, points to the disastrous developments of the past period and the ensuing bad prognosis. Tom’s lungs are damaged beyond repair so that he no longer can breathe without artificial respiration. Given the fact that he engages in a constant struggle with the respirator tube, this makes a future hooked on to this device into an endless agony. Everyone agrees that this can no longer go on. There is

The Moral Load 171

no chance of recovery whatsoever and Tom seems to have no more energy to brave another crisis. It is decided that on Monday there will be a talk with the parents. For Tuesday a staff meeting is scheduled to determine the policy. Monday: The attending and a resident are sitting in the parents’ room with Tom’s parents. The doctors extensively explain his current situation and the past months pass in review again. Even if the list of problems addressed is impressive, his parents’ feelings of hope are not so easily brushed aside. Tom’s father: ‘we cannot handle the final decision. If we say “just go on treating,” we may perhaps keep him alive three more months while to him it only means more suffering. If we now say “just stop with it,” we would be left with a feeling that perhaps something could have been done after all.’ The physicians, when asked for advice, propose to change to a non-reanimation policy, to which Tom’s parents consent. Tuesday: At the staff meeting all neonatologists are present, as well as the nurse and residents involved and a social worker. It is decided that they no longer intervene in case of a crisis. Later that day the attending informs Tom’s parents about the meeting and they agree once more with the new policy. In the staff meeting on Tom it is decided to switch to a non-reanimation policy, meaning that he will not be actively treated anymore and that any new medical problem, such as a bradycardia*, will not be responded to. He will continue to receive daily care, though. With this policy the staff enters a new treatment trajectory, one without checks and tests. While the nursing staff becomes more central, the medical technology is relegated to the margin. For an answer to the question of how Tom is doing one no longer gazes at the monitor or at the flow chart but turns to the child in the incubator. This trajectory is marked by serenity rather than hectic activity. The acceptance that Tom is better off with death than with life, however, also marks the beginning of a period of anxious waiting for his parents, as each phone call can be the one with the fatal message. How should they get through the day, after all the exhaustion of the previous months? The nursing staff too feels more anxious, now that each moment Tom may work out his tube and enter into a bradycardia. Although all tools for reanimation are within reach, the policy now is not to use them in his

172 Uncertainty in Medical Innovation

case. Non-intervention basically implies that their hands are tied and can only stand idly by also when a crisis comes up. If the waiting has begun, things soon prove to be slightly more complicated than foreseen. The staff meeting in which the non-reanimation policy on Tom was adopted took place three weeks ago. Since it was agreed to withhold treatment, he is left alone, and this appears to be beneficial. He sleeps well and responds less panicky. He is very active, clear and eats well. Tom improves and his condition painfully reveals the effects of the activities that were central in the previous part of the treatment trajectory. It was marked by maximal intervention, implying a search for the right course of medical action, and this called for a detailed picture of his condition. For collecting all the relevant information Tom had to undergo a sheer endless series of interventions in the form of checks and tests. Now, in retrospect, it turned out that these interventions also have contributed to his deteriorating condition. If his condition seems better, doubts about the correctness of the current policy become stronger. As his father comments: ‘Tom is much better now than two weeks ago. He looks better and is much more relaxed. Occasionally he makes contact. Sometimes he is looking at me for as many as ten minutes. When he is looking at me that way I begin to doubt whether we made the right decision. You have no solid basis. The more I delve into the matter on the Internet, the more I find that they still know very little about this kind of children.’ The parents of Tom are no longer sure of the correctness of their decision. They feel they have more contact with him than ever and hardly believe he is in such bad shape, so that they start having doubts about the DNR policy. To end their doubts, they ask the physicians to redetermine Tom’s condition. Although they feel that his prognosis has not changed, they agree and have new testing done the same day. This leads to uncertainty among those in the night shift on what to do in case of an emergency. What is the current policy? In the light of the adopted DNR policy, a new request for a physical examination causes confusion. Which trajectory are they on? To end this uncertainty, a nurse raises the issue at the daily medical round. Academic Medical Center The nurse: ‘what is actually the point of the tests performed on Tom? Last night there was some confusion among those in the night shift. They didn’t know anymore what to do if something were to go wrong with him.’

The Moral Load 173

The resident joins her side, for she also has trouble with the tests because what is their point? The neonatologists do not see the issue. As one of them comments: ‘our agreement is clear, isn’t it? And if our policy will change, everyone will be informed. Moreover, we have no reason to assume that the results will be better.’ Both the nurse and the resident are not convinced: ‘what is the use of those tests? And what are we to do when they show improvements? It was our policy to do nothing all along. This includes doing tests, isn’t it?’ The neonatologists stick to their guns. The attending: ‘The tests were done on the parents’ request and our policy will not be changed. If there will be reason to do so, it will be formally discussed with all parties. I do not see why the night shift has a problem. The policy is clearly spelled out in the record and it says that the tests were done at the request of the parents.’ One of the neonatologists sides with him: ‘We have to see the parents’ question as an expression of doubt. They doubt the prognosis that underpins the policy.’ A nurse comments that some of them share this doubt and that ‘many nurses have trouble with that policy. That it was agreed upon in the presence of three staff nurses does not automatically mean that the other fifty agree.’ The attending: ‘those three nurses represent the nursing staff. You cannot expect us to discuss the issue subsequently with each of you individually.’ ‘I feel that there should be room to articulate our feelings,’ one nurse reacts. One of the neonatologists tries to calm the nurse: ‘Of course we take you into account and nothing will be decided without you knowing. If there is anything, we are always willing to listen. You know it; it has always been that way. But what do you think? That because his blood results are better his lungs will improve again as well? Impossible, his lungs are entirely destroyed. What is not there anymore can no longer be cured.’ They agree that the DNR policy will not be changed. The shared basis for the adopted DNR policy appears subject to erosion. With his clear gaze and relaxed behaviour Tom has undermined the robustness of the decision to withhold treatment. The unexpected change in his behaviour puts pressure on the prognosis and his parents and

174 Uncertainty in Medical Innovation

several nurses begin to doubt the correctness of the choice for a DNR policy. The consensus on this new treatment trajectory is slowly disintegrating. In the NICU context the DNR decision proves to be less manageable. The physicians try to turn the tide before doubts gain the upper hand and point to the irreparability of the damaged lungs. A policy change, then, is not at issue. If the adopted course is not changed, new problems have arisen. A DNR policy does not mean that a child is disconnected from the respirator. Such intervention belongs to another trajectory, the one of active life-ending action. Prior to the staff meeting, it was the damaging effect of this machine that was foregrounded and seen as defining the direction of the treatment’s development. After this meeting the respiration pressure is as harmful, but it is no longer a focal point. This is not to say, however, that the respirator has a marginal position at this point. Also on the DNR trajectory this device is one of the key players. Given its script, it is still required at regular intervals to tape the respirator tube onto Tom’s upper lip, so as to prevent the tube from slipping out of his throat. But when this is done, he tends to panic and strongly resist. This creates the risk that when the old tape is removed the tube may slip out. The DNR policy, however, prohibits re-intubation, which means that Tom will die at that point. A nurse who cares for him: ‘The policy involving Tom is hard to carry out for us. We all fear to be the one to have to face the fact of the tube slipping out of his throat. This fear has even become stronger now that his parents have indicated that they absolutely want to know who’s on duty when something happens to Tom. The way they put it seems to suggest that they do no want to talk it over; it rather seems they are out to file a complaint against you. Because of his parents’ attitude, no one really dares to care for him. It all shouldn’t last too long anymore. Some nurses even have indicated already that one day their finger will ‘accidentally’ be caught on the tube, implying active intervention on the part of the nursing staff. Then this horrible situation will have come to an end at last. It’s all lasting much too long. If this policy was agreed with three nurses representing us, the fact that the other fifty have trouble with it is not considered.’ The decision to switch to a DNR policy has transformed a routine act, such as the taping of the tube, into a risky act. On another trajectory with another objective the same medical act has another meaning.

The Moral Load 175

With this child at this stage of his life, the taping of the tube leads to other circumstances than those foreseen in the staff meeting. All parties involved share the responsibility for the decision to stop treatment. However, the burden of the consequences of this decision ends up inordinately on the shoulders of some: the nurses who take care of Tom. They may accidentally cause his death. Also the physicians on duty belong to the select company that may directly feel the consequences of this decision: in case of a crisis they are only allowed to watch it. Academic Medical Center Monday night: It is around ten when Tom’s heart rate drops to 100 beats per minute. The physicians and nurses worry that he will no longer be able to get it up again alone. In line with the policy no one intervenes. The physician phones his parents in the hope that Tom will still live when they arrive in the unit. His heart is ever slower now. When after half an hour his parents come in, he is still alive. After some talk his mother cautiously takes him into her arms. Not long afterward Tom dies, on his mother’s lap. Tom was six months old. After his death the equipment is turned off and the monitor falls silent.14 But not all is said and done yet. One more time those involved change their trajectory and switch to the set of actions that are tied to handling things after a child has died. This trajectory too has its own order. It is marked by doing things in slow motion, by respect and another language: that of compassion and the absence of numbers. There are no more toiling ventilators, no blinking monitors or beeping alarms. Here there is tranquillity and all technology is brushed aside. If Tom’s life in the NICU was always marked by a constant entwining with the technology around his incubator, only after his death these ties can be severed, which ends his hybrid existence. Once he is released from all lines and tubes he can wear regular baby clothes while for the first time his parents can hold him in their arms. The death of the child marks the transition to a new trajectory and therein the parents are central.15 In this trajectory the emotional bond with their child plays a major role. As much as they can, staff members try to help parents to cope with their loss. They start from the view that parents will better cope with their loss if they have had a good emotional bond with their child. Bonding in order to let go is the underlying idea here. Good contact with their child would reduce parents’ sense of guilt and improve their self-worth. It is based on this premise that Tom’s

176 Uncertainty in Medical Innovation

mother gets him on her lap so that he can die in her arms.16 After his death his parents spend much time with him alone in the parents’ room. The parents of Esther also had her with them after she died. Her mother: ‘After Esther died I did not want to see her anymore. But the nurse recommended otherwise because we could regret it later on. I then said: Fine, we do it. When I saw her wheeling the bed towards us I felt so scared that I wanted to run away immediately. But when I saw her and saw how serene she was, it instantly made me feel calm.’ Her father: ‘This is the best memory we have of Esther. Only then I could hold her for the first time. When I held her in my arms and hugged her, this triggered a deep sense of fatherhood. Very strongly I felt that afterward I wanted to have another child. She had such a sweet face. I would recommend everyone to do it. It seemed scary at first, but it did a lot of good to us.’ One tries to help parents with their grief in other ways as well. With this aim in mind, in the Academic Medical Center they take pictures of the child after s/he died. If parents do not wish to have them at that moment, they are stored in the record so they can always ask for them later on. Pictures were made of Tom as well. They show him in a cute blue baby suit, white socks and with no respirator tube in his nose. Especially the latter makes a big difference. In hindsight it shows how the tube and the tape with which he was attached defined how he looked. Without tube and tape he almost seems another child. Now his incubator is empty the staff realizes how they have grown accustomed to his presence in the NICU. It is an odd sight on the day after he died: this empty incubator on the spot that over the months had become his whole domain. Yet a sense of relief prevails among the staff. The resident who was on duty when Tom died: ‘I feared it would happen in my shift. I already worried about something going wrong with the tube. Fortunately it did not go this way. He fell asleep very calmly. The parents spent quite some time with Tom on their lap in the parents’ room. Too bad I hardly had time to talk to them, because at the same time a new admission came in.

The Moral Load 177

This was a premature of twenty-six weeks who needed all our attention.’ One of the nurses from the Academic Medical Center: ‘It couldn’t have gone more beautifully. Now there is no one who will be left with a sense of guilt. Tom did it entirely alone. It all was very peaceful. Like a candle that slowly went out.’ Relief is but one of the emotions that occur after a child’s death in the NICU. In their study on coping with grief among nurses in a paediatric IC, Rashotte et al. (1997) describe that relief mainly occurs among experienced nurses. The death of a child may also lead to guilt feelings, anger, and deep sorrow. Nurses have various ways of dealing with feelings of grief. For instance, by talking a lot about it with each other, by controlling their emotion, or by dissociating when it is clear that the child will die. The relationship with the child’s parents plays a major role in how nurses cope with grief. A good relationship with them and the opportunity to support them help nurses to cope with their own feelings. A bad relationship with the parents, so Rashotte et al. argue, also has a negative impact on how nurses cope with their own grief.17 With Tom’s death the tension that had been building for months around his incubator suddenly ebbs away – a tension that originated in a confluence of events, decisions, actions and interactions among the people involved, as well as between them and a host of technologies. Inadvertently, this led to a situation in which Tom’s quality of life was increasingly considered insufficient, while improvement of his condition could no longer be expected. The discussions around Tom’s incubator illustrate clearly how in case of a bad prognosis the self-evidence of the treatment’s continuation may be questioned and it may even be decided to stop treatment. A careful examination of these discussions shows in detail the interlinked process of establishing medical facts and moral concerns in life-ending decisions. The involved staff members and parents have to deal with difficult dilemmas. Each of the options has serious ramifications. For instance, if they decide to discontinue Tom’s treatment they may rob him of his final chance of recovery, but the opposite decision implies the risk that their effort only contributes to the child’s senseless suffering. This chapter demonstrates how the testing of prognostic facts and the testing of values go hand in hand. Questions addressing the prognostic nature of a case and how to act properly cannot be dealt with in

178 Uncertainty in Medical Innovation

isolation: a specific answer to one question implies a particular answer to another, and vice versa. Facts, values, and medical actions are mutually interdependent. Furthermore, this chapter illustrates in detail how formal regulations are reshaped in the swirl of medical practice. Regulations and guidelines, in the form of assessment schemes or decision procedures, aim to guide and direct the NICU staff in taking the proper decision. These guidelines originate from both a societal call and medical need to regulate decisions on life-ending action. However, an analysis of situations in which guidelines apply show that existing rules come with a specific notion of the practice involved and that this notion is not always in line with the actual practice. These guidelines start from a problem or situation that is not always identical to the situation at hand. In everyday NICU practice, general moral principles always have to be applied in idiosyncratic ways: they have to be ‘custom-made’ so as to fit the concrete situation. Basically, the issue of morality in the NICU context involves a decision about fixing a demarcation point, a boundary beyond which intervention is discontinued. To generate a custom-made morality in specific cases, several activities at various moments are required. The challenges of NICU life, in other words, require a situated morality in which the distinction between ‘general moral category’ and ‘the child’s unique situation’ is resolved.

7 The End of the Journey

The NICU is a setting where decisions have to be made on a permanent basis. Some are directly or indirectly related to the quality of the life of the child and/or the parents. These moral decisions, however, do not form a separate category, next to factual or medical decisions. In theory morality applies to the good life and how to act properly. In the cases addressed in this book a standard ethical approach would focus on the relation between ethical principles and the child’s specific situation. Through deductive reasoning, ethical principles can be applied to a child’s situation. But, as my argument has revealed, this deductive model is not used in the NICU when it comes to making moral decisions. In this context, these decisions do not involve the application of general ethical rules to a unique situation, but call for another activity. Significantly, morality has to be custom-made, tailored to fit in the NICU practice. In fact, morality mainly appears to involve the fixing of a moral turning point: the boundary that cannot be crossed. Here morality means demarcation. This demarcation is construed on the basis of a range of divergent activities by several actors in various sites and at different moments. Custom-made morality requires testing the robustness of facts and the elasticity of moral categories, exploring the room for medical action, searching the right words, creating a sense of shared concern, exploring disciplinary boundaries and distributing responsibility, and learning to deal with the consequences of decisions. In the NICU one encounters a situated morality in which the distinction between general moral category and the unique case (the child’s situation) is resolved. Yet moral decisions are often seen as a separate category next to medical decisions. How is this possible? It has to do with reflection. The reason for this distinction is found in the difference between action and reflection. It is only afterwards that we can 179

180 Uncertainty in Medical Innovation

see the reality of deductive moral decisions. It is only in retrospect, in reflection, that we perceive and recognize deductive moral decisions. If Tom’s death marks the end of his journey, it also gives rise to a process of reflection and reviewing in order for those in the NICU practice to move forward again. In this concluding chapter this retrospective gaze is centerstage. As such it also applies to this study as a whole and the various treatment trajectories addressed. Therefore, I too will reconsider the questions and concerns articulated at the beginning of this study. In the introductory chapter I announced that this book would provide insights into both the reshaping of the societal responses to health innovations in the concrete work setting of medical care practice, and to open up the interface between diagnosis and prognosis, between men and machine, and between medical facts and moral concerns and to examine these interlinked yet discrete processes. In the second part of this chapter I will return to these issues, but first I will focus on the NICU as a reflective practice.

7.1

Moments of reflection

The neonatology practice, with its checks and balances and various inputs ranging from technologies and numbers to emotions and uncertainties, is a reflective practice – one in which moral considerations play a major role. This is the case not just in terms of formal ethics or in the retrospective reflection of those explicitly named and recognized as actors, but it is also interwoven with the dynamic of the daily course of affairs and the ordering mechanisms in this practice. The cases described show that a decision that is taken in a staff meeting is only one of many. By situating such decisions within a wider temporal frame they appear in another light and it becomes possible to reflect on the nature of the moral choices involved. In the case of Maureen the choice for maximal intervention proves to be part of a process that had already started years before her birth. Being an IVF child, her existence was already entwined with medical action, technological interventions and statistical chances prior to her conception. Hooking her up to the respirator in the delivery room thus became a logical step in a process of a year or more – and one that was all but finished. Maureen represented ‘an investment’ that had already taken much ‘effort and money.’ Her parents’ age, moreover, made chances of another successful IVF treatment slight. Thus her treatment became ‘a matter of saving what could be saved.’ Not only the specificity of a child’s situation deserves our attention, but also that of the process in which treatment takes shape. This makes it understandable that decisions to start or discontinue treatment of

The End of the Journey 181

children with a similar medical problem may have different outcomes. Such decisions of course are closely intertwined with previous events and decisions as well as with future prospects. This puts pressure on a conceptualization of moral choices as decisions that are made at a certain time and in a specific place. The notion of moral choice as tied to a specific moment only emerges in our reflection on such decisions. However, this moral reflection usually concentrates on one aspect or moment of the treatment trajectory, for instance, a staff meeting in which staff members collaboratively discuss the decision whether to go on treating in a specific case. This reflection tends to assume a unit that is neatly divided into a world of facts and a world of values; a domain in which there is a clear distinction between main issues and side-issues. Such temporal and local fixation, however, automatically ignores major dimensions of the treatment trajectory, such as the process in which the situation that determines particular decisions has emerged, including the actors’ interplay, the contingent events along the trajectory and the power structures in the NICU. In reflective fixation, moral issues are detached from the history of the situation in which they occurred and reduced to conclusions of more or less rational considerations. The various sorts of noise – such as mutual conflict, frustrations about a child who is struggling all the time with a tube, lab results that are flawed or cannot be found, parents who refuse to come – have been filtered out. Ethical reflection arises from the order that a retrospective gaze establishes. Once the story’s ending is known, acts can be interpreted ethically, which in turn provides ground for discussion. This is precisely characteristic of reconstructions. This is why an ethical handbook can propose that first facts have to be clear, only after which the moral considerations can be addressed. As has become clear in this study, the world of the NICU is anything but orderly. Here values do not come before or after the facts, but the two categories constantly mingle. In the NICU one has to deal with organic, process-like developments, some of which may follow textbook descriptions while others evolve much more rapidly or end in a deadlock. In these situations the context of decisions tends to be shadowy and changeable by definition. This complicates one’s overview of the situation. As a neonatologist from the Academic Medical Center puts it: ‘Sometimes you are only busy solving separate problems. You refuse to add up. You’re absorbed by your relentless search for possibilities to solve problems, one after the other. This takes up so much of your time that you lose sight of the overall picture. You do not even want to think about it. There is still room, after all, for doing something. But

182 Uncertainty in Medical Innovation

when later on you look back you realize that things were wrong at that point already.’ Evaluation assumes a specific result that makes it possible to see what one did not see before. In an unfolding treatment trajectory many elements that are hard to integrate in protocols prove to be crucial. Each situation, each detail can have a normative influence on that treatment. Are those involved merely carried along with the flow of the ongoing dynamic of the treatment process and should we assume that only outsiders can reflect on what takes place in a NICU? This conclusion seems premature because the NICU practice itself has reflective moments built in. The staff meetings on Esther and Tom offer clear examples, as is true of the attending physician’s meeting with the parents – several weeks after their child’s death – that specifically aimed to look back at the trajectory’s course. Thus the staff tries to help parents in coping with their loss: a form of retrospection aimed at allowing them to move on with their lives. Two days after Esther died, her father talks about how he and his wife plan to get through the first days: ‘Currently it feels as if we’re in a movie. It all went by so fast that we cannot comprehend our reality. Tomorrow is Esther’s funeral and afterwards we go home. We are not looking forward to it. It will be the end of the movie, after which we will find ourselves in a void. My wife plans to go to her mother’s place for a few days first, so that I have time to dismantle what was to be Esther’s room at home. We keep a small shoebox with the cap she wore in hospital, the label from her incubator and that sort of thing. Everything else will be stored in a large box in a far corner of the attic. And if we are to have a new baby, we buy everything again. The contents of this box are Esther’s belongings and they will remain so. You do learn who your friends are in a period like this. There seem to be fewer of them than we thought. And you also have people who say: “Oh well, you’re still young. You may have so many other children.” They really don’t have a clue. But, sure enough, we have to pull through.’ There are also moments of reflection that apply in a more structural way to the NICU practice. For instance, the medical staff in both the Academic Medical Center and the Northeastern General regularly meets to reflect on treatments that were not concluded successfully. In retrospect the physicians evaluate their actions and decisions. In the Academic Medical Center there are also multidisciplinary deliberations in which physicians and nurses discuss children’s situations together. These

The End of the Journey 183

moments of reflection, however, tend to be the first to be sacrificed; in fact, work pressure and understaffing caused these meetings to be cancelled for some months in a row. As a result some nurses feel that they had no possibility to present their views on specific treatment policies. For the nurses who cared for Tom the actual implementation of the DNR policy proved to be a burden, but they felt that the doctors were not listening to them. This is why they asked for a meeting, but this meeting, too, was cancelled because Tom died the day before. One of the nurses complained about this at a nursing staff meeting: ‘I feel it is wrong that the meeting about Tom was cancelled, as if it was only going to be about him. It was also meant to deal with our policies. Parents receive enough support here, but I feel that there is not enough support for the nursing staff. This is why I think we still have to schedule a meeting.’ This nurse raises the issue of standards that are important for the constitution of the practice. These standards also affect how specific problems are interpreted. In the Academic Medical Center, staff members do not formally discuss each treatment afterward. Many decisions on intervention are routine and need no further discussion. When, however, does a routine case change into an exceptional case? It is not always clear, as one nurse explains: ‘In Tom’s case it was all very gradual. After a month of treatment you have no overview yet and you do not anticipate that six months later he will still be in the unit. One morning you come in and realize how long he has been around. We always have a child like Tom here, and then one day it is over. We actually talk too little about what happens to them, also among the nurses. Occasionally you may overhear someone talk about Tom and how sad his case was. But what about his parents and doctors? You do not know what they feel at the side of the incubator or when they are busy with him. At one point you reach a point at which nothing works anymore. Then you have reached the boundary between treatment and maltreatment. Some time ago I have been away from work for six months in order to clear my mind. It is precisely this kind of case that causes you to feel the need to do so.’ As this nurse suggests, it is crucial to reflect on the past in order to muster new energy to be able to go on.

184 Uncertainty in Medical Innovation

Reflection also occurs in informal settings. Over coffee and lunch breaks nursing staff members chat with each other about children such as Tom. Although in his case relief prevailed after his death, it also triggered questions such as: Why did we start treatment in his case to begin with? Why didn’t we decide earlier to withdraw treatment? Didn’t we know all along that Tom had no chance of survival? Knowing how a particular story is concluded makes a world of difference. In their reflection on a particular course of treatment, those involved have the opportunity to see a neatly ordered world of facts and values, of main issues and side issues. This allows them to review the actions and re-assess the choices that were made. Moral reflection emerges out of the order that is retrospectively construed. The intervention trajectory – once so full of unexpected surprises – is now a logical narrative, one that can be interpreted and debated in terms of the moral nature of the acts involved. How did those involved in Tom’s treatment looked back upon his life? Rewriting the past in reflection In looking back on Tom’s treatment, the actors reflect on the interventions applied and the decisions made. What was once an erratic treatment trajectory is now a consistent story that is subject to interpretation. If right after his birth maximal intervention seemed proper medical action, it certainly lost its shine or heroism in light of the outcome. The actors, however, have different views on the cause of Tom’s misery. Again there is a redistribution of responsibility. One of the nurses, for instance, puts blame on the physicians: ‘The doctors have painted a much too rosy picture to the parents, which caused them to still have hopes and therefore they continued to insist on reanimation. Moreover, we have the sense that doctors increasingly hide behind theoretical chances and use everything at their disposal in terms of state-of-the-art technology, so that they do not have to take on the responsibility for stopping a child’s treatment. It seems as if physicians no longer dare to take such a decision and hence postpone it all the time. Thus a child gets no more chance to die.’ This nurse views Tom as a child who was not allowed to die rather than as a child who had a chance to survive. In her opinion ‘doing well’ in this case is to admit one’s powerlessness and have the courage to give up.

The End of the Journey 185

The attending acknowledged this principle but he has another opinion on the proper moment for the decision to refrain from further treatment: ‘It was the presence of real possibilities that made us decide to go on treating Tom, rather than exclusively defining his situation in terms of pain and suffering. Emotionality should never be the basis for such decisions. It could cause medicine to slide down to the level of “we cannot handle it, so let us just end the patient’s life.” What is more, much of the suffering was not caused by us but was a given we had to deal with, whether we wanted to or not. In addition, we cannot ignore the judgement of the parents. His parents viewed it all much too positively. During our last effort to reanimate him, therefore, we had to pull out all the stops to “bring him back.” That a child recovers does not say anything about what will happen next, though. But his parents seemed not to be willing to hear this. Whatever happened they continued to be hopeful and therefore each time they insisted on reanimation.’ In a similar vein, two residents discuss Tom’s treatment during a lunch break: ‘In Tom’s case we kept going on too long. We knew that he was going to die.’ ‘Well, we are not the only ones who have a say. There are also the parents.’ ‘But we all had quite a lot of difficulty with it. We go on even if we do not always see why. It is only that our superiors have so decided. They are quite clear on this issue: we will always intervene. They say we can’t judge on someone else’s quality of life and hence we have to act. Yet I have great trouble with this.’ ‘Heck no, in ten years you talk just like our superiors and will treat all cases. This is how we are trained after all.’ This last remark calls up an image of neonatologists as doctors who always want to try and act. Maximal intervention does not preclude reflection, however. One neonatologist from the Academic Medical Center looks back at Tom’s treatment as follows: ‘When you see these six months of suffering and the final result is that the child dies, you start questioning what you are doing. Neonatology is developing very rapidly. But is society better off with it, or does it

186 Uncertainty in Medical Innovation

make parents happier? You do wonder in such situations. But you cannot tell in advance how things will play out. You do not know in advance whether a child will make it or not or when it will die. This is why you have to get the most out of its treatment. You try the available options until you have exhausted them all and cannot go on anymore – until you are standing with your back against the wall. Only then you say “there is nothing we can do anymore.” But you do not know this when such a child is born and you embark on your effort. Of course, you look at what you are holding in your hands. And if the child is really very underdeveloped, you do not go on. But usually you first act, only to start thinking later on.’ Tom’s parents have doubts on whether those involved acted properly. To them, every decision is too guilt-ridden to allow for clear and unambiguous choices. Tom’s father: ‘If we had said “keep on treating,” we perhaps could have kept him alive three more months, perhaps three more months of suffering. And if we had decided to stop treatment we could have been left with the sense that perhaps something could have been done after all and that we had robbed Tom of that chance. Whatever we decided, it would have left us with doubts and a sense of guilt. This is why we could not handle the decision and left it up to the physicians. But now that we know he is dead, we wonder sometimes whether his twin brother has not been better off. After all, he died right after the delivery.’ For Tom’s father the ultimate treatment result has put pressure on the moral correctness of the decision to intervene at the start of Tom’s life. His doubt reveals the relativity of moral correctness. After all, physicians cannot entirely control the consequences of a decision to start treatment. Additional effects, unforeseen risks and contingent events may turn a decision into unanticipated directions. This renders the correctness of a decision fragile.1 Whether a decision to intervene will be qualified as correct at the end of the treatment trajectory involves a certain dose of ‘moral luck’ (Williams, 1981). After all, children can always respond in ways that are not found in the handbooks; medication may not work, respirators prove hard to set, equipment may limit treatment options and their negative side-effects may gain the upper hand. Moreover, maximal technological deployment and a high chance of recovery can still not guarantee that suffering is averted in each and every case.

The End of the Journey 187

Tom’s treatment reveals the paradoxical effect of medical work in its most harrowing form: each effort to help Tom got him deeper and deeper into the morass. In this child’s treatment we recognize the structure of a tragedy: whatever was being done and despite all good intentions, the child failed to do better.

7.2 Situating knowledge and morality in treatment processes At the start of this study I posed the question how social responses to health innovations are reshaped in the concrete work setting inside the hospital walls, and what actually happens at the interface of the interlinked processes involved in the treatment of newborns, like the process of diagnosing and prognostication, and of establishing facts and values. Throughout this book I approach both questions by focusing on complexities of decision-making and acting in the face of uncertainty. The cases in the Academic Medical Center and Northeastern General Hospital demonstrate how intricate the treatment of NICU children can be: their condition has to be defined and redefined all the time; parents have to (be) fit into the unit’s established everyday patterns; the clinical picture’s course has to be controlled; decisions have to be made on deploying devices, interpreting data, continuing treatment and parental authority. Issues of a varied nature and urgency thereby play a role: Who is the child? Will its treatment catch on? What is this child’s future like? Does it have a chance to be taken off the respirator? In most cases the staff manages to ensure a positive outcome of the intervention process. The book follows doctors and nurses while they prevent the treatment process from stalling. To concentrate on the way actors succeed in acting promptly and adequately in situations of medical and ethical uncertainty this book articulates the intermediate zone where multiple lines of action and reasoning intersect. What did I find while following these lines of action and reasoning in the day-to-day experience of health care workers and how do these activities intersect? Opening up the in-between zone By opening up the interface between diagnostic processes and processes of prognostication; the general and the particular; actors and technologies; formal protocols and the swirl of the treatment trajectory; public and local accountability; facts and values; expectation and experience, it becomes clear that good intentions and a gamut of data and guidelines

188 Uncertainty in Medical Innovation

can never fully preclude problems from occurring. In the treatment of children in the NICU there is always a marked tension between the familiar and the unusual, control and contingency, certainty and doubt, cohesion and conflict. In some cases the knowledge or technologies needed may simply not be available. But even if they are, it is not yet taken for granted which knowledge or experience matters or which technology is effective. Time and again, the value of the available knowledge has to be weighed, or it has to be decided which guidelines apply or which perspective is most valuable. These aspects are always, directly or indirectly, objects of negotiation. To keep complex treatment processes going more appears to be needed than merely the presence of knowledge, experience, and technology. Their application is not automatic but requires active involvement from doctors, nurses and parents. My analytical focus on the in-between zone – on the hinge between the known and the unknown, the risk and responsibility, and the collective and the individual – enables me to exnovate an available yet hidden form of competence that allows the staff to prevent the treatment process from stalling. One of the key insights of this book is the identification of the coordination between the individual, collective and systemic resources that are used to enhance the staff’s competence to produce workable knowledge and reliable performance while acting within a complex critical care unit such as a NICU. For this coordination the actors have several styles of ordering and interpreting at their disposal: reasoning and legitimizing (repertoires), speaking and interpreting (vocabularies), and more or less explicit directions for action (scripts). My usage of these second-order notions such as repertoire, vocabulary and script serves to elucidate the practice involved, yet what, exactly, has been their specific value in this context? With the introduction of new medical technologies the intervention process has become a much more complicated trajectory, with an everincreasing density of lines of action and reasoning, involving all kinds of decision moments with their own specific options and related risks, possible tensions, dilemmas and uncertainties. By focusing on moments of uncertainty the previous chapters have shown that for the alignment of the different constituents of medical practice as to ensure the continuation of the various processes that are part of a treatment trajectory, more is needed than expertise, experience, and technology. It turns out that in addition to the more tangible and concrete actions and interventions of the NICU staff members, more abstract processes are at work as well, and prominently so. Specifically, various styles of arranging and interpretation play a major role in decisions on treatment: repertoires

The End of the Journey 189

guide processes of reasoning and legitimating, vocabularies suggest ways of speaking and interpreting, and scripts imply more or less explicit directions for action. Their application is not automatic but requires active involvement from doctors, nurses and parents. Exnovating the competence of staff members to select and combine the different styles of ordering provides insights into the way they actually deploy their knowledge, experience, and technology in the treatment trajectory. What does such a distinction render visible at the second-order level? If specific repertoires, vocabularies and scripts allow those involved to face up to uncertainties, neutralize tensions, handle random incidents and solve dilemmas, thus to ensure the continuation of the treatment process, the second-order concepts allow us to identify the involved processes of ordering. A finegrained analysis of several case studies has shown what actually takes place at the interface of diagnosis and prognosis, actors and technology, medical facts and moral concerns. To examine what happens in the intermediate zone of diagnosis and prognostication, and of establishing facts and values, we need to recapitulate and evaluate the analytical focus of the book on styles of ordering as an act of exnovation. First, as my argument has demonstrated, careful analysis of a decision’s genesis discloses what actually happens at the interface of the different constituents of medical practice. Exnovation of styles of ordering reveals how facts, values, and medical actions are interdependent and what actually takes place at their intersections. Questions addressing the nature of a case and how to act properly cannot be dealt with in isolation: a specific answer to one question implies a particular answer to another, and vice versa. With the help of various repertoires and vocabularies it becomes possible to provide case descriptions. This exceeds, however, the domain of prognostic facts. Case descriptions suggest directions for action and have consequences for the ways in which one deals with moral issues in the NICU. Second, close analysis of the process in which case descriptions emerge, then, offers insight into how moral choices are made in the NICU. It becomes clear, for example, that actors must build a consensus about the interpretation of the facts and chances (the prognosis), as well as about the definition of the boundaries of the moral category involved (‘unliveable’). Among other things this calls for a coordinative effort on their part to bring individual cases and the specific moral rules involved into line with each other. Study of this effort reveals that the NICU practice implies not so much an ongoing struggle over the discrepancies between general moral categories and unique cases or situations; rather,

190 Uncertainty in Medical Innovation

situated knowledge and morality are produced in one and the same movement, whereby various kinds of repertoires can be deployed. Third, if we consider the decision process at the first-order level of the individual case it becomes clear that in some cases second-order level coordination is necessary. Specific styles may clash, which implies that the ultimate decision on a specific intervention needs to be negotiated. In the staff room discussion about Esther’s treatment, the participants managed to arrive at a decision only by combining their different vocabularies. This coordination effort has to be performed each time anew. The two major vocabularies differ in terms of issues of equality and this affects who has the right to speak and who is responsible. This also applies to repertoires and scripts. A case repertoire, for one, foregrounds other arguments than a situational repertoire. These repertoires differently define the boundary between what belongs to the unit’s responsibility and what does not. A cyclical repertoire implies another significance of past, present and future than a linear repertoire. Likewise, scripts foreground specific treatment options at the expense of others, if not blocking other options. The various ways of reasoning, speaking, and acting intervene differently in the constitution of facts, values, and actions. The performative character of the various ways of ordering gives direction to how the treatment process takes shape. This may lead to frictions in the decision process regarding responsibility, the timing of when something has to be done, or when enough is really enough. To resolve such frictions, the actors look for a way out by combining vocabularies: they speak in terms of ‘compelling advice’; they in part move away from the script and add elements to a treatment guideline; they combine repertoires or merely opt for a confirmation of the radiological report. The various styles of reasoning, therefore, will rarely be encountered in their pure form. This is why the situating of knowledge and morality requires coordination not only at the level of the individual case (the construction of facts, the robustness of a moral turning point or the choice of a treatment trajectory), but also at the level of repertoires, vocabularies and scripts. Securing the treatment trajectory’s progress calls for coordination at the level of the individual case as well as at a meta-level. To foreground these acts of coordination offers not only a new perspective on the competence of the actors involved, but also on the ordering character of reasoning, speaking, and acting. Experienced pioneers My analyses in this study elucidate not only the complexity of treatment trajectories and the inventiveness of those involved, but also

The End of the Journey 191

how guidelines and regulations, perceptions and expectations, legal frames and social roles, triggered by societal processes, shape medical practices while at the same time are reshaped in the vortex of these practices. For example, the growing amount of administrative work for physicians and nurses can be considered as a consequence of the medical domain’s public accountability for the actual intervention process. It contributes to a level of transparency that society asks for. This transparency was called for due to a growing concern about the increasing power of medicine over life and death. However, some ideas about the potential of medical innovation are exaggerated and result in distorted expectations of medical possibilities. Entering a NICU ward with these expectations and deceived by its hi-tech surroundings, families might become very disappointed. In worst case this frustration is re-framed in anger and distrust and results into a legal claim. The danger to be indicted for malpractice has set in motion an avalanche of numbers. Therefore, the tremendous amount of numbers needs to be interpreted as an activity that involves more than merely diagnostic and prognostic confirmation. It shows how legal frames affect hospital practices. Outside the hospital walls these kinds of numerical data is considered as hard facts, and as such they are considered as important evidence in court. Especially in the United States a high degree of juridicalization forces doctors to found their decisions more solidly on quantitative data. To have evidence in advance, diagnostic and prognostic tests are executed and recorded. In these cases the legal system turns medicine into a ‘defensive medicine’ in which their interventions are not always motivated by the child’s medical condition but aim to solidifying the professional’s juridical position. A fine-grained analysis of cases like Esther, Tom, and Maureen also demonstrates that protocols and medical technologies have only a limited potential when it comes to facilitating a sound decision process in the NICU context. Whether technologies and regulations will fail those involved can only be determined in hindsight. To address this, we return one last time to the cases of Tom and Robert. Both Tom and Robert fully depend on medical technology from the outset. Within several weeks Robert manages to rid himself of all technology to become the baby his parents hoped for – a baby they can hold and hug. By contrast, Tom proves to depend on technology and medication his entire life. After six months of having been observed, intravenously fed and hooked up to a respirator he dies. The maximal deployment of care, knowledge and technology turned out to be in vain in his case and led to a brief artificial life rather than a dignified human existence.

192 Uncertainty in Medical Innovation

The deployment of all the care and technology resulted in life for Robert and in suffering and death for Tom. If in these cases it is too simple to use the labels correct treatment versus incorrect treatment, how would we have reacted if Robert had not made it, as happens to many children in the same situation? We would have talked and judged differently about the application of medical-technological options. We would have asked why earlier cases in which this kind of operation failed were apparently insufficient reason to abstain from treating him. He was operated on because there was a chance of success. In a similar vein, Tom’s treatment might have been successful. Medical technology, then, is never intrinsically good or bad, but we only know a posteriori in which cases its deployment was useful. In hindsight we clearly see what technology does and what it does not, but initially the intrinsic promise of technology is still fully at work. From this perspective we say: physicians have to broaden and refine their technological arsenal even further. Technological development thus becomes an answer to each and every problem; with better equipment the problems encountered by today’s physicians will disappear. This would be a valid argument if medical innovation was not evolving simultaneously with our horizon of expectation and takes doctors and nurses into yet unexplored territories. Like fully-fledged pioneers, NICU staff has to find its way in these high-density zones of uncertainty. Missing the safe beacons on the horizon they have to navigate on the basis of less reliable clues. At first glance the situation of the NICU staff resembles that of scientists working in their lab settings while trying to push back the frontiers of knowledge. Both scientists and NICU staff members embark on passages of practice that lack identifiable beacons. Instead, they have to explore, in real time practices, what the contours of their journey might be. While tuning their activities carefully in the ‘dialectic of resistance and accommodation’ (Pickering, 1995) they are able to solve the problems they encounter. This delicate fine-tuning of actions and re-actions enables them to develop a constructive (medical) intervention.2 In medicine, however, technological innovations do not only create new treatment opportunities, new questions and dilemmas, but also new tragic cases; sooner or later, new tragedies and uncertainties will catch up with technological advances. Progress, then, should not be equated with a structural reduction of problems and uncertainties. The short lives of Tom and Esther illustrate that technological advance itself gives rise to new (moral) problems. This conclusion can be interpreted as a plea for strict regulation, for guidelines that set clear limits

The End of the Journey 193

for intervention, thus properly guiding the NICU practice. Yet my analyses of situations in which guidelines apply also show that existing rules come with a specific notion of the practice involved and that this notion is not always in line with the actual practice. Will other, more stringent rules perhaps provide a solution? This question assumes that morality can only be guaranteed when people apply clear rules unambiguously. My study reveals another source of morality, however. In the NICU practice, morality refers not so much to a deductive process of applying moral rules, but to a process of deliberation aimed at delineating stringent ultimate boundaries. This hardly means that facts dictate morality and that guidelines can be abolished. That medical progress causes the boundaries of medical intervention to shift all the time underscores the significance of having protocols and guidelines that allow space for adjustments based on what goes on in actual practice. They do play a functional role, yet not in an instrumental or rationalist way, as is often assumed, but as matters to which one has to relate. They may, for instance, invite reflection, compel those involved to discuss the issue, or give parents a voice. Thus guidelines also imply scripts. Specific guidelines foreground a specific repertoire or a certain vocabulary. In the NICU they exist side by side with many other scripts. Some guidelines are at odds with other scripts, some complement each other and always specific ways of reasoning, speaking and acting are at issue. In the interaction with other scripts the boundaries of the rules may be defined in more detail. By situating morality in the abovementioned demarcation process, the opposition between the formal world of the guidelines and the informal world of the everyday NICU practice is cancelled. Guidelines have their own part in the treatment trajectory and thus give shape to the practice, while the specific medical acts in their turn lead to new knowledge and new guidelines. But there is much more involved than guidelines and knowledge alone. Time and again it has to be decided what the value is of specific knowledge, whether the guidelines apply, which perspective has to be weighed most and so on. If one actor claims that studies ‘show that this child has a five percent chance, so we should go on,’ another actor may say with as much conviction that those same studies indicate that ‘nineteen out of twenty children do not make it, so let us stop.’ This issue of who or what determines the morality of the NICU practice and its official rules completely ignores the dynamic of general/specific or formal/informal. Workable rules are codified experiences. Guidelines can only offer a hold when they are integrally linked to the practice. The arrival of new medical knowledge and options causes boundaries of

194 Uncertainty in Medical Innovation

intervention to shift. This implies that guidelines should leave room for adjustments based on experiences in practice. This does not mean necessarily that the treatment domain is broadened. Recently, for instance, one of the neonatology centres in the Netherlands indicated that the lower limit for treatment of newborns in their unit was raised from a term of 24 weeks to a term of 25 weeks. The centre’s staff did so not because the rules forced it to do so, nor because doctors have fewer technological options at their disposal to keep a child alive. On the contrary, the number of tools and devices is only increasing. However, it is experience that told the staff to do so. Follow-up study, for instance, has shown that it matters quite a bit where a bleeding in the head is localized or what it implies to keep a severely premature child alive. Based on study, reflection and deliberation, then, it was decided to move up the lower limit again.3 Compared to ten years ago one has much more insight into the effects of treatment on premature children. Based on this raised prognostic certainty, NICU staff – as reflective practitioners – may decide to refrain from treatment, even if there are more possibilities for keeping the child alive.4 Behind the dynamic of permanently changing facts, knowledge, standards and experience, the horizon is shifting along (and not always in a forward direction). The NICU practice, in other words, will always retain major elements of uncertainty because breaking new ground belongs to its core business. Its staff members will have to go on exploring uncharted territories. Despite their growing experience, they are bound to be pioneers. It cannot be denied that as a medical practice the NICU – given the particular dynamic of its task and the changing nature of medical facts, norms, expertise, and experience – will always have an uncertain future, just like its staff members will always continue to be pioneers. This is why they are best and most respectfully identified, perhaps, as experienced pioneers.

Notes Chapter 1

Neonatology: A Permanent Dynamic of Change

1 In the Netherlands, about 200,000 babies are born annually, 40,000 of whom are admitted to the hospital while 4500 of them need intensive care. Over the past decades the need for IC-beds has sharply increased as a result of medical-technological advances. Children who previously had no chance of survival are now eligible for treatment. Demographic changes, notably the rise in the numbers of teenage migrant women and of women who decide to have children at a later age, also contributed to a higher need for IC-beds for newborns (Health Council for the Netherlands, 2000). 2 The asterisk refers to the explanatory word list at the end of this book. 3 In the Netherlands there has been a striking rise in the number of premature births in the 1990s. In some Dutch hospitals this rate went up as much as 30 percent (Stoelhorst et al., 2005). 4 Today’s fast changes in medical technology has been a reason for the Dutch Government to ask the Health Council to develop an early warning system for innovations in care that are either just entering the market or that will become available within the foreseeable future. In their advisory report the Council proposes ‘a “national system”, which includes scientific process supervision (collecting and filtering information, further evaluation); providing a framework for consultation; and initiating a modular production process that will result in periodic reports’ (Health Council of the Netherlands, 2005). 5 For example, Franklin and Roberts (2006); Lock et al. (2000). 6 Examples are Brown and Webster (2004); Conrad and Gabe (1999); Elston (1997); Franklin and Lock (2003); Webster (2006). 7 In this respect the NICU is similar to other health care domains, such as the Coronary Care Unit (CCU), the Emergency Department or Psychiatric Emergency Services (PES). 8 In a comparative study of styles of participant research – semiotics versus symbolic interactionism – Mol and Mesman (1996) argue the politics of methodological choices. 9 Although the field notes were collected in the 1990s, the data did not loose its relevance. On the contrary, field research on a NICU in 2006 and 2007 confirms the claims I make in this book. 10 For example Anspach (1993), Bosk (1979), Fox (1957), Freidson (1970), Frohock (1986), Guillemin and Holmstrom (1986), Layne (1996), Schlomann (1994), Strauss et al. (1985) and Zussman (1992). 11 Examples of health care studies that rightfully claim to be full ethnographies are Pool (1996), The (2002) and Vermeulen (2001). Even if these studies aim to describe and explain the regular course of affairs of a particular practice, they also suggest that there is no single true story about the practice addressed and that there are always more stories to tell. 195

196 Notes 12 See Geertz (1988). 13 For example, The (2002). 14 This is why the field of Science and Technology Studies gratefully makes use of the available insights from medical anthropology and sociology. Since the 1990s there is a growing collaboration between medical anthropology and science and technology studies. See for instance, Casper and Koenig (1996), Kaufert and Kaufert (1996), Layne (1998). This same trend is visible in respect to medical sociology and science and technology studies. See Bartley (1990), Bury (1986), Casper and Berg (1995), Lock and Gordon (1988), and Nicolson and McLaughlin (1988). 15 Classic examples in ethnographic studies in the field of Science and Technology Studies are Collins (1985), Knorr-Cetina (1981), Latour and Woolgar (1979) and Traweek (1988). 16 Especially in micro-sociologically oriented directions like ethnomethodology. Put briefly, ethnomethodology focuses on everyday rules, on the recipes people rely on for organizing their life on a day-to-day basis. Garfinkel (1967) and Lynch, Livingston and Garfinkel (1983) are examples of ethnomethodological studies, whereby the latter is specifically geared towards scientific practices. 17 This empirical ethics is expressed in divergent forms. Some, such as Jonsen and Toulmin (1988), resort to the casuistry method whereby a situation’s specificity serves as a starting point. It is also possible, however, to emphasize interpersonal relationships and the related ethics of care. Examples are Manschot and Verkerk (1994) and Tronto (1993). Others, like Nussbaum (1986), criticize the ‘moral mathematics’ of normative ethics and refer in this context to the importance of narrativity (DuBose, Hamel and O’Conell, 1994; Widdershoven, 2000). 18 Despite the fact that in this book ethical issues are explored on the basis of ethnographic research, it would be incorrect to characterize this study as a ‘normative ethnography.’ Thus I follow the qualification used in the field of ethics, which considers ethnographic study as a major yet merely preliminary step. To be able to define an ethnography as normative, Dutch ethicist Ten Have identified four subsequent research steps: ethnographic research of the internal morality of the (medical) practice involved, a sociological analysis of the social values and norms regarding medical-ethical themes, the development of a new theoretical perspective regarding the health care practice involved, and, finally, a new conceptualization of bioethics with which the interactions between internal and external morality (as explored in steps one and two) can be explained (ten HAVE and Lelie, 1998). For a discussion of the relevancy of ethnographic research for ethics, see also Arnold and Forrow (1993), Brody (1993), Crigger (1995), Pearlman, Miles and Arnold (1993), and Pellegrino (1995). 19 This issue, however, will be addressed explicitly in Chapter 6, where I will argue in detail how case descriptions have consequences for the ways in which moral questions are solved in the NICU. 20 Already 20 years ago we could find reports in the literature on treatment of children with extreme low birth weight, such as a child of 440 grams (Koops, 1984) or even 280 grams (Muraskas, Carlson and Halset, 1991). In 2004 a baby girl weighing only 244 grams at birth was treated. These three cases occurred

Notes 197

21

22

23 24 25 26

27 28

in the United States. At the time of reporting, these children were 24, 18 and six months, respectively. At the turn of the 21st century even children with an exceptionally low weight at birth (below 500 grams) may survive. Recent studies show a survival rate in this specific patient population of 60 percent (Rieger-Fackeldey et al., 2005). In the past decade the study of handicaps that emerge later in life has moved beyond physical disorders; emotional and social handicaps and behavioural effects – also called ‘new morbidity’ – are now studied in follow-up research as well (Speechley & Avison, 1995). This trend in mortality and morbidity rates is not limited to the USA or the Netherlands but can be identified in most western societies. Similar quantitative studies were conducted in the Netherlands in the 1990s by van der Heide (1997); de Kleine et al. (1993) and de Leeuw et al. (1996). For a more extensive account of this discussion in the United States, see Frohock (1986) and Weir (1984). ‘Baby Doe’ was born with Down syndrome and with oesophageal atresia, a separation of the oesophagus from the stomach. Because of the atresia the child was not able to absorb any food. Although atresia is a repairable birth defect, the attending obstetrician advised the parents to withhold treatment because their child would have only a 50 percent chance to survive the surgery. Moreover, he argued, even if the operation would be a success, their son would still suffer a severe mental handicap. His advice implied that the baby would die because it was unable to take any food and water by mouth. However, a paediatrician at the same hospital and the family’s physician disagreed with their medical colleague about the bad prognosis of the child and advised the parents to give their consent to surgery. Nevertheless, the parents refused to consent to the operation. Now the hospital went to a family court to have a court order for medical treatment on the basis that the boy would be considered a neglected child. In this way the hospital was able to bypass parental authority for consent. However, the court ruled that parents, when confronted with contradictory medical views, are the ones who decide. Now the hospital took their case to U.S. Supreme Court on the basis that non-treatment was a form of discrimination against handicapped persons. Before this court could rule on the appeal the baby died of dehydration and pneumonia. However, while the case was pending, it had attracted a lot of media attention. When the baby died it was the beginning of a fierce debate about the right of handicapped newborns and the position of the government to intervene in these kinds of decisions. Some groups, like Conservatives, pro-life activists, and a large number of physicians and other health care providers considered the decision not to operate on Baby Doe as a form of infanticide and demanded formal government regulations. Others, notably the American Academy of Pediatrics and liberals, considered these regulations as an undesirable interference of government with medical practice. This discussion significantly relied on the publication by Molenaar, Gill and Dupuis (1988). The inaugural lecture by Van Goudoever (2004) provided the occasion for this discussion.

198 Notes

Chapter 2

Newly Born and Indeterminate

1 Payot et al. (2007) describe two models neonatologists use: they remain as neutral as possible and allow parents to make their own decision or formulate a proposal to which parents can choose. In this way, neonatologists attempt to ensure an informed decision. 2 The data for this book was collected in the 1990s. At that time American neonatologists considered a gestational age of 24 weeks as being eligible for treatment. However, since the turn of the 21st century they also take action in cases of 22 weeks. Although the current number of weeks differs, the involved principle and dilemmas are still the same. Note: In Dutch hospitals the limit of active intervention is still fixed on a gestational age of 25 weeks. 3 The significance of gender becomes painfully clear when it is not yet possible to say whether a newborn is a boy or a girl. In the case of androgyny, parents and others involved miss a major basis for bonding with the child, while it also complicates the process of defining the child. A good DNA test takes a couple of weeks. Until then, the child resides in a hybrid no man’s land and much of the birth ritual halts. The child’s formal registration with the municipality has to be postponed, as is true of the printing of the birth announcements. 4 Bruno Latour (1987) defined ‘immutable mobiles’ as ‘objects that can be replaced and combined without them changing, disintegrating or becoming corrupt. Examples include texts, maps, tables, illustrations, diagrams, formulas, and schemes’ (p. 227). Law’s notion of ‘ordering center’ is quite similar to ‘center of calculation’ as used by Latour (1987, p. 232). 5 Compare the notion ‘center of coordination’ as introduced by Garfinkel (1967). 6 In the literature there are many notions that refer to ordering processes, such as ‘language game’ (Wittgenstein, 1958), ‘paradigm’ (Kuhn, 1962), ‘framework’ (Goffman, 1974), ‘styles of reasoning’ (Hacking, 1992), ‘modes of ordering’ (Law, 1994) and ‘logics’ (Mol and Berg, 1994). I use the notion of ‘repertoire’, thereby emphasizing styles of reasoning as ordering strategies. For more on the usage of this notion, see also Bal (1998), Gilbert and Mulkay (1984), Mol (1997), Rip (1995), and de Wilde (1998). 7 Compare Heimer (2001) on styles of analysis regarding the aggregate (cases) and the particular (biographies). She defines case analysis as the dominant mode of reasoning among staff members and biographical analysis as that among family members. My findings differ in this respect. In my data staff members also use biographical analysis (similar to a situational repertoire). The implications of Heimer’s specific positioning of styles of reasoning among staff and relatives become visible in the social construction of the child as a person. The child becomes a person in a biographical analysis. By attributing this style of reasoning to relatives, she implies that staff members exclusively approach the child as case. In this regard, too, my observations differ from those of Heimer. 8 In this case ‘three’ and ‘four’ refer to the seriousness of the brain haemorrhage, the latter figure being the most severe category.

Chapter 3

Co-travellers

1 Therefore research of the NICU practice is not limited to purely medical concerns; the parents of neonates are a possible focus as well. Studies have

Notes 199

2

3

4

5 6

7

8

addressed, for example, stress factors for parents (Seideman et al., 1997; Shields-Poë and Pinelli, 1997), the communication between staff and parents (Bogdan, Brown and Bannerman Foster, 1982), the emotional processes that parents with a child in the NICU go through (Honig Bachman and Furlong Lind, 1997; Koppe, 1989; de Leeuw et al., 1993; de Leeuw and Bakker, 1995), and the role of parents in the decision processes involved (van der Heide et al., 1998). It is perhaps relevant to point out again that these examples are not representative of the contact between staff and parents. In most cases there appears to be a good relationship. I give preference to ‘vocabulary’ over ‘perspective’ because the latter concept implicitly assumes just one reality that can be looked at from various angles; by contrast, pragmatism emphasizes that different realities always exist side by side (de Wilde, 1987). Various psychological theories and studies indicate the necessity of parent/ child contact directly after birth. In a study about the so-called Harlow monkeys, the American psychologists Harlow and Suomi demonstrated that baby monkeys prefer a safe and soft environment over food. Bowlby (1971) studied this theory of bonding in human beings. Studies of the implications of hospital admission of newborns generally start from the bonding theory. See, for instance, the classic study of Klaus and Kennel (1970), and de Leeuw, Cuttini and Reid (1993). The theory of bonding underlies the open visiting policy in the NICU, but also the system of Ronald McDonald Homes, where parents can stay to be near to their hospitalized child all the time. See also Bell (1997), Honig Bachman and Furlong Lind (1997) and Heimer and Staffen (1995). Anspach (1993) indicates that parents with a low level of education are involved in the decision process to a lesser degree. I did not identify this phenomenon in the Academic Medical Center nor in Northeastern General Hospital. The measure in which parents took part in the decision process, as well as the ways in which they participated, seemed to be more dependent on whether or not the staff had a sense that they were in contact with the child’s parents. For an ethnographic study of the role of nurses in the decision process involving the refraining from further treatment in adult patients (euthanasia), see The (1997). This concept should not be confused with ‘internal morality’ which refers to the ward’s specific moral order. Endogenous normativity does not limit itself to moral standards.

Chapter 4

Uncertain Trajectories

1 Fox’s study Training for Uncertainty (1957) marks the start of a series of medicalsociological studies on the subject of uncertainty among physicians. Other studies are those by Atkinson (1984), Bosk (1979; 1980), Fox (1988; 2000), Hunter (1991), and Zussman (1992). In this literature several sources of uncertainty are identified. The first arises from lack of knowledge with the physician, who can never cover the entire medical knowledge domain. Next, medical knowledge itself has boundaries. In some areas medicine is still

200 Notes

2

3 4

5 6

7

8 9

10

fumbling in the dark. It is up to the doctor to determine whether his uncertainty has a personal or professional ground, and this itself is another source of uncertainty. This distinction between an uncertain world and uncertain knowledge about that world is at odds with the constructivist view of knowledge that I use in this study. In this view it is useless to talk about the situation of the patient otherwise than in terms of the knowledge we have about that situation. Thus there are no medical situations that are intrinsically more uncertain than other situations; there are situations about which physicians know less and that hence are less predictable. The choice whether or not one should intervene after birth is discussed in Chapter 2. The start of Robert’s treatment trajectory is also discussed in Chapter 3. On this dilemma, see also Frohock (1986); Guillemin and Holmstrom (1986); Strauss et al. (1985); Zussman (1992); and Vermeulen (2001). Although notions like ‘protocol,’ ‘standard’ and ‘guideline’ tend to be used interchangeably in the literature, I use the first to refer to a standardized (and hence commonly shared and broadly supported) form of medical acting. The ‘treatment plan’ mentioned in this text, then, does not belong to this category, but in practice it functions in the same way. For other examples of the notion of ‘certainification’, see van Asselt, Mesman, and van ’t Klooster (2007). Although I consider the material setting (equipment, architecture) as an actively structuring element of a practice, in my analysis I hold on – unlike actor-network theorists – to making a distinction between people and objects. In the debate in science and technology studies on the nature of the relation between the social and material order, the common actor-network theory approach, in which the world is viewed as one whole of human and non-human actors, is criticized for its functionalist character. My holding on to the distinction between humans and objects allows me to devote attention to the frames of accountability for acting. It offers the (conceptual) possibility to acknowledge and study people’s own experience, decisions, and responsibility. See also Oudshoorn (1996) and Verbeek (2005). However, a protocol not only creates a practical order, but its use also assumes an already existing order. Moreover, a protocol can only work in such an order. For a description of medical records and protocols, see Berg (1997b). For other examples of normativity in medical technology, see Berg and van der Lyke (1997) and Oudshoorn (1996). There are various ways of conceiving the relationship between the social order and the material order. Interactive materialism, for instance, starts from the notion that both the social order and the material order consist of people as well as things and that both have an active part. As such the oppositional relation between the world of people and the world of objects disappears. For the discussion on this issue in the Netherlands, see Harbers (2001), Harbers and Koenis (1999), Mol and Harbers (1994) and de Vries (2001). On this particular point, see also Law and Mol (1995). The latter approach of numbers falls under enumerology. This field focuses not so much on the value of numbers, but on the social processes in which they are constructed and the ways in which they give shape to our reality (Bogdan and Ksander, 1980).

Notes 201 11 But senior staff members themselves do not always know what is going on, even if others expect them to know everything. 12 The question arises what would remain of medical practice if we were to take out the numbers. A medicine without numbers seems possible, but is a very different medical practice – different in style of thinking, identity, and medical interventions. For such medical practice without numbers we have to return to the early 19th century, to encounter the remnants of a qualitative regular medicine. Warner (1986) describes the rise of quantitative thinking in medical practice. Nineteenth-century doctors viewed illness as a disturbance of the natural balance. They were supposed to restore the homeostasis. Thereby the subjective experience of the patient and the information on his specific situation were essential. Each patient was unique and comparison was therefore seen as useless. To restore the balance, bloodletting and ample alcoholic drinks belonged to the standard package of medical treatment. In the course of the 19th century, physicians increasingly began to rely on numbers. Gradually the old way of medical thinking and acting was replaced by a new medical practice, a quantified medicine in which health as ‘natural’ is replaced by health as ‘normal’. One looked no longer for a disturbed balance of the overall condition, but for deviations related to a specific physiological system or organ. The focus was not on excessive stimulation of the natural balance, but on increased numbers of disease-entities. Instead of the patient’s subjective experience, objectively measured results became central. All this led to a medical world that ‘becomes scientific.’ Other branches of science, as well as everyday life, have increasingly fallen under the spell of numbers as well of course. For a historical description of this change, see Gigerenzer (1990), Hacking (1990) and Porter (1995).

Chapter 5

Beacons on the Horizon

1 I opt for the notion ‘position’ rather than ‘perspective’ because perspectivism assumes a single reality, whereby each perspective provides insight into only one part of that reality. By combining various perspectives, the perspectivist claims, a more complete picture of reality would emerge (de Wilde, 1992, p. 31 ff.). 2 This not only applies to prognostic views. Doctors and nurses also have divergent views on other subjects, which may trigger conflicts. For a historical survey of studies of the doctor-nurse relationship and its conflicts, see McMahan, Hoffman and McGee (1989). 3 Unlike Anspach’s analysis (1993), my analysis also takes into account parents as producers of prognostic knowledge. See also the critique on Anspach by Layne (1996). 4 This methodological choice and its development are based on the study by Anspach (1993), who has shown in detail how the distribution of tasks in a NICU plays a role in the perception of a child’s prognosis. 5 See Anspach and Beeson (2001) for a detailed analysis of the way different levels of emotional engagement are shaped by institutional arrangements in their study on the relation between emotions and morality. 6 For an elaborate argument for this position, see Benner, Tanner and Chesla (1996).

202 Notes 7 For more on the boundary between the human and the monstrous, see Schlomann (1994). 8 The effects of a long-term hands-off policy on a child’s emotional development are unclear, but the prevailing view is that it does not benefit a child (Sparshott, 1997). 9 For an extensive overview, see Sparshott (1997). 10 Not so long ago the prevailing view was that newborns – premature babies in particular – would have no experience of pain. This view has been left behind for years and pain in neonates now counts as a major field of research. See Bours et al. (1997). 11 Examples of outcome studies are population-based studies; studies based on follow-up clinics of individual NICUs, and pooled dates from the follow-up clinics of several nurseries. For an example of an outcome study on the followup of very-low-birth-weight infants into early adulthood, see Hack et al. (2002). 12 In medical practice the reference to unexpected outcomes in the form of anecdotes is even a common part of the knowledge transfer (Benner, Tanner and Chesla, 1996). 13 The notion of ‘repertoire’ is described in the second chapter. Van Vree (1998) refers to the totality of manifestations of historical awareness as historical culture, thereby relying on work by Koselleck. Van Vree points to the fact that the implications of the work of Koselleck reach beyond the boundaries of historiography. The various forms of historical awareness, as well as the shifts therein, have specific implications for how reality is perceived. 14 Cf. Koselleck (1985) ‘rational prognosis’. 15 A classic example of a study of the temporal orders in a hospital is the one by Zerubavel (1979) that focuses on the temporal order based on, among other things, staff rotation and work schedules. 16 Cf. the notion of temporal matrix as developed by Strauss et al. (1985).

Chapter 6

The Moral Load

1 Of every 200,000 children that are born annually, 1000 die within their first year. In 600 of them a medical decision on life-ending action was involved (Verhagen and Sauer, 2005). 2 See also Chapter 1. 3 The report from the KNMG commission also makes this distinction. It refers to a ‘non-real chance of survival’ prognosis and an ‘unliveable life’ prognosis. 4 In the discussion on the parents the psychosocial and juridical vocabularies dominated. For an analysis of another part of this staff meeting, see Chapter 3. 5 This also makes it impossible to discuss morality without involving facts. One cannot enter into a moral conflict with each other and still agree about the facts. Thus moral dissent excludes factual consensus. 6 On universalizing and localizing processes, see Gabriëls (2000) and Timmermans and Berg (1997). 7 Nussbaum (1986) localizes ethics in the ways in which one deals with the consequences of the decision.

Notes 203 8 The notion ‘moral cost’ I derive from Williams (1981). Moral costs are the outcome of a moral dilemma. In a choice between two evils there is always the pain, the moral burden of the decision. These are the moral costs bound up with each choice in a moral dilemma. 9 In the United States the Supreme Court of Texas decided in 2003 that parents have no right to refuse the resuscitation of their child before birth in the case of extremely premature infants (Cf. Robertson, 2004). This decision was based upon the premise that decisions before birth could not be fully informed. However, this rejection of the parental right to determine whether care should be provided to their prematurely born child is inconsistent with the guidelines of both the American College of Obstetricians and Gynecology and the Committee of the Fetus and the Newborn of the Academy of Pediatrics, as well as with neonatology practices. The uncertainties that characterize the prognosis in such situations are for them precisely the reason to give the parents a voice in the decision-making process (Lorenz, 2005). 10 For an elaborate description of the distribution of responsibility between staff and parents, see Chapter 3. 11 Dodd et al. (2004) proposes nurses to engage more in ethical activism and ethical assertiveness. The first implies making hospitals more receptive to nurse participation, while the latter refers to nurses’ participation in ethical deliberations without being formally invited. See also Becker and Grunwald (2000). 12 The relationship of power between doctors and nurses is a major object of study in various disciplines such as medical sociology and nursing science. In the discussion on the relation between physicians and nurses a shift can be identified. Initially this relation is described in the classic study of Freidson (1970) from a dominance model in which the physician is the one who decides and nurses obey. There is also another picture in which nurses are seen to influence treatment, but they do so in a hidden fashion. This is why the physician/nurse relationship is characterized as a game – the doctor-nurse game – in which maintaining the social order comes first (Stein 1967). This means that open dispute is absolutely forbidden. This is avoided because nurses ‘package’ their advice to physicians in such a way that without loss of face they can still put it forward as their own idea. Thus the nurses’ knowledge forms no threat for the hierarchical order. On account of various social changes such as the increase of female doctors, the social status of physicians and the professionalization of nurses, neither the dominance model nor the ‘doctor-nurse game’ will persist. The representation of the practice in terms of power blocks is replaced by a description in terms of a ‘negotiated order’. See Hughes (1988), Porter (1991), Stein (1969), and Stein, Watts and Howell (1990). In this negotiation practice nurses have received ever more room as a result of the recognition of the importance of the social dimension in the treatment of patients and the introduction of team nursing and collective patient deliberations (Svensson, 1996). For a historical overview of this development, see McMahan, Hoffman and McGee (1989) and Pillitteri and Ackerman (1993). 13 See Chapter 3. 14 For an analysis of the controversy on the definition of and criteria for death, see Lock (1996) and Wackers (1994).

204 Notes 15 A classic study of the social organization of dying in a hospital is the one by Sudnow (1967). 16 This is no requirement. Some parents do not want it and the child is simply left in the incubator. 17 See also Downey et al. (1995).

Chapter 7

The End of the Journey

1 Compare ‘the fragility of goodness’ by Nussbaum (1986). 2 Whereas traditional accounts of science depicts its end as the production of a set of facts that represent how the world is. In contrast, science can also be thought of as a realm of instruments, devices and substances that do something in the world (Pickering, 1995). Understanding science in this way, it becomes, just like medicine, a performative field. 3 The Leiden University Medical Center (LUMC) took this decision in part based on the results of a British longitudinal study in which premature children (born before the 26th week) were followed for 30 months (Wood et al., 2000). Two out of three died shortly after birth, while most others proved seriously handicapped (van der Broek, 2001). When the LUMC announced its decision, the results of a longitudinal Dutch study, the so-called POPS-study, came out as well. Since 1983 this project has followed children born after a term of 31 weeks or less. These children prove to have serious learning and behavioural problems that only increase as they grow older (Verloove-Vanhorick, den Ouden and Walther, 2001). The results of the POPS-study also played a role in the decision of the neonatologists from Leiden to no longer treat children born before a term of 24 weeks. 4 See Schön (1983) for the notion ‘reflective practitioner’.

Bibliography Abe, N., Catlin, A. and Mihara, D. (2001) End of life in the NICU: a study of ventilator withdrawal. American Journal of Maternal Child Nursing 26: 141–146. Akrich, M. (1992) The de-scription of technical objects, in W. E. Bijker and J. Law (eds) Shaping Technology, Building Society. Studies in Sociotechnical Change. Cambridge, MA, The MIT Press. Allen, A. M. (1995) Stressors to neonates in the neonatal unit. Midwives 108: 139–141. Amann, K. and Knorr-Cetina, K. D. (1989) Thinking trough talk: An ethnographic study of a molecular biology laboratory. Knowledge and Society 8: 3–26. Anspach, R. R. (1987) Prognostic conflict in life-and-death decisions: the organization as ecology of knowledge. Journal of Health and Social Behavior 28: 215–231. Anspach, R. R. (1993) Deciding Who Lives: Faithful Choices in the Intensive-Care Nursery. Berkeley, University of California Press. Anspach, R. R. and Beeson, D. (2001) Emotions in medical and moral life, in B. Hoffmaster (ed) Bioethics in Social Context. Philadelphia, Temple University Press. Arnold, R. M. and Forrow, L. (1993) Empirical research in medical ethics: an introduction. Theoretical Medicine 14: 195–196. Asselt van, M., Mesman, J. and van ’t Klooster, S. (2007) Dealing with Prognostic Uncertainty. Futures 39: 669–684. Atkinson, P. (1984) Training for certainty. Social Science and Medicine 19: 949–956. Bal, R. (1998) Grenzenwerk. Over het Organiseren van Normstelling Voor de Werkplek. Enschede, Twente University Press. Bartley, M. (1990) Do we need a strong programme in medical sociology? Sociology of Health & Illness 12: 371–390. Becker, P. T. and Grunwald, P. C. (2000) Contextual dynamics of ethical decision making in the NICU. Journal of Perinatal and Neonatal Nursing 14: 58–72. Bell, P. L. (1997) Adolescent mothers’ perceptions of the neonatal intensive care unit environment. The Journal of Perinatal & Neonatal Nursing 11: 77–84. Benner, P., Tanner, Ch. and Chesla, C. (1996) Expertise in Nursing Practice. Caring, Clinical Judgement, and Ethics. New York, Springer Publishing Company. Berg, M. (1996) Practices of reading and writing: the constitutive role of the patient record in medical work. Sociology of Health & Illness 18: 499–524. Berg, M. (1997a) Rationalizing Medical Work. Decision-Support Techniques and Medical Practices. Cambridge, MA, The MIT Press. Berg, M. (1997b) Problems and promises of the protocol. Social Science and Medicine 44 (8): 1081–1088. Berg, M. (1998) Order(s) and disorder(S): of protocols and medical practices, in M. Berg and A. Mol. (eds) Differences in Medicine: Unraveling Practices, Techniques, and Bodies. Durham, Duke University Press. Berg, M. and van der Lyke, S. (1997) De normativiteit van een infuuspomp. Over zorgpraktijken en technologie, in M. Verkerk (ed) Denken over zorg. Concepten en praktijken. Utrecht, Elsevier/De Tijdstroom. 205

206 Bibliography Berg, M. and Mol, A. (eds) (1998) Differences in Medicine: Unraveling Practices, Techniques, and Bodies. Durham, Duke University Press. Bogdan, R., Brown, M. A. and Bannerman Foster, S. (1982) Be honest but not cruel: staff/parent communication on a neonatal unit. Human Organization 41: 6–15. Bogdan, R. and Ksander, M. (1980) Policy Data as a Social Process: A Qualitative Approach to Quantitative Data. Human Organization 39 (4): 302–309. Bosk, Ch. L. (1979) Forgive and Remember: Managing medical failure. Chicago, The University of Chicago Press. Bosk, Ch. L. (1980) Occupational Rituals in Patient Management. The New England Journal of Medicine 303: 71–76. Bours, G. J. J. W., Huijer Abu-Saad, H., Hamers, J. P. H. and van Dongen, R. T. M. (1997). Pijnbeoordeling bij neonaten. Tijdschrift voor Kindergeneeskunde 65: 1–8. Bowlby, J. (1971) Attachment. second ed. 3 vols. Vol. 1, Attachment and Loss. London: Penguin Group. Brody, B. A. (1993) Assessing empirical research in bioethics. Theoretical Medicine 14: 211–219. Brodwin, P. (ed) (2000) Biotechnology and Culture: Bodies, Anxieties, Ethics. Bloomington, Indiana University Press. Broek van der, M. (2001) Begrip voor verleggen grens hulp aan vroeggeborenen; Zelfs baby’s tot 32 weken vaak al gehandicapt. De Volkskrant, 29 mei 2001. Brown, N. and Webster, A. (2004) New Medical Technologies and Society: Reordering Life. Cambridge, Polity Press/Blackwell. Bury, M. R. (1986) Social construction and the development of medical sociology. Sociology of Health & Illness 8: 137–169. Carson, D. P. (1996) The socially complex family. New dilemmas for the neonatal social worker. Clinics in Perinatology 23: 609–620. Casper, M. J. and Berg, M. (1995) Constructivist perspectives on medical work: medical practices and science and technology studies. Science, Technology & Human Values 20: 395–407. Casper, M. J. and Koenig, B. (1996) Reconfiguring nature and culture: intersections of medical anthropology and technoscience studies. Medical Anthropology Quarterly 10: 523–536. Collins, H. M. (1985) Changing order. Replication and Induction in Scientific Practice. London, SAGE Publications. Conrad, P. and Gabe, J. (1999) Sociological Perspectives on the New Genetics. Malden, MA, Blackwell Publishers. Crigger, B. J. (1995) Bioethnography: fieldwork in the land of medical ethics. Medical Anthropology Quarterly 9: 400–417. Davidson, A. I. (1999) Styles of Reasoning: conceptual history, and the emergence of psychiatry, in M. Biagioli (ed) The Science Studies Reader. New York, Routledge. Davies, K. (1994) The tension between process time and clock time in care-work. The example of day nurseries. Time & Society 3: 277–303. Dodd, S. J., Jansson, B., Brown-Saltzman, K., Shirk, M. and Wunch, K. (2004) Expanding nurses’ participation in ethics: an empirical examination of ethical activism and ethical assertiveness. Nursing Ethics 11: 15–27. Downey, V., Bengiamin, M., Heuer, L. and Juhl, N. (1995) Dying babies and associated stress in NICU nurses. Neonatal Network 14: 41–45.

Bibliography 207 Dubbelman, K. A. and Hermans, H. E. G. M. (1990) Besluitvorming in de neonatologie. ‘s-Gravenhage, Vuga/De Tijdstroom. DuBose, E. R., Hamel, R. P. and O’Conell, L. J. (eds) (1994) A Matter of Principles? Ferment in U.S. Bioethics. New York, Trinity International Press. Duff, R. S. and Campbell, A. G. M. (1973) Moral and ethical dilemmas in the special care nursery. The New England Journal of Medicine 289: 890–894. Eliasson, A. H., Howard, R. S., Torrington, K. G., Dillard, T. A. and Phillips, Y. Y. (1997) Do-not-resuscitate decisions in the medical ICU. Comparing physician and nurse opinions. CHEST 11: 1106–1111. Elston, M. A. (1997) The Sociology of Medical Science and Technology. Oxford, Blackwell. Fox, R. C. (1957) Training for uncertainty, in R. K. Merton, G. G. Reader and P. L. Kendall (eds) The Student-Physician. Introductory studies in the sociology of medical education. Cambridge, MA, Harvard University Press. Fox, R. C. (1988) Essays in Medical Sociology. Journeys into the Field. 2nd edition. Oxford, Transaction Books. Fox, R. C. (2000). Medical uncertainty revisited, in G. L. Albrecht, R. Fitzpatrick and S. C. Scrimshaw (eds) Handbook of Social Studies in Health and Medicine. London, SAGE Publications. Frader, J. (1979) Difficulties in providing intensive care. Pediatrics 64: 10–16. Franklin, S. and Roberts, C. (2006) Born and Made: An Ethnography of Preimplantation Genetic Diagnosis. Princeton, Princeton University Press. Franklin, S. and Lock, M. (eds) (2003) Remaking Life & Death: Toward an Anthropology of the Biosciences. School of American Research Press. Oxford, James Currey Ltd. Freidson, E. (1970) Profession of Medicine. A Study of Applied Knowledge. New York, Harper and Row. Frohock, F. M. (1986) Special Care: Medical Decisions at the Beginning of Life. Chicago, The University of Chicago Press. Gabriëls, R. (2000) Simsalabimbambasaladusaladim. Over grote woorden. Krisis. Tijdschrift voor empirsche filosofie 1: 89–107. Garfinkel, H. (1967) Studies in Ethnomethodology. Englewood Cliffs, NJ, PrenticeHall. Geertz, C. (1988) Works and Lives. The Anthropologist as Author. Stanford, Stanford University Press. Gigerenzer, G. (1990) The Empire of Change: How Probability Changed Science and Everyday Life. Cambridge, Cambridge University Press. Gilbert, N. and Mulkay, M. (1984) Opening Pandora’s Box. A Sociological Analysis of Scientists’ Discourse. Cambridge, Cambridge University Press. Goffman, E. (1974) Frame Analysis. An Essay on the Organization of Experience. Cambridge, Harvard University Press. Goodman, N. (1978) Ways of Worldmaking. Indianapolis, Hackett Publishing Company. Goudoever van, H. (2004) Te jong om te sterven. Inaugural lecture. Rotterdam, Erasmus Medisch Centrum. Griffin, T., Kavanaugh, K., Soto, C. F. and White, M. (1997) Parental evaluation of a tour of the neonatal intensive care unit during a high-risk pregnancy. Journal of Obstetrics, Gynecology and Neonatal Nursing 26: 59–65. Guillemin, J. H. and Holmstrom, L. L. (1986) Mixed Blessings: Intensive Care for Newborns. Oxford, Oxford University Press.

208 Bibliography Guyer, B., Freeman, M. A., Strobino, D. M. and Sondik, E. J. (2000) Annual Summary of Vital Statistics: Trends in the health of Americans during the 20th century. Pediatrics 16: 1307–1317. Hack, M. and Fanaroff, A. A. (2000) Outcomes of children of extremely low birth weight and gestational age in the 1990s. Seminars in Neonatology 5: 89–106. Hack, M., Flannery, D. J., Schluchter, M., Cartar, L., Borawski, E. and Klein, N. (2002) Outcomes in young adulthood for very-low-birth-weight infants. The New England Journal of Medicine 346: 149–157. Hacking, I. (1982) Language, truth and reason, in M. Hollis and S. Lukes (eds) Rationality and Relativism. Oxford, Basil Blackwell. Hacking, I. (1990) The Taming of Chance. Cambridge, MA: Cambridge University Press. Hacking, I. (1992) The self-vindication of the laboratory sciences, in A. Pickering (ed) Science as Practice and Culture. Chicago, University of Chicago Press. Hammersley, M. and Atkinson, P. (1983) Ethnography: Principles in practice. London, Routledge. Harbers, H. (2001) De geest is uit de fles. Krisis 2: 79–86. Harbers, H. and Koenis, S. (1997) De wonderbaarlijke terugkeer van de nietmensen. In Museum Aarde. Natuur: criterium of constructie?, edited by J. Keulartz and M. Korthals. Meppel: Boom. pp. 107–122. Harbers, H. and Koenis, S. (1999) Ter Inleiding. Themanummer De bindkracht der dingen. Kennis en Methode 23 (1): 3–10. Have ten, H. and Lelie, A. (1998) Medical ethics research between theory and practice. Theoretical Medicine and Bioethics 19: 263–276. Health Council of the Netherlands (2000) Perinatal Intensive Care. The Hague, Health Council of the Netherlands, no. 2000/08E. Health Council of the Netherlands (2005) Providing information on significant developments in health care; identification of new and changing health technology. The Hague, Health Council of the Netherlands, no. 2005/08. Heath, Ch. (1982) Preserving the consultation: medical record cards and professional conduct. Sociology of Health & Illness 4: 56–74. Heide van der, A. (1997) Medical end-of-life decisions made for neonates and infants in the Netherlands. The Lancet 350: 251–255. Heide van der, A., van der Maas, P. J., van der Wal, G., Kollée, L., de Leeuw, R. and Holl R. A. (1998) The role of parents in end-of-life decisions in neonatology: physicians’ views and practices. Pediatrics 101: 413–418. Heimer, C. A. (2001) Cases and Biographies: an essay on routinization and the nature of Comparison. Annual Review of Sociology 27: 47–76. Heimer, C. A. and Staffen, L. R. (1995) Interdependence and reintegrative social control: labeling and reforming ‘inappropriate’ parents in neonatal intensive care units. American Sociological Review 60: 635–654. Honig Bachman, D. and Furlong Lind, R. (1997) Perinatal social work and the family of the newborn intensive care infant. Social Work in Health Care 24: 21–37. Hughes, D. (1988) When nurse knows best: some aspects of nurse/doctor interaction in a casualty department. Sociology of Health & Illness 10: 1–22. Hunter, K. M. (1991) Doctor’s Stories, The Narrative Structure of Medical Knowledge. Princeton, NJ, Princeton University Press. Ihde, D. (ed) (1990) Technology and the Lifeworld. From Gardens to Earth. Indianapolis, Indiana University Press.

Bibliography 209 Jonsen, A. R. and Toulmin, S. (1988) The Abuse of Casuistry. A History of Moral Reasoning. Berkeley, University of California Press. Kaufert, P. and Kaufert, J. (1996) Anthropology and Technoscience Studies: Prospects for synthesis and ambiguity. Medical Anthropology Quarterly 10: 686–690. Klaus, M. J. and Kennel, J. H. (1970) Mothers separated from their newborn infant. Pediatrics Clinics of North America 17: 1015–1037. Kleine de, M. J. K., de Leeuw, R., Kollee, L. A. A. and Berger, H. M. (1993) Voortzetten of staken van levensverlengend handelen bij pasgeborenen. Een onderzoek in 4 centra voor neonatale intensieve zorg. Nederlandsch Tijdschrift voor Geneeskunde 137: 496–500. KNMG-Commissie Aanvaardbaarheid Levensbeëindgend Handelen (1990) Interimrapport inzake levensbeëindigend handelen bij wil-onbekwame patiënten. Deel 1: zwaar-defecte pasgeborenen. Utrecht, KNMG. Knorr-Cetina, K. D. (1981) The Manufacture of Knowledge: An Essay on the Constructivist and Contextual Nature of Science. Oxford, Pergamon Press. Koops, B. (1984) Extreme immaturity: a frontier in Neonatology. American Journal of diseases of children 138: 713–714. Koppe, J. G. (ed) (1989) Neonatologie. Utrecht, Bohn, Scheltema & Holkema. Koppelman, L. M. (2005) Are the 21-Year old Baby Doe rules misunderstood or mistaken? Pediatrics 115: 797–802. Koselleck, R. (1985) Futures Past. On the Semantics of Historical Time. Cambridge, MA, The MIT Press. Kuhn, Th. S. (1962) The Structure of Scientific Revolutions. 2nd edition. Chicago, University of Chicago Press. Latour, B. and Woolgar, S. (1979) Laboratory life. The Social Construction of Scientific Facts. Princeton, NJ, Princeton University Press. Latour, B. (1987) Science in Action: How to Follow Scientists and Engineers through Society. Cambridge, Harvard University Press. Law, J. (1994) Organizing Modernity. Oxford UK, Blackwell. Law, J. and Mol, A. (1995) Notes on materiality and sociality. The Sociological Review 43 (2): 274–294. Layne, L. (1996) ‘How’s the baby doing?’ Struggling with narratives of progress in a neonatal intensive care unit. Medical Anthropology Quarterly 10: 624–656. Layne, L. (1998) Anthropological approaches in Science and Technology Studies. Science, Technology & Human Values 23 Special Issue. Leeuw de, R., Cuttini, M. and Reid, M. (1993) Staff views on parental visiting of very low birth weight infants in neonatal intensive care units. Results of a staff enquiry across eleven countries of the European Community. Journal of Perinatal Medicine 21: 43–52. Leeuw de, R. and Bakker, L. (1995) Te vroeg geboren. Utrecht, Kosmos-Z&K Uitgevers. Leeuw de, R., De Beaufort, A. J., de Kleine, M. J., van Harrewijn, K. and Kollée, L. A. (1996) Foregoing intensive care treatment in newborn infants with extremely poor prognoses. A study in four neonatal intensive care units in the Netherlands. The Journal of Pediatrics 129: 661–666. Lock, M. and Gordon, D. (1988) Biomedicine examined. Dordrecht, Kluwer Academic Publishers. Lock, M. (1996) Death in a technological time: locating the end of a meaningful life. Medical Anthropology Quarterly 10: 575–600.

210 Bibliography Lock, M., Young, A., Cambrosio, A. and Harwood, A. (eds) (2000) Living and working with the new medical technologies: intersections of inquiry. Cambridge, Cambridge University Press. Lorenz, J. M. (2005) Ethical dilemmas in the care of the most premature infants: the waters are murkier than ever. Current Opinion in Pediatrics 17: 186–190. Lynch, M., Livingston, E. and Garfinkel, H. (1983) Temporal Order in Laboratory Work, in K. D. Knorr-Cetina and M. Mulkay (eds) Science Observed. Beverly Hills, Sage. Manschot, H. and Verkerk, M. (eds) (1994) Ethiek van de zorg: Een discussie. Amsterdam, Boom. Mattingly, Ch. (1998). In search of the good: narrative reasoning in clinical practice. Medical Anthropology Quarterly 12: 273–297. McMahan, E., Hoffman, K. and McGee, G. (1989) Physician-Nurse Relationships in Clinical Settings: A Review and Critique of the Literature, 1966–1992. Medical Care Review 51: 83–112. Meulen ter, R. and Berg, M. (1999) Het normatieve karakter van richtlijnen onderzocht. Tijdschrift voor Sociale Gezondheidzorg 77: 376–379. Michael, M. (1996) Constructing Identities. The Social, the Nonhuman and Change. London, Sage Publications Ltd. Mol, A. (1997) Klant, burger, zieke. Het goede in drie talen, in M. Verkerk (ed) Denken over Zorg. Concepten en praktijken. Lochem, Elsevier/De Tijdstroom. Mol, A. (2002) The Body Multiple. Ontology in Medical Practice. Durham, Duke University Press. Mol, A. and Berg, M. (1994) Principles and practices of medicine. Culture, Medicine and Psychiatry 18: 247–265. Mol, A. and Harbers, H. (1994) Inleiding. In: De aard der dingen, Special Issue Kennis en Method, Vol. XVII, nr. 1, pp. 5–11. Mol, A. and Mesman, J. (1996) Neonatal food and the politics of theory. Some questions of method. Social Studies of Science 26: 419–444. Molenaar, J. C., Gill, K. and Dupuis, H. M. (1988) Geneeskunde, dienares der barmhartigheid. Nederlands Tijdschrift Voor Geneeskunde 132: 1913–1917. Muraskas, J. K., Carlson, N. J. and Halset, C. (1991) Survival of a 280-g infant. The New England Journal of Medicine 324: 1598–1599. Nardi, B. A. and Engeström, Y. (1999) A web on the wind: the structure of invisible work. Computer Supported Cooperative Work 8: 1–8. Nederlandse Vereniging voor Kindergeneeskunde (1992) Doen of laten? Grenzen van het medisch handelen in de neonatologie. Utrecht, Nederlandse Vereniging voor Kindergeneeskunde. Nicolson, M. and McLaughlin, C. (1988) Social constructionism and medical sociology: a study of the vascular theory of multiple sclerosis. Sociology of Health & Illness 10: 234–261. Nussbaum, M. C. (1986) The Fragility of Goodness. Luck and Ethics in Greek Tragedy and Philosophy. Cambridge, Cambridge University Press. Ouden den, A. L., Anthony, S., van Baar, A. L. and Dorrepaal, C. A. (2000) Overlevingskans van zeer immature pasgeboren in Nederland. Tijdschrift voor Kindergeneeskunde, 68: 241–246. Oudshoorn, N. E. J. (1996) Genderscripts in Technologie. Noodlot of Uitdaging? Enschede: Universiteit Twente.

Bibliography 211 Payot, A., Gendron, S., Lefebvre, F. and Doucet, H. (2007) Deciding to resuscitate extremely premature babies: How do parents and neonatologists engage in the decision? Social Science & Medicine 64: 1487–1500. Pearlman, R. A., Miles, S. H. and Arnold, R. M. (1993) Contributions of empirical research to medical ethics. Theoretical Medicine 14: 197–210. Pellegrino, E. D. (1995) The limitation of empirical research in Ethics. The Journal of Clinical Ethics 6: 161–162. Pickering, A. (1995) The Mangle of Practice: time, agency, and science. Chicago: University of Chicago Press. Pillitteri, A. and Ackerman, M. (1993) The ‘Doctor-Nurse Game’: a comparison of 100 years – 1880–1990. Nursing Outlook 41: 113–116. Polin, R. (2005) Neonatology and Perinatology. Editorial comment. Current Opinion in Neonatology 17: 165–166. Pols, J. and Mol, A. (1994) Dus alles is in keurig overleg gedaan. Over patiënten en de overdracht van informatie. Gezondheid. Theorie in praktijk 2: 165–179. Pool, R. (1996) Vragen om te sterven. Euthanasie in een Nederlands ziekenhuis. Rotterdam, WYT uitgeefgroep. Porter, S. (1991) A participant observation study of power relations between nurses and doctors in a general hospital. Journal of Advanced Nursing 16: 728–735. Porter, Th. M. (1995) Trust in Numbers. The pursuit of objectivity in science and public life. Princeton, NJ: Princeton University Press. Raines, D. (1996) Parents’ values: a missing link in the Neonatal Intensive Care Equation. Neonatal Network 15 (3): 7–14. Rapp, R. (2000) Testing women, Testing the Fetus: the Social Impact of Amniocentesis in America. New York, Routledge. Rashotte, J., Fothergill-Bourbonnais, F. and Chamberlain, M. (1997) Pediatric intensive care nurses and their grief experiences: a phenomenological study. Heart & Lung 26: 372–386. Rheinberger, H. J. (2000) Beyond Nature and Culture: modes of reasoning in the age of molecular biology and medicine, in M. Lock, A. Young, A. Cambrosio and A. Harwood (eds) Living and Working with the New Medical Technologies: Intersections of Inquiry. Cambridge, Cambridge University Press. Rieger-Fackeldey, E., Schulze, A., Pohlandt, F., Schwarze, R., Dinger, J. and Lindner, W. (2005) Short-term outcome in infants with a birth weight less than 501 grams. Acta Pædiatrica 94: 211–216. Rip, A. (1995) De sociale constructie van wetenschap, in W. van der Meeren and L. van Langenhove (eds) Management en Wetenschap. Utrecht, Uitgeverij Lemma BV. Robertson, J. A. (2004) Extreme prematurity and parental rights after Baby Doe. Hasting Center Report 34: 32–39. Schlomann, P. (1994) Initiating and sustaining ethical discourse in a neonatal intensive care unit: an ethnographic study. Lexington, The University of Kentucky. Schön, D. A. (1983) The Reflective Practitioner: How professionals think in actions. New York: Basic Books. Seideman, R., Watson, M. A., Corff, K. E., Odle, P., Haase, J. and Bowerman, J. L. (1997) Parent stress and coping in the NICU and PICU. Journal of Pediatric Nursing 12: 169–177. Shields-Poë, D. and Pinelli, J. (1997) Variables associated with parental stress in neonatal intensive care units. Neonatal Network 16 (1): 29–37.

212 Bibliography Sparshott, M. (1997) Pain, Distress and the Newborn Baby. Oxford, Blackwell Science Ltd. Speechley, K. N. and Avison, W. R. (1995) Admission to a neonatal intensive care unit as a predictor of long-term health: a 12-year follow-up. Developmental and Behavioral Pediatrics 16: 397–405. Star. S. L. (1995) Ecologies of Knowledge. Work and Politics in Science and Technology. New York, State University of New York Press. Star, S. L. and Strauss, A. (1999) Layers of silence, arenas of voice: the ecology of visible and invisible work. Computer Supported Cooperative Work 8: 9–30. Stein, L. (1967) The Doctor-Nurse Game. Archives of General Psychiatry 16: 699–703. Stein, L. (1969) The Doctor-Nurse Game. Etc.: a review of general semantics 26: 205–216. Stein, L., Watts, D. and Howell, Th. (1990) The doctor-nurse game revisited. The New England Journal of Medicine 322: 546–549. Stoelhorst, G. M. S. J., Rijken, M., Martens, S. E., Brand, R., den Ouden, A. L., Wit, J. M. and Veen, S. (2005) Changes in Neonatology: Comparison of two cohorts of very preterm infants (Gestational age *bl32 weeks): The Project on Preterm and Small for Gestational Age Infants 1983 and The Leiden FollowUp Project on Prematurity 1996–1997. Pediatrics 115: 396–405. Strauss, A., Fagerhaugh, S., Suczek, B. and Wiener, C. (1985) Social Organization of Medical Work. Chicago, University of Chicago Press. Sudnow, D. (1967) Passing On: the Social Organization of Dying. Englewood-Cliffs, Prentice-Hall. Svensson, R. (1996) The interplay between doctors and nurses – a negotiated order perspective. Sociology of Health & Illness 18: 379–398. Swierstra, T. (1994) Over de waarden der feiten, in D. Pels and G. de Vries (eds) Burgers en vreemdelingen. Opstellen over filosofie en politiek. Amsterdam, Van Gennep. The, A. M. (1997) ‘Vanavond om 8 uur …’ Verpleegkundige dilemma’s bij euthanasie en andere beslissingen rond het levenseinde. Houten/Diegem: Bohn Stafleu Van Loghum. The, A. M. (2002) Palliative care and communication. Experiences in the clinic. Buckingham, Open University Press. Tijmstra, T. (1987) Het imperatieve karakter van medische technologie en de betekenis van ‘geanticipeerde beslissingsspijt’. Nederlands tijdschrift voor Geneeskunde 131: 1128–1131. Timmermans, S. and Berg, M. (1997) Standardization in action: achieving local universality through medical protocols. Social Studies of Science 27: 273–305. Traweek, S. (1988) Beamtimes and Lifetimes. The World of High Energy Physicists. Cambridge, MA, Harvard University Press. Tronto, J. (1993) Moral Boundaries: a Political Argument for an Ethic of Care. New York, Routledge. Verbeek, P. P. (2005) What Things Do – Philosophical Reflections on Technology, Agency, and Design. Penn State: Penn State University Press. Verhagen, E. and Sauer, P. J. J. (2005) The Groningen Protocol – Euthanasia in Severely Ill Newborns. The New England Journal of Medicine 352: 959–962.

Bibliography 213 Verloove-Vanhorick, S. P., den Ouden, A. L. and Walther, F. J. (2001) Uitkomsten van een Nederlands cohort van zeer vroeg geboren kinderen uit 1983. Nederlands Tijdschrift voor Geneeskunde 145: 989–997. Vermeulen, E. (2001) Een Proeve van Leven. Praten en Beslissen over Extreem te Vroeg Geboren Kinderen. Amsterdam: Aksant. Versluys, C. (1987) Doen en laten sterven aan het begin van het leven. Een persoonlijk standpunt van een kinderarts-neonatoloog. Medisch Contact 42: 1083–1085. Vrakking, A., van der Heide, A., Onwuteaka-Philipsen, A., Keij-Deerenberg, I., van der Maas, P. and van der Wal, G. (2005) Medical end-of-life decisions made for neonates and infants in the Netherlands, 1995–2001. The Lancet 365: 1329–1331. Vree van, F. (1998) De scherven van de geschiedenis. Over crisisverschijnselen in de hedendaagse historische cultuur. Amsterdam, Universiteit van Amsterdam. de Vries, G. (2001) Wetenschaps – en techniekonderzoekers, waar is de geest gebleven? Krisis 2 (1): 62–78. Wackers, G. (1994) Constructivist Medicine. Maastricht, Universitaire Pers Maastricht. Warner, J. H. (1986) The therapeutic perspective: medical practice, knowledge and identity in America 1820–1885. Boston: Harvard University Press. Webster, A. (ed) (2006) New Technologies in Health Care. Challenge, Change and Innovation. Hampshire, Palgrave Macmillan. Weir, R. (1984) Selective Non-Treatment of Handicapped Newborns: Moral Dilemmas in Neonatal Medicine. New York, Oxford University Press. Widdershoven, G. (2000) Ethiek in de Kliniek. Hedendaagse Benaderingen in de Gezondheidsethiek. Maastricht, Boom. Widdershoven, G. and Verkerk, M. (1995) Technologie en zorg. Inleiding op het thema. Gezondheid: theorie in praktijk 3: 250–253. Wilde de, R. (1987) Tegen het perspectivisme. Kennis en Methode 11 (4): 354–363. Wilde de, R. (1992) Discipline en Legende. De Identiteit van de Sociologie in Duitsland en de Verenigde Staten 1870–1930. Amsterdam, Van Gennep. Wilde de, R. (1997) Voorbij de utopische waarheid. Een essay over de relatie tussen waarheid, wetenschap en utopie. Kennis & Methode. Tijdschrift voor empirische filosofie 1: 39–53. Wilde de, R. (1998) Reflexiviteit als ideologie. Krisis: Tijdschrift voor filosofie 71: 59–70. Wilde de, R. (2000) Innovating Innovation: a contribution to the philosophy of the future. Paper read at the POSTI Conference ‘Policy Agendas for Sustainable Technological Innovation’, 1–3 December 2000, London. Williams, B. (1981) Moral Luck. Philosophical Papers: 1973–1980. Cambridge, Cambridge University Press. Wilson-Costello, D., Friedman, H., Minich, H., Siner, B., Taylor, G., Schluchter, M. and Hack, M. (2007) Improved Neurodevelopmental Outcomes for Extremely Low Birth Weight Infants in 2000–2002. Pediatrics 119: 37–45. Wittgenstein, L. (1958) Philosophical Investigations. New York: Macmillan Publishing. Wood, N. S., Marlow, N., Costeloe, K., Gibson, A. T. and Wilkinson, A. R. (2000) Neurologic and developmental disability after extremely preterm birth. The New England Journal of Medicine 343: 378–384.

214 Bibliography Young, A. (2000) History, hystery and psychiatric styles of reasoning, in M. Lock, A. Young, A. Cambrosio and A. Harwood (eds) Living and Working with the New Medical Technologies: Intersections of Inquiry. Cambridge, Cambridge University Press. Zahr, L. K. and Balian, S. (1995) Responses of premature infants to routine nursing interventions and noise in the NICU. Nursing Research 44: 179–185. Zerubavel, E. (1979) Patterns of Time in Hospital Life. Chicago, The University of Chicago Press. Zussman, R. (1992) Intensive Care: Medical Ethics and the Medical Profession. Chicago, The University of Chicago Press.

Index Admission, 27–28, 57 Aggressive intervention, 23–25 Limits set upon, 14, 25, 36, 44 Akrich, Madeline, 98, 100 Anspach, Renée, 52, 59–60, 73, 118, 122–123, 195n10, 199n6, 201n3–5 Assent model, 73 Attachment see bonding Baby Doe, 18, 64 Baby Ross, 18 Berg, Marc, 9, 31, 33–34, 78, 86–87, 196n14, 198n6, 200n7–8, 202n6 Bonding, 57–59 Boundaries Professional, 39–42, 44 Technological, 104 BPD, 101–102, 117 CAL-report, 15, 154 Certainification, 89 Co-production, 11 Child – NICU, 33–35, 39 Facts – value, 78, 165–166, 177, 181, 189 Codified experience, 193 Collaboration, 84, 91, 126 Communication Barriers for, 160–161 Errors due to, 91 Staff – parents, 49–50, 52–55, 72, 103 Competence, 95, 100, 108 Consultation, 8, 77, 110, 169 Control, 81, 93, 115 Coordination resources, 188–190 Current past, 136 Custom-made morality see situated morality Death, 175–177 Management of family concerning, 175–176

Staff reactions to, 176–177 Decision-making, 20–26, 35–44, 61–79, 87, 152 As collective process, 25–26, 34–44, 73, 76, 160–166 As negotiation process, 37–44, 161–166 Distribution of responsibility in, 166–169, 175, 184 Role of consultants in, 14, 39–41, 163–165 Role of parents in, 14, 20–26, 45, 54, 56, 61–67, 69, 71–79, 166 see also life saving decisions, withholding treatment Diagnosis, 83, 108 Diagnostic test, 27–28, 33, 83 Delivery room, 19–26 Demarcation, 179, 193 DNR code, 63–68, 170–174 Emotions Of parents, 25, 47, 52–54, 58–59, 61, 129 Of staff, 47, 51, 59, 61, 63, 65, 69–70, 127–129, 177 Errors, 91–92 Esther (baby), 19–22, 25–26, 28–29, 35–45, 58, 71–79, 142, 152, 154, 156–167, 170, 176, 182 Ethical dilemmas, 10–13, 117, 153 In making prognosis, 14, 22, 117, 157–158, 171 Procedures to resolve, 10, 14–17, 23, 36, 39–40, 154–156, 159, 161–165, 189 Technologies creating, 3, 13, 17, 22, 180 see also life saving decisions, withholding treatment Ethics, 10, 180–181 Empirical turn in, 10 Rule-base, 10 215

216 Index Ethnography, 6–8 Normative, 196n18 Theory-guided, 8 Exnovation, 5–6, 9, 188–189 Expectation, 3, 5, 7, 26, 48–57, 123–125, 129, 136, 140–142, 146, 150–151, 187, 191 Foetal condition, 25 Frohock, Fred, 195n10, 197n25, 200n3 Future present, 139 Grief, 175–177 Groningen Protocol, 15–16 Guidelines see protocols Guillemin, Jeanne and Holmstrom, Lynda, 51, 59, 70, 86, 104, 195n10, 200n3 Hands-off policy, 97, 132 Hands-on policy, 132 Handicapped perspective, 21–22, 37–38, 40, 42, 44 see also quality of life Horizon of expectation, 124, 136, 140, 151, 192, 194 Hygienic rules, 46–47 Identity of the infant, 26–33 Idhe, Don, 94 Informed consent, 62, 68–70, 73, 86 Interactive cues, 131 Intermediate zone, 6–7, 187–188 Intersection of activities, 3, 9, 188 Intuition, 130 Invisible work, 9 Knowledge, medical Distribution of, 95, 100, 108, 111–112 Ecology of, 123 Incomplete, 83, 110 Kinds of, 122 Narrative structure of, 83 Sources of, 122, 126–127 Specialized, 95 Koselleck, Reinhart, 124, 136, 140–141, 202nn13–14

Law, John, 5, 9, 31–33, 198n4, 198n6, 200n9 Legal system, 24, 39–40, 63–64, 67–71, 106–107, 114, 191 Life and death decisions see life saving decisions, withholding treatment Life ending decisions see withholding treatment Life saving decisions, 14, 16, 20–26, 62–63, 65–67, 157–158, 184–187 As collective process, 25–26 Balanced against prognosis, 16, 20–26, 65–67, 157 Justification of, 14, 22–25, 157, 184–187 Role of parents, 21–22, 25–26, 62–63, 157 Role of quality of life, 14, 21–22, 66 Role of statistics, 21–25, 157–158 Mathematics of certainty, 105 Maureen (baby), 64–71, 77–79, 119–120, 126, 138, 146, 169, 180 Medical innovation, 2–4, 10, 12–13, 54 Effects on mortality and morbidity of, 12–13, 16 Societal responses to, 3–4, 10, 40, 106–107, 154, 191 Medical round, 34, 85 Method see ethnography Monitor, 94, 96, 98, 109 Monitoring, 2, 34, 93–97, 113, 130 Moral choices, 153, 156–159, 180–181, 184–186 Moral costs, 166 Moral luck, 186 Moral turning point, 161, 178–179 see also demarcation Morality, 10–11, 179, 193 Morality-in-the-making see situated morality Mortality rates, 12–13, 21, 36, 93 Near-misses, 89–91 Neonatologist Responsibilities of, 31, 39, 62, 69, 77, 86, 91, 125, 148–156

Index 217 Spatial awareness of, 148–149 Training of see resident Neonatology As cutting edge medicine, 2, 12–13 History of, 12–13 Professional boundaries of, 39–42, 44 NICU, 1–2 American, 23, 60 As quantitative practice, 114 As reflective practice, 180–184, 194 Dutch, 19 Moral identity of, 167 Practical order of, 33–35, 99, 106 Social order of, 46–48, 61, 70, 74, 77–78, 99–100 Spatial order of, 147–150 Temporal order of, 144–147 NICU staff Professional position, 121–123, 137 Relation with parents, 47–52, 59–60 Rotation, 91, 95–96 Numbers see quantitative data Nurses And patient care, 93, 95, 98–99 Ethical dilemmas of, 69, 117, 168–169, 172, 183 Legal position of, 68–69, 174 Role in decision-making, 68–69, 168–169 Spatial awareness of, 150 Supportive role of, 59, 175 Nursing Primary care, 126 Team, 127 NVK-report, 15, 154–160 Ordering process, 5, 7, 9, 28–36, 98, 188–189

Role in decision-making see decision-making Social-psychological situation of, 51, 60 Parent – infant relationship, 58–59, 125–129, 131, 175–177 see also bonding Paternalism, 52, 61, 73, 77 Patient care, 96, 99, 132–133 Role of physical appearance, 131, 176 Patient population American, 123 Dutch, 19 Patient record, 31–34, 98 Physical examination, 27–31 Present future, 136 Prognosis, 20–26, 38–45, 116–151, 154–157, 172–173, 201n4, 202n3, 202n14, 203n9 Different, 117–121, 124 Prognostic markers, 136–150 Interactive markers, 132–133 Observational markers in, 130–132 Spatial markers, 147–150 Statistical markers, 137–138 Technological markers, 141–147 Temporal markers, 144–147 Prognostication process, 122–152 Interactive dimensions of, 125–136, 138 Role of comparison in, 137–139 Role of intuition in, 130 Spatial dimension of, 125, 130, 139 Temporal dimension of, 135, 139, 140 Protocol, 86–88, 92–93, 178, 193 Critique on, 193 Public debate about Neonatology, 13–17

Parental authority, 26, 54, 56, 61–64, 68, 71–79, 166–167 Parents, 20–26, 54, 61–79, 157, 165, 171 Communication with, 52, 103 As knowledge producers, 123, 138–140, 142–144, 147, 172 As object of care, 55, 59, 63, 71–79, 175

Qualitative studies, 2–3, 6 Quantitative data, 29–32, 104–105, 113–114 Interpretation of, 107–114, 137–138 Juridical significance of, 106–107, 114, 191 Objectivity of, 105–106, 109–110, 113–114 see also statistics

218 Index Quantitative language, 105–106, 113–114, 168 Reflection, moral, 179–186, 193 Reflective practitioner, 194 Repertoire, 36, 41–45, 139–142, 144–145, 149, 157, 163–164, 188–190 Case, 38–39, 42–45, 158, 163–164, 166, 168 Cyclic, 139–145 Linear, 139–145 Situational, 37, 39, 42–45, 149, 157, 163–164, 166 Resident, 29–31, 85, 107–109, 134–135, 185 Training of, 24, 29, 59, 135 Respirator, 97–99, 101–104, 117, 141, 174 Responsibility Distribution of see decision-making, withholding treatment Locus of, 78 Risk, 83, 86, 88–92, 101, 115, 174 Calculated, 101 Robert (baby), 57, 83–84, 87–92, 95–97, 100, 114–115, 126, 131–132, 135, 147, 191 Round see medical round Science and Technology Studies, 8–10 Script, 98–100, 156, 174, 188–190, 193 Situated morality, 11, 162–166, 179, 187, 190, 193 Societal responses see also public debate Consequences for medical practices of, 3–4, 10, 40, 106–107, 154, 191 Space of experience, 124 Standardizing medical work, 86 Statistics, 12–13, 21–23, 38, 40, 137–138, 140, 157–158 Uncertainty of, 157–158 Stereotypes, 51, 120 Styles of ordering, 7, 9, 77–79, 98, 139–140, 163, 188–189

Surgery Pre-surgical care, 84, 87–88 Post-surgical care, 88–92 Survival rates, 12–13, 21–23 Technology, 93–94, 191–192 As source of uncertainty, 95 Disciplining effect of, 99–100 Imperative character of, 13, 160, 180 In relation to care, 96–97, 99, 103, 171 Unintended effects of, 101–104, 191–192 Time, 75–77, 82–86, 88, 93, 136, 144–147, 151 Creation of additional, 84, 159–160 Tom (baby), 1, 27–31, 94, 98–99, 101–104, 115–126, 129–131, 133, 138–140, 144, 147, 152, 168, 170–177, 183–187, 191 Trajectory, 82, 97–104, 115, 136, 158–160, 170–172, 175 Transitional moment, 87–91, 95–96, 105, 142 Transposition, 84, 87–88 Uncertainty Communication with parents of, 40, 71–72, 76 Coping with, 81, 84–89, 93, 105, 108–109, 113 Diagnostic, 4, 26, 82–85, 111–112 Ethical, 15, 152–178 Medical, 4, 23, 80–84, 89, 108 Prognostic, 4, 19–22, 26, 38–43, 122, 157, 171–174, 177 Role of statistics, 157 Sources of, 80–83, 87, 89, 96, 108–109, 112, 172, 199–200n1 Ventilator see respirator Vocabulary, 55–56, 72–74, 77–79, 188–190 Juridical, 56, 61–74, 77–79 Psycho-social, 55, 57, 59–61, 70, 72–74, 77, 79

Index 219 Williams, Bernard, 166, 186, 203n8 Withholding treatment, 14–16, 35–45, 154, 160 As collective process, 36, 73, 76, 160, 166–167 Balanced against diagnosis, 155 Balanced against prognosis, 35–45, 76, 154–155, 161–165, 170–71, 194 Distribution of responsibility in, 166–169, 175–176, 184 Justification of, 37–45, 65, 69, 76, 155, 159, 161–167, 194

Position of nurses in, 68–69, 171, 174 Role of parents in, 37, 45, 62–64, 68, 71–79, 165–167 Role of quality of life in, 37–42, 44, 66, 69, 75, 170–171, 177 Role of statistics, 23–25, 38, 40, 42, 159, 194 Zussman, Robert, 120, 195n10, 199n1, 200n3