Spirituality, Religiousness, and Health: From Research to Clinical Practice Paul J. Mills, Ph.D. University of California, San Diego
special issue title, the articles cover topics of relevance to research (including reviews of the existing literature and methodologies and an examination of new instruments) and to clinical practice. Carl E. Thoresen and Alex H. S. Harris present an overview of the field, examine evidence, and address possible mechanisms that may underlie the potential effects of religious and spiritual factors on health outcomes. The authors also discuss implications for health care professionals. Richard P. Sloan and Emelia Bagiella present a critical review of the literature linking religious involvement and health outcomes. The authors call into question the methodological soundness of these studies and illuminate a tendency of review articles to rely on inappropriate secondary sources. These articles serve to advance the field by calling into question the prior tendency of “uncritical positivism” when reporting research findings (7). Three articles address the critical issue of the development and testing of instruments. As in any field of research, true progress can only be built on the foundation of valid and reliable data. These instruments have been designed for topics of spirituality, religiousness, and health among healthy and chronically ill populations. They have had to address some rather basic and difficult questions as well. Exactly what spirituality is, for example, has been a vexing issue in the literature and somewhat of a stumbling block (8,9). Concepts such as sense of peace, faith, compassion, religious behavior, and belief in God have needed to be addressed and incorporated into definitions of spirituality and religiousness. Other challenges have included determination of the most appropriate way to assess religiousness and spirituality in respondents from different religious traditions and how to assess spirituality without the use of terms typically of a religious nature.
There are several high-profile and controversial topics in medicine today (cloning and stem cell research come to mind). A less well known yet controversial area receiving increasing federal and private funding is spirituality, religiousness, and health. Attention to topics of spirituality, religiousness, and health has increased substantially in medical and graduate school curricula, clinical practice, and research (1–3). Of the many interesting aspects of this phenomenon, perhaps the most remarkable is the observation that medical science, the field of inquiry that initially separated mind from body (e.g., the ghost from the machine) (4), now finds it compelling and perhaps even necessary to reexamine relationships among spirit, mind, and body. The number of empirical studies on spirituality, religiousness, and health has proliferated in the scientific literature. Using the keywords religion and health and spiritual/spirituality and health, a Medline search from the year 1975 to the present reveals a striking trend, especially in the last 5 to 6 years (Figure 1). In recent years, every major medical, psychiatric, and behavioral medicine journal has published on the topic. In 1999, the National Institutes of Health Office of Behavioral and Social Sciences Research created an expert panel of scientists to critically examine this growing body of literature. Usually, such an increase in interest and publications on a topic follows a new discovery, such as that of a novel gene; the development of a new medical instrument; or perhaps a more sensitive and reliable assay. This is obviously not the case in this instance because spirituality and religion have been relative constants of cultures. The word spirituality did not even appear in Medline until the 1980s. Reasons for this relatively dramatic increase likely include the growing field of complementary and alternative medicine as well as one of the more unpalatable characteristics of managed care: the impersonal nature of assessment and treatment. There have been efforts on numerous fronts, the impetus for a majority of which originated from patients themselves, to bring acknowledgment of the “whole person” back into medicine (5). The impetus for this special issue of Annals of Behavioral Medicine arose from a panel presentation1 on spirituality, religiousness, and health at the annual meeting of the Society of Behavioral Medicine, held in March 2000 in Nashville, Tennessee (6). Six of the eight articles are from members of the Society, and two were solicited from nonmembers. As indicated by the
1Supported
by the Fetzer Institute.
Reprint Address: P. J. Mills, Ph.D., Department of Psychiatry, UCSD Medical Center, 200 West Arbor Drive, San Diego, CA 92103–0804. E-mail:
[email protected]
FIGURE 1 Number of published studies having the keywords religion and health and spiritual/spirituality and health appearing in the Medline database from 1975 to December 2001.
© 2002 by The Society of Behavioral Medicine.
1
2
Mills
It is obvious from these articles that enormous efforts have gone into the development of meaningful, valid, and reliable instruments to assess attitudes toward religiousness and spirituality as well as spiritual experience. Lynn G. Underwood and Jeanne A. Teresi describe the development, testing, and health-related outcomes of the Daily Spiritual Experience Scale. This scale fills an important need because it provides assessment of everyday ordinary experience rather than assessment of particular religious beliefs or behaviors. Gail Ironson and colleagues examine relations among spirituality, religiousness, health outcomes, and survival in people living with HIV using the Ironson–Woods Spirituality/Religiousness Index, an instrument designed to assess private and public aspects of both spirituality and religiousness. Amy H. Peterman and colleagues report on the development and testing of a measure of spiritual well-being in people with chronic illness. The authors present findings from the Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being Scale in two samples of cancer patients. Leila Shahabi and colleagues used data from several instruments to determine the sociodemographic and behavioral correlates of self-perceptions of spirituality and religiousness in a large sample of American adults. They included in their questionnaires the Daily Spiritual Experience Scale (see Underwood and Teresi, this issue) and questions on religious denomination and public and private religious behavior. These investigators found that attitudes toward spirituality and religiousness were different across disease-risk relevant demographic characteristics (e.g., age, sex, education). These data have significant implications when the data linking religiousness or spirituality with morbidity and mortality outcomes are considered, and they lend support to prior calls for the need to more thoroughly consider potentially confounding variables in research design and interpretation (10–12). Although most of the research has been devoted to the determination of whether spirituality or religious behavior actually has significant effects on health (positive or negative), it is in the other half of the endeavor, the examination of the potential for translation into clinical practice, where the more substantive controversies exist. Archives of Internal Medicine recently published an article that examined the effects of remote intercessory prayer on outcomes in patients admitted to a coronary care unit (13). A beefy 15 letters to the editor were subsequently published in response to that article (Vol. 160, June 26, 2000). The letters addressed a host of topics, including methodological and statistical concerns and questions regarding ethical issues, informed consent, and implications for medical practice. The final two articles of this issue address these latter, clinically oriented questions, including whether physicians should take a spiritual history (i.e., whether they should inquire about a patient’s faith or religious commitment) and whether physicians should prescribe religious activities as adjunctive medical treatment (14–16). Walter L. Larimore and colleagues write in favor of these efforts, whereas Raymond J. Lawrence argues against them. Together, these eight articles should provide the readers of Annals of Behavioral Medicine with an appreciation of the numerous research and clinical challenges in the field today and of
Annals of Behavioral Medicine the creative approaches being taken to address them. With time, the literature will undoubtedly provide a more conclusive answer to the question of whether and through what mechanisms spirituality and religiousness influence health. Resolving issues related to the clinical translation of such research findings will be more difficult. In the meantime, a more immediate and certainly welcome outcome is that the medical community has extended an invitation for the spirit, mind, and body to once again reside together when the physical health and well-being of the individual are considered. REFERENCES (1) Anandarajah G, Long R, Smith M: Integrating spirituality into the family medicine residency curriculum. Academic Medicine. 2001, 76:519–520. (2) Post SG, Puchalski CM, Larson DB: Physicians and patient spirituality: Professional boundaries, competency, and ethics. Annals of Internal Medicine. 2000, 132:578–583. (3) Astrow AB, Puchalski CM, Sulmasy DP: Religion, spirituality, and health care: Social, ethical, and practical considerations. American Journal of Medicine. 2001, 110:283–287. (4) Cotterill RMJ: No Ghost in the Machine: Modern Science and the Brain, the Mind and the Soul. London: Heinemann, 1989. (5) Cangialose CB, Cary SJ, Hoffman LH, Ballard DJ: Impact of managed care on quality of healthcare: Theory and evidence. American Journal of Managed Care. 1997, 3:1153–1170. (6) Mills PJ, Bagiella E, George LK, Sloan RP, Thoresen CE: Spirituality, religiousness and health: From research to clinical practice. Annals of Behavioral Medicine. 2000, 22:xviii–xix. (7) Lederberg MS, Fitchett G: Can you measure a sunbeam with a ruler? Psycho-Oncology. 1999, 8:375–377. (8) Hawks SR, Hull ML, Thalman RL, Richins PM: Review of spiritual health: Definition, role, and intervention strategies in health promotion. American Journal of Health Promotion. 1995, 9:371–378. (9) Fetzer Institute, National Institute on Aging Working Group: Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research. Kalamazoo, MI: Fetzer Institute Publications, 1999. (10) Sloan, RP, Bagiella E, Powell T: Religion, spirituality and medicine. Lancet. 1999, 353:664–667. (11) Weaver AJ, Flannelly LT, Flannelly KJ, Koenig HG, Larson DB: An analysis of research on religious and spiritual variables in three major mental health nursing journals, 1991–1995. Issues in Mental Health Nursing. 1998, 19:263–276. (12) Cotton SP, Levine EG, Fitzpatrick CM, Dold KH, Targ E: Exploring the relationships among spiritual well-being, quality of life, and psychological adjustment in women with breast cancer. Psycho-Oncology. 1999, 8:429–438. (13) Harris WS, Gowda M, Kolb JW, et al.: A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Archives of Internal Medicine. 1999, 159:2273–2278. (14) Larimore WL: Providing basic spiritual care for patients: Should it be the exclusive domain of pastoral professionals? American Family Physician. 2001, 63:36–39. (15) Koenig HG: Spiritual assessment in medical practice. American Family Physician. 2001, 63:30–31. (16) Sloan RP, Bagiella E, VandeCreek L, et al.: Should physicians prescribe religious activities? New England Journal of Medicine. 2000, 342:1913–1916.
Spirituality and Health: What’s the Evidence and What’s Needed? Carl E. Thoresen, Ph.D. and Alex H. S. Harris, M.S. Stanford University
physical activities, and quiet reflection and prayer (1,2). Ancient wisdom, situated within major religions, echo much of the message voiced by Hippocratic medicine in ancient Greece, the notion that lifestyle (including beliefs, emotions, and behaviors) powerfully influences health and risk of disease. Indeed, one can view Engel’s seminal work in the 1970s calling for a biopsychosocial model (vs. one only focusing on biological factors) as a contemporary expression of ancient conceptions of health (3). (Note that we use the abbreviation RS throughout the article to signify “religious and/or spiritual” to save space. Spirituality, religiousness, and religion are often used interchangeably yet for many represent somewhat distinct, if not at times independent, constructs.) Interest in possible relations between religion, spirituality, and health has attained the status of a “hot topic.” Scientific audiences have become interested, with special issues on spirituality and health scheduled or published in scholarly journals (e.g., American Psychologist, Journal of Health Psychology). In 1997, the John Templeton Foundation in collaboration with the National Institute of Healthcare Research (NIHR) sponsored a series of conferences with medical and behavioral scientists to review the scientific evidence relating RS variables with physical health, mental health, alcoholism and other addictive disorders, and neurobiological factors (4). That effort helped trigger the Office of Behavioral and Social Sciences Research (OBSSR) at the National Institutes of Health (NIH) to establish a panel of behavioral scientists to critically review existing evidence presumably linking RS factors with health. Reactions to the popularity of RS issues and to the notion that RS factors influence health outcomes have been varied. They tend to range from cynical skepticism on one hand (e.g., 5) to supportive advocacy on the other (e.g., 6). Benson (7), for example, contended that faith in God has health-promoting effects. Others, however, have argued that suggestions linking RS factors with better health are clearly unwarranted and warn against physicians involving spiritual and religious matters in their practice (8,9). Others have offered a less critical perspective on the quality of existing evidence linking RS factors to health and believe that health professionals need to consider this area in meeting ethical responsibilities to their patients (e.g., 10,11). In this article, we seek to familiarize readers with some recent empirical evidence about possible associations between RS factors and health outcomes. In considering this evidence, we believe a healthy skepticism is in order. One needs to remain open to the possibility that RS-related beliefs and behaviors may indeed be important to health processes, in promoting and in endangering health. Yet one needs rigorous empirical evidence based on well-controlled studies that support these claims and conclusions. We hope to introduce the dismissing critic to suggestive data that may create tempered doubt. We also hope to in-
ABSTRACT In this article, we familiarize readers with some recent empirical evidence about possible associations between religious and/or spiritual (RS) factors and health outcomes. In considering this evidence, we believe a healthy skepticism is in order. One needs to remain open to the possibility that RS-related beliefs and behaviors may influence health, yet one needs empirical evidence based on well-controlled studies that support these claims and conclusions. We hope to introduce the dismissing critic to suggestive data that may create tempered doubt and to introduce the uncritical advocate to issues and concerns that will encourage greater modesty in the making of claims and drawing of conclusions. We comment on the following questions: Do specific RS factors influence health outcomes? What possible mechanisms might explain a relation, if one exists? Are there any implications for health professionals at this point in time? Recommendations concern the need to improve research designs and measurement strategies and to clarify conceptualizations of RS factors. RS factors appear to be associated with physical and overall health, but the relation appears far more complex and modest than some contend. Which specific RS factors enhance or endanger health and well-being remains unclear. (Ann Behav Med
2002, 24(1):3–13)
INTRODUCTION Cultures throughout history have viewed health and disease as directly related to a variety of religious beliefs and practices, as evidenced by specific religious prescriptions concerning diet, Preparation of this article was supported, in part, by grants to Carl E. Thoresen from the Fetzer Institute, John Templeton Foundation, and from participation on the Spirituality, Religiousness and Health: State of the Science Panel, Office of Behavioral and Social Sciences Research (OBSSR/NIH). Work by Alex H. S. Harris was partially supported by a Stanford Presidential Graduate Fellowship. Portions of this article were presented at the Annual Meeting of the Society of Behavioral Medicine, Nashville, Tennessee (2000). Part of this article will appear in the forthcoming Handbook of Health Psychology (Vol. 2), edited by J. Raczynski, L. L. Leviton, and L. Bradley (American Psychological Association), and do appear in Faith and Healing: Psychological Perspectives, edited by T. Plante and A. Sherman (Guilford, 2001). The major contributions of several researchers in this area are gratefully acknowledged, including Harold Koenig, David Larson, Michael McCullough, William R. Miller, Ken Pargament, and Everett Worthington, Jr. Reprint Address: C. E. Thoresen, Ph.D., School of Education, Stanford University, Stanford, CA 94305–3096. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
3
4
Thoresen and Harris
troduce the uncritical advocate (some might say the “cheerleader”) to issues and concerns that will encourage greater modesty in the making of claims and drawing of conclusions. To facilitate more informed perspectives on both sides of the debate, we address briefly the following questions: Do specific RS factors influence health outcomes? Is there empirical evidence of sufficient quality to justify an answer to this question? What possible mechanisms might explain or account for a relation if one exists? Are there any implications for health professionals based on the evidence at this point in time? We do not discuss here several possible explanations for the growing interest in the RS factor–health connection among the general population and among health professionals, and we do not comment on the common reluctance of many scientists to study RS factors, such as the belief that they cannot or should not be studied (see 12). We do offer a brief comment on how the terms spirituality and religion are used, but we omit a fuller discussion of definitional and conceptual issues. The background of earlier empirical research on RS factors and health, essentially underway since the mid 20th century, is also not presented. For extended treatments of these topics, the reader is referred elsewhere (12–19). Comments on the scientific evidence that RS factors are associated with health status include the following:
• Some tentative conclusions about physical health from • • •
•
Scientific Research on Spirituality and Health: A Consensus Report (4). A consideration of four recent well-conducted epidemiological studies to provide the reader with a more concrete sense of the state of the science on RS factors and health. A comment on the few available experimental interventions. A brief consideration of some issues and concerns emerging from the current OBSSR/NIH panel on the state of scientific evidence concerning RS factors and health. A discussion of some research issues deserving attention.
Before considering the evidence, we briefly comment on the terms spirituality and religion, given the confusion often surrounding these constructs, with suggestions for further reading. RELIGION AND SPIRITUALITY: SAME, SIMILAR, OR SEPARATE CONCEPTS? Broadly speaking, the term religion is often viewed as a societal phenomenon, involving social institutions composed of members who abide by various beliefs and adhere to certain rules, rituals, covenants, and formal procedures. By contrast, a typical view of spirituality refers to the individual’s personal experience, commonly seen as connected to some formal religion but increasingly viewed as independent of any organized religion (13). However, the term religiousness is often used to convey the individual’s personal experience as part of an organized religion. William James (20), for example, used the term reli-
Annals of Behavioral Medicine giousness in this way with focus on the personal attitudes, emotions, and personality factors. The Oxford English Dictionary offers 10 pages of reference material on the concept of spirituality (21). Two related themes dominate this material. First, spirituality refers to life’s most animating or vital issues and concerns (e.g., the term spiritus in Latin means “the breath,” that which is most vital to life). Second, spirituality is seen as the more immaterial or subjective features of life, as distinct from the body or other more tangible and material things, including the senses, such as sight and hearing. Some contend that religion is the more inclusive concept, with spirituality as its major focus: “Religion is a search for significance in ways related to the sacred” (22, p. 11). Others contend that spirituality is the more inclusive term of which religion and religiousness may or may not be a part. We opt for the view that these constructs may be viewed as two overlapping circles (Venn diagrams), with spirituality being the larger circle yet sharing with religion many overlapping areas, but each having nonoverlapping areas (23). As such, spiritual matters often concern seeking meaning and purpose in life, as well as those experiences which provide, for example, a greater sense of inner peace, harmony, hopefulness, and compassion for others and oneself. We suspect, however, that spirituality, if it is to mean something more than any idiosyncratic personal belief, involves seeking a sense of being or becoming connected to something greater than just oneself (e.g., “beyond the ego”) (24), something that provides a sense of the sacred or holy. For further discussion of these cultural, social, psychological, and theological issues, see Pargament (25). Also see Woods and Ironson (26) and Shahabi and colleagues (this issue/27) for empirical studies of health-related differences between those identifying themselves as only spiritual or only religious or as both or neither. We comment on prevalence in the United States of religiousness and spirituality later in this article. WHAT IS KNOWN ABOUT THE RS FACTOR–HEALTH CONNECTION? Interesting if not surprising associations between religious involvement, broadly defined, and health has been reported. Until very recently, the vast majority of the research examining potential relations between religious factors and physical health status, including mortality, has been cross-sectional in nature. Note that almost all of the research conducted to date has focused on religion or religiousness, not on spirituality seen as somewhat or completely independent from religion. Essentially, there is at present no well-controlled data on spirituality, seen as independent of religion, and health. Furthermore, assessment of religious involvement has been almost always limited to a person’s reported affiliation with any organized religion (or the person’s particular denomination within a religion) or to the frequency of attendance at religious services. Occasionally, researchers have asked questions about the importance or meaningfulness to the person of religion or religious beliefs. Furthermore, assessments have been typically limited to one or a few questionnaire items administered on one occasion. We return to this highly restrictive perspective of RS factors, com-
Volume 24, Number 1, 2002 menting on ways to improve assessment and evaluation in research. Scientific Research on Spirituality and Health: A Consensus Report: Physical Health (4) In the NIHR consensus report, Matthews, Koenig, Thoresen, and Friedman (28) cited studies providing some evidence to link religious involvement, usually frequency of religious service attendance, with physical health factors, such as the following:
• Lower rates of coronary disease, emphysema, cirrhosis, • • •
• • •
and suicide (29). Lower blood pressure (30). Lower rates of myocardial infarction (31). Improved physical functioning, medical regime compliance, and self-esteem and lower anxiety and health-related worries 1 year after surgery in heart transplant patients (32). Reduced levels of pain in cancer patients (33). Better perceived health and less medical service utilization (34). Decreased functional disability in the nursing-homedwelling elderly (35–37).
Keep in mind that many of the studies cited were either cross-sectional in design or prospective (i.e., longitudinal) based on selective samples in terms of participants’ characteristics (e.g., ethnicity, education, health status) and the area (e.g., the southeast region of the United States) from where they were drawn. In addition, these studies seldom used sufficient control measures or covariates known to influence health. Lessons From Recent Well-Controlled Epidemiological Studies: Religious Attendance and Mortality Because these issues, like other possible variables that could explain the reported relations linking more frequent religious service attendance with less disease, were not adequately taken into account, the results reported previously remain difficult to interpret. For example, cigarette smoking is clearly related to morbidity and mortality and is possibly associated with various markers of RS involvement. Failure to adequately control for smoking in any study involving RS factors and physical health raises doubt about what can be concluded about the role played by RS factors. Recently, researchers have focused on the use of state of the art epidemiological designs, often with fairly large population-based samples, to see if RS factors predicts health status, especially mortality, when many other factors related to health outcomes are included in the analysis. Table 1 describes four such recently completed studies (38–41). Several things are noteworthy about these four studies, including the national sampling used in two of them and the effort in all to include many covariates (potential confounders) that could compete, so to speak, with the religious factors in the prediction of health outcomes. As can be seen from these studies, even with 12 or more control variables used in the analysis,
Spirituality and Health
5
the higher frequency of attending services still independently predicted less all-cause mortality. Note that some of the covariates also independently predicted mortality (e.g., being married, better physical functioning of daily tasks, higher levels of social support, being female). That these factors also independently predicted mortality highlights an important point about a specific RS variable: It may be one of many significant influential team players, but one not always on the field, in the influence of health and disease processes (42). Indeed, some of these independently significant factors may over time interact with each other to eventually alter disease risk (12). To illustrate, RS variables may influence opportunities to develop and maintain socially supportive relations, which in turn may reduce depression and moderate alcohol use, which in turn may help reduce undesirable physiological states, such as chronically elevated cortisol and norepinephrine levels. At present, this indirect or distal role remains speculative because empirical evidence is not available but could become available if changes are made in future research designs, statistical analyses, and assessment strategies. Note also that all four of these studies were conducted with reasonably healthy people, not patients with particular medical diagnoses. At present, the mechanisms linking religious variables and mortality for patients in large, well-controlled prospective studies remain unclear. Hummer, Rogers, Nam, and Ellison (38) found that not attending religious service independently predicted a significantly higher risk of respiratory-related deaths as well as higher residual causes of mortality and a marginally higher chance of infectious and circulatory diseases. This study also demonstrated that compared to other significant independent risk factors, seldom or not attending services was associated with 50% more mortality than weekly attending services. The risk of not attending was only surpassed by the risks of heavy smoking (but not moderate or light smoking), being male, and self-reported poor overall health (63, 60, and 258% increased mortality rate, respectively). Note also that without any of the control factors used, nonattenders experienced 87% more mortality (hazard ratio [HR] = 1.87) than those attending services more than once a week, but this dropped to 50% more mortality in the model that used all of the control variables. The study by Musick, House, and Williams (39) commendably attempted to unpack or disaggregate broad, complex dimensions such as “religious involvement.” Perhaps the “tip of the iceberg” was revealed in this study by the use of various combinations of variables to best clarify possible RS factor and mortality relations. First, they found that religious involvement independently predicted mortality when social demographic factors were controlled. Then they discovered that once health behaviors (e.g., smoking, exercise) were entered in the analysis, religious involvement (a combination of attending services, praying, listening to religious TV programs, and reading scripture from the Bible) no longer independently predicted mortality. They then separated religious involvement into two categories: private and public religiousness. This resulted in religious service attendance (a public type of religiousness) predicting
6
Thoresen and Harris
Annals of Behavioral Medicine
TABLE 1 Recent Controlled Studies of Relations Between Religious or Spiritual Factors and Physical Health Study and Sample Characteristics
Control Variables
Hummer, Rogers, Nam, and Ellison (38): Epidemiological study of religious involvement and mortality, with an 8-year follow-up, in a sample of 21,204 U.S. adults. Primary predictor variable: Frequency of religious service attendance. Outcome variable: Mortality analyzed by cause of death.
1, 2, 3, 4, 5, 6, 7, 8, 9, 11
Musick, House, and Williams (39): Epidemiological study of religious involvement and mortality, with a 7.5-year follow-up, in a sample of 3,617 U.S. adults. Primary predictor variable: Frequency of religious service attendance. Outcome variable: Mortality
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, and other religious factors (theodicy/ beliefs)
Oman and Reed (40): A prospective study of religious service attendance and all-cause mortality over 5 years in a sample of 1,931 older residents of Marin County, CA. Primary predictor variable: Frequency of religious service attendance. Outcome variable: Mortality Strawbridge, Cohen, Shema, and Kaplan (41): Epidemiological study of religious involvement and mortality, with multiple assessments over 28 years, in a regional sample of 5,286 Alameda County, CA residents. Primary predictor variable: Frequency of religious service attendance. Outcome variable: Mortality, improvement in health practices, increased social contacts, and stable marriages.
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, and years of residence in county
1, 2, 3, 4, 6, 8, 9, 10, and religious affiliation
Results and Conclusions In the full model (all covariates included) across all cause-of-death categories, those who never attended services had an HR of 1.50 (p < .01); less than once a week had 1.24 (p < .05); and the weekly group had 1.21 (p < .05) over the more than once a week RG. In the full model at p < .05 level, those who never attended services were at greater risk for death from respiratory disease (HR = 2.11) and residual causes (HR = 2.42) than RG, and at marginally greater risk (p < .10) for circulatory and infectious diseases, cancer, diabetes, and external causes. In the full model, compared to the less than once a month RG, those who attended services one to three times a month had an HR of 0.75 (25% less); those who attended weekly had 0.65 (35% less); those who attended more than once per week had 0.61 (39% less). This relation was unexpectedly stronger among persons younger than 60 years of age. Other religious behaviors and beliefs did not explain, and sometimes suppressed, the inverse association between service attendance and mortality. For example, private religiousness suppressed the effects of public religiousness (attending services) on mortality. The need for more specific religious and spiritual factors was demonstrated. In the full model, weekly attendees of religious services had lower mortality than nonattendees, RR = .72 (95% CI = 0.55, 0.93). Contrary to their hypothesis, religious attendance tended to be slightly more protective for those with high social support. In the full model, frequent attendees of religious services had lower mortality than infrequent attendees (RH = 0.77; 95% CI = 0.64, 0.93); the effect was significant for women (RH = 0.66; 95% CI = 0.51, 0.86) but not for men (RH = 0.90; 0.95; CI = 0.78–1.15). At follow-up, frequent attendees were more likely to have stopped smoking, OR = 1.90 (95% CI = 1.27, 2.85); increased exercise, OR = 1.38 (95% CI = 1.08, 1.77); increased social contacts, OR = 1.50 (95% CI = 1.02, 2.21); and stable marriages, OR = 1.79 (95% CI = 1.36, 2.35).
Note. 1 = sex, 2 = age, 3 = health status, 4 = health behaviors, 5 = income, 6 = education, 7 = marital status, 8 = race/ethnicity, 9 = social support, 10 = mental health, 11 = geographic region, 12 = employment status; HR = hazard ratio; RG = reference group; RR = relative risk; CI = confidence interval; OR = odds ratio.
less death but private activities, such as watching religious TV and reading scripture, not being predictive. Instead, these private activities seemed to diminish the power of overall religious involvement (public and private types combined) to predict less mortality. Stated differently, private religiousness suppressed the association of public religiousness (service attendance) in predicting mortality. How this might be explained raises impor-
tant questions. One possibility might be that those engaging in private activities are less healthy physically, perhaps unable to attend services, and thus at greater risk of mortality. Another might be that those engaging in more private activities may differ from others in personality-related factors or other psychological characteristics (17,27).
Volume 24, Number 1, 2002 Finally, the study of Strawbridge, Cohen, Shema, and Kaplan (41) demonstrated at least two major points. First, when both men and women were grouped together in the model including all control variables, frequent attenders had lower mortality rates than infrequent attenders (HR = 0.77, or 23% less mortality). However, this changed dramatically when men and women were considered separately: Frequent attendance was significantly associated with less death only in women (HR = 0.66; 95% confidence interval [CI] = 0.51, 0.86) but not in men (HR = 0.90; 95% CI = 0.70, 1.15). Second, this represents the only well-controlled study to date that has provided evidence using repeated measures revealing that changes in specific health behaviors (e.g., smoking, exercise) and social factors (e.g., frequency of social contacts, stability of marriages) were associated with a religious variable, in this case frequent attendance at religious services. A recent follow-up study (43) examined gender differences in the maintenance of positive changes in physical and mental health factors. Men and women attending services weekly or more often were more likely to quit smoking, become more physically active, not get depressed, and increase their number of personal relationships. In terms of preventive actions, high attenders were less likely to avoid physical checkups, divorce, and decrease personal friendships and marginally less likely to become depressed. In general, what is the evidence that RS variables predict health in controlled prospective studies that account for other possible factors? McCullough, Hoyt, Larson, Koenig, and Thoresen (42) provided a tentative answer. They conducted a meta-analytic review of religious involvement (including the four studies cited previously), mostly based on religious service attendance, and all-cause mortality. Twenty-nine different samples were used, involving 42 effect sizes and 18 control variables or covariates. These studies involved more than 126,000 participants. Although the independent effect size was modest (0.10), it was of the same magnitude in predicting all-cause mortality as depression, social support, excessive alcohol consumption, and the use of cholesterol-lowering drugs for those at high risk. Stated differently, those frequently attending religious services had approximately 29% fewer deaths from all causes when compared to those who were not religiously active (odds ratio [OR] = 1.29; CI = 1.20–1.39). Several other factors in the meta-analysis also independently predicted mortality risk, such as marital status, social support, perceived overall health status, and whether public or private measures were used (e.g., OR = 1.43 for public religiousness but only 1.04 for private religiousness). Evidence from any meta-analytic study, no matter how large the number of independent studies or total number of participants, is not without possible shortcomings. For an interesting exchange on how this meta-analysis on RS factors and health can be viewed in very different ways, see Sloan and Bagiella (44) and a reply by McCullough, Hoyt, and Larson (45). Although these studies illustrate that frequency of religious service attendance has a nontrivial and nonrandom association with mortality, they do not demonstrate that attendance or religion in general caused mortality less often. Furthermore, a signif-
Spirituality and Health
7
icant statistical relation in itself, such as one between greater attendance and reduced all-cause mortality, even if found within a prospective, well-controlled design, does not mean that religious attendance benefits everyone or even benefits most people (46). It does tell us that, on average, religious service attendance is associated with less mortality from all causes. It is not a random relation. However, such a relation in itself seldom sheds light on which persons may benefit more or less, and it does not clarify possible mechanisms or processes by which this happens. These findings, although in many ways representing state of the art epidemiological studies, suggest several issues that deserve attention: First, more specific factors within the broader RS concepts currently used, such as religious involvement, must be assessed. One can readily conjecture, given the way constructs have been measured, that some RS factors and health relations may actually turn out to be much stronger, or much weaker, than current evidence suggests. Indeed, McCullough et al. (42) found significant variability among the 29 samples studied, such that generalizations need to be cautiously viewed. It is important that the examination of more specific features of RS factors may help clarify possible negative or harmful relations of religious factors and health, a topic that has received far too little attention in the empirical literature (this issue/27). Second, the measurement of additional psychological factors within these studies is needed. Essentially, the use of well-established psychological concepts in this field of study has been “missing in action” (19,47). Assessing for more psychological factors would permit examination of possible interaction effects of certain person factors with certain religious factors (e.g., how narcissistic personality characteristics interact with religious attendance or with a person’s beliefs about God, such as a strict and punishing God vs. a loving and forgiving God). Beyond Simple Baseline Predictions Miller and Thoresen (13) observed that when predictor or control variables are related to each other, too often they fail to be considered for further study because they are found to be statistically insignificant in epidemiological studies. They suggested that experimental and repeated measures designs can answer many of the questions raised by epidemiological studies that are limited to the assessment of RS factors on a single occasion. Chatters (48) also made this point in discussing the existing literature on contextual factors in RS and health. Typically, most measures of context are used with the assumption that they will not change over time and that broad concepts, such as being religious, African American, or elderly, validly capture all persons assigned that label. Clearly, individual differences among any social, cultural, or ethnic grouping exist and may moderate any RS factor and health relation for better or for worse. Results and Lessons From Experimental Studies Unfortunately, relatively few experimental studies of RS factors and health exist. Worthington, Kurusu, McCullough, and Sandage (49) reviewed almost 150 studies that focused on RS
8
Thoresen and Harris
factors, counseling, and mental health. Roughly, only 7% involved experimental (vs. correlational or descriptive) designs. Experimental studies have been conducted primarily in the areas of meditation, almost always conducted within a secular rather than religious framework (e.g., 50), and prayer, mostly intercessory prayer in which one or more persons pray for the recovery of someone suffering from a serious chronic disease or disorder. Here, we comment on experimental studies examining the health-related effects of intercessory prayer. For a review of other RS-related health intervention studies, such as those examining meditation, religiously framed cognitive behavioral therapy, 12-step fellowship, and forgiveness interventions, the reader is referred to Harris, Thoresen, McCullough, and Larson (51) and Koenig, McCullough, and Larson (18). Intercessory prayer and distant healing. In recent years, the effects of praying for others, often people unknown personally to those praying for them and living at a distance, have been studied with the use of randomized experimental designs. The results of these studies are intriguing and worthy of careful consideration. Here, we can only touch the surface (see 52). Byrd (53) and Sicher, Targ, Moore, and Smith (54) reported double blind studies on the effect of intercessory prayer on mortality and morbidity outcomes. For example, among patients recovering from acute myocardial infarction, Byrd found that patients in the prayer condition did substantially better than control patients on a number of health-related outcome categories, such as having 7% fewer antibiotics required at discharge (p < .005) and 6% less need for intubations (p < .002). In addition, they had 6% less pulmonary edema (p < .03), 6% less congestive heart failure (p < .03), and 5% less cardiopulmonary arrest (p < .02), although these differences were less significant when adjusted at the experiment-wide p < .05 level. In a replication of the Byrd study (53), published in Archives of Internal Medicine and involving 990 patients, Harris et al. (55) found that the prayed-for group in the coronary care unit (CCU) had significantly lower (11%) CCU Overall Course Scores than those with usual care. Course Score is an index of several major in-hospital procedures and outcomes, ranging from need for specific medications to bypass surgery, reinfarction, and death. However, the length and number of hospital stays did not differ significantly. It is important that researchers in this study controlled for response expectancy effects (56), which are often very powerful, by obtaining permission from their institutional review board not to inform anyone, including patients, about the prayer intervention. Thus, the attending physicians, nurses, and patients themselves all remained uninformed about the study. Results such as these deserve attention. Although they do not shed light on how intercessory prayer works, they clearly provide evidence that the effects of prayer can be studied with empirical methods and can include objectively measurable and clinically important health outcomes. Not all intercessory studies, however, have demonstrated significant effects when experimental designs were used. Needed at this point is replication of such effects by other researchers using very similar procedures
Annals of Behavioral Medicine and examining a broader range of person, health, and sociodemographic factors. STATE OF THE SCIENTIFIC EVIDENCE: THE OBSSR/NIH PANEL In 1999, the OBSSR in the NIH created an expert panel of social and behavioral scientists under the leadership of Norman Anderson and William R. Miller to report on the state of the science concerning RS factors and health. Here, we comment briefly on some current issues and concerns raised in the panel’s work. (The panel’s full report, composed of several articles, will be available in several months as a special issue of American Psychologist.) The panel is organized into several working groups, covering areas such as physical health, measurement, possible psychosocial mediators, neurobiological pathways, and contextual factors. Of the many issues raised by the panel, a primary concern remains the limited quality of available studies in terms of research designs and assessments. Some of these issues have already been voiced in this article and by several others (4,8,19). For example, although cross-sectional studies can provide useful data in some areas, especially in the early stages of inquiry, prospective or longitudinal studies are essential to understand what changes over time. Cross-sectional snapshots are no substitute for real-time films. One’s religious affiliation, such as Christianity or Judaism, may remain stable, but many specific factors associated with spirituality and religiousness may vary with circumstances and contexts over time. A related issue cited by the panel is the absence, with rare exception, of experimental research in this area. Exceptions to date have been, as noted, in the area of intercessory prayer and meditation. Understanding RS factors more fully will require conducting more experimental studies that probe, for example, what factors when altered influence or mediate RS and health factors. Do RS factors, such as religious coping or beliefs about the nature of God, change as a result of a psychosocial intervention? If so, does the magnitude of a change in a RS factor relate to health changes? What conclusions will the panel likely reach? Although not yet finalized, we suspect a modestly positive relation between RS factors and health will be reported, along with a host of caveats and concerns. The evidence to date will probably support the view that the association between some RS factors, especially frequent attendance at services, and health is neither trivial nor a random artifact. That is, the relation is not readily explained by several other known health-related factors, such as health behaviors (e.g., smoking), perceived social support or social networks, and various demographic factors. However, this relation could prove to be less positive when more sensitive and robust measures of various psychosocial and contextual variables, such as personality and specific dimension of spirituality or religiousness, are used. RESEARCH ISSUES AND RECOMMENDATIONS Several conceptual, methodological, and analytic issues are relevant to the improvement of RS factor research. Here, we build on what has been presented previously and elsewhere, briefly discussing some important issues and next steps in the
Volume 24, Number 1, 2002 clarification of RS factor–health relations, including the possible mechanisms that underlie these relations. The Value of a Behavior–Belief–Motivation– Experience Framework Why does weekly religious service attendance predict less overall mortality in well-controlled studies (42)? How do we explain, for example, that service to others (“selfless service,” or volunteering in the community) predicts less mortality even when conventional risk factors, including social support, are controlled (57)? Answers to such questions based on empirical data require more detailed information regarding what persons involved in such studies are doing, thinking, and feeling (12). Here are two examples that illustrate recent efforts to move away from large macro-level concepts (e.g., religious affiliation, denomination) toward the kind of specificity needed to unravel these relations. Pargament et al. (58) asked if certain kinds of religious coping might be associated with mental health and religious outcomes. Religious outcomes consisted of three items assessing the individual’s perceived changes in closeness to God, spiritual growth, and closeness that occurred as a result of coping with the impending surgery of a family member. They found that among family members waiting for a relative to undergo major surgery, those who used a “collaborating with God” style of religious coping, compared with a more self-directing or pleading-with-God style, had better coping outcomes and better religious outcomes. They also found that this collaborative coping style appeared to mediate the effects of depression and anxiety on psychological and religious outcomes, whereas other religious and secular kinds of coping (e.g., planning, instrumental social support) failed to do so. Keefe et al. (59) examined RS factors among patients with arthritis and explored how they were related to daily positive and negative emotions (moods) as well as to experienced pain. They used an unusual research design in which each participant completed daily ratings of RS factors, mood and emotional states, social and emotional support, and pain level over 30 consecutive days. RS factors were studied in considerable detail in terms of daily spiritual experiences (e.g., “feeling deep inner peace or harmony”), RS coping (e.g., “looked to God for strength, support, and guidance”), RS efficacy (e.g., “extent my religious or spiritual coping allowed me to control my pain today”), and perceived salience of religion each day to pain management. The combination of daily spiritual experiences with perceived social support predicted the greatest increases in positive mood and the largest reductions in negative mood. Significantly, RS factors were clearly related to daily mood (positive and negative) and to RS efficacy. When viewed independently, these factors may be misrepresented in how they indirectly exert influence on the experiences of people, such as pain. Through the use of a research design that revealed not only between-person differences but also within-person variability (in this case, from day to day over 30 days), a more complete and informative picture was produced. Interestingly, pain was at its lowest on those days when religious coping efficacy and daily spiritual experiences were higher,
Spirituality and Health
9
even though daily spiritual experiences by themselves were not directly related to pain. Perhaps the changes in positive and negative emotions along with perceived RS efficacy were the active ingredients that led to reduced pain. These two studies illustrate the value of working with more specific concepts, such as particular behaviors and beliefs, in more intensive ways than have been assessed in the past, to yield more useful data in terms of what may account for RS and health relations. Health Hazards of RS Factors? Are RS involvements or experiences associated with any undesirable physical or mental health outcomes? Although the preponderance of studies to date have reported health-enhancing findings (18,42), some researchers have found negative evidence (e.g., 60–62). For example, Exline, Yali, and Lobel (63) found that among believers of God (who were often religiously active), those who were unforgiving of God for some hurt or offense suffered more anxiety and depression. Galanter (64) reported serious mental health problems among those from a charismatic religious sect (“Moonies”). Pargament, Smith, Koenig, and Perez (65) looked at “negative religious coping” (e.g., expecting God to solve one’s problems) and found it to be associated with greater stress, depression, and suicidality. Under some conditions, it seems likely that certain RS factors, as with social support or particular medications, may indeed be hazardous to health and well-being. Certain persons, for example, may experience serious physical and mental health problems that are associated with particular religious beliefs and practices. This may be especially true for beliefs and practices that are presented within a strict authoritarian framework, one that is intolerant if not hostile to any other perspective, religious or otherwise. Such frameworks may employ coercive and harsh forms of punishment, including social ridicule and shaming, and may foster pervasive feelings of guilt for any deviation or failure to comply. Booth (66) noted anecdotally that religious beliefs and practices, unfortunately, can take on serious addictive qualities, with persons becoming excessively anxious and dependent, which leads to a variety of health problems. Unfortunately, a clear and comprehensive picture based on well-controlled studies of the possible health-endangering correlates of RS factors is not yet available. Such a portrait will need to be provided with careful and sensitive attention to the many factors within and outside of RS factors that could explain possible health hazards and risks. Implications for Professional Practice Currently, with rare exception, health professionals receive little if any education in the possible role of RS factors in health and disease (67). This situation may be changing because over 50 medical schools are currently offering elective courses on religion and spirituality for medical students (see 68), and several nursing training programs offer some RS-related training, often centered around hospice care. This lack may be especially true in psychological training at any level (13,69). What training, if any, should psychologists and other related health professionals
10
Thoresen and Harris
receive? As noted, Sloan, Bagiella, and Powell (9) spoke out against including RS-related preparation for physicians. We believe, however, that the topic deserves at least some attention in professional training, given the evidence currently available. Others have written extensively on this topic, primarily on grounds that one’s RS beliefs and practices can prove to be a potent factor in a person’s lifestyle, influencing important choices, strivings, attitudes, and values, which in turn can influence health and disease risk (69,70). At a minimum, we argue that health-related psychologists need to be at least introduced to the evidence linking RS-related factors to particular health outcomes, such as all-cause mortality. Furthermore, all should be informed of the prevalence of RS factors in the population, such as national and community survey data on RS beliefs and practices in the general population. This seems especially pertinent given the growing sensitivity in health care to multicultural issues in which RS factors often play an important role (60). Also, the discrepancy between health professionals and the general public in terms of their RS involvement, beliefs, and practices appears often to be substantial (67). As with several other issues related to RS factors, those of professional involvement need to be carefully considered and clarified. Through What Mechanisms Might RS Factors Influence Health? Different models have tried to explain the relation between RS factors and health outcomes. Four points deserve comment. First, several factors may be involved in mediating or moderating the pathways possibly connecting RS factors with health. The examination of univariate associations, or even single multivariate pathways, may be too simplistic. Second, any particular study would likely focus on only a slice of this RS factor–health conceptual pie (e.g., looking at the relation of a spiritual factor with selected health behaviors and physiological indicators). Third, some RS factors, as noted, under certain conditions could diminish or endanger health. It is unlikely that all RS factors would benefit health under all conditions for all people. Finally, any model at this point is highly likely to be flawed in several ways, given limitations in our current knowledge. Any conceptual model is at best a work in progress, subject to continual revision. For example, Powell, Shahabi, and Thoresen (17) proposed a pathways model involving 10 major constructs, each with more specific subconstructs, including sociodemographic, person, cultural, religious–spiritual, and neurological and neuroendocrine processes leading to healing processes and, finally, to health as the end outcome. Unfortunately, empirical studies in this area to date have not been designed to test conceptual models or theory but have focused on whether relations exist between a RS factor and some health outcome The Importance of Linking Conceptualization and Measurement Religion and spirituality are complex, multidimensional constructs (71). In the design of studies and the use of relevant measurement tools, it is important to specify which features of
Annals of Behavioral Medicine the constructs are being measured and to specify a theoretical rationale for doing so. Ellison and Levin (72) made a useful theoretical distinction between the functional and behavioral (or structural) features of religious involvement. They argued that it is the identification and measurement of the possible functional roles of religion, such as the provision of an existential framework of life meaning and purpose, specific coping strategies, or support for specific health behaviors, that will pay dividends in terms of the understanding of the mechanisms through which health is influenced. Most empirical studies have focused on behavioral or structural aspects of religious involvement, such as church attendance or rituals, and have not examined the functions that religious involvement may serve in people’s lives. Pargament et al. (65) offered an example of a functional approach by examining different types of religious coping. Other Research Design Issues Research questions should dictate, as much as possible, the use of research designs. As our questions have become more sophisticated and precise, the need has developed to employ a greater variety of study designs. Other designs and methods, such as single-participant (N = 1) experiments (73); a variety of interview designs, including qualitative methods (see 74); daily monitoring methods (75); and controlled intervention studies have been largely missing. Particularly for health professionals interested in applying culturally sensitive, empirically validated treatments, intervention research offers the double benefit of the simultaneous development of effective treatment strategies and opportunities to test theoretical propositions concerning underlying mechanisms. CONJECTURES AND CONCERNS Conclusions about the emerging area of RS factors and health need to be highly tentative and stated in the spirit of reasoned conjectures based on suggestive evidence. The following conclusions seem justified at this point in time. We recognize, however, that some readers may find them overly cautious, whereas others may see them as excessively enthusiastic. We hope that they appear reasonable enough to encourage colleagues to take questions concerning the associations of RS factors and health seriously:
• A large number of Americans currently profess a belief in God (96%), attend religious services regularly (42%), consider their RS beliefs as very important in their lives (67%), associate frequent religious involvement with greater happiness (47%), and express the need for greater spiritual growth (82%). Given such data, RS factors deserve careful and critical consideration by health care professionals in research and training and in conventional and alternative or complementary health care practice (76,77). Keep in mind, however, that marked differences exist in various parts of the United States concerning most RS factors. For instance, whereas 42% of Americans nationwide attend services regularly, that figure drops to less than 20% in the San Francisco area and is over 60% in parts of the Southeast (78).
Volume 24, Number 1, 2002
• RS factors appear to be associated with physical and overall health, but the relation may be far more complex and modest than some contend. Which specific RS factors enhance or endanger health remains unclear. • Few studies have explored how spiritual factors may differ from religious factors with respect to health outcomes. Clarifying this distinction is a high priority because confusion about these terms obscures understanding and impedes research. • Evidence linking frequent attendance at religious services to reduced all-cause mortality has been impressive, but the nature of and explanation for this relation remains unclear. In general, evidence from well-controlled prospective studies linking religious attendance to specific major diseases, such as various cancers (vs. all-cause mortality), remains insufficient. • Missing in almost all major studies has been a more careful examination of person and psychological factors (except perhaps for social support, smoking, and perceived general health). Topics such as agentic and self-evaluative processes, including self- and collective efficacy beliefs, value orientations and personal strivings, and chronic emotional states ( positive as well as negative emotions) in particular social situations deserve study. Such factors may prove to be powerful moderating or mediating variables in any RS factor and health relation. • The need to use a much greater variety of research designs and assessment strategies and to combine more qualitative with quantitative methods seems imperative. Problems and questions need to dictate the selection of particular research designs and methods, not the other way around. • Several RS-related factors that may benefit health or reduce disease risk, when engaged in as part of one’s spiritual and religious orientation (rather than as secular activities), have yet to be studied. These include such topics as volunteering to help others, forgiveness, hope, and meditation. For example, over 12 forgiveness intervention studies have all used a secular perspective (80). What are, for example, the health advantages, if any, of framing a forgiveness intervention (or hope, meditation, etc.) within a spiritual or religious orientation compared to a secular one? Perhaps the new millennium represents a fitting occasion for scholars and practicing professionals concerned with health to reconsider their perspectives on the role of RS factors in health. Very few, as noted, have received professional training on this topic. Yet it remains a major concern for many of those we serve and study. As with some issues of Church and State, matters of religion and spirituality often elicit strong involuntary, knee jerk reactions. We believe, however, that avoidance is not productive and that the time has come to address this topic candidly, with solid rigor and sensitive respect. Advocacy has its place as does skepticism, but extremes of either seldom clarify the complexities. Too often, topics involving RS issues and health have fallen prey to reductionistic concepts, dichotomous thinking, and stereotyped images. The issues are indeed complex and challenging, but they are not insurmountable. We face the prospect, at least potentially, of reaping significant improvements in health care effectiveness, quality of life, and well-being if we proceed with sensitivity, pa-
Spirituality and Health
11
tience, and perseverance. Einstein (80) may have been right when he noted that science without religion is lame, whereas religion without science is blind. Perhaps the same may also be true for spirituality and health. REFERENCES Portions of this article are from Thoresen CE, Harris AHS, Oman D: Spirituality, religion, and health: Evidence, issues, and concern. In Plante TG and Sherman AC (eds), Faith and Healing: Psychological Perspectives. New York: Guilford, 2001, 15–52. Reprinted with permission from Guilford. (1) Rosen G: A History of Public Health. New York: Johns Hopkins University Press, 1993. (2) Frank J: The faith that heals. Johns Hopkins Medical Journal. 1975, 28:306–319. (3) Thoresen CE: Spirituality, health, and science: The coming revival? In Roth-Roemer S, Robinson SK, Carmin C (eds), The Emerging Role of Counseling Psychology in Health Care. New York: Norton, 1998, 409–431. (4) Larson DB, Sawyers JP, McCullough ME (eds): Scientific Research on Spirituality and Health: A Consensus Report. Rockville, MD: National Institute for Healthcare Research, 1998. (5) Dawkins R: You can’t have it both ways: Irreconcilable differences. Skeptical Inquirer. 1999, 23:62–64. (6) Koenig HG: Is Religion Good for Your Health? Binghamton, NY: Haworth Pastoral, 1997. (7) Benson H: Timeless Healing: The Power and Biology of Belief. New York: Scribner’s, 1996. (8) Sloan RP, Bagiella E, VandeCreek L, et al.: Should physicians prescribe religious activities? New England Journal of Medicine. 2000, 342:1913–1916. (9) Sloan RP, Bagiella E, Powell T: Religion, spirituality, and medicine. Lancet. 1999, 353:664–667. (10) Post SG, Puchalski CM, Larson DB: Physicians and patient spirituality: Professional boundaries, competency, and ethics. Annals of Internal Medicine. 2000, 132:578–583. (11) Shafranske EP, Malony HN: Religion and the clinical practice of psychology: A case for inclusion. In Shafranske EP (ed), Religion and the Clinical Practice of Psychology. Washington, DC: American Psychological Association, 1996, 561–586. (12) Miller WR, Thoresen CE: Spirituality, religion, and health: An emerging research field. American Psychologist (in press). (13) Miller WR, Thoresen CE: Spirituality and health. In Miller WR (ed), Integrating Spirituality Into Treatment: Resources for Practitioners. Washington, DC: American Psychological Association, 1999, 3–18. (14) Levin JS: Religion and health: Is there an association, is it valid, and is it causal? Social Science and Medicine. 1994, 38:1475–1482. (15) Levin JS, Chatters LM: Research on religion and mental health: An overview of empirical findings and theoretical issues. In Koenig HG (ed), Handbook of Religion and Mental Health. New York: Academic, 1998, 34–47. (16) George LK, Larson DB, Koenig HG, McCullough ME: Spirituality and health: What we know, what we need to know. Journal of Social and Clinical Psychology. 2000, 19:102–116. (17) Powell LH, Shahabi L, Thoresen CE: Religion and spirituality: Linkages to physical health. American Psychologist (in press).
12
Thoresen and Harris
(18) Koenig HG, McCullough M, Larson DB: Religion and Health: A Century of Research Reviewed. New York: Oxford University Press, 2000. (19) Thoresen CE, Harris AHS: Spirituality, religion, and health. In Raczynski JM, Leviton LL, Bradley L (eds), Handbook of Health Psychology (Vol. 2). Washington, DC: American Psychological Association (in press). (20) James W: Varieties of Religious Experience. New York: Random House, 1902. (21) Simpson J, Weiner E (eds), The Oxford English Dictionary (2nd Ed.). New York: Oxford University Press, 1989. (22) Pargament KI: Psychology of religion and spirituality. International Journal for the Psychology of Religion. 1999, 9:3–16. (23) Thoresen CE: Spirituality and health: Is there a relationship? Journal of Health Psychology. 1999, 4:291–300. (24) Walsh R, Vaughn F: Paths Beyond Ego: The Transpersonal Vision. New York: Plenum, 1995. (25) Pargament KI: The Psychology of Religion and Coping: Theory, Research, and Practice. New York: Guilford, 1997. (26) Woods TE, Ironson GH: Religion and spirituality in the face of illness: How cancer, cardiac, and HIV patients describe their spirituality and religion. Journal of Health Psychology. 1999, 4:393–412. (27) Shahabi L, Powell L, Musick MA, et al.: Correlates of self-perceptions of spirituality in American adults. Annals of Behavioral Medicine. 2002, 24:59–68. (28) Matthews DA, Koenig HG, Thoresen CE, Friedman, R: Physical health. In Larson DB, Swyers JP, McCullough ME (eds), Scientific Research on Spirituality and Health: A Consensus Report. Rockville, MD: National Institute for Healthcare Research, 1998, 31–54. (29) Comstock GW, Partridge KB: Church attendance and health. Journal of Chronic Diseases. 1972, 25:665–672. (30) Larson DB, Koenig HG, Kaplan BH: The impact of religion on men’s blood pressure. Journal of Religion and Health. 1989, 28:265–278. (31) Madalie JH, Kahn HA, Neufeld HN: Five-year myocardial infarction incidence: II. Association of single variables to age and birthplace. Journal of Chronic Disease. 1973, 26:329–349. (32) Harris RC, Dew MA, Lee A, et al.: The role of religion in heart-transplant recipients’ long-term health and well being. Journal of Religion and Health. 1995, 34:17–32. (33) Yates JW, Chalmer BJ, St. James P: Religion in patients with advanced cancer. Medical and Pediatric Oncology. 1981, 9:121–128. (34) Frankel BG, Hewitt WE: Religion and well-being among Canadian university students: The role of faith groups on campus. Journal for the Scientific Study of Religion. 1994, 33:62–73. (35) Idler EL, Kasl SV: Religion among disabled and nondisabled persons: II. Attendance at religious services as a predictor of the course of disability. Journals of Gerontology: Series B: Psychological Sciences and Social Sciences. 1997, 52:S306–S316. (36) Idler EL, Kasl SV: Religion among disabled and nondisabled persons: I. Cross-sectional patterns in health practices, social activities, and well being. Journals of Gerontology: Series B: Psychological Sciences and Social Sciences. 1997, 52:S294–S305. (37) Idler EL, Kasl SV: Religion, disability, depression, and the timing of death. American Journal of Sociology. 1992, 97:1052–1079.
Annals of Behavioral Medicine (38) Hummer RA, Rogers RG, Nam CB, Ellison CG: Religious involvement and U.S. adult morality. Demography. 1999, 36:272–285. (39) Musick MA, House JS, Williams DR: Attendance at religious services and mortality in a national sample. Annual Meeting of the American Sociological Association. Chicago: 1999. (40) Oman D, Reed D: Religion and mortality among the community-dwelling elderly. American Journal of Public Health. 1998, 88:1469–1475. (41) Strawbridge WJ, Cohen RD, Shema SJ, Kaplan GA: Frequent attendance at religious services and mortality over 28 years. American Journal of Public Health. 1997, 87:957–961. (42) McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoresen C: Religious involvement and mortality: A meta-analytic review. Health Psychology. 2000, 19:211–222. (43) Strawbridge W, Shema S, Cohen RD, Kaplan G: Religious attendance increases survival by improving and maintaining good health behaviors mental health and social relationships. Annals of Behavioral Medicine. 2001, 23:68–74. (44) Sloan RP, Bagiella E: Religion and health [Letter to the editor]. Health Psychology. 2001, 20:228. (45) McCullough ME, Hoyt WT, Larson DB: Small, robust and important: Reply to Sloan and Bagiella [Letter to the editor]. Health Psychology. 2001, 20:228–229. (46) Ellison CG, Levin JS: The religion–health connection: Evidence, theory and future directions. Health Education and Behavior. 1998, 25:700–720. (47) Thoresen CE, Harris AHS, Oman D: Spirituality, religion, and health: Evidence, issues, and concerns. In Plante TG, Sherman AC (eds), Faith and Health. New York: Guilford, 15–52. (48) Chatters LM: Religion and health: Public health research and practice. Annual Review of Public Health. 2000, 21:335–367. (49) Worthington Jr. EL, Kurusu TA, McCullough ME, Sandage SJ: Empirical research on religion and psychotherapeutic processes and outcomes: A 10-year review and research prospectus. Psychological Bulletin. 1996, 119:448–487. (50) Kabat-Zinn J, Wheeler E, Light T, et al.: Influence of a mindfulness meditation-based stress reduction intervention on rates of skin clearing in patients with moderate to severe psoriasis undergoing phototherapy (UVB) and photochemotherapy (PUVA). Psychosomatic Medicine. 1998, 60:625–632. (51) Harris AHS, Thoresen CE, McCullough ME, Larson DB: Spiritually and religiously-oriented health interventions. Journal of Health Psychology. 1999, 4:413–434. (52) Dossey L: Prayer and medical science: A commentary on the prayer study by Harris et al and a response to critics. Archives of Internal Medicine. 2000, 160:1735–1737. (53) Byrd RB: Positive therapeutic effects of intercessory prayer in a coronary care unit population. Southern Medical Journal. 1988, 81:826–829. (54) Sicher F, Targ E, Moore D, Smith H: A randomized double-blind study of the effect of distant healing in an advanced AIDS population. Western Journal of Medicine. 1998, 169:356–363. (55) Harris WS, Gowda M, Kolb JW, et al.: A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Archives of Internal Medicine. 1999, 159:2273–2278. (56) Kirsch I, Lynn SJ: Automaticity in clinical psychology. American Psychologist. 1999, 54:504–515. (57) Oman D, Thoresen CE, McMahon K: Volunteerism and mortality. Journal of Health Psychology. 1999, 4:301–316.
Volume 24, Number 1, 2002 (58) Pargament KI, VandeCreek L, Cole B, et al.: The vigil: Religion and the search for control in the hospital waiting room. Journal of Health Psychology. 1999, 4:327–342. (59) Keefe FJ, Affleck G, Lefebvre J, et al.: Living with rheumatoid arthritis: The role of daily spirituality and daily religious and spiritual coping. Pain (in press). (60) Bergin AE: Religiosity and mental health: A critical reevaluation and meta-analysis. Professional Psychology: Research and Practice. 1983, 14:170–184. (61) Gartner J, Larson DB, Allen GD: Religious commitment and mental health: A review of the empirical literature. Spirituality: Perspectives in theory and research. Journal of Psychology and Theology. 1991, 19:6–25. (62) Asser S, Swan R: Child fatalities from religion-motivated medical neglect. Pediatrics. 1998, 101:625–629. (63) Exline JJ, Yali AM, Lobel M: When God disappoints: Difficulty forgiving God and its role in negative emotion. Journal of Health Psychology. 1999, 4:365–380. (64) Galanter M: “Moonies” get married: A psychiatric follow-up study of a charismatic religious sect. American Journal of Psychiatry. 1986, 143:1245–1249. (65) Pargament KI, Smith BW, Koenig HG, Perez L: Patterns of positive and negative religious coping with major life stressors. Journal for the Scientific Study of Religion. 1998, 37:710–724. (66) Booth L: When God Becomes a Drug: Breaking the Chains of Religious Addiction and Abuse. Los Angeles: JP Tarcher, 1992. (67) Miller WR: Diversity training in spiritual and religious issues. In Miller WR (ed), Integrating Spirituality Into Treatment: Resources for Practitioners. Washington, DC: American Psychological Association, 1999, 253–264. (68) Puchalski CM, Larson DB: Developing curricula in spirituality and medicine. Academic Medicine. 1998, 73:970–974.
Spirituality and Health
13
(69) Plante TG, Sherman AC (eds): Faith and Health. New York: Guilford, 2001. (70) Miller WR (ed): Integrating Spirituality Into Treatment. Washington, DC: American Psychological Association, 1999. (71) Miller WR, Thoresen CE: Spirituality, religion, and health: Definitions and constructs [Special issue]. American Psychologist (in press). (72) Ellison CG, Levin JS: The religion–health connection: Evidence, theory, and future directions. Health Education and Behavior. 1998, 25:700–720. (73) Hilliard RB: Single-case methodology in psychotherapy process and outcome research. Journal of Consulting and Clinical Psychology. 1993, 61:373–380. (74) Richard TA, Folkman S: Spiritual aspects of loss at the time of a partner’s death from AIDS. Death Studies. 1997, 21:527–552. (75) Keefe FJ, Affleck G, Lefebvre JC, et al.: Pain coping strategies and coping efficacy in rheumatoid arthritis: A daily process analysis. Pain. 1997, 69:35–42. (76) Gallup G: The Gallup Poll: Public opinion 1995. Wilmington, DE: Scholarly Resources, 1995. (77) Myers D: Spiritual Hunger in a Land of Plenty. New Haven, CT: Yale University Press, 2000. (78) Putnam R: Social Capital Community Benchmark Survey, 2001. Retrieved from http://www.cfsv.org/communitysurvey. (79) McCullough ME, Pargament KI, Thoresen CE: The psychology of forgiveness: History, conceptual issues, and overview. In McCullough ME, Pargament KI, Thoresen CE (eds), Forgiveness: Theory, research, and practice. New York: Guilford, 2000, 1–16. (80) Calaprice, A. The Quotable Einstein. Princeton NJ: Princeton University Press, 1996.
Claims About Religious Involvement and Health Outcomes Richard P. Sloan, Ph.D. Columbia-Presbyterian Medical Center; Department of Psychiatry, Columbia University; and New York State Psychiatric Institute
Emilia Bagiella, Ph.D. Mailman School of Public Health Columbia University
These studies, it has been suggested, have demonstrated the beneficial impact of religious activities on those who practice them. Corresponding to the rise in these beliefs have been increasing calls to incorporate religious and spiritual activities into clinical practice. For example, in December 2000, the Harvard Medical School Department of Continuing Education offered a program on spirituality and healing entitled “Mainstreaming Spirituality.” Harris et al. recommended exploration of the potential of prayer as an adjunct to standard medical care (4). Nearly 30 U.S. medical schools now include courses on religion, spirituality, and health for medical students (5). According to a recent article in USA Today, a Denver-based HMO offers spiritual counseling. Numerous authors have claimed an abundance of literature supporting the positive health effects or religious involvement. Koenig reported that there are more than 850 articles on religious involvement and mental health, with more than two thirds showing an advantage to the religiously active, and more than 350 articles on religious involvement and physical health, with over half showing an advantage to the religious (2). An earlier claim asserted the existence of 325 studies in the area, of which over 75% showed benefits of religious involvement (6). Luskin reported that “almost all studies that evaluate the effect of religious experience show a positive health value” (7). Puchalski wrote that “a number of studies show that having spiritual beliefs is beneficial to patients, particularly those with serious illness” (8). Ellison and Levin asserted that “a substantial body of literature reports what appear to be generally desirable effects of other aspects of religious involvement (e.g., frequency of attendance, subjective religiosity) on a wide range of health outcomes” (1). In this article, we examine two elements of these assertions: (a) there are a substantial number of articles that address the relation between religious involvement and health and (b) of those articles relevant to this issue, a substantial fraction demonstrate positive associations between religious involvement and beneficial health outcomes. For the former, we describe how studies about religion and health nevertheless may be about topics unrelated to a relation between religious involvement and health. For the latter, we selected two recent literature reviews of this relation and examined the data cited in these reviews to determine if they justified claims of beneficial associations.
ABSTRACT Claims about religion, spirituality, and health have recently appeared with increasing frequency, in both the popular media and professional journals. These claims have asserted that there are a great many studies in the literature that have examined relations between religious involvement and health outcomes and that the majority of them have shown that religious people are healthier. We examined the validity of these claims in two ways: (a) To determine the percentage of articles in the literature that were potentially relevant to such a claim, we identified all English-language articles with published abstracts identified by a Medline search using the search term religion in the year 2000, and (b) to examine the quality of the data in articles cited as providing support for such a claim, we examined all articles in the area of cardiovascular disease and hypertension cited by two comprehensive reviews of the literature. Of the 266 articles published in the year 2000 and identified by the Medline search, only 17% were relevant to claims of health benefits associated with religious involvement. About half of the articles cited in the comprehensive reviews were irrelevant to these claims. Of those that actually were relevant, many either had significant methodological flaws or were misrepresented, leaving only a few articles that could truly be described as demonstrating beneficial effects of religious involvement. We conclude that there is little empirical basis for assertions that religious involvement or activity is associated with beneficial health outcomes. (Ann Behav Med
2002, 24(1):14–21)
INTRODUCTION Claims about religion, spirituality, and health have recently appeared with increasing frequency, in both the popular media and professional journals. This trend is based in part on evidence that patients want to consider broad alternatives to conventional medical practice and on published studies linking religion and health outcomes. Regarding the latter, it has been asserted frequently that there have been a great many studies in the literature that have examined relations between religious involvement and health outcomes and that the majority of them have shown that religious people are healthier (1–3). Reprint Address: R. P. Sloan, Ph.D., Columbia University, Box 427, 622 West 168th Street, New York, NY 10032. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
14
Volume 24, Number 1, 2002 STUDIES OF RELIGION AND HEALTH It is indeed true that in the literature, at least the literature available through Medline, there are numerous articles in which a variables measuring religion or religious activity and health appear together. Nonetheless, many are irrelevant to claims of a health advantage associated with religious involvement. They include studies of denominational differences, articles only remotely about religious matters, studies examining how religious factors influence medical decision making, and studies in which religious activity functions as a dependent variable. Denominational Differences Many articles about religion and health have been those about denominational differences in health. Authors of these articles have examined the differences between Protestants and Catholics, Jews and Christians, and Orthodox and secular Jews on various indexes of health and illness. For example, the incidence of myocardial infarction was greater among Protestant than Catholic men in Middlesex County, Connecticut (9). Walden, Schaefer, Lemon, Sunshine, and Wynder (10) compared serum lipid levels of Seventh-Day Adventists (SDAs) with an age-matched sample of healthy men and women from New York City. Eklund, Belchak, Lapidos, Raha-Chowdhury, and Ober recently reported on genetic polymorphisms in Hutterites (11). For two reasons, these studies of denominational differences in health have conveyed no information on the health value of religious involvement. First, most denomination studies have been conducted to take advantage of denominational differences in genetic endowment, health behaviors, or ethnicity. For example, studies of SDAs often have been conducted precisely because of behavioral codes proscribing smoking and alcohol consumption. Thus, for example, total cholesterol levels were lower across all age groups for a cohort of SDAs compared to age-matched healthy New York City men and women (10). Similarly, a study of the Amish was conducted to examine genetic contributions to congenital glaucoma (12). Second, studies of denominational differences cannot be used to demonstrate the beneficial effects of religion on health because they have always contrasted the prevalence of disease in one group to the prevalence in another. Thus, comparisons of Catholics and Protestants in Middlesex County, Connecticut, revealed an advantage to Catholic men in rates of myocardial infarction (9). However, unless it was implied that these denominational differences were associated with differences in religiosity, these studies are irrelevant to assertions of health advantages of religious involvement. Studies Only Remotely About Religion Some articles in which measures of health and religion appeared were only remotely about religious matters. For example, among the articles cited in comprehensive reviews of religion and health was one by Friedman and Hellerstein (13). In this study, the fraction of the study sample that was Jewish was measured, but this was only one of many variables collected, and the title of the article was “Occupational Stress, Law School Hierarchy, and
Claims About Involvement and Outcomes
15
Coronary Artery Diseases in Cleveland Attorneys.” As the title suggests, this was not an article about religion and health. Similarly, Comstock reported data on church attendance in an article entitled “Fatal Arteriosclerotic Heart Disease, Water Hardness at Home, and Socioeconomic Characteristics” (14). Again, church attendance was only one of many variables measured. The same was true of many of these articles. Several specific problems arise from attempting to draw conclusions from articles not explicitly about religion and health. First, because the interests of the authors were elsewhere, an examination of an association between a religious variable and a health outcome was only one of many comparisons reported, even if tests of statistical significance were conducted. This raises the problem of the failure to control for multiple comparisons. Searching through articles to find a positive statistical test for a religion–health comparison among dozens of tests conducted violates accepted standards of methodology. Similarly, because an association between a religious variable and a health outcome was not the primary focus of these articles, these analyses lacked controls for confounders and covariates. As we indicated (15), this feature is common in many studies. Even if, for example, a study showed a univariate association between church attendance and a health outcome, without a control for the important covariate of functional status, it is impossible to rule out the possibility that church attendance was a product of, rather than the cause of, poor health. Studies of Medical Decision Making Many articles addressed such issues as attitudes toward organ donation in the Sikh community (16), ethical issues in transfusions of children of Jehovah’s Witnesses (17), and religious attitudes toward organ transplantation (18). Studies such as these are relevant to an examination of the relation between religion and health only insofar as they documented the influence of religious beliefs on making difficult medical decisions. Articles on the influence of religious orientation toward health practices or procedures (e.g., transfusion and organ transplantation) and articles demonstrating the religious consequences of having medical problems have no direct bearing on the relation between religious involvement and health outcomes. Religious Activity as a Dependent Variable Others studies about religion and health examined the religious consequences of having medical problems. For example, Levin, Lyons, and Larson examined the association between subjective assessments of poorer health and frequency of prayer during pregnancy (19). In studies such as these, religious activities function as a dependent variable, the product of difficult life circumstances. Another study examined the changes in spiritual beliefs after traumatic disability (20). Although articles such as these illustrate how many people turn to religious practices under the threat of medical problems, they are irrelevant to an empirical foundation for the introduction of religious interventions in medical practice. Therefore, even if there were many studies in the literature in which both religious and health variables appeared, many had
16
Sloan and Bagiella
no relevance to putative beneficial effects of religious involvement in health outcomes. The question then becomes, What fraction of the great many articles on religion and health are relevant to this matter? Literature on Religion and Health in the Year 2000 One way to examine the fraction of articles on religion and health that are relevant to claims about the benefits of religious activities is to search the current literature. We conducted a Medline search for all articles identified by the keyword religion. We limited the search to English-language articles with abstracts in the year 2000. This search yielded 266 articles. Each of us then reviewed all of the abstracts independently and classified them according to whether they were relevant to claims of health benefits of religious activities. After reconciling differences between the two sets of reviews, we determined that only 42 of 266 (i.e., 17%) were relevant to these claims. Many studies were about denominational differences. Some were about the impact of health conditions on religious activities. Others were about physician behavior. Some described health fairs conducted in churches, and even this low number was inflated by the inclusion as relevant studies of sexual behavior and yoga. EXAMINATION OF STUDIES THAT SUPPORTED CLAIMS OF BENEFITS ASSOCIATED WITH RELIGIOUS INVOLVEMENT We examined two recent reviews of the literature (3,7). To make this task manageable, we focused on only cardiovascular disease and hypertension and evaluated them for relevance to recommendations to make religious activities adjunctive medical treatments. Luskin (7) reported that “evidence continues to mount that demonstrates the positive value of spiritual and religious factors in maintaining health” (p. 8). In the section reviewing cardiovascular disease, 12 studies were cited. One (21) was itself a review of other studies. Two cited articles were denominational studies (22,23). One was published only as an abstract (24). Generally, it is inappropriate to consider abstracts of presentations at scientific meetings because as abstracts only, sufficient information generally is unavailable for critical review. In this case, however, even the abstract made it clear that there was no control for multiple comparisons. Many studies in the literature on religion and health failed to make adjustments for the greater likelihood of finding a statistically significant result when conducting multiple statistical tests. Two other cited studies also failed to control for multiple comparisons. The famous Byrd study (25) of the impact of intercessory prayer on coronary care unit (CCU) patients has been roundly criticized (15). In this double blind study, patients in a CCU were assigned randomly either to standard care or to receive daily intercessory prayer by three to seven born-again Christians. Patients and their doctors did not know which patients were receiving prayer. Twenty-nine outcome variables
Annals of Behavioral Medicine were measured, and on 6, the prayer group had fewer newly diagnosed problems, but there was no control for multiple comparisons. Moreover, the 6 significant outcomes were not independent: The prayer group had less newly diagnosed heart failure and fewer newly prescribed diuretics and less newly diagnosed pneumonia and fewer newly prescribed antibiotics. Similarly, in an article by Koenig et al. (26), at least 126 statistical tests were conducted with no control whatsoever for multiple comparisons. A simple Bonferroni correction for multiple comparisons (i.e., dividing the standard 5% level by the number of comparisons) yielded a critical significance level of .0004. None of the comparisons in this article achieved this level of statistical significance. Luskin reported that an article by Harris et al. (27) “found that heart transplant patients who had stronger religious beliefs and religious commitment showed improved functioning, adherence to treatment, and diminished health concerns a year after surgery” (p. 10). Several problems with this study existed. For example, these findings were selected from among 42 correlation coefficients computed by the authors, again with no control for multiple comparisons. In addition, improved functioning and adherence to treatment were based solely on self-report, with no independent assessment. Another article was cited as showing that religious coping was “protective” (28), suggesting some documented benefit to patients. In fact, the primary focus of this article was about the relation between health locus of control and responses on the Helpfulness of Prayer Scale. Contrary to the hypotheses, there was no relation between these two indexes. There was nothing whatsoever in the article about protection. Two more studies (29,30) were cited as demonstrating beneficial effects of organized religious activity on blood pressure (BP). As we indicated previously (15), both these articles had significant methodological flaws. Graham et al. reported that religious attendance and importance of religion interacted such that those high in both had lower diastolic blood pressure (DBP; but not systolic blood pressure [SBP]) than those low in both. The report contained no statistics on this interaction, and although potential covariates such as socioeconomic status, age, smoking status, and body mass index were measured, the absence of the statistical model made it difficult to evaluate the findings. Based on the same data set, Larson et al. (30) reported that the frequency of church attendance was positively related to SBP (but not DBP) after controlling for Quetelet Index, smoking and socioeconomic status. The data, the multivariate model, and the amount of variance accounted for by the religious variable were not presented. Moreover, neither study contained a measure of health status, which, as many have indicated, can influence the capacity to attend church. In an article by McSherry, Ciulla, Salisbury, and Tsuang (31), patients admitted for coronary artery bypass graft surgery and acute spinal cord injury were evaluated for religiosity and followed throughout their hospital stay. Those rating themselves as moderately to highly religious had significantly shorter lengths of stay, but no analysis of the effects of confounders or covariates was conducted.
Volume 24, Number 1, 2002 Finally, an article by Oxman, Freeman, and Manheimer (32) was cited. In this article, the impact of religious activity on mortality after elective cardiac surgery was examined in 232 patients. In a multivariate analysis, one item (“strength or comfort from religion”) from a 5-item scale of religiosity was found to be associated with mortality independent of history of cardiac surgery, functional status, age, and a measure of social participation. However, neither the other items of this scale nor the composite score itself, which has been used in many other studies (e.g., 33), was related to mortality. Thus, of the 12 studies considered by Luskin to support the claim of a beneficial impact of religious activity on cardiovascular health, all had significant problems. This evidence provided no basis for such a claim. A much more comprehensive review of the literature appeared in the recently published Handbook of Religion and Health (3). The comprehensive list of studies on heart disease and hypertension (pp. 555–558) included 89 different citations. Of these, 33 were studies of denominational differences in health. An additional 11 were reviews of other studies, case reports, or mere descriptions of projects. Three were published only in abstract form and, as indicated previously, could not be critically reviewed. By the criteria of the Handbook, 8 additional studies showed no association between religious activity and health. This left 34 (38%) articles that could be the basis of claims about the direction and strength of the religion–health relation. In the comprehensive summary of studies on heart disease and hypertension, 39 of the 89 (44%) articles listed carried the indication “positive.” That is, there was at least one positive association with a better health outcome. The difference between the 39 studies rated as positive and our determination of 34 relevant to the question at hand arose because some of those we called irrelevant were rated as positive. Like the studies cited by Luskin, most had significant problems that limited their value as evidence of associations between religious activities and beneficial health outcomes. Indeed, five of the same studies (25,26,29,30,32) are cited by the Handbook. Studies With Methodological Problems Among the 34 positive studies not cited by Luskin, studies by Scotch (34) and Comstock (14) failed to control for multiple comparisons. Of course, as we pointed out previously, because neither Scotch nor Comstock were principally concerned with the health benefits of religious activities, there was no reason for them to make adjustments of alpha levels. However, reviewers using these studies to claim associations between religious involvement and health must do so. One such “positive” article, by Leserman, Stuart, Mamish, and Benson (35), reported the results of a small trial of the relaxation response on cardiac surgery patients. Postoperative supraventricular tachycardia (SVT) was lower in the treatment group than in the control group (p = .04). However, this was the only difference between the groups. They did not differ in postoperative SBP or DBP, heart rate, incidence of ventricular arrhythmias greater than Grade 3 on the Lown scale, incisional pain, incisional distress, or length of postoperative stay. If one
Claims About Involvement and Outcomes
17
corrected for multiple comparisons, the significant SVT finding would disappear. Several of the articles failed to control for confounding. The treatment of the study by Sudsaung, Chentanez, and Veluvan (36) was especially enlightening. The Handbook reported that 52 male college students were taught Buddhist meditation and were compared to 30 control students who were not. Meditation participants but not controls had lower BP at 3- and 6-week follow-ups compared to their levels at study entry. What the original article made clear, however, and the Handbook did not, was that the groups were not randomly assigned. Rather, they were self-selected, with meditation participants volunteering to be cloistered with monks for 2 months during their summer vacation, during which time they engaged in no activities other than “walking about 1 km to receive food from people in the morning” (36, p. 544). Control participants stayed at home for summer vacation. Similarly, Timio et al. (37) found that Roman Catholic nuns had lower BP than a control group but these nuns were cloistered for 20 years. Needless to say, the control group was not. Both of these studies were hopelessly confounded with respect to drawing conclusions about the health benefits of a religious practice. Comstock’s finding of an inverse association between church attendance and mortality (14) also appeared to be confounded by functional status (38). Two articles by Medalie and colleagues (39,41) were cited. Both the text of the chapter on heart disease and the summary table in chapter 34 indicated that these articles compared orthodox versus secular Jews and that the results showed an advantage to the orthodox. In fact, these articles, although reporting on data collected in Israel, had nothing to do with religion, except for a single sentence in which the authors reported, with no supporting evidence, a finding of an inverse relation between religiosity and incidence of myocardial infarction. Two other articles had inadequate or nonexistent control groups. Cooper and Aygen (41) reported on the effect of transcendental meditation (TM) on blood lipids. Only the intervention participants showed a significant decline in blood cholesterol. However, the number of participants was small (12 in the treatment group and 11 in the control group), and assignment to treatment condition was not random. Treatment participants volunteered to participate in the study while attending lectures on TM. Control participants were recruited from a medical outpatient clinic. Four of the participants originally in the treatment group deemed insufficiently active in meditation were reassigned to the control group. Another study by Blackwell et al. (42) also had no control group. In another article cited as positive (43), a multivariate model predicting coronary heart disease mortality included standard risk factors but omitted religion, and no information on risk ratio or confidence intervals or even level of statistical significance was provided. Thus, like the articles by Graham and Larson, this article lacked complete statistical information. In a study by Wenneberg et al. (44), the impact of TM on reactivity to laboratory stressors (mental arithmetic, mirror tracing, and public speaking) and ambulatory blood pressure (ABP) was examined. In this clinical trial, although participants were
18
Sloan and Bagiella
randomly assigned to a TM or health education condition, there was substantial dropout. Among those who remained in the study, no differences were found between pretreatment and posttreatment reactivity to laboratory stressors. When only high-compliance participants (in both conditions) were examined, the TM participants had higher SBP reactivity to the preparation for the speech task and to the speech task itself. Similarly, there were no treatment differences in ABP when all participants were considered. When high compliers only were analyzed, TM participants had lower posttreatment diastolic ABP than control participants. This study, even as presented, was at best equivocal with respect to an advantage to the TM group because analyses of data from all participants showed no differences between groups in the laboratory or the field. When analyses were restricted to high-compliance participants, the TM group had greater posttreatment SBP reactivity to the speech task and lower diastolic ABP. More broadly, however, restricting analyses only to high-compliance participants or to those who did not drop out violated the intention-to-treat requirement of randomized clinical trials. Finally, a study by Miller on the impact of remote healing was cited (45). This study reported that hypertensive patients who were the recipients of remote healing had a greater reduction in SBP than those in the control group. No information about the patients, except that there were 96 of them, was presented. There were eight “healers.” The interval length between preintervention and postintervention data collection was unclear. Moreover, the SBP finding was the only one that achieved statistical significance. DBP, pulse rate, weight, and health status, a poorly defined composite measure, did not change. In addition, data were presented only for the patients of four of the healers “who had the highest number of returned patients” (45, p. 486). Assuming that “returned patients” means patients who did not drop out, this means that statistical analysis was conducted on an incomplete data set. Studies That Were Misrepresented Several of the studies cited as positive by the Handbook were methodologically adequate. They failed to provide support for claims about the health benefits of religious activities because they were not fundamentally about religion or, contrary to claims, had no significant effects. For example, the three studies by Patel and colleagues described a multicomponent intervention including health education, breathing exercises, deep muscle relaxation, biofeedback, and yoga (46–48). Like the studies by Patel, the report of Burell (49) was about a multicomponent intervention delivered to patients after coronary artery bypass graft surgery. It was impossible to attribute the benefits of these programs to any one component. The weight control program described by Kumanyika and Charleston qualified as a study of religion and health only because it was conducted in a church (50). The program itself had nothing to do with religion. The articles by Scotch and by Stavig, Igra, and Leonard were studies of the impact of assimilation into new societies, either by immigration or by moving
Annals of Behavioral Medicine from rural to urban life; religious affiliation was used merely as an index of assimilation into the new society (34,51). One article was a study of denominational comparisons (52). Two studies, although about religious activities, were cited as positive when in fact they reported no significant effects. Hixson, Gruchow, and Morgan (53) was cited as a study with one or more positive associations between a religious variable and a health outcome with p < .05. In this study of religiosity, health behaviors and BP data were obtained for 112 women. There was not a single p value less than .05 in the article. In fact, the authors reported that the results of the one-way analyses of variance that they conducted examining the influence of religiosity on BP for all participants did not reveal any statistically significant findings. An article by Koenig (54) was cited in the summaries as marginally positive. In fact, this article reported no significant associations between a religious variable and a health outcome. An article by Alexander et al. (55) was included in the summary table as a positive study, although it was described as a supplementary analysis of data presented in another article already in the table (56). Listing both studies was misleading. The Handbook cited as a positive study a report by Friedlander, Kark, and Stein (57) in which the religiosity of parents was related to the blood lipids of their adolescent children in Israel. Even after adjustment for sex, ethnic origin, social class, body mass, and season of the year, total cholesterol, LDL cholesterol, and triglycerides were lowest in the children of orthodox parents and highest in those of secular parents. This study did not examine the relations between religiosity among adolescents and their risk of disease. A study by Lapane, Lasater, Allan, and Carleton (58) cited as positive by the Handbook examined two cross-sectional surveys of the Pawtucket Heart Health Program. Multivariate analyses of the differences between church members and nonmembers were carried out, but no statistical models were presented. After controls for age, sex, and Portuguese ethnicity, there was a significant advantage in cigarette smoking to church members. However, church members also were more likely to have greater body mass indexes. After adjustment, there was a small DBP advantage for church members. The authors concluded, “Overall, we found that church members were not different from non-members with respect to most CVD [cardiovascular disease] risk factors. With the exception of cigarette-smoking status, majority-church members may actually have more adverse CVD-risk–factor profiles” (58, p. 162). Two articles by Walsh were cited as independent, but they appeared to be reports from overlapping samples of participants (59,60). In the first, unspecified relations between BP and church attendance were reported to be .17 and .09, with neither achieving statistical significance (60). In the second, church attendance and importance of religion were combined in some unspecified manner to create an index of “religious commitment” and then entered into several analyses. Although it was unclear in the first article, it became clear in the second that all data were collected by Walsh himself, raising concerns about lack of blindness (59).
Volume 24, Number 1, 2002 Some studies cited as positive at best present equivocal evidence. For example, Hutchinson (61) examined the association of stress and BP in a Caribbean population. She found that men but not women who never attended church had higher BPs than frequent attenders. However, these analyses were at the univariate level, and there was no control for confounders or covariates. In another positive study, Gupta, Prakash, Gupta, and Gupta (62) found that prayer habits were associated with lower prevalence of coronary heart disease among men (p = .04) in a multivariate analysis. Prayer habit was entirely undefined, and no relation was found among women.
Unpublished Studies Three studies cited in the summary as positive were available only in abstract form (63–65). Until they are published, they are impossible to review critically. However, the abstract by Thoresen was about the Recurrent Coronary Prevention Program (RCPP), a multicomponent intervention based originally on the Type A behavior pattern literature and designed to reduce recurrent coronary events (65). Because several articles on the RCPP have been published, there is no need to rely on a 1990 abstract presented at a meeting in Sweden. None of the published reports from the RCPP indicated anything about religion.
Articles That Could Reasonably Be Called Positive The strongest study among those identified as positive was by Schneider et al. (56), who, in a randomized clinical trial, showed that both a TM and a progressive relaxation intervention significantly reduced BP in a group of African Americans, 55 years of age and older. The TM condition led to a greater BP reduction than the progressive relaxation condition. In a study, the aim of which was primarily to explore information about social integration and BP in Black Americans, Livingston et al. (66) found that church affiliation (undefined) was associated with lower SBP and DBP in both men and women in a multivariate model. With the qualification that the meaning of church affiliation was unclear and that it was used as an index of social integration, this appeared indeed to be a positive study showing an advantage to the religiously active. In a case control study by Friedlander, Kark, and Stein (67), the risk of myocardial infarction was significantly greater among secular compared to orthodox Jews in Israel, even after control for covariates. A study by Zamarra, Schneider, Besseghini, Robinson, and Salerno (68) was cited as demonstrating the benefits of TM in patients with documented coronary heart disease. In this study, 10 intervention participants showed significantly greater improvements in exercise duration, maximal workload, and delay of onset of ST-segment depression compared to 6 wait-list control participants. This study was unambiguously positive, but the number of participants was small, and because data from 2 experimental participants and 3 control participants who withdrew from the study were not included in the analysis, the intention-to-treat principle was not followed.
Claims About Involvement and Outcomes
19
CONCLUSIONS Using two different approaches, we examined the claim that the empirical literature on religion and health provides a basis for the assertion of the existence of substantial support for a beneficial effect of religious activities on health outcomes. After reviewing the abstracts from 266 articles identified by a Medline search of publications appearing in 2000 and meeting criteria for a search of religion as a major MESH term, we determined that only 17% of these studies were relevant to an association of religious activities and positive health outcomes. Next, we examined two secondary sources reviewing the empirical literature on religion and health, focusing on the literature on heart disease and hypertension. One review was dedicated to this topic (7). The second reviewed the literature on heart disease and hypertension as part of a comprehensive review of the entire literature on religion and health (3). Although both secondary sources claimed to identify empirical evidence supporting the benefits of religious activities on heart disease and hypertension, a review of the articles cited revealed that the majority were either flawed or misinterpreted. Virtually none of the 12 articles cited by Luskin could provide support for the claim that religious activity or religious involvement is associated with beneficial heart disease outcomes. Of the 39 studies on heart disease and hypertension listed as positive by the Handbook, only 4 could reasonably be claimed to support this assertion. Thus, there is little evidence to support claims that health benefits derive from religious activity. However, this analysis was limited in several potentially significant ways. With regard to the review of studies appearing in Medline, we established their relevance to claims about beneficial effects of religious involvement based only on the contents of the abstracts. Although it is difficult to evaluate the methodology or conclusions of a study from material presented only in the abstract, the relevance of a study to claims about the benefits of religious involvement generally can be ascertained. However, it is conceivable that reviewing the entire articles would lead to different conclusions. Restricting the examination of secondary sources to heart disease and hypertension is justified only if the treatment of the literature on these topics is representative of the larger literature examining religious activity as it relates to other health outcomes. We know of no studies that shed light on this matter but have no reason to believe that the literature in general is any more supportive of claims of benefits of religious involvement for other health outcomes. To conclude, it is indeed true that there are many studies in which religious variables and health outcomes appear together. It is not true, however, that most of these studies are relevant to putative health benefits deriving from religious involvement. In fact, most are irrelevant to this claim. Among those that are relevant, two general problems arise: methodological deficiencies of the studies themselves and inaccurate accounts appearing in secondary sources. Examination of these studies in the area of heart disease and hypertension revealed that there is little empirical support for claims of health
20
Sloan and Bagiella
benefits deriving from religious involvement. To suggest otherwise is inconsistent with the literature. REFERENCES (1) Ellison CG, Levin JS: The religion–health connection: Evidence, theory, and future directions. Health Education and Behavior. 1998, 25:700–720. (2) Koenig HG: Religion, spirituality, and medicine: Application to clinical practice. Journal of the American Medical Association. 2000, 284:1708. (3) Koenig HG, McCullough ME, Larson DB: Handbook of Religion and Health. New York: Oxford University Press, 2001. (4) Harris WS, Gowda M, Kolb JW, et al.: A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Archives of Internal Medicine. 1999, 159:2273–2278. (5) Levin JS, Larson DB, Puchalski CM: Religion and spirituality in medicine: Research and education. Journal of the American Medical Association. 1997, 278:792–793. (6) Matthews DA, Clark C: The Faith Factor. New York: Viking, 1998. (7) Luskin F: Review of the effect of spiritual and religious factors on mortality and morbidity with a focus on cardiovascular and pulmonary disease. Journal of Cardiopulmonary Rehabilitation. 2000, 20:8–15. (8) Puchalski CM: Spirituality and health: The art of compassionate medicine. Hospital Physician. 2001, Mar.:30–36. (9) Wardwell WI, Bahnson CB, Caron HS: Social and psychological factors in coronary heart disease. Journal of Health and Human Behavior. 1963, 4:154–165. (10) Walden RT, Schaefer LE, Lemon FR, Sunshine A, Wynder EL: Effect of environment on the serum cholesterol–triglyceride distribution among Seventh-Day Adventists. American Journal of Medicine. 1964, 36:269–276. (11) Eklund AC, Belchak MM, Lapidos K, Raha-Chowdhury R, Ober C: Polymorphisms in the HLA-linked olfactory receptor genes in the Hutterites. Human Immunology. 2000, 61:711–717. (12) Martin SN, Sutherland J, Levin AV, et al.: Molecular characterisation of congenital glaucoma in a consanguineous Canadian community: A step towards preventing glaucoma related blindness. Journal of Medical Genetics. 2000, 37:422–427. (13) Friedman EH, Hellerstein HK: Occupational stress, law school hierarchy, and coronary artery diseases in Cleveland attorneys. Psychosomatic Medicine. 1968, 30:72–86. (14) Comstock GW: Fatal arteriosclerotic heart disease, water hardness at home, and socioeconomic characteristics. American Journal of Epidemiology. 1971, 94:1–10. (15) Sloan R, Bagiella E, Powell T: Religion, spirituality, and medicine. Lancet. 1999, 353:664–667. (16) Exley C, Sim J, Reid N, Jackson S, West N: Attitudes and beliefs within the Sikh community regarding organ donation: A pilot study. Social Science and Medicine. 1996, 43:23–28. (17) Sheldon M: Ethical issues in the forced transfusion of Jehovah’s Witness children. Journal of Emergency Medicine. 1996, 14:251–257. (18) Gallagher C: Religious attitudes regarding organ donation. Journal of Transplantation Coordinators. 1996, 6:186–190. (19) Levin JS, Lyons JS, Larson DB: Prayer and health during pregnancy: Findings from the Galveston Low Birthweight Survey. Southern Medical Journal. 1993, 86:1022–1027.
Annals of Behavioral Medicine (20) McColl MA, Bickenbach J, Johnston J, et al.: Spiritual issues associated with traumatic-onset disability. Disability and Rehabilitation. 2000, 22:555–564. (21) Levin JS, Vanderpool HY: Is religion therapeutically significant for hypertension? Social Science and Medicine. 1989, 29:69–78. (22) Berkel J, de Waard F: Mortality pattern and life expectancy of Seventh-Day Adventists in the Netherlands. International Journal of Epidemiology. 1983, 12:455–459. (23) Enstrom JE: Health practices and cancer mortality among active California Mormons. Journal of the National Cancer Institute. 1989, 81:1807–1814. (24) Krucoff MW, Crater SW, Green CL, et al.: Randomized integrative therapies in interventional patients with unstable angina: The Monitoring & Actualization of Noetic Training (MANTRA) feasibility pilot. Circulation. 1998, 98:I–280. (25) Byrd RC: Positive therapeutic effects of intercessory prayer in a coronary care unit population. Southern Medical Journal. 1988, 81:826–829. (26) Koenig HG, George LK, Hays JC, et al.: The relationship between religious activities and blood pressure in older adults. International Journal of Psychiatry in Medicine. 1998, 28:189–213. (27) Harris RC, Dew MA, Lee A, et al.: The role of religion in heart-transplant recipients’ long-term health and well-being. Journal of Religion and Health. 1995, 34:17–32. (28) Saudia TL, Kinney MR, Brown KC, Young-Ward L: Health locus of control and helpfulness of prayer. Heart and Lung. 1991, 20:60–65. (29) Graham TW, Kaplan BH, Cornoni-Huntley J, et al.: Frequency of church attendance and blood pressure elevation. Journal of Behavioral Medicine. 1978, 1:37–43. (30) Larson DB, Koenig HG, Kaplan BH, et al.: The impact of religion on men’s blood pressure. Journal of Religion and Health. 1989, 28:265–278. (31) McSherry E, Ciulla M, Salisbury S, Tsuang D: Spiritual resources in older hospitalized men. Social Compass. 1987, 35:515–537. (32) Oxman TE, Freeman DH, Manheimer ED: Lack of social participation or religious strength and comfort as risk factors for death after cardiac surgery in the elderly. Psychosomatic Medicine. 1995, 57:5–15. (33) Idler EL, Kasl SV: Religion among disabled and nondisabled persons: II. Attendance at religious services as a predictor of the course of disability. Journal of Gerontology. 1997, 52B:S306–S316. (34) Scotch NA: Sociocultural factors in the epidemiology of Zulu hypertension. American Journal of Public Health. 1963, 53:1205–1213. (35) Leserman J, Stuart EM, Mamish ME, Benson H: The efficacy of the relaxation response in preparing for cardiac surgery. Behavioral Medicine. 1989, 15:111–117. (36) Sudsuang R, Chentanez V, Veluvan K: Effect of Buddhist meditation on serum cortisol and total protein levels, blood pressure, pulse rate, lung volume and reaction time. Physiology and Behavior. 1991, 50:543–548. (37) Timio M, Verdecchia P, Venanzi S, et al.: Age and blood pressure changes. A 20-year follow-up study in nuns in a secluded order. Hypertension. 1988, 12:457–461. (38) Comstock GW, Tonascia JA: Education and mortality in Washington County, Maryland. Journal of Health and Social Behavior. 1977, 18:54–61.
Volume 24, Number 1, 2002 (39) Medalie JH, Kahn HA, Neufeld HN, et al.: Myocardial infarction over a five-year period—I. Prevalence, incidence and mortality experience. Journal of Chronic Diseases. 1973, 26:63–81. (40) Medalie JH, Kahn HA, Neufeld HN, Riss E, Goldbourt U: Five-year myocardial infarction incidence—II. Association of single variables to age and birthplace. Journal of Chronic Diseases. 1973, 26:329–349. (41) Cooper MJ, Aygen MM: A relaxation technique in the management of hypercholesterolemia. Journal of Human Stress. 1979, 5:24–27. (42) Blackwell B, Bloomfield S, Gartside P, et al.: Transcendental meditation in hypertension. Individual response patterns. Lancet. 1976, 1:223–226. (43) Goldbourt U, Yaari S, Medalie JH: Factors predictive of long-term coronary heart disease mortality among 10,059 male Israeli civil servants and municipal employees. Cardiology. 1993, 82:100–121. (44) Wenneberg SR, Schneider RH, Walton KG, et al.: A controlled study of the effects of the transcendental meditation program on cardiovascular reactivity and ambulatory blood pressure. International Journal of Neuroscience. 1997, 89:15–28. (45) Miller RN: Study on the effectiveness of remote mental healing. Medical Hypotheses. 1982, 8:481–490. (46) Patel C, North WR: Randomised controlled trial of yoga and bio-feedback in management of hypertension. Lancet. 1975, 2:93–95. (47) Patel C, Carruthers M: Coronary risk factor reduction through biofeedback-aided relaxation and meditation. Journal of the Royal College of General Practitioners. 1977, 27:401–405. (48) Patel C, Datey KK: Relaxation and biofeedback techniques in the management of hypertension. Angiology. 1976, 27:106–113. (49) Burell G: Group psychotherapy in project New Life: Treatment of coronary-prone behaviors for patients who have had coronary artery bypass graft surgery. In Allan R, Scheidt S (eds), Heart and Mind. Washington DC: American Psychological Association, 1996, 291–310. (50) Kumanyika SK, Charleston JB: Lose weight and win: A church-based weight loss program for blood pressure control among black women. Patient Education and Counseling. 1992, 19:19–32. (51) Stavig GR, Igra A, Leonard AR: Hypertension among Asians and Pacific islanders in California. American Journal of Epidemiology. 1984, 119:677–691. (52) Wardwell WI, Hyman M, Bahnson CB: Socio-environmental antecedents to coronary heart disease in 87 white males. Social Science and Medicine. 1968, 2:165–183.
Claims About Involvement and Outcomes
21
(53) Hixson KA, Gruchow HW, Morgan DW: The relation between religiosity, selected health behaviors, and blood pressure among adult females. Preventive Medicine. 1998, 27:545–552. (54) Koenig HG: Religion and death anxiety in later life. Hospice Journal. 1988, 4:3–24. (55) Alexander CN, Schneider RH, Staggers F, et al.: Trial of stress reduction for hypertension in older African Americans. Hypertension. 1996, 28:228–237. (56) Schneider RH, Staggers F, Alexander CN, et al.: A randomised controlled trial of stress reduction for hypertension in older African Americans. Hypertension. 1995, 26:820–827. (57) Friedlander Y, Kark JD, Stein Y: Religious observance and plasma lipids and lipoproteins among 17-year- old Jewish residents of Jerusalem. Preventive Medicine. 1987, 16:70–79. (58) Lapane KL, Lasater TM, Allan C, Carleton RA: Religion and cardiovascular disease risk. Journal of Religion and Health. 1997, 36:155–163. (59) Walsh A: Religion and hypertension: Testing alternative explanations among immigrants. Behavioral Medicine. 1998, 24:122–130. (60) Walsh A: The prophylactic effect of religion on blood pressure levels among a sample of immigrants. Social Science and Medicine [Medical Anthropology]. 1980, 14B:59–63. (61) Hutchinson J: Association between stress and blood pressure variation in a Caribbean population. American Journal of Physical Anthropology. 1986, 71:69–79. (62) Gupta R, Prakash H, Gupta VP, Gupta KD: Prevalence and determinants of coronary heart disease in a rural population of India. Journal of Clinical Epidemiology. 1997, 50:203–209. (63) Steffen PR, Alumenthal J, Sherwood A: Religious coping, ethnicity, and ambulatory blood pressure. Psychosomatic Medicine. 2000, 62:148. (64) Merritt MM, Bennett GG, Williams RB: Religiosity enhances cardiovascular reactivity among Black males with low education. Psychosomatic Medicine. 2000, 62:139. (65) Thoresen CE: Long-term, 8-year follow-up of Recurrent Coronary Prevention Program. Meeting of the International Society of Behavioral Medicine. Uppsala, Sweden: 1990. (66) Livingston IL, Levine DM, Moore RD: Social integration and Black intraracial variation in blood pressure. Ethnicity & Disease. 1991, 1:135–149. (67) Friedlander Y, Kark JD, Stein Y: Religious orthodoxy and myocardial infarction in Jerusalem: A case control study. International Journal of Cardiology. 1986, 10:33–41. (68) Zamarra JW, Schneider RH, Besseghini I, Robinson DK, Salerno JW: Usefulness of the transcendental meditation program in the treatment of patients with coronary artery disease. American Journal of Cardiology. 1996, 77:867–870.
The Daily Spiritual Experience Scale: Development, Theoretical Description, Reliability, Exploratory Factor Analysis, and Preliminary Construct Validity Using Health-Related Data Lynn G. Underwood, Ph.D. Fetzer Institute
Jeanne A. Teresi, Ed.D., Ph.D. Hebrew Home for the Aged at Riverdale and Columbia University Stroud Center
many individuals. As far back as James’s The Varieties of Religious Experience: A Study in Human Nature (1), there has been interest in this experience of the individual from a psychological perspective. This article presents the development of a Daily Spiritual Experiences Scale (DSES), the items of which attempt to measure everyday ordinary experience rather than particular beliefs or behaviors; although developed for the predominately Judeo-Christian U.S. population, it is intended to transcend the boundaries of particular religions. Spirituality and religiousness have received increasing attention as potential health research variables. Frequent reference has been made to the body of data linking religious variables to mental and physical health outcomes (2–5). The National Institute on Alcohol Abuse and Alcoholism just funded a set of seven proposals after a request for applications to solicit research examining the relationship of spirituality to alcoholism. However, the particular aspects of religiousness and spirituality that have been examined vary across studies, which has resulted in a lack of clarity regarding the construct measured and an accompanying lack of clarity as to the implications each study has for action. The results of a recent meta-analysis of studies examining the relationship between religiousness and medical outcomes (6) have underscored the need for adequate measurement. Aligned with the concern that studies of this subject need more rigor and thoughtfulness, the National Institute on Aging and the Fetzer Institute cosponsored a meeting at the National Institutes of Health in March 1995, where participants examined conceptual and methodological issues at the interface of religion, health, and aging. After the meeting, a working group received support to develop a multidimensional approach to the measurement of religious and spiritual variables that could be used in health studies (7). This approach was particularly helpful due to the plethora of ways that exist to measure these variables, often with little or no justification given for the particular method used. Conceptual overlap was the rule, with scales that measured religious preference (8), attendance (reviewed in 9), or intrinsic versus extrinsic religiosity (10) and multidimensional sections of surveys (11). Also, scales either have tended to be solidly based in a single religious tradition (12) or, in trying to avoid that approach, have examined issues such as meaning and values from an existential perspective (13). The working group identified daily spiritual experience (DSE) as one aspect of religiousness and spirituality that had never been fully addressed despite its anecdotal importance in individuals’ lives and its potential connection to health. Underwood, in response to that recommendation, developed the
ABSTRACT Spirituality and religiousness are gaining increasing attention as health research variables. However, the particular aspects examined vary from study to study, ranging from church attendance to religious coping to meaning in life. This frequently results in a lack of clarity regarding what is being measured, the meaning of the relationships between health variables and spirituality, and implications for action. This article describes the Daily Spiritual Experience Scale (DSES) and its development, reliability, exploratory factor analyses, and preliminary construct validity. Normative data from random samples and preliminary relationships of health-related data with the DSES also are included. Detailed data for the 16-item DSES are provided from two studies; a third study provided data on a subset of 6 items, and a fourth study was done on the interrater reliability of the item subset. A 6-item version was used in the General Social Survey because of the need to shorten the measure for the survey. A rationale for the conceptual underpinnings and item selection is provided, as are suggested pathways for linkages to health and well-being. This scale addresses reported ordinary experiences of spirituality such as awe, joy that lifts one out of the mundane, and a sense of deep inner peace. Studies using the DSES may identify ways in which this element of life may influence emotion, cognition and behavior, and health or ways in which this element may be treated as an outcome in itself, a particular component of well-being. The DSES evidenced good reliability across several studies with internal consistency estimates in the .90s. Preliminary evidence showed that daily spiritual experience is related to decreased total alcohol intake, improved quality of life, and positive psychosocial status. (Ann Behav Med
2002, 24(1):22–33)
INTRODUCTION The inner experience of spiritual feelings and awareness are an integral part of the everyday religious and spiritual lives of
This research was funded, in part, by the Fetzer Institute. We thank Dr. Frank Keefe, Dr. Jeanne Zechmeister, Leila Shahabi, Dr. Lynda Powell, and Dr. Scott Tonigan for data collection and use of that data in this article, and Katja Ocepek-Welikson for data programming. We thank Wayne Ramsey and Heidi Matteo for their help with the article. Reprint Address: L. G. Underwood, Ph.D., Fetzer Institute, 9292 West KL Avenue, Kalamazoo, MI 49009. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
22
Volume 24, Number 1, 2002 DSES. After the development of the instrument, the scale was included, with the author’s permission, in a variety of studies. The data were made available to the author. The main purpose of this article is to describe the development of the DSES, including its reliability, exploratory factor analyses (EFAs), and preliminary construct validity. The relationship of DSE to health-relevant data is examined. Theoretical Orientation and Definition of DSE This scale is intended to measure a person’s perception of the transcendent (God, the divine) in daily life and his or her perception of his or her interaction with or involvement of the transcendent in life. The items attempt to measure experience rather than particular beliefs or behaviors; therefore, they are intended to transcend the boundaries of any particular religion. Many characterizations of spirituality involve such an inner dimension (14). Development of this measure began by the examination of what constitutes the substantive feelings and thoughts that describe the interface of faith with daily life. It appeared that here might lie some of the proximal connections of spirituality with health. Through reflection on the aspects of the spiritual or religious perspective that weave through thought processes and feelings in daily events, an attempt was made to develop questions that would elicit those inner qualities as they express themselves at specific moments in the midst of daily life events. The intention was to determine the extent to which spiritual feelings and inner experiences might constitute an integral part of the life of the ordinary person and, ultimately, to examine the relation of these factors to health and well-being. The decision was made at the outset to use the word spiritual rather than religious in the definition of the collection of items in this measure. Although there are different understandings of the distinction between religiousness and spirituality, the following clarifying statement has been useful in a variety of medical research settings: Religiousness has specific behavioral, social, doctrinal, and denominational characteristics because it involves a system of worship and doctrine that is shared within a group. Spirituality is concerned with the transcendent, addressing ultimate questions about life’s meaning, with the assumption that there is more to life than what we see or fully understand. Spirituality can call us beyond self to concern and compassion for others. While religions aim to foster and nourish the spiritual life—and spirituality is often a salient aspect of religious participation—it is possible to adopt the outward forms of religious worship and doctrine without having a strong relationship to the transcendent. (7, p. 2)
The spiritual, for the ordinary person, is most often and most easily described in language that has religious connotations. Religious language can be stated in such a way that it is more amenable to translation; for example, the word God, although not acceptable for some, can be interpreted by a person responding to a questionnaire to include various notions of the divine or a transcendent aspect of life, without losing its meaning to those for whom it has significance.
Daily Spiritual Experience Scale
23
In preparation for a meeting on the role of spirituality and religiousness in disability at the National Institute of Child Health and Human Development in 1994, a model was developed (15) to reflect the integration of the variety of aspects of an individual’s life. In addition to an integrative core, the model had four different dimensions of life: the vital (physical and emotional), the functional (intellectual and physical), the interpersonal (social and cultural), and the transcendent. The model examined how we influence others and they influence us; how we are shaped by our physical endowment, our environment; our emotional dispositions; and our orientation to the transcendent. Drawing from this model, the DSES assesses features that can affect physical and mental health, social and interpersonal interactions, and functional abilities. In turn, the physical and emotional can have effects on DSE, as do intellectual interpretations of meaning and belief, cultural environment and experiences, and interpersonal interactions. Recent developments in cognitive neuroscience have encouraged the adoption of such an integrative model. For example, the issue of an integrative core is compatible with a variety of neuroscientific understandings, whether the core is seen as residing within a specific neural network (16) or as not necessarily synonymous with a physical location (17). Work by Damasio (18) has shown the incorrectness of our common assumption that if only we could get the emotions out of the way, our intellect could function more clearly. People with neurological deficits in the emotional area are actually incapacitated in much decision making. The driving force for decision making is somehow dependent on an integrative activity using emotions and more rational thought. In the same way, it is very possible that the integration of the transcendent sphere also may be crucial to decision making, behaviors, and attitudes. The DSES assesses features that, in this model (15), pass through the core of the person to affect physical and mental health, social and interpersonal interactions, and functional abilities. The DSE construct represents those aspects of life that make up day-to-day spiritual experience for many people, a more direct assessment of some of the more common processes through which the larger concepts of religiousness and spirituality are involved in everyday life, grounding them in specifics. The items are designed to assess aspects of day-to-day spiritual experience for an ordinary person and should not be confused with measures of extraordinary experiences (e.g., near-death or out-of-body experiences and other more dramatic mystical experiences) that may tap something quite different and have a different relation to health outcomes. This choice was deliberate. Scales exist that measure these more extraordinary experiences (19). The experiences reflected in the DSES may be evoked by a religious context or by other events of daily life or by the individual’s religious history or religious or spiritual beliefs. Underhill (20), a British theologian of the early 1900s, referred to this kind of experience as “practical” rather than what usually is thought of as “mystical,” emphasizing the ordinariness of these experiences. The scale differs from other measures of religiousness such as religious coping (21), as it is not necessarily dealing with stressful life events. It also differs from religious motiva-
24
Underwood and Teresi
tion measures (10,22) that tap whether people are motivated by intrinsic or more socially driven religious factors and from the Spiritual Well-Being Scale (23,24), which examines existential and religious quality of life issues. Religious commitment or salience items tap the importance of religion and cognitive assessment of application of religious principles in daily life (25). An important point is that there is no assumption that the more of these daily spiritual experiences (DSEs) you have, the better you are in spiritual terms. The intent is to capture a set of experiences that may play a strong role in the lives of many; such measures, to date, have been absent from our attempts to assess what factors might play important roles in the lives of individuals and their actions, thoughts, and attitudes. Implicit in the model presented here is the assumption that there is a type of DSE that can contribute positively to health and that can be defined broadly to include spiritual, psychological, and social well-being as well as physical health. Analogously, although psychological stress has been extensively linked to health problems through specific physiologic effects, emotional and physical dispositions can buffer this stress (26). Positive emotional experiences have also been connected with positive effects on the immune system, independent of the negative effects of stress (27). Likewise, positive expectations for outcomes have been linked to positive immune effects (28,29). There may also be overlap between the endorsement of a “sense of deep peace” and the condition that leads to or emanates from direct neurologic and endocrine effects similar to those identified during meditation (30). Despite work linking church attendance with health outcomes (31,32), this association has many potential confounding and mediating factors, such as social support effects, need for reasonable health to participate in public activities, and links with behavioral dictums of religious groups. Very little empirical work has sought to link the spiritual experiences of daily life with health outcomes. One suggestive study was that of Oxman, Freeman, and Manheimer (33), in which one of the items most strongly predictive of positive health outcome in cardiovascular disease was “I obtain strength and comfort from my religion” (elements that were incorporated into the DSES). The inclusion of the DSES in health studies has the potential for the establishment of a pathway by which religiousness and spirituality might influence physical and mental health.
METHODS Development of the DSES Content validity. To begin development of the scale, Lynn G. Underwood held in-depth interviews and focus groups with individuals from many religious perspectives. This process provided basic qualitative information regarding the spiritual experiences of a wide range of individuals. She also conducted a review of scales that attempt to measure some aspects of spiritual experience (19,32,34) and drew as well on a variety of theological, spiritual, and religious writings (35–40 are representative). The writings helped to categorize experiences to develop a concise set of items.
Annals of Behavioral Medicine Refinement of the instrument involved several stages. First, Underwood elicited individual interpretations of the questions through semistructured interviews and refined the items in the light of the responses. Then, individual, open-ended interviews were conducted to confirm what the items actually meant to people responding to them. In this process, efforts were made to ground the questions in the specific whenever possible, although still keeping them broad enough to encompass a variety of perspectives and situations. Finally, she further revised the items based on a review of the instrument by representatives of a variety of spiritual orientations at a meeting of the World Health Organization Working Group on Spiritual Aspects of Quality-of-Life. This group included agnostics, atheists, Buddhists, Christians, Hindus, Jews, and Muslims. (Details of the qualitative methodology used in instrument development can be obtained from Underwood by request.) The development process can be illustrated by the consideration of how to handle the question whether to include the word God in some of the items. Initial interviews with Christians, Jews, Muslims, agnostics, and atheists indicated a variety of words used to refer to the transcendent or divine, but the dominant word used was God. Although some aspects of spiritual experience could be addressed without this word, for many specific aspects a single word for the divine was necessary. In the subsequent testing of the instrument, in which Underwood explored with the respondents what they took the items to mean, most found the word God to be easily understood and the best word for them. Those outside the Judeo-Christian orientation, including Muslims, people from indigenous religious perspectives, and agnostics, were generally comfortable with the word, being able to “translate” it into their concept of the divine. The only group for which this wording did not translate easily was the Buddhists. There are a number of other items that do address Buddhist spiritual experience. The introduction to the instrument was subsequently designed to encourage people who are not comfortable with the word God to “substitute another idea which calls to mind the divine or holy for you.” Conceptual orientation and rationale for specific items. DSE is composed of a variety of concepts. It was expected that some aspects of DSE would be more important for some people than for others and that it was possible that specific components would be particularly important for mental and physical health. The interviews revealed that connection was an important concept. Western spirituality emphasizes a personal connection with God, whereas Eastern and Native American spirituality, for example, place more emphasis on a connection with all of life and on connection as being part of a greater whole (41). Two items were developed to address both people whose experience of relationship with the transcendent is one of personal intimacy (“I feel God’s presence”) and those who describe a more general sense of unity as their connection with the transcendent (“I experience a connection to all of life”). Many people experience frequent interaction with the transcendent as a fundamental part of life, an active involvement of the divine in the nitty-gritty of life, and not only in moments of
Volume 24, Number 1, 2002 stress. “Social support from the divine” can be experienced as instrumental or emotional. Perception of a supportive interaction with the transcendent is measured in this instrument in three ways. The first way is that of strength and comfort (“I find strength in my religion or spirituality” and “I find comfort in my religion or spirituality”). These were initially separated into two items; however, they were highly intercorrelated so that, although conceptually different, a case can be made for use of only the strength item. Second, perceived divine love (“I feel God’s love for me directly” and “I feel God’s love for me through others”). These items measure whether the individual experiences God’s love rather than whether one just believes that God loves people generically or that God and love are related conceptually. Feeling loved may prove important in seeking to quantify the relation of religious or spiritual issues to health outcomes. The quality of love imputed to God can be different from the many kinds of love shared by people, with a particular kind of love from others that many attribute to God. Divine love, directly and through others, can be experienced as affirming and can contribute to self-confidence and a sense of self-worth, independent of actions. The third aspect of interaction with the transcendent is when support from the divine may be experienced as inspiration or discernment (“I ask for God’s help in the midst of daily activities” and “I feel guided by God in the midst of daily activities”). These items address the expectation of divine intervention or inspiration and a sense that a divine force has intervened or inspired. The guidance item was most often described similar to a “nudge” from God and more rarely as a more dramatic action. The perception that life consists of more than physical states, psychological feelings, and social roles may help one in transcending the difficulties of present physical ills or psychological situations. (“During worship, or at other times when connecting with God, I feel intense joy which lifts me out of my daily concerns”). The language of this item translates from metaphysical terminology into more practical lay language by focusing on concrete examples that might occur in the context of a lively worship service or a walk in nature. This has been described by many as an important feature in dealing with chronic disease, pain, and disability (15). A sense of wholeness and internal integration is reflected frequently in the spiritual literature of both Eastern and Western traditions, with an accompanying sense of inner harmony (“I feel deep inner peace or harmony”). During interviews, persons who had experienced depression insisted that they could have this experience even in the midst of feeling very distressed. This sense of peace has a transcendent dimension that may be affected but not determined by events and affect, and this question was designed to elicit something other than positive mood or psychological well-being. Van Kaam (41) suggested that awe is the central quality of spiritual life and has an ability to elicit experience of the spiritual that crosses religious boundaries and affects people with no religious connections (“I am spiritually touched by the beauty of creation”) (41). Gratefulness (“I feel thankful for my blessings”) is considered a central component of spirituality by many (42). Because
Daily Spiritual Experience Scale
25
of the potential connections between gratitude and circumstances of life, external life circumstances or stressors may modify a respondent’s feelings of gratefulness; however, some people find blessings even in the most dire circumstances. The attitudes of compassion and mercy are more active and less passive than the qualities of experience just mentioned, but they still result in inner spiritual experience and, therefore, are included in this measure. Unconditional love, agape or compassionate love, is central to many spiritual traditions (41). The item “I feel a selfless caring for others,” which may seem unwieldy, was easily understood by individuals representing a spectrum of educational levels. It describes a love centered on the good of the other and did not generally connote self-abnegation in the interviews assessing content validity. An excellent examination of this concept is found in Vacek (43). Mercy, giving others the benefit of the doubt, dealing with others’ faults in light of one’s own, and being generous of heart, describe inner experiences in which the spiritual can be evident in everyday life. The item “I accept others even when they do things I think are wrong” addresses the felt sense of mercy rather than the mere cognitive awareness that mercy is a good quality. This fundamental acceptance of others is not the same as forgiveness, which is based on response to a particular act. Vanier’s writings (44), which use the example of attitudes toward those with developmental disabilities, address this concept thoughtfully. The final two items assess spiritual longing (“I desire to be closer to God or in union with the divine”). This is a key concept in the Muslim tradition (45) and may be more relevant for those who are seeking interaction with the divine. The final item (“How close do you feel to God?”) was originally included as a way of calibrating the previous question. Item format. The measure includes 16 items described in the following paragraphs; the first 15 are scored using a modified Likert scale, in which response categories are many times a day, every day, most days, some days, once in a while, and never or almost never. (Use to date as reflected in this article has been such that lower scores reflect more frequent DSE: e.g., many times a day = 1, never or almost never = 6.) The introduction to the items states that The list that follows includes items which you may or may not experience, please consider how often you directly have this experience, and try to disregard whether you feel you should or should not have these experiences. A number of items use the word God. If this word is not a comfortable one for you, please substitute another idea which calls to mind the divine or holy for you. Each item is cast in positive terms. Initially, some items were cast in negative terms, but it became clear that this mode of assessment measured something other than the opposite to the concept being addressed, perhaps anomie or alienation. The 16th item, “In general, how close do you feel to God?,” has four response categories: not close at all, somewhat close, very close, and as close as possible.
26
Underwood and Teresi
Annals of Behavioral Medicine
In addition to the 16-item scale, a 6-item version was developed for incorporation into surveys. The 6-item version of the scale, used in the Brief Multidimensional Measure of Religiosity and Spirituality (7), was developed by selection of items representing a few key aspects of DSE from the 16-item version. (This is not a recommended short form for this instrument.) The selection of these items was developed in conjunction with inputs from the National Institute on Aging/Fetzer working group, the goal being to have a set of items to complement other domains in the multidimensional instrument. The main reason for the presentation here of these 6 items is that they were embedded in the General Social Survey (GSS) for 1997–1998 (46) and therefore allowed us to examine some normative population data for a subset of DSES items. The strength and comfort questions were combined to read “I find
strength and comfort in my religion,” as the psychometric properties of the combined item were known, and it has been extensively used with some predictive data for health. The 2 items regarding love directly from God and love from God through others were also combined into 1 item, although it is recognized that it would be preferable to maintain the 2 separate items, as correlations with social support may vary between the 2. The additional items are those on presence, touched by beauty, and desire to be in union. Examination of the 6-item version in this article provides a basic measure of spiritual experience and allows us to examine the distributions of some of these experiences in the general population. Only extensive testing of the 16-item version in health studies will permit identification of the shorter list of items most predictive of positive health outcomes. See Table 1 for a list of items.
TABLE 1 Summary Statistics for the Daily Spiritual Experience Scale Across Four Sites: Item Means, Standard Deviations, Scale Alphas, and Intraclass Reliability Coefficients Chicago SWANa Item Content 1. I feel God’s presence. 2. I experience a connection to all life. 3. During worship, or at other times when connecting with God, I feel joy, which lifts me out of my daily concerns. 4. I find strength in my religion or spirituality 5. I find comfort in my religion or spirituality. 6. I feel deep inner peace or harmony. 7. I ask for God’s help in the midst of daily activities. 8. I feel guided by God in the midst of daily activities. 9. I feel God’s love for me, directly. 10. I feel God’s love for me, through others. 11. I am spiritually touched by the beauty of creation. 12. I feel thankful for my blessings. 13. I feel a selfless caring for others. 14. I accept others even when they do things I think are wrong. 15. I desire to be closer to God or in union with Him 16. In general, how close do you feel to God?
Loyola Universityb
Corrected Item-Total Correlations
GSSc
M
SD
M
SD
M
SD
Chicago SWANd
Loyola Universitye
GSSf
ICC Reliability (Ohio)g
2.76 2.96
1.66 1.48
3.00 3.03
1.35 1.17
3.23 —
1.67 —
.86 .69
.79 .83
.77 —
.71 —
3.48
1.64
3.39
1.22
—
—
.85
.76
—
—
2.76
1.54
2.94
1.31
3.23
1.66
.88
.82
.82
.74
2.79
1.51
2.83
1.32
3.23
1.66
.88
.82
.82
.74
3.22
1.47
3.38
1.00
3.26
1.40
.81
.67
.70
.64
2.91
1.70
3.39
1.39
—
—
.83
.75
—
—
3.22
1.73
3.60
1.26
—
—
.89
.82
—
—
3.06
1.74
3.33
1.37
3.11
1.59
.89
.83
.83
.67
3.03
1.57
3.22
1.28
3.11
1.59
.83
.76
.83
.67
2.58
1.34
2.51
1.26
2.71
1.51
.68
.57
.63
.75
1.97 2.94 2.85
1.01 1.26 .99
2.27 2.80 2.70
1.08 1.05 1.04
— — —
— — —
.73 .37 .33
.66 .49 .36
— — —
— — —
2.63
1.50
2.75
1.35
3.14
1.62
.72
.75
.80
.78
2.69
.89
2.24
.97
—
—
.72
.76
—
—
Note. For the GSS site, Items 4 and 5 are equivalent to Item 14b from the Short Form, and Items 9 and 10 are equivalent to Item 14e, also from the Short Form. SWAN = Study of Women Across the Nation; GSS = General Social Survey; ICC = Intraclass Correlation Coefficient. an = 233. bn = 122. cn = 1,445. dScale α = .95. eScale α = .94. fScale α = .91. gn = 50.
Volume 24, Number 1, 2002
Daily Spiritual Experience Scale
Samples. Data from several samples are included in this article, as follows: 1. Rush-Presbyterian–St. Luke’s Medical Center, Chicago, conducted a series of psychometric analyses of the 16-item DSES as part of the Study of Women Across the Nation (SWAN), a multisite, multiethnic, mulifactorial study of midlife (L. Shahabi & L. Powell, personal communication, March 2, 1999). The Chicago site contributed 233 cases for these analyses. All were women; 60% were White, 53% were Catholic, 18% were Protestant, 21% were Baptist, and 8% belonged to other religions. The mean age was 46.76 (SD = 2.74). 2. The Ohio University Medical Center examined the spiritual and religious dimensions of daily life of 45 patients with arthritis pain (47). The Ohio study contributed the interrater reliability estimates for the six DSES items contained in the GSS. 3. Loyola University (J. Zechmeister, personal communication, March 24, 1999) administered the 16-item DSES to a sample of 122 individuals from the University of Chicago area; 58% were full-time students. The sample was female (61%), male (49%), White (72%), non-White (28%), Catholic (49%), and non-Catholic (51%). Although the range in age was from 15 to 88, the mean was 27.7 (SD = 13.4). 4. The GSS for 1997–1998 (46) used the 6-item version of the DSES among 1,445 individuals nationally. This survey was designed to constitute a random, representative sample of the U.S. population geographically, socioeconomically, and racially. The sample was 79% White and 45% female, with a religious distribution representative of the U.S. population. The mean age was 45.64 (SD = 17.06). Approach to the analyses. Presented next are (a) descriptive and normative data for the DSES, (b) estimates of reliability, (c) results of EFAs, and (d) preliminary evidence for conTABLE 2 Exploratory Factor Analyses of the Daily Spiritual Experience Scale: Factor Loadings from the Structure Matrix Short Item Wording 1. Presence 2. Connection 3. Joy when connecting 4. Strength in R/S 5. Comfort in R/S 6. Deep inner peace 7. God for help 8. Guided by God 9. Love through others 10. Love directly 11. Touched by beauty 12. Thankful for blessings 13. Selfless caring 14. Accept others 15. Desires to be in union 16. Close Note.
R/S = Religiousness/Spirituality.
Factor 1
Factor 2
.90 .69 .88 — .89 .82 .88 .93 .87 .93 .68 .74 .33 .27 .77 .77
.33 .62 .39 — .38 .48 .26 .29 .33 .33 .60 .52 .77 .78 .26 .27
27
struct validity. Scale and item means and standard deviations are presented for several samples and within each sample for different sex, racial–ethnic, and religion subgroups. Estimates of internal consistency (Cronbach’s α) are presented across three samples; a fourth study contributed estimates of interrater reliability for a subset of DSES items, with the intraclass correlation coefficient. Estimates of test–retest reliability were also provided for the six items contained in the GSS version of the DSES. The results of EFAs are presented, and preliminary evidence for construct validity is discussed in terms of differences in DSES scores for different demographic and religion subgroups and correlations of DSE with psychological and other health-related variables. RESULTS Descriptive Statistics for the DSES Table 1 shows means and standard deviations for the 16 DSES items across three studies: Chicago SWAN, Loyola University, and the GSS (the Ohio study provided interrater reliability estimates with respect to items contained in the GSS but no item frequencies). From a psychometric standpoint, most items, although somewhat skewed toward the more frequent tail of the distribution, demonstrate adequate variability. However, few respondents endorsed the never or almost never category for Items 11 through 14. Examination of item frequencies (available from Underwood) and means (Table 1) showed that several statements were more frequently endorsed across sites. These were “I am spiritually touched by the beauty of creation,” “I am thankful for my blessings,” and “I desire to be closer to God or in union with him.” Those items less frequently endorsed were “During worship, or at other times when connecting with God, I feel joy which lifts me out of my daily concerns,” “I feel guided by God in the midst of daily activities,” and “I feel God’s love for me, directly.” Correlations Among Items Examination of the zero-order correlations among items (not shown here) for the SWAN study showed that most items were moderately to highly intercorrelated (average range of correlations = .60–.80). Two items, “I feel a selfless caring for others” and “I accept others even when they do things I think are wrong,” had lower correlations (in the .20s) with all other items. However, two items, “finds strength in religion, spirituality” and “finds comfort in religion, spirituality” were collinear. The correlation was .96; all responses were almost identical for both items. The item wording of the two items is very similar, and respondents did not appear to distinguish between the terms comfort and strength. If this pattern is observed in other samples, it is recommended that in future work one of the two collinear items be omitted. Although the GSS (46) version combines the items, from a psychometric standpoint, double-barreled items are to be avoided. Because of the problems of collinearity in this data set, EFAs were conducted for both a 16-item version and a 15-item version; however, results are shown only for the 15-item version (see Table 2) of the DSES.
formed somewhat more consistently and yielded results more similar to those with the continuous response format. The internal consistency for this latter dichotomous version of the 16-item scale was .93. Additional exploratory analyses of two ways of dichotomizing items provided preliminary evidence that the items could be treated as binary if necessary for populations in which a Likert-type response continuum may be problematic (e.g., the frail or the very old). Details of these analyses are available from Underwood.
Psychometric Analyses of DSES: Classical Test Theory Results Test–retest of six items. It is expected that DSE is relatively stable over the short term. However, because this construct measures perceptions and feelings, scores may vary according to external stressors and emotional state. Therefore, assessing response stability over a relatively brief period is appropriate. Test–retest reliability currently is being assessed with the entire 16 items; however, the 6 items were incorporated into a test–retest of separate subscales of the Brief Multidimensional Measure of Religiousness/Spirituality, the results of which follow (48). Forty-seven treatment-seeking substance users were tested for 2-day response stability. The Spiritual Experience subscale had good response stability (Pearson product–moment correlation = .85; intraclass correlation coefficient =.73). The Cronbach’s alpha estimate of internal consistency was .88 for test and .92 for retest.
Preliminary Construct Validity: Differences in DSES Scores for Different Demographic and Religion Subgroups Summary statistics for the DSES were examined across several samples and subgroups (see Table 3). Means on the DSES 16-item version are about 47 for both the SWAN (SD = 18.69) and Loyola (SD = 13.81) studies. The mean score for the 6-item GSS version was 18.68 (SD = 7.91). Of note are the lower mean scores for African American women in the SWAN study (37.78, SD = 14.87) in contrast to Whites (52.79, SD = 18.58), t = 6.82, p < .01. This indicates that the responding African American women reported a significantly greater degree of DSE than did Whites. This pattern was repeated for the GSS (46) 6-item version of the scale, t = 8.44, p < .01. This was consistent with other findings from the GSS (46) data and with studies indicating high levels of religiousness among African American women as measured by religious involvement (both organizational and nonorganizational) and subjective ratings (49). As would be expected, the GSS (46) data showed that those who reported “no religion” had the highest GSS mean scores (25.91, SD = 7.30), that is, the least frequent daily spiritual experiences. Comparison of those with no religion with those who claimed to be either Protestant or Catholic showed that the former had significantly different mean scores, reflecting less frequent DSEs, F = 126.60, p < .01. Scheffé multiple range tests indicated that individuals with no religion had significantly less frequent daily spiritual experiences than did those who claimed to be Catholic or Protestant. (The score for those of self-proclaimed Jewish faith showed high frequency daily spiritual experiences; however, the subsample size was too small [26] to permit reliable inferences.) Women also reported significantly more frequent daily spiritual experiences than did men, t = 6.26, p < .01. This was also consistent with other GSS (46) data, with women scoring significantly higher than men on virtually every item in the other domains of religiousness or spirituality: public activity, private activity, coping, religious intensity, forgiveness, and beliefs.
Interrater reliability. Interrater reliability is not a concern for most applications of the DSES because it is usually self-administered. However, if administered by an interviewer, say, to a frail or very old population, interrater reliability would be of concern. Table 1 presents interrater reliability estimates calculated by the Ohio site, with the intraclass correlation coefficient for 6 items representing 8 of the 16 DSES items (2 DSES items were combined in two analyses). As shown, the reliability coefficients were adequate, ranging from .64 to .78. Internal consistency reliability. The internal consistency reliability estimates with Cronbach’s alpha were very high, .94 and .95 for the 16-item version of the scale and .91 for the 6-item scale used in the GSS (46). EFA. Several EFAs were performed for the Chicago SWAN study. An exploratory principal components analysis was first performed to examine the dimensionality of the DSES. The item set tended to be unidimensional for this sample. This interpretation was supported by the fact that the first eigenvalue was about 10 times that of the second; this can be demonstrated graphically by the scree test, a plot of the eigenvalues against the factor rank. An EFA with an oblique rotation was then performed. As shown in Table 2, nearly all items loaded highly on the first factor, with loadings ranging from .69 to .93, except for two items, which loaded at .33 and .27; these items loaded more highly on a second factor (these items were “feels selfless caring for others” and “accepts others even when they do wrong things”). However, a 2-item scale is generally undesirable and, in this case, not meaningful in terms of explained variance (about 8%). It is also to be noted that the items with explicit reference to God did not factor out separately from those without such reference. Two other EFAs (not shown here) were performed with the dichotomization of items at two different points: (a) the combination of never and once in a while versus some days to many times a day and (b) the combination of never, once in a while, and some days versus most days to many times a day. The results of these analyses indicated that the second dichotomization per-
Correlations With Psychosocial and Other Health-Related Variables In the Chicago SWAN study, frequency of DSE (scored positively for this analysis) was significantly negatively correlated with a variety of psychosocial factors: anxiety assessed with the State–Trait Anxiety Inventory (50), depression measured with the Center for Epidemologic Studies–Depression F (51), and the Cohen Perceived Stress Scale (52) (see Table 4). It
28
29 48.65 14.51 46 37.7
20.14 8.25 654 45.3
17.49 7.42 791 54.7
— — — —
46.45 13.48 74 60.7
46.87 18.69 233 100
Male
19.28 7.91 1,143 79.1
48.85 14.17 88 72.1
52.79 18.58 141 60.3
White
14.93 6.37 198 13.4
NAb NA NAb NA
37.78 14.87 93 39.7
African American
Ethnicity
— — — —
40.00 12.40 12 9.8
— — — —
Hispanic
19.24 8.59 104 7.2
— — — —
— — — —
Other
16.53 7.10 783 55.3
40.65 10.02 13 10.7
46.79 19.99 39 17.7
Protestant
19.96 7.45 370 26.1
43.75 14.25 60 49.2
49.09 17.04 117 53.2
Catholic
25.08 8.39 26 1.8
— — — —
NAb NA NAb NA
Jewish
— — — —
— — — —
37.94 15.61 47 21.4
Baptist
Religion
16.35 6.79 23 1.6
— — — —
— — — —
Christian
25.91 7.30 198 14.0
NAb NA NAb NA
NAb NA NAb NA
None
17.06 7.37 17 1.2
46.93 7.04 14 11.5
NAb NA NAb NA
Other
Note. Individual sample sizes do not sum to total sample size because some sample sizes were too small for inclusion. SWAN = Study of Women Across the Nation; GSS = General Social Survey. an = 233; theoretical range = 16–94, observed range = 17–89; bSample size was too small for this group (< 10). cThe percentage of the column category in each sample; for example, 100% of the SWAN study respondents were female. dn = 122; theoretical range = 16–79, observed range = 16–75. en = 1,445; theoretical range = 6–36, observed range = 6–36.
Chicago SWANa M SD n %c Loyola Universityd M SD n %c GSSe M SD n %c
Female
Sex
TABLE 3 Mean Scores on the Daily Spiritual Experience Scale for Different Demographic Subgroups by Site
30
Underwood and Teresi
Annals of Behavioral Medicine
TABLE 4 Correlations Between the Daily Spiritual Experience Scale and Psychosocial and Other Health-Related Factors (SWAN Study) Factor Quality of Life (SF-36) Sleep Problems Physical Ailments Alcohol Consumption Anxiety Center for Epidemiological Studies–Depression Speilberger Anger–Coping Scale Cohen Perceived Stress Cook Medley Hostility Scheirer Optimism Berkman Perceived Social Support
Correlation .240** –.060 –.110 –.200** –.394** –.220** –.303** –.197** –.157* .352** .183**
Note. The Daily Spiritual Experience Scale was scored in the positive direction for these analyses. The following scales were scored such that a high score reflects more positive outcomes: Quality of Life, Optimism, Social Support. SWAN = Study of Women Across the Nation; SF = Short Form. *p < .51, two tailed. **p < .01, two tailed.
was positively significantly correlated with Scheirer’s Optimism Scale (53) and Berkman’s scale of Perceived Social Support (54). The DSES was significantly negatively correlated with alcohol consumption. This reflected the sum alcohol intake, combining wine, beer, and liquor, with more DSE linked with less daily alcohol intake. Also, the more DSE, the higher the Short Form-36 rating (55) of quality of life. No significant correlations were observed with self-reported sleep problems or with a self-report of physical symptoms. In the Loyola study, more frequent DSE was correlated with more positive affect (Pearson’s correlation = .29, p < .01, two-tailed) when measured with the Watson and Clark Positive and Negative Affect Scale (56). No significant correlation occurred with negative affect when that scale was used.
DISCUSSION In general, the findings reported here support the use of the DSES to measure DSE. The DSES demonstrated good internal consistency reliability across all samples. The high internal consistency estimates for the DSES suggest that the items function together to consistently measure the spiritual experience construct. Preliminary interrater reliability data showed acceptable agreement for the subset of items examined. Preliminary construct validity was established through examination of the mean scale scores across sociodemographic subgroups. The DSES also appeared to discriminate between religion, sex, and racial subgroups in a fashion consistent with and predicted from the literature. Evidence of construct validity was also provided by examination of correlations of the DSES with health and quality of life variables. Significant associations in the expected direction were observed for most variables. Examination of the item distributions across several samples indicated that the items and the scale have adequate item
distributions and are not badly skewed. The highest cross-population mean was for “being spiritually touched by the beauty of creation.” This item was designed to address a broad population. The central position of awe in spiritual experience was reinforced by these results. If ultimate links with health outcomes are found for this item, this might mean that exposure to nature may be a way of encouraging and enriching the transcendent dimension of life in a particular accessible way. Another item with a high mean was “experiencing deep inner peace.” However, this was more rarely reported as occurring many times a day or every day for the different studies than was the “spiritually touched by beauty” item. Although this aspect of spiritual experience is highly prized, it seems to be less accessible. Although these two items had mean scores indicative of a higher frequency of report, there were relatively large proportions of the samples that endorsed the categories many times a day or never or almost never. This pattern was true of most DSES items. In future work, it may be important to investigate whether endorsement of responses at the extreme ends of the spectrum is associated with particular health outcomes, attitudes, and behaviors. Another interesting finding from the GSS, based on cross-sectional data, is that 42 to 43% of the sample reported that they “experience God’s presence,” “feel strength and comfort from their religion,” and “feel God’s love” everyday or many times a day. The role that these daily experiences may play in informing decision making, shaping motivations, and influencing health outcomes needs to be investigated. Preliminary EFAs suggests that this scale is unidimensional; however, two items did not load as highly on the first factor and, if combined with additional like items, might form a separate factor in other analyses. Further work will investigate this result. One caveat is that only an EFA was performed. It is acknowledged that the ideal situation would be to conduct EFAs on more data sets and to use other data sets or to use random subsamples of data sets for the confirmatory factor analyses. However, this sample did not permit such subdivision. Confirmatory factor analysis is planned as more data emerge from the many health studies in which this instrument has been embedded. The items with explicit reference to God did not factor out separately from the others, which supports achievement of one of the goals of the instrument: to compose an instrument that addresses a possible common ground that transcends many religious boundaries. One issue that cannot be avoided in any psychometric assessment is the possibility of bias borne of self-report. Many of the items require a certain kind of discrimination between events. “I accept others even when they do things I think are wrong,” for example, requires an inner judgment to be made that acceptance is indeed taking place, in such a way that it will feed into feelings and attitudes and behaviors. It may be that there are some people who take a very critical view of their own inner experiences and, therefore, rate themselves as having the experience less frequently than someone who might be less discerning of their own real attitude. For elaborations of some of these issues, see Underwood (57). For this reason, the combination of
Volume 24, Number 1, 2002 this measure with other religious measures and measures of values and behaviors may be warranted as further work develops. Implications for Health and Well-Being Better social support has been connected to improved health in a variety of settings and with a variety of measures (58). Although cross-sectional, the links shown with perceived social support may identify possible ways in which DSE can enrich our experiences of relationships with others, creating stronger, more supportive bonds with others. Also the “social support from the divine” element in this measure may tap an additional source of social support for the many individuals who report significant DSE of an intimate nature. Recent work shows better health in those with more types of social relationships (59), and this measure looks at an additional type of relationship or connection on which to draw. The potential that DSE might behave as a stress buffer is reinforced by the positive association with psychosocial variables such as optimism and positive affect and the negative association with perceived stress. Further longitudinal studies of health outcomes could clarify these associations. The negative associations with anxiety and depression in this cross-sectional work could merely show that those who are depressed and anxious are less likely to have frequent daily spiritual experiences. Ways to determine whether there is a buffering effect of DSE on depression and anxiety might include the use of prospective studies or ecological momentary assessment. Understanding deeply seated emotional factors people define in spiritual terms may help to identify ways to effect behavior and attitude changes that can be beneficial to physical and emotional health. Although based on cross-sectional data, the connection of alcohol intake with DSE shown in this study gives us an indication of the kind of research work that may be possible. DSE may identify something that buffers one during the stresses and strains of lives and relationships. In this context, DSE may play a role in the creation of an internal environment in which alcohol may not be perceived as a need. The finding that DSE is significantly associated with quality of life is suggestive. The feelings of joy, comfort, and connection that are tapped by this instrument could provide understanding of the potential benefit of encouraging spiritual aspects of life for individuals experiencing various forms of illness, for example, people with disabilities and chronic pain. Information on the influence of DSE on well-being and health could provide a resource for dealing with illness or a source of resilience for those at risk. There may be a variety of ways to enrich this aspect of life, from choral singing to hiking in nature to natural views from hospital rooms to private reading. This aspect of life could be enriched through referral to a religious setting, if that is a part of a particular person’s life, or the recommendation of something as basic as writing about one’s life story from the perspective of meaning, or the “more than” perspective. Defining the self in a way that does not depend on physical functioning can be helpful when experiencing chronic disease and disability (60). A number of religious and spiritual activities could encourage more frequent DSE. Just the acknowledgment of the potential
Daily Spiritual Experience Scale
31
importance of this aspect of life may help us to better design social support interventions or psychosocial approaches to depression and pain. Further work is being conducted on the DSES; for example, the DSES has been used in three large studies of physical health outcomes and in other smaller studies. In addition to the studies outlined next, the DSES is being used in an ongoing study at Duke University examining the effect of DSE on health care utilization and length of hospital stay. A version has been added to ENRICHED, a National Heart, Lung, and Blood Institute multicenter trial of a social support intervention for post myocardial infarction patients. Qualitative and quantitative evaluation on a non-Judeo-Christian, Asian population is also underway at the University of California, San Francisco. The results of the construct validity analyses seem to suggest that higher DSE may be positive; these conclusions are based on positive correlations of the DSES with variables such as quality of life and negative correlations with anxiety, depression, and alcohol consumption. However, it is not possible with cross-sectional data to make a definitive statement regarding the impact of DSE. These preliminary findings support the use of the scale to measure DSE and its use in health studies. It has been incorporated into a variety of health-related research, where an experiential scale of this type has been well received. As well as appealing to the less religious, it also addresses aspects of spirituality that resonate with the most deeply religious and spiritual. It holds promise as a measure of features of daily life, with possible implications for physical and mental health and well-being. REFERENCES (1) James W: The Varieties of Religious Experience: A Study in Human Nature. New York: Modern Library, 1994 (1902). (2) Strawbridge WJ, Cohen RD, Shema SJ, Kaplan GA: Frequent attendance at religious services and mortality over 28 years. American Journal of Public Health. 1997, 8:957–961. (3) Levin JS: How religion influences morbidity and health: Reflections on natural history, salutogenesis, and host resistance. Social Science and Medicine. 1996, 43:849–864. (4) Pargament KI, Smith BW, Koenig HG, Perez L: Patterns of positive and negative religious coping with major life stressors. Journal for the Scientific Study of Religion. 1998, 37:710–724. (5) Koenig HG: Is Religion Good for Your Health? Binghamton, NY: Haworth, 1997. (6) Sloan RP, Bagiella E, Powell T: Religion, spirituality, and medicine. Lancet. 1999, 353:664–667. (7) Fetzer Institute, National Institute on Aging Working Group: Multidimensional Measurement of Religiousness, Spirituality for Use in Health Research. A Report of a National Working Group Supported by the Fetzer Institute in Collaboration With the National Institute on Aging. Kalamazoo, MI: Fetzer Institute, 1999. (8) Bradley MB, Green NM, Jones DE, Lynn M, McNeil L: Churches and Church Membership in the United States: 1990. Atlanta, GA: Glenmary Research Institute, 1992. (9) Koenig H, Futterman A: Religion and Health Outcomes: A Review and Synthesis of the Literature. Methodological Ap-
32
(10) (11)
(12)
(13)
(14) (15)
(16) (17) (18) (19)
(20) (21) (22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
Underwood and Teresi proaches to the Study of Religion, Aging, and Health. Washington, DC: 1995. Hoge D: A validated intrinsic religious motivation scale. Journal for the Scientific Study of Religion. 1972, 11:369–376. Idler EL, Kasl SV: Religion among disabled and nondisabled elderly persons: I. Cross-sectional patterns in health practices, social activities, and well-being. Journal of Gerontology: Social Sciences. 1997, 52B:S294–S305. Benson PL, Donahue MJ, Erickson JA: The Faith Maturity Scale: Conceptualization, measurement, and empirical validation. In Lynn ML, Moberg DO (eds), Research in the Social Scientific Study of Religion (Vol. 5). Greenwich, CT: JAI, 1993, 1–26. Schwartz SH, Bilsky W: Toward a universal psychological structure of human values. Journal of Personality and Social Psychology. 1987, 53:550–562. McGinn B: The letter and the spirit: Spirituality as an academic discipline. Christian Spirituality Bulletin. 1993, 1(2):2–9. Underwood L: A working model of health: Spirituality and religiousness as resources: Applications to persons with disability. Journal of Religion, Disability and Health. 1999, 3(3):55–71. Tononi G, Edelman G: Consciousness and complexity. Science. 1998, 282:1846–1851. Bennett WJ: Neuroscience and the human spirit. National Review. 1998, 50(25):32–35. Damasio A: Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Avon, 1994. Hood R: The construction and preliminary validation of a measure of reported mystical experience. Journal for the Scientific Study of Religion. 1975, 22:353–365. Underhill E: Practical Mysticism. New York: Dutton, 1914. Pargament KI: The Psychology of Religion and Coping: Theory, Research, Practice. New York: Guilford, 1997. Allport G, Ross J: Personal religious orientation and prejudice. Journal of Personality and Social Psychology. 1967, 5:447–457. Paloutzian RF, Ellison CW: Loneliness, spiritual well-being and quality of life. In Peplau LA, Perlman D (eds), Loneliness: A Sourcebook of Current Theory, Research, Research and Therapy. New York: Wiley Interscience, 1982, 224–237. Ellison CW: Spiritual well-being: Conceptualization and measurement. Journal of Psychology and Theology. 1983, 11:330–340. Williams DR: Commitment. In Fetzer Institute, National Institute on Aging Working Group: Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research, A Report of a National Working Group Supported by the Fetzer Institute in Collaboration with the National Institute on Aging. Kalamazoo, MI: Fetzer Institute, 1999, 72. Cohen S, Kessler R, Underwood-Gordon L (eds): Measuring Stress: A Guide for Health and Social Scientists. New York: Oxford University Press, 1995. Stone AA, Bovbjerg DH: Stress and humoral immunity: A review of the human studies. Advanced Neuroimmunology. 1994, 4:49–56. Flood AB, Lorence DP, Ding J, et al.: The role of expectations in patients’ reports of post-operative outcomes and improvement following therapy. Medical Care. 1993, 31:1043–1056. Roberts AH, Kewman DG, Mercier L, et al.: The power of nonspecific effects in healing: Implications for psychosocial and biological treatments. Clinical Psychology Review. 1995, 12:375–391.
Annals of Behavioral Medicine (30) Davidson R, Kabat-Zinn J, Schumacher J, et al.: Alterations in Brain and Immune Function Produced by Mindfulness Meditation. Manuscript submitted for publication, 2002. (31) Hummer RA, Rogers RG, Nam CB, et al.: Religious involvement and U.S. adult mortality. Demography. 1999, 36:273–285. (32) Idler EL, Kasl SV: Religion among disabled and nondisabled elderly persons: II. Attendance at religious services as a predictor of the course of disability. Journal of Gerontology: Social Sciences. 1997, 52B:S306–S316. (33) Oxman TE, Freeman Jr. DH, Manheimer ED: Lack of social participation or religious strength and comfort as risk factors for death after cardiac surgery in the elderly. Psychosomatic Medicine. 1995, 57:5–15. (34) Elkins DN, Hedstrom LJ, Hughes LL, Leaf JA, Saunders C: Toward a humanistic–phenomenological spirituality: Definition, description, and measurement. Journal of Humanistic Psychology. 1988, 28(4):5–18. (35) Buber M: I and Thou (Kaufmann W, trans.). New York: Touchstone, 1996 (1937). (36) De Wit H: Contemplative Psychology. Pittsburgh, PA: Duquesne, 1991. (37) Hanh TN: A Joyful Path: Community, Transformation and Peace. Berkeley, CA: Parallax, 1994. (38) Merton T: Life & Holiness. New York: Doubleday, 1969. (39) van Kaam A: Formation of the Human Heart (Formative Spirituality Series, Vol. 3). New York: Crossroads, 1991. (40) Smith H: The World’s Religions: Our Great Wisdom Tradition (Rev. Ed.). San Francisco: Harper, 1991. (41) van Kaam A: Fundamental Formation (Formative Spirituality Series, Vol. 1). New York: Crossroads, 1986. (42) Steindl-Rast D: Gratefulness, the Heart of Prayer—An Approach to Life in Fullness. New York: Paulist, 1984. (43) Vacek EC: Love, Human & Divine: The Heart of Christian Ethics. Washington, DC: Georgetown University Press, 1994. (44) Vanier J: Becoming Human. Mahwah, NJ: Paulist, 1999. (45) Lofty M: WHO and Spirituality, Religiousness and Personal Beliefs (SRPB): Report on WHO Consultation June 22–24, 1998. Unpublished report. World Health Organization, Division of Mental Health and Prevention of Substance Abuse. 1998. (46) David JA, Smith TW, Marsden PV: General Social Surveys, 1972–2000 Cumulative Codebook. Chicago: National Opinion Research Center, 2001. (47) Keefe FF, Affleck G, Lefebvre J, et al.: Living with rheumatoid arthritis: The role of daily spirituality and daily religious and spiritual coping. Journal of Pain. 2001, 2:101–110. (48) Tonigan J, Walter S, Underwood L: Test–Retest Study of the Brief Measure of Religiousness and Spirituality Among Treatment Seeking Substance Users. Manuscript submitted for publication, 2002. (49) Ellison C: Religious involvement and self perception among Black Americans. Social Forces. 1993, 71:1027–1055. (50) Spielberger CD, Gorush R, Lushene R, Vagg P, Jacobs G: Manual: For the State–Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press, 1983. (51) Radloff LS: The CES–D scale: A self report depression scale for research in the general population. Applied Psychological Measurement. 1977, 1:385–401. (52) Cohen S, Karmarck T, Mermelstein R: A global measure of perceived stress. Journal of Health and Social Behavior. 1983, 24:385–396. (53) Scheirer MF, Carver CS, Bridges MW: Distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and
Volume 24, Number 1, 2002 self-esteem): A reevaluation of the Life Orientation Test. Journal of Personality and Social Psychology. 1994, 67:1063–1078. (54) Seeman TE, Berkman LF: Structural characteristics of social networks and their relationship with support in the elderly. Social Science and Medicine. 1988, 7:737–749. (55) McHorney CA, Ware Jr. JE, Lu JF, Sherbourne CD: The MOS 36-item Short-Form Health Survey (SF–36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Medical Care. 1994, 32:40–66. (56) Watson D, Clark LA, Tellegen A: Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology. 1988, 54: 1063–1070.
Daily Spiritual Experience Scale
33
(57) Underwood LG: The human experience of compassionate love: Conceptual mapping and data from selected studies. In Post SG, Underwood LG, Schloss JP, Hurlbut WB (eds), Altruism and Altruistic Love: Science, Philosophy, and Religion in Dialogue. New York: Oxford University Press, 2002. (58) Cohen S, Underwood L, Gottlieb B (eds): Social Support Measurement and Intervention: A Guide for Health and Social Scientists. New York: Oxford University Press, 2000. (59) Cohen S, Doyle WJ, Skoner DP, Rabin BS, Gwaltney Jr. JM: Social ties and susceptibility to the common cold. Journal of the American Medical Association. 1997, 227:1940–1944. (60) Brown J: The Self. Boston: McGraw-Hill, 1998.
The Ironson–Woods Spirituality/Religiousness Index Is Associated With Long Survival, Health Behaviors, Less Distress, and Low Cortisol in People With HIV/AIDS Gail Ironson, M.D., Ph.D. Department of Psychology and Psychiatry University of Miami
George F. Solomon, M.D. University of California Los Angeles
Elizabeth G. Balbin, B.S., Conall O’Cleirigh, M.S., and Annie George, M.A. Department of Psychology and Behavioral Medicine University of Miami
Mahendra Kumar, Ph.D. Department of Psychiatry University of Miami
David Larson, M.D., M.S.P.H. National Institute for Healthcare Research
Teresa E. Woods, Ph.D. University of Wisconsin, Madison
concentrations and altruistic behavior as mediators of the relation between SR and long survival.
ABSTRACT The purpose of this study was to determine the reliability and validity of an instrument that measures both spirituality and religiousness, to examine the relation between spirituality and religiousness and important health outcomes for people living with HIV, and to examine the potential mediators of these relations. One aim was to determine whether subscales of spirituality, religiousness, or both would be independently related to long survival in people living with AIDS. The Ironson–Woods Spirituality/Religiousness (SR) Index is presented with evidence for its reliability and validity. Four factors were identified on the Ironson–Woods SR Index (Sense of Peace, Faith in God, Religious Behavior, and Compassionate View of Others). Each subscale was significantly related to long survival with AIDS. That is, the long-term survivor (LTS) group (n = 79) scored significantly higher on these factors than did the HIV-positive comparison (COMP) group (n = 200). Long survival was also significantly related to both frequency of prayer (positively) and judgmental attitude (negatively). In addition, the Ironson–Woods SR Index yielded strong and significant correlations with less distress, more hope, social support, health behaviors, helping others, and lower cortisol levels. The relation between religious behavior and health outcomes was not due to social support. Further analyses were conducted, which identified urinary cortisol
(Ann Behav Med
2002, 24(1):34–48)
INTRODUCTION The notion that spirituality, religiousness, or both may be related to health has been with us for centuries. Illness has often been seen as associated with sin (1). Conversely, reconnection to the sacred has represented a possible approach to healing (2). “Spirituality and a sense of the sacred, after being banished from medicine, are making a comeback” (3). There have been several major reviews showing a link between religion and health. Levin and Schiller (4) and Levin (5) reviewed hundreds of epidemiologic studies reporting relations between religion and health and concluded that there was an association, the association was probably valid, and there may be a causal link. These authors called for more systematic research. McCullough, Hoyt, Larson, Koenig, and Koenig (6) conducted a meta-analysis of 42 independent studies and found that people high in religious involvement were more likely to be alive at follow-up than people lower in religious involvement (odds ratio = 1.29). In their conclusions, they recommended researchers use more reliable measures of multiple dimensions of religious involvement (including public religious involvement, private religious activities, and religious beliefs). Koenig, McCullough, and Larson (7) and Larson, McCullough, and Swyers (8) pointed out the need for a clearer conceptualization of religiousness and for differentiation between spirituality and religiousness. We, therefore, felt that a measure that would capture both spirituality and religiousness was needed. In our previous work, we found that a large proportion of people with HIV identified themselves as spiritual but not religious or as both (9). This self-identification was also true of cancer patients (but to a lesser
This research was graciously funded by National Institutes of Health Grant R01MH53791 and National Institute for Healthcare Research. Thanks to Sandy Romero, Tamika Bailey, and Peter Vitaliano Reprint Address: G. Ironson, M.D., Ph.D., University of Miami, Behavioral Medicine Research Program, P.O. Box 248185, Coral Gables, FL 33124–2070. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
34
Volume 24, Number 1, 2002 extent), who identified themselves evenly between being spiritual and religious. A related question was raised by the McCullough et al. (6) meta-analytic review: Is it the positive psychological states that may be fostered by more private forms of worship or measures of public religious involvement that provide the protective health effects? Thus, we wanted to create a measure that would include items relevant to private as well as public SR and that would capture spirituality as well as religiosity. Second, no one has looked at the association between religiousness/spirituality and health in an HIV population. The HIV population provides some unique opportunities for the study of religiousness/spirituality for a number of reasons. As noted previously, many in this population identify as spiritual rather than religious (9). Next, many have been stigmatized by their churches (10,11,12). In their ambivalence toward organized religion (13), some have been able to redefine their connection with the sacred in new terms outside of traditional religion and, therefore, have had to struggle to find a way to do so. Finally, the studied group represents wide ethnic diversity and socioeconomic backgrounds. Traditional religious institutions are places of importance both to ethnic majorities and minorities and remain an important site for community social gatherings where minorities may be supported (14). Third, if a relation does exist between religiousness/spirituality and health in HIV or AIDS, then the exploration of potential pathways (mediators) would be needed. We previously hypothesized (15) that mediators for long survival might be due to physiologic protection (i.e., low cortisol and those associated psychological states expected to be related to reductions in stress hormones) or changes in health-related behaviors (e.g., adherence to medications, safe sex, smoking, alcohol use). Measures of psychobiological variables related to proposed benefits of spiritual and religious beliefs outlined by Covey (16) were also considered as mediators. Potential mediators suggested include social support; avoidance of drugs, cigarettes, and alcohol; and reasons to be of service and help to others. Based on the aforementioned considerations, this study had five separate aims: 1.
To develop a short form of a scale (from an existing long form of 89 items, the Ironson–Woods Spirituality/Religiousness [SR] Index, which is too long for most practical purposes) that would capture dimensions of both spirituality and religiousness. 2. To provide evidence of this scale’s reliability and validity. 3. To determine whether religiousness/spirituality is related to health (in particular, long survival with AIDS). 4. To explore what the mediators of the relation between spirituality and health might be and to include an examination of whether a religious behavior–health association may be driven by social support. 5. To determine whether the measure of religiousness/spirituality would be significant beyond religious behavior alone.
Ironson–Woods SR Index and HIV/AIDS Survival
35
Based on subsequent data analysis, we also became interested in the possibility of negative health effects of judgmental religiosity (i.e., condemning or judging others harshly). METHODS Participants Two groups of HIV-seropositive participants were recruited: Long-term survivors (LTSs) of AIDS (n = 79) and an HIV-positive comparison (COMP) group (n = 200). LTSs of AIDS are defined by the Centers for Disease Control as persons who have survived twice the median survival time expected for those with AIDS. Before 1996, when this study began (which was before protease inhibitors [PIs] became available to the general public), the median survival time following the appearance of a Category C symptom (i.e., a very serious AIDS-defining opportunistic infection or neoplasm such as Pneumocystis carinii pneumonia [PCP] or Kaposi’s sarcoma [KS]) was 18 to 20 months. To be conservative, to qualify for our LTS group, participants had to be at least 4 years past an opportunistic infection or neoplasm before starting on PIs. (This was possible before 1996 when PIs became widely available.) The second group was an HIV-positive COMP group. Participants had to have CD4+ (T-helper/inducer) cell counts between 150 and 500/mm3 at entry to the study (and had no history of CD4 counts under 75/mm3) and never had any AIDS Category C symptoms (e.g., KS, PCP, toxoplasmosis, non-Hodgkin’s lymphoma). This group was chosen for comparison because of its equivalence to the LTS group on HIV serostatus and CD4 counts. Because some of the COMP group, when followed longitudinally, may become long survivors, this represents a conservative COMP group. Exclusion criteria. Participants were excluded if they were under 18, had another life threatening illness (e.g., cancer), were on medications thought to affect stress hormones (e.g., steroids, propranolol), had taken street drugs within the past month, were actively psychotic or suicidal, or had current alcohol or drug dependence. Measures Overview. In addition to the Ironson–Woods SR Index, to accomplish the goals of the study, we included measures to establish convergent validity (other established measures of religiousness/spirituality) and measures to determine the psychosocial variables (e.g., distress, optimism) and behaviors to which spirituality might be related. Furthermore, this second set of measures could help to establish discriminant validity (i.e., the Ironson–Woods SR Index should correlate lower with these than with measures of religiousness) and to explore mediators of the relation between the Ironson–Woods SR Index and long survival (i.e., pathway by which religion could affect long survival). Measures for the exploration of correlates, discriminant validity, and mediators (described in the Psychosocial Measures and Behaviors sections) included perceived stress, less affective distress (anxiety, depression), physiologic stress (cortisol), more hope (less hopelessness, more optimism), social support
36
Ironson et al.
(Enhancing Recovery in Coronary Heart Disease Social Support Instrument [ESSI]), health behaviors (report of practicing safe sex, adherence, telling partner HIV status, and less smoking and drinking alcohol), and helping others with HIV. Participants filled out questionnaires for all of the measures described except for the following questions, which were asked by interview: (a) “Do you consider yourself to be religious, spiritual, both, or neither?” (b) “Do you believe in GOD?” and (c) the safe sex questions noted in the Behaviors section.
Religiousness/spirituality measures: Ironson–Woods SR Index short form. Twenty-two items chosen from the longer 89-item form of the Ironson–Woods SR Index (17) were selected to represent the most salient proposed dimensions from the 89-item version and were grouped into seven conceptual groups of items (see Table 2). Items included those derived from interviews of 60 medically ill patients, 20 cancer, 20 HIV, and 20 cardiac, who represented people identifying themselves as spiritual, religious, or both, from a previous study by Woods and Ironson (9). They were asked a variety of questions pertaining to their spiritual or religious beliefs, behaviors, attitudes, and feelings. A key purpose of the development of this scale was to include items that were both pertinent to traditional religion and relevant for those who described themselves as spiritual only or as both religious and spiritual. We wanted to determine what people meant when they said they were spiritual or religious (i.e., to determine or dissect the dimensions of religiousness/spirituality). A secondary purpose was to capture both public and private religious or spiritual beliefs, behaviors, attitudes, and feelings. The seven subscales identified by the interviews conceptually are as follows: (a) Comfort, Strength, Meaning; (b) Feeling a Connection, Less Alone; (c) Existential/Afterlife; (d) View of God; (e) Somatic, which is related to recovery from illness; (f) Religious Behavior; and (g) View of Others/Compassion for Others. These components of religiousness/spirituality are listed in Table 2 by original cluster and by the four factors derived by the factor analysis described later in the Results section of this article.
Other measures of religiousness. The Hoge Intrinsic Religious Motivation Scale (18) is a 10-item scale. Higher scores indicate greater intrinsic religiosity. (Actually, there are 7 intrinsic and 3 extrinsic items.) Because alpha was adequate only for the Intrinsic scale, only the Intrinsic scale was used in the analysis (αs = .89 and .35 for the Intrinsic and Extrinsic scales, respectively). The Duke Religion Index (19) is a 5-item scale designed to capture the three major dimensions of religiousness as described by Koenig and Futterman (20). The three subscales are Organizational (frequency of attending church or other religious activities), Non-Organizational (time spent in private religious activities, e.g., prayer, meditation, or Bible study), and Intrinsic Religiosity (3 items from the Hoge, α = .75). The COPE (21) is a series of items used to measure state or trait coping. In this study, only the subscale Turning to Religion to cope was used. The scale has 4 items and a high reliability (α = .92). Finally, a face
Annals of Behavioral Medicine valid question, “Do you believe in God?,” was used to examine the validity of the Ironson–Woods SR Index. Psychosocial measures. Perceived stress was measured by the Perceived Stress Scale (22,23), which measures “the degree to which situations in one’s life are appraised as stressful” (22, p. 385). The State–Trait Anxiety Inventory (24), State version with 20 items was used for this study. The Beck Depression Inventory (25,26) 21-item version was given to participants as the measure of depression. The Beck Hopelessness Scale (27), a 20-item scale, was used to measure hopelessness. Optimism was measured by the Life Orientation Test (28). The measure of social support used was the ESSI (29). It is composed of 5 items with documented predictive validity in cardiovascular patients (α = .86). The Cook Medley Hostility scale was derived from items from the Minnesota Multiphasic Personality Inventory (30,31). Behaviors. Safe sex was measured by the answer to two questions (α = .88) “Have you been practicing safe sex?” and “Have you been using condoms?” (answers were scored 0 = no, 1 = sometimes, and 2 = yes). (These questions were only used if a person answered yes to the question “Have you been sexually active in the last six months?”) Adherence was measured by the AIDS Clinical Trials Group (ACTG) Questionnaire (32), which asks about medications prescribed and missed doses. This instrument was developed by the Adherence and Retention Subcommittee of the ACTG Outcomes Committee and has been used in at least nine clinical trials of combination therapy. The measure we used was the proportion of missed doses in the last 3 days. Telling partner your HIV status was determined by the answer to the interview question “ Do your past and current partners know your HIV status?” (0 = no, 1 = some, and 2 = yes). Smoking was determined by a demographics question “How many packs of cigarettes do you smoke per day?” Finally, alcohol use was determined by the answer to the question “How much alcohol do you consume in a typical week?” The variable “helping others with HIV” was created from three items that related to helping others. As each of the three items were independently related to both group membership and Ironson–Woods SR Index total score, they were combined into one measure. The three items were subjected to a reliability analysis that yielded an alpha reliability coefficient of .62 Cortisol. Each participant collected urine for 15 hr (6:00 p.m.–9:00 a.m.) in plastic containers containing 1 g of sodium metabisulfate preservative. Fifteen-hr urines rather than 24-hr urines were used because compliance was much better among those who work, this collection time has been used in past stress research (33), and because it captures the most important time during which stress-related differences are observed (i.e., overnight). Participants were instructed to keep the urine refrigerated during the collection time. On delivery to the laboratory, the volume of urine was measured, and 10 mL of urine was collected in tubes and stored at –70 °C until it was assayed. Urinary cortisol was determined by radioimmunoassay with Diagnostic
Volume 24, Number 1, 2002 Products (Los Angeles) kits with 50 mL of a 500-mL sample extracted with 10 mL of dichloromethin. The 50 mL of urine was evaporated to dryness under nitrogen. I125 (1 mL)-labeled cortisol was added to tubes coded with antibodies, incubated for 45 min, decanted, and quantified for 1 min with a gamma counter calibrated for I125. Levels of cortisol in the sample were calculated with a standard calibration curve. Cortisol levels were expressed through a log10 transformation of the value expressed as milligrams per 100 mL of urine. The log transformation was utilized to normalize the distribution. Statistical Methodology Selecting covariates. We performed analyses controlling for background variables on which the LTS and COMP groups differed: age, socioeconomic status (SES; education, employment, and income), and route of infection (straight sex vs. other route). Sex, a variable often related to religiousness, was not controlled for because the groups (LTS and COMP) were equivalent. To limit the loss in degrees of freedom, covariates were only retained if they contributed uniquely to the prediction of group membership (LTS vs. COMP). Because age and income did not contribute unique variance to the model, they were dropped as covariates. The remaining covariates (education, employment, and route of infection: heterosexual vs. other) were significantly related to both the Ironson–Woods SR Index (–.13, p < .05; .15, p < .05; .16, p < .01) and LTS status (.18, p < .01; .17, p < .01; –.15, p < .05; n = 279). RESULTS Characterization of the Sample: Demographics The sample was diverse with good representation from different ethnic groups, sexes, and risk categories. As can be seen from Table 1, our sample was approximately 75% male, 25% female, 33% African American, and slightly less than 33% each White and Hispanic. There were no differences between the LTS group (n = 79) and the HIV-positive COMP group (n = 200) on sex, ethnicity, or sexual orientation, which was approximately 55% homosexual, 40% heterosexual, and 5% bisexual. There were significant differences on age and SES (education, employment, and income). The LTS group was significantly older by 3 years, t(277) = –2.72, p < .05; had more education, χ2(3, N = 279) = 10.66, p < .02, were more likely to be on disability and less likely to be working full time, employment χ2(4, N = 279) = 24.02, p < .01; and had a correspondingly lower income, χ2(3, N = 279) = 10.41, p < .02. Medical Information The LTS and COMP groups were comparable on CD4 number (t = .69, ns) and on viral load (t = –1.34, ns). The groups were also equivalent on past sexually transmitted diseases (t = –.33), averaging just over one, and did not differ on proportion of missed doses over the previous 3 days, a measure of adherence to medications (t = –.33). However, the groups did differ on prescribed medications, χ2(3, N = 279) = 14.22, p < .01, with more of the LTS group on PIs and more of
Ironson–Woods SR Index and HIV/AIDS Survival
37
the COMP group not taking medication. This difference probably is a reflection of the entry criteria to the study: LTS members must have had a Category C symptom, whereas COMP group members must not have had a Category C symptom at entry. In terms of past drug use, the LTS group had used more hallucinogens, whereas the COMP group had used more cocaine and cannabis. For the group as a whole, roughly 50% acquired HIV by gay sex, 33% by heterosexual sex, and 5% by intravenous drug use. Acquisition by intravenous drug use was probably underrepresented in this sample because anyone with present drug dependence or any use in the last month was excluded. The two groups differed significantly in that a higher proportion of the COMP group acquired HIV by heterosexual sex, although a higher proportion of the LTS group acquired it by multiple means, χ2(4, N = 279) = 14.73, p < .01; both groups had approximately equal proportions acquiring HIV by gay sex. Overall Religious or Spiritual Information Background information was obtained during the interview on a number of questions relating to religiousness/spirituality. Participants were asked if they were spiritual, religious, both, or neither. Roughly 50% self-identified as spiritual and not religious, roughly 33% identified as both spiritual and religious, 8% identified as primarily religious, and 10% identified as neither. A large proportion (roughly 66%) reported a belief in God; however, another 25% reported being unsure of the existence of God. Approximately 66% believed in heaven or both heaven and hell, and a little more than 66% believed in an afterlife. In terms of organized religious background (religion in family of origin), participants were 37% Catholic, 19% Protestant, 19% Baptist, 5% Jewish, 6% no identification, and 14% other. There were no significant differences between the LTS and COMP groups on any of these background questions. Ironson–Woods SR Index (Short Form) The items on the short form of the Ironson–Woods SR Index together with the loadings (principal components with varimax rotation) on the factors to which each item pertains is presented in Table 2. Twenty-two items chosen from the longer 89-item form of the Ironson–Woods SR Index to reflect the composition of the longer scale were grouped a priori into seven conceptual groupings; these are presented in Table 2 (the a priori conceptual groupings are noted by the indented labels over each a priori grouping). For each item referring to their religious or spiritual beliefs and views, respondents were asked to indicate how strongly they agreed or disagreed with that statement by circling a number from 1 (strongly disagree) to 5 (strongly agree). As noted, the 22 items on the short form were given to 279 people with HIV infection (79 LTS and 200 COMP). The responses on the short form were subjected to a principal components factor analysis with a varimax rotation. The analysis yielded four factors with eigenvalues greater than 1.0, which together accounted for 73% of the variance (54.3, 8.1, 5.4, and 4.9%, respectively). The four factors were named Sense of Peace, Faith in God, Religious Behavior, and Compassionate
38
Ironson et al.
Annals of Behavioral Medicine
TABLE 1 Background Information Demographics Variable Sex Male Female Age* M SD Ethnicity White African American Hispanic Other Education* HS or less HS graduate Some college/Trade College graduate Employment* Full time Part time Unemployed Disability Other Income* < $5k/year $5k–10k/year $10k–20k/year > $20k/year Sexual orientation Homosexual Heterosexual Other
Medical Information
LTS
COMP
74.4% 25.3%
72.4% 27.6%
40.43 7.76
37.72 8.52
31.6% 31.6% 27.8% 9.0%
29.6% 36.2% 29.6% 4.5%
6.3% 7.6% 50.6% 35.5%
16.6% 16.7% 39.9% 26.8%
3.8% 11.4% 7.6% 74.7% 2.5%
17.6% 12.1% 16.6% 44.7% 9.0%
19.0% 48.1% 22.8% 9.1%
30.5% 30.5% 19.5% 12.0%
58.2% 35.4% 6.3%
54.0% 42.4% 3.5%
Variable Immune measures CD4# M SD Viral load M SD Antiretroviral medication* None 1 or no PI ≥ 2, no PI ≥ 1, with PI Past STDs M SD Medications doses missed in past 3 days M SD History of drug use, abuse, or dependence* Sedatives Cannabis* Cocaine* Opioids Hallucinogens* Other drugs History of alcohol abuse or dependence Abuse/Dependence Route of infection* Gay/Bisexual sex Heterosexual sex IV drug use Multiple Other
LTS
276 236.87 73,277 150,555
COMP
300 104.15 45,464 121,812
6.3% 2.5% 20.3% 70.9%
19.5% 1.0% 30.5% 49.0%
1.205 1.28
1.152 1.18
10.61% 0.23
8.80% 0.21
14.6% 43.7% 29.2% 14.6% 23.2% 20.2%
22.7% 62.1% 55.0% 16.7% 12.6% 26.9%
16.7%
26.9%
55.7% 20.3% 5.1% 11.4% 7.6%
53.3% 36.2% 4.5% 3.0% 3.0%
Religion and Spirituality Information Variable Belief in God Yes No Unsure Belief in Heaven or Hell Heaven Heaven and Hell Neither Unsure
LTS
COMP
68.6% 8.6% 22.9%
70.9% 3.9% 25.1%
27.6% 44.8% 17.2% 10.3%
19.9% 41.0% 16.7% 22.4%
Variable Belief in afterlife Yes No Unsure Religious/Spiritual Spiritual Religious Both Neither
LTS
COMP
69.2% 7.71% 23.1%
71.7% 9.0% 19.3%
43.8% 0.0% 43.8% 12.5%
50.0% 10.7% 28.6% 10.7%
Note. LTS = long-term survivor group; COMP = comparison group; CD4 = T-helper/inducer cell counts/mm3; PI = protease inhibitor; STD = sexually transmitted disease; HS = high school; IV = intravenous; k = thousand. *A significant difference between the groups was observed as indicated by either t test or chi-square test significant at p < .1.
Volume 24, Number 1, 2002
Ironson–Woods SR Index and HIV/AIDS Survival
TABLE 2 Items in Ironson–Woods Spirituality/Religiousness Scale With Loading on the Factor to Which Each Item Pertains Factor 1: Sense of Peace Comfort, Strength, Meaning My beliefs give me a sense of peace. My beliefs help me to know everything will be fine. My beliefs give meaning to my life. My beliefs help me to be relaxed. Feeling a Connection, Less Alone My beliefs help me feel protected. My beliefs help me to feel I am not alone. My beliefs help me feel I have a relationship or a connection with a higher form of being. Existential/Afterlife My beliefs help me be less afraid of death. I believe my soul will live on in some form after my body dies. Factor 2: Faith in God View of God I believe God created all things in the universe. God will not turn his back on me no matter what I do. Somatic/Illness Recovery When I am ill, God gives me courage to cope with my illness. When I am ill, God will answer my prayers for a recovery. My beliefs are very influential in my recovery when I am ill. When I am ill, my faith gives me optimism that I will recover. Factor 3: Religious Behavior Religious Behavior I attend religious services. I participate in religious rituals. I pray or meditate to get in touch with God. I discuss my beliefs with others who share my belief. My beliefs give me a set of rules I must obey. Factor 4: Compassionate View of Others View of Others/Compassion for Others My beliefs teach me to help other people who are in need. My beliefs help me feel compassion/love/respect for others. I have a responsibility to help others.a My beliefs increase my acceptance and tolerance of others.a I feel I am connected to all humanity.a
.78 .78 .76 .69 .80 .78 .60
.77 .56
39
tor to which they were assigned a priori, three items did not fall cleanly on one factor. The first of these, “I pray or meditate to get in touch with God,” loaded on both Factor 2 (.58 on Faith in God) and Factor 3 (.46 on Religious Behavior). It was retained on Religious Behavior because it fit better conceptually there and had a reasonably high item-to-total subscale correlation, although it makes sense that it loaded on both factors. Another item, “When I am ill my faith gives me optimism that I will recover,” loaded on both Factor 1 (.57 on Sense of Peace) and Factor 2 (.55 on Faith in God). It was retained on Factor 2 because conceptually it fit better with Faith in God and because it was derived from that portion of the interview that asked how participants’ beliefs influenced their recoveries. The third item, “My beliefs help me feel I have a relationship or a connection with a higher form of being,” loaded on both Factor 1 (.60 on Sense of Peace) and Factor 2 (.50 on Faith in God). It was retained on Factor 1 because of its higher loading and because it did not clearly fit better conceptually on either. Four Factors: Items and Reliabilities
.82 .78
Peace I leave with you. … Do not let your hearts be troubled and do not be afraid. (New International Version; John 14:27)
.78
When [man’s] soul is in peace, he is in peace, and then his soul is in God. In cold or in heat, in pleasure or in pain, in glory or disgrace, he is ever in Him. (Bhagavad Gita 6:7)
.76 .62 .57
.85 .85 .58 .56 .49
.86 (.85) .84 (.81) (.82) (.71) (.68)
Note. Numbers in parentheses indicate factor loadings for Expanded Comparison subscale with resampled participants. aItems added after resampling of participants.
View of Others. Only 1 item had been misplaced by a priori assignment: “My beliefs help me to be relaxed” was originally derived from the somatic or illness recovery portion of the interview (9) but had a higher loading on the Sense of Peace factor. Although for the most part items loaded much higher on the fac-
The first factor was labeled Sense of Peace. Other names considered for this factor were Serenity, Spiritual Comfort, Security, and Sense of Well-Being. One of the definitions of peace from Webster’s dictionary (34) is “freedom from disquieting feelings and thoughts: serenity” (p. 516). This factor includes items from three a priori subscales: Comfort, Strength, Meaning; Feeling a Connection, Less Alone; and Existential/Afterlife. This factor contains notions that life has meaning and that everything will be all right. As can be seen in Table 2, factor loadings on the first factor were all above .56, and six of the nine items had primary loadings above .76. Except for the “connection” item mentioned in the previous paragraph, which loaded on two factors, the highest secondary loading was .33. Now faith is being sure of what we hope for and certain of what we do not see. (Hebrews 11:1). Faith gives the wisdom and grace of knowing that, however conditions and appearances may change, the trueness of life remains always unchanged. (The Teaching of Buddha) You shall love the Lord your God with all your heart, and with all your soul, and with all your might. (Deuteronomy 6:5) God is the source of your strength. (Psalm 68:29) The second factor, Faith in God, was composed from two a priori subscales: View of God and Somatic (how God plays a
40
Ironson et al.
role in recovery). The lowest primary factor loading was .55, and four out of seven items had factor loadings above .76. Except for the faith item mentioned previously, all secondary loadings were below .37. Pray continually … give thanks in all circumstances. (1 Thessalonians 5:17, 18) Observe the Sabbath day and keep it holy. (Deuteronomy 5:12) The third factor, Religious Behavior, was defined by the two items with very high loadings: participating in religious rituals and attending religious services. As noted previously, the prayer item loaded on two factors. The remaining two items also loaded somewhat on other factors, although their highest loading was on Factor 3. Deeds of kindness are equal in weight to all the commandments. (Talmud) Finally, all of you, live in harmony with one another; be sympathetic; love as brothers; be compassionate and humble. (Peter 3:8) Deal kindly with your parents and your kinsfolk, and the poor, as well with the neighbor near of kin as the neighbor not your kin. (Koran 4:36) The final fourth factor, Factor 4, was labeled Compassionate View of Others. Although there were only two items on this factor, the alpha reliability was high (.88), and the items clearly had high loadings (> .84) on this factor and lower loadings (maximum = .32) on secondary factors. Reliability An overall total was also calculated (the sum of all 22 items). Alpha for the total was .96. Alphas for each of the factors were quite high as well: .94 for Sense of Peace, .93 for Faith in God, .85 for Religious Behavior, and .87 for Compassionate View of Others. Robinson, Shaver, and Wrightsman (35) noted that alphas of this size are considered “exemplary” (p. 12). Because the fourth factor only contained two items on the factor analysis, further data collection was done (described in the Expansion of Factor 4, Very Long Survival, and Additional Judgment Factor section). For completeness, the extra items added are listed in Table 2 followed by a superscript a, and their factor loadings are listed in parentheses. Test–retest reliability calculated from initial assessment to the first assessment after the appearance of a Category C AIDS symptom (average of 18 months) for participants developing a Category C symptom (n = 20) was .88** for the total scale, .78** for Sense of Peace, .76** for Faith in God, .62* for Religious Behavior, and .54* for Compassionate View of Others. Robinson et al. (35) rated test–retest reliabilities of this size as excellent.
Annals of Behavioral Medicine Convergent–Discriminant Validity and Correlation With Associated Variables The Ironson–Woods SR Index total score and each of the four factors were correlated with the Hoge and Duke Religiosity scales and the Use of Religion to Cope subscale from the COPE (see Table 3). First, an omnibus test was done to control for overall error rate through the correlation of the Ironson–Woods SR Index Total score with a composite of the Hoge, Duke, and Cope–Religion; a significant correlation of .50** was obtained. The Ironson–Woods SR Index total was significantly correlated with the Hoge Intrinsic scale, the Duke scale, and the Use of Religion to Cope separately as well. In all three cases, the most highly correlated factors from the Ironson–Woods SR Index with the other religion scales were Religious Behavior and Faith in God. (Note that through examination of the correlation between the Ironson–Woods SR Index and the subscales on the Duke scale, it became apparent that organized religion on the Duke was most related to religious behavior on the Ironson–Woods SR Index. Nonorganized religion on the Duke [private religious activities, e.g., prayer, meditation, or Bible study] was correlated much more evenly with the four factors on the Ironson–Woods SR Index.) Finally, the Intrinsic subscale on the Duke was most correlated with Religious Behavior on the Ironson–Woods SR Index. The Hoge Intrinsic subscale correlated most highly with Faith in God followed by Religious Behavior , Sense of Peace, and Compassionate View of Others. Belief in God was most highly correlated with Factor 2 (Faith in God) on the Ironson–Woods SR Index and next with Religious Behavior. Thus, Factors 2 (Faith in God) and 3 (Religious Behavior) appeared to be more closely associated with traditional measures of religiousness, including a belief in God, than did Factors 1 (Sense of Peace) and 4 (Compassionate View of Others). Discriminant Validity and Associated Variables The Ironson–Woods SR Index and its subscales were then correlated with other variables that religiousness/spirituality might be associated with both to determine discriminant validity and to determine what variables might serve as mediators between religiousness/spirituality and health (i.e., long survival). Convergent or discriminant validity was supported by the higher correlations of the Ironson–Woods SR Index with the three measures of religiousness noted previously as compared with correlations of the Ironson–Woods SR Index with other variables. Because of the large number of psychosocial measures (six), an omnibus test was done first to correlate the Ironson–Woods SR Index total score with a psychosocial composite score (the sum of all of the psychosocial measures), r = –.31, p < .01, which was then followed by a test of individual correlations. The Ironson–Woods SR Index did correlate significantly and fairly strongly with less distress (perceived stress, less affective distress [anxiety, depression], less physiologic stress [cortisol]), more hope (less hopelessness, more optimism), more social support (ESSI), better health behaviors (report of practicing safe sex, telling partner HIV status [Factor 4 only], and less smoking and drinking alcohol), and help-
Volume 24, Number 1, 2002
Ironson–Woods SR Index and HIV/AIDS Survival
41
TABLE 3 Correlations of Ironson–Woods Spirituality/Religiousness Index With Religious Measures, Psychosocial Measures, Health Behaviors, and Cortisol Ironson–Woods Spirituality/Religiousness Index Variable Religious Measures Hogea Dukea Organized Rel. Non-organized Rel. Intrinsic COPE–Religion Belief in God Psychosocial measures Perceived stress Hopelessness Optimism Anxiety Depression Social support (ESSI) Behaviors Helping others Safe sex Tell partner HIV+ Smoking Drinking alcohol Biologic Cortisol
Total .66** .60** (.34*) (.48**) (.57**) .70** .27**
F1: Peace .47** .41** (.15) (.42**) (.37*) .62** .19**
F2: Faith .67** .42** (.18) (.36*) (.40**) .63** .33**
F3: Behavior .61** .71** (.68**) (.41**) (.64**) .67** .27**
F4: Compassion .44** .35** (.10) (.37*) (.32*) .39** .00
Multiple R
Factors
.71 .74
2, 3 3
.72 .40
3, 1, 2 2, 4
–.21** –.45** –.38** –.30** –.29** .33**
–.28** –.48** –.43** –.36** –.34** .33**
–.10 –.40** –.30** –.20** –.21** .32**
–.14** –.30** –.25** –.23** –.18** .24**
–.16** –.21** –.22** –.21** –.16** .16*
.34 .49 .44 .39 .35 .34
1, 2 1 1 1, 2 1 1
.24** .25** .04 –.44** –.24**
.21** .25** .01 –.43** –.26**
.17** .21* .06 –.53** –.22**
.28** .21* .01 –.27** –.10
.32** .12 .14* .01 –.22**
.37 .29 .18 .59 .34
4 1 4 2, 4 1
–.19*
–.27**
–.16*
–.14
–.13
.22
1
Note. Numbers in parentheses represent subscale correlations. F = factor; Hoge = Hoge Intrinsic Religious Motivation Scale; Duke = Duke Religion Index; Intrinsic = Intrinsic Religiosity; ESSI = Enhancing Recovery in Coronary Heart Disease Social Support Instrument. an = 49 for correlations between Ironson–Woods Spirituality/Religiousness Index, Hoge, and Duke. All other correlations, n = 279, except cortisol n = 179 and safe sex n =160. *p < .05. **p < .01.
ing others with HIV. The Ironson–Woods SR Index was not correlated significantly with medication adherence (not included in table). The unique contribution of each different factor is discussed later. Differences Between the LTS and COMP Groups on Ironson–Woods SR Index and Religious Behaviors Table 4 shows the comparison of the two groups (LTS and COMP) on the Ironson–Woods SR Index and on religious behaviors and addresses the question of whether the long survivors were higher on SR than the HIV-positive COMP group. The LTS group was significantly higher on total score (p = .02), Faith in God (p = .04), Ironson–Woods SR Index Religious Behavior (p = .01), and Compassionate View of Others (p = .02) on two- tailed tests and on Sense of Peace on a one-tailed test (p = .05). (We felt justified in using a one-tailed test because the direction of the hypothesis tested—namely, higher religiousness—would be related to longer survival and could be stated a priori based on the McCullough et al. [6] review.) We performed these analyses controlling for background variables on which the LTS and COMP groups differed: SES (education, employment) and route of infection
(straight sex vs. other route). The groups were also compared on the frequency (in the last month) of religious behaviors (praying, meditating, and going to services). The LTS group was significantly higher on the frequency of these reported behaviors. A post hoc analysis of each behavior separately showed the LTS group was significantly higher than the COMP group on reported prayer (23.54 vs. 18.94; partial r = .15; t = 2.37, p = .02). There was a nonsignificant trend for meditating (14.94 vs. 11.14, p = .15) and attending services (5.14 vs. 2.82; t = 1.81, p = .07, two-tailed, significant on a one-tailed test at p = .035). The remainder of Table 4 is covered in the Expansion of Factor 4, Very Long Survival, and Additional Judgment Factor section. Consideration of Mediators Next, mediators of the relation between the Ironson–Woods SR Index and long survival (i.e., group membership, LTS vs. COMP) were explored with regression analysis as detailed by Baron and Kenny (36). We previously hypothesized that mediators for long survival might be due to physiologic protection (i.e., low cortisol and those psychological states that might be expected to be related to reductions in stress hormones) or changes in health-related behaviors (adherence to medications,
42
Ironson et al.
Annals of Behavioral Medicine
TABLE 4 Means and Standard Deviations for LTSa and COMPb Groups on Ironson–Woods Spirituality/Religiousness Index and Specific Reported Religious/Spiritual Behaviors No. of Items Index total M (SD) Factor 1: Sense of Peace M (SD) Factor 2: Faith in God M (SD) Factor 3: Religious Behavior M (SD) Factor 4 (short): Compassionate View of Others M (SD) Factor 4 (long): Compassionate View of Otherse M (SD) Reported religious behaviors last monthf M (SD)
LTS
COMP
89.33 (19.38)
84.04 (20.77)
37.50 (7.63)
35.78 (9.35)
29.63 (6.63)
23.64 (6.75)
17.91 (5.80)
15.94 (6.00)
9.19 (1.49)
8.69 (1.78)
22.10 (2.97)
19.89 (4.31)
44.33 (41.36)
32.77 (35.67)
22
9
6
5
2
5
3
tc
p
–2.43*
.02
–1.67
.05d
–2.07*
.04
–2.81**
.01
–2.36*
.02
–2.08*
.04
–2.13*
.04
Note. LTS = long-term survivor group; COMP = comparison group. an = 79. bn = 200. cControlling for background variables (education, employment, route of infection). dOne-tailed test (all others are two-tailed). eData from the very-long-term survivor subsample (n = 21) and the COMP subsample (n = 49) obtained 1 to 3 years later (October–November 2000). fPrayer/meditation/services.
safe sex, smoking, and alcohol use). We also considered as mediators measures of variables related to proposed benefits of spiritual and religious beliefs outlined by Covey (16). These included social support; avoidance of drugs, cigarettes, and alcohol; and reasons to be of service and to help others. The first prerequisite for consideration as a mediator was that the mediator had to be related to both the predictor (in this case, the Ironson–Woods SR Index) and to the outcome (long survival operationalized as group membership: LTS = 1 vs. COMP = 0). Table 3 lists 12 of the variables (psychosocial measures, behaviors, and cortisol) included as part of the examination of convergent or discriminant validity. These 12 variables were also considered as potential mediators. Of the 10 variables significantly related to the Ironson–Woods SR Index, only 3 were also significantly related to long survival. Thus, only 3 variables (urinary cortisol concentration, helping others with HIV, and optimism) were significantly correlated with both long survivor status (LTS vs. COMP group membership) and total score on the Ironson–Woods SR Index and therefore were examined as putative mediators of that relation (see Figure 1). After it is determined that a mediator is significantly related both to the predictor (Ironson–Woods SR Index) and the dependent variable (long survival status), there is the final test of a mediator: When the independent variable (Ironson–Woods SR Index) and the mediator are both in the regression equation predicting the dependent variable (long survival), the independent
variable (Ironson–Woods SR Index) is no longer significant (36). When group membership was regressed on cortisol concentration and the Ironson–Woods SR Index, the relation between the Ironson–Woods SR Index and group membership became nonsignificant, t(176) = .21, p > .05, and the relation between cortisol and group membership maintained its significance, t(176) = –2.36, p < .05. Therefore, cortisol was a mediator. When group membership was regressed on both helping others with HIV and the Ironson–Woods SR Index total score, the direct relation between Ironson–Woods SR Index and group membership was no longer significant, t(276) = 1.10, p > .05, and the significance of the relation between helping others with HIV and group membership was maintained, t(276) = 3.51, p < .01. Therefore, helping others with HIV also met the criteria for a mediator. When group membership was regressed on optimism and Ironson–Woods SR Index total score, the direct relation between the Ironson–Woods SR Index and long survival status was no longer significant, t(223) = 1.79, .05 < p < .10, but the relation between optimism and group membership was also no longer significant, t(223) = 1.31, p > .05. Consequently, optimism did not meet the criteria for mediation. We repeated the previous analyses controlling for background variables (SES [education and employment] and route of infection) on which the groups differed significantly (see Methods section, Statistical Methodology: Selecting Covariates). Measures of the direct relation between the
Volume 24, Number 1, 2002
Ironson–Woods SR Index and HIV/AIDS Survival
43
FIGURE 1 (a) Cortisol, (b) helping others, and (c) optimism as mediators of the relationship between spirituality and long survivor status (i.e., long-term survivor [LTS] group or the HIV-positive comparison [Comp] group) when significant background variables (education, employment, and route of infection) were controlled. The numbers in parentheses are partial correlations indicating the unique contribution of the independent and mediator variables when the variance associated with background variables has been removed. The numbers outside parentheses are standardized Beta coefficients in the model with Long Survival (Group Membership LTS vs. Comp) as the dependent variable and spirituality and the respective mediator as the independent variables.
Ironson–Woods SR Index and group membership were provided with the standardized beta coefficients from the regression equation when the control variables (education, employment, and route of infection) were entered as the first block and the Ironson–Woods SR Index was entered as the second block. The significance of the direct relation between the Ironson–Woods SR Index and group membership was maintained, t(272) = 2.43, p < .05. Similarly, the significance of the relations between the mediators and group membership was also maintained for cortisol, t(174) = –2.33, p < .05, helping others with HIV, t(272) = 3.09, p < .01, and optimism, t(220) = 2.14, p < .05. For the final test of mediation, the standardized beta coefficients were calculated separately when group membership was regressed on the socioeconomic and route of infection variables, the Ironson–Woods SR Index, and the respective mediator variable. Urinary cortisol met the criteria for mediation as the direct relation between the Ironson–Woods SR Index and group membership was no longer significant, t(174) = .46, p > .05, and the significance of the relation between cortisol and group membership was maintained, t(174) = –2.33, p < .05. Similarly, helping others with HIV also met the criteria for mediation; the direct relation between the Ironson–Woods SR Index and group membership was no longer significant, t(271) = 1.07, p >.05, and the relation
between helping others with HIV and group membership maintained its significance, t(272) = 3.09, p < .01. Optimism failed to meet the requirements for mediation. When group membership was regressed on the demographic and medical variables, optimism, and the Ironson–Woods SR Index, the direct relation between the Ironson–Woods SR Index and group membership was maintained, t(219) = 2.18, p < .05, and the relation between optimism and group membership was no longer significant, t(220) = 1.08, p > .05. The results of these analyses confirmed that both urinary cortisol concentration and helping others with HIV mediated the relation between the Ironson–Woods SR Index total score and group membership. It was not possible to specify the directionality of these findings. Expansion of Factor 4, Very Long Survival, and Additional Judgment Factor Because the factor analysis showed a fourth factor (Compassionate View of Others) with only 2 items, the items relevant to a view of others from the long Ironson–Woods SR Index were given to a subsample of LTS and COMP groups who could be relocated in October 2000. This was roughly 3.5 years after the initial administration, which began in February 1997. There were
44
Ironson et al.
13 items on the long Ironson–Woods SR Index Compassionate View of Others scale and an additional 8 items relevant to a view of others. To do a factor analysis on 21 items, we reasoned that we needed a minimum of 70 people (roughly 3.5 times the number of items). We contacted as many of the original LTS participants as we could (n = 21) and 49 members of the COMP group to fill out the longer questionnaire. (The recontacted LTS participants are referred to as very-long-term survivors [VLTSs] because the timing represented an average of 3 years [38.6 months] after the original assessment, so they were now all at least 7 years past their initial opportunistic infection or neoplasm.) We expanded the original 2-item Compassionate View of Others scale to 5 items by choosing items with high item-to-total correlations. Because the 2 items were also both heavily endorsed and high on social desirability, an additional consideration was to add items with lower means to maintain reliability and increase variability. The difference between the LTS and the COMP groups on the 5-item measure was then retested with the VLTS (n = 21) versus COMP (n = 49) group comparison and found to maintain its significance (t = –2.08, p = .04; see Table 4). The factor analysis of the 21 items revealed a strong first factor explaining 39% of the variance and a second factor explaining 11% of the variance. The first factor corresponded to Compassionate View of Others; the factor loadings for the 5-item scale with the new 3 items are presented in Table 2. “Judge not lest ye be judged” (Matthew 7:1). The second factor was named Judgmental. The Judgmental items were as follows, with their respective factor loadings: “Some people will go to hell when this life is over” (.83), “Only those who believe what I believe will go to heaven when they die” (.85), and “God will condemn those who do wrong someday” (.83). This factor was also significantly related to long survival in a negative direction; respective means for the VLTS and COMP groups were 6.14 (SD = 3.28) and 8.57 (SD = 3.52); t = 2.67, p = .01. Thus, the VLTS group was significantly lower on being judgmental than the HIV-positive COMP group. The Judgmental factor had good internal consistency (α = .82). Interestingly, the factors (Compassionate View of Others and Judgmental) were not correlated (r = –.05), indicating it is possible to be compassionate toward some people and condemning and judgmental of others. Of additional interest, the Judgmental factor was highly correlated with hostility (r = .49, p < .01, n = 70), whereas Compassionate View of Others was negatively correlated with hostility (r = –.14, ns). An additional item from the long form of the Ironson–Woods SR Index was also related to judgment, but judgment of self rather than of others. This item reads, “God will judge me harshly one day.” Its correlation with the three-item Judgmental scale was .58. Adding this statement to the three-item scale resulted in a four-item Judgmental scale with an alpha of .83. This four-item Judgmental factor was also significantly related to long survival in a negative direction; respective means for the VLTS and COMP groups were 8.38 (SD = 4.80) and 10.96 (SD = 4.38), t = 2.17, p = .03. Thus, religiously based harsh self-judgment may also have negative health consequences.
Annals of Behavioral Medicine Which Comes First: The Chicken or the Egg? Are people who are more spiritual or religious more likely to become long survivors, or do people become more religious when they get sick? Because the data presented are cross-sectional in nature, one might raise the question as to whether when one develops a serious AIDS symptom (i.e., Category C), one becomes more religious. (Recall our long survivors all had a Category C symptom, whereas none of the COMP group did.) This question can only be answered longitudinally. Fortunately, the COMP group is being followed longitudinally. Twenty-two of the 200 people in the COMP group (who initially had no past or present Category C symptoms) did develop Category C symptoms over the course of the study. We have post-Category-C data on 20 of these people. There was no significant change from initial assessment to the post-Category-C assessment for the total Ironson–Woods SR Index score (t = –.61, ns) or the Sense of Peace (t = –.66, ns), Faith in God (t = –.39, ns) Religious Behavior (t = –.34, ns), or Compassionate View of others (t = –.72, ns) subscales. Thus, it is not likely that the major findings of higher Ironson–Woods SR Index and its factors (Faith in God, Religious Behavior, and Compassionate View of Others on two-tailed tests and Sense of Peace on a one-tailed test) were due to changes in participants after they got sick. Are Four Separate Factors Necessary on the Ironson–Woods SR Index? Two factors capture spirituality (Sense of Peace and Compassionate View of Others), and two factors capture religiousness in the more traditional sense (Faith in God and Religious Behavior). In addition, one of the religious factors is more private (Faith in God), whereas the other (Religious Behavior) is more public, and both of the other factors (Sense of Peace and Compassionate View of Others) can be viewed as more private and relevant for people high in either spirituality or religiousness. In addition to content differences, differences in relevance to spiritual or more traditionally religious orientations, and differences in private versus public orientation, the factors are also related uniquely to different outcomes. Factor 1 (Sense of Peace) is the factor most strongly related to physiologic stress (low cortisol), affective distress (perceived stress, anxiety), and hope (optimism, hopelessness). Factor 2 (Faith in God) is most strongly related to a traditional belief in God and is also related to less distress, more hope, lower cortisol (although less strongly than Factor 1 but through a more traditional religious orientation) and long survival. Factor 3 (Religious Behavior) is most strongly traditionally religious and is significantly related to long survival. Factor 4 (Compassionate View of Others), a dimension that can be regarded as relevant to being either spiritual or religious, is significantly related to long survival (and to very long survival) and is most strongly related to helping others and to telling one’s partner of one’s HIV status. The same pattern of factors being related to outcomes is reflected in the Regression Factors column of Table 3. We obtained a forward regression for each outcome, adding in only those factors contributing significantly to the model. What is most striking about the data is that in 9 of 12 psychological and health out-
Volume 24, Number 1, 2002 comes, Factor 1 (Sense of Peace) was the first factor entered. A second interesting observation is that religious behavior did not contribute uniquely to the relation with outcomes when the other factors (Sense of Peace, Faith in God, and Compassionate View of Others) were present. However, this analysis was repeated for long survival status (data from Table 4). Here, the factors contributing significantly to the model were Factors 3 (Religious Behavior) and 4 (Compassionate View of Others). Next, a series of regression analyses were conducted to determine if the individual factors added any variance over the total score. In 12 of the 16 regressions, the individual factors added significant variance over what was explained by the total Ironson–Woods SR Index score alone (compare the Multiple R column in Table 3 to the Total column to estimate the size of the effect). (No variance was added for social support, safe sex, drinking alcohol, or cortisol.) Thus, the individual factors did contribute, but for many variables, the total Ironson–Woods SR Index captured most of the variance. Is Asking About Religious Behavior Alone Sufficient? Another question that arises from the literature is whether asking about religious behavior alone is sufficient (6,8). For the 12 psychosocial, behavioral, and physiologic (cortisol) variables in Table 3, regressions were done to determine whether the three other factors (Factors 1, 2, and 4) added beyond religious behavior in the prediction of the 12 variables. For 9 of the 12 regressions, at least one factor added to religious behavior, and in 8 of these it was Factor 1 (Sense of Peace). Thus, asking about religious behavior alone left out significant variance, and that variance particularly seemed to be related to the sense of peace with which spiritual or religious beliefs may be associated. Are the Benefits of Religious Behavior Accounted for by Social Support? One might hypothesize that attending services and participating in the activities of a religious community could be associated with more social support than private aspects of spirituality (peace, faith, compassion).To test whether the religious behavior–health relations might be driven by social support, we examined the association of Religious Behavior with the 12 outcomes in Table 3 controlling for social support. Most (6 of the 8 that were significant before, not including social support) of these relations were not driven by social support (i.e., they remained significant when social support was controlled; partial rs = –.25** for hopelessness, –.19** for optimism, –.16* for anxiety, .25** for helping others, .19* for safe sex, and –.24 for smoking). Only perceived stress (–.07, ns) and depression (–.11, ns) were driven at least partially by social support. In addition, the relation between Religious Behavior and long survival remained significant when social support (partial r = .17**) was controlled. Thus, the social support associated with religious behavior may help people to be less distressed, but social support associated with religious behavior does not appear to drive most of the religious behavior–health relations studied.
Ironson–Woods SR Index and HIV/AIDS Survival
45
DISCUSSION One main purpose of the study was to attempt to develop a measure that would capture both spirituality and religiousness and that would include both private and public aspects of spirituality and religiousness. The four dimensions appear conceptually and statistically to achieve this objective. Of the four separate factors on the Ironson–Woods SR Index instrument, two capture aspects of spirituality and/or religiousness (Sense of Peace and Compassionate View of Others) and two capture aspects of religiousness in the more traditional sense (Faith in God, Religious Behavior). In addition, one of the religious factors is more private (Faith in God), whereas the other (Religious Behavior) is more public, and both of the other factors (Sense of Peace, Compassionate View of Others) can be viewed as private and are relevant for those who might consider themselves both spiritual and religious. Our finding that long survivors were significantly higher than the HIV-positive COMP group on religious behavior is very consistent with prior reviews of the literature. However, our study extended this finding to an AIDS population (6) and to factors other than religious behavior alone. This prior literature has looked at the religion–mortality association primarily in healthy, community-dwelling populations. Thus, those with private religious or spiritual feelings, such as faith or compassion for others, also had an associated long survival. In addition, having a sense of peace was strongly related to (lower) cortisol, suggesting that physiologic benefits might accrue from beliefs that are not necessarily a part of any organized religion. Thus, health benefits might come from both organized and nonorganized spiritual or religious beliefs and behaviors. In regard to the question raised by the McCullough et al. (6) meta-analytic review, “Is it the positive psychological states that may be fostered by more private forms of worship or measures of public religious involvement that provide the protective health effects?,” our answer may be “both.” In particular, peace was associated with both more positive psychological states (optimism, lower hopelessness, less distress and anxiety) and with lower cortisol (and with long survival on a one-tailed test). Faith was also associated with more positive psychological states, lower cortisol, and long survival. Public religious involvement was associated with both long survival and one of the mediators (i.e., helping others). In conclusion, our correlations, the regression analyses, and the analysis of mean differences between the long survivors and the COMP group made it clear that much is added by the measurement of other factors besides religious behavior alone and that protective effects on health are associated with other dimensions of religiousness/spirituality besides religious behavior. This is particularly noteworthy because many prior studies have used a single-item measure of church attendance. Another related question from the literature raises the question of whether social support (i.e., a behavioral aspect of religiousness) versus non-social-support-related religiousness/spirituality (e.g., faith) is a critical component of the longevity or health outcomes. If social support were a critical component of religious or spiritual effects on health, one would expect the more private aspects of religion (faith) not to relate to health,
46
Ironson et al.
whereas the more public aspect would. The observation that both public and private aspects of religion relate equally strongly suggests that the religion–health effects are not primarily accounted for by social support. Our additional analysis, showing that the religious behavior–health (long survival and other outcomes) association remained significant even when social support was controlled, lends further support to the notion that religious or spiritual effects on health do not appear to be driven primarily by social support. However, our analyses also showed that the social support associated with religious behavior may help people to be less distressed. How might greater religiousness/spirituality be affecting health in those with HIV? We explored a number of possible mediators. Only three were related to both religiousness/spirituality and long survival status: cortisol, helping others with HIV, and optimism. The strongest support was found for low cortisol and helping others as links between religiousness/spirituality and long survival. This link is supported by literature showing that cortisol, a neurohormone associated with stress, enhances the ability of HIV to infect normal human lymphocytes (37) and is also associated with down-regulation of the immune system (38,39). The second mediator, helping others with HIV, is very consistent with literature showing that both strength and comfort from religion and social participation are associated with health outcomes (40). The combination of social participation and religious comfort may be a particularly potent one, especially for people with HIV who are often stigmatized and may have to redefine new social networks and “families.” Helping others with HIV/AIDS may get one involved in a network of supportive people where one is accepted for who one is. There are many studies supporting an association between social support and better health outcomes. Religion may also provide another route for increased social participation because religious coping is frequent among people with HIV (41). Finally, although optimism was significantly related to both SR and long survival status, it did not meet the statistical criteria for a mediator. Additionally, hostility remains a potential mediator (negatively) to be investigated by future studies. In our sample, harsh judgment was significantly correlated both with long survival (negatively) and higher hostility. A similar construct, punishing God reappraisals, was found to be associated with poorer adjustment in another sample (42). Furthermore, although hostility was not directly (negatively) related to long survival, we previously showed that in those with distressing life events, high hostility was longitudinally (over 6 months) related to faster disease progression (greater decline in CD4 cells and greater increase in viral load [43]). Thus, hostility does remain a viable candidate for a mediator, and both this variable and the others should continue to be considered and tested as mediators in longitudinal studies. Beyond hostility is the related Judgmental factor, through which religion, perhaps better termed in this case religiosity, may be negatively associated with health outcomes. Others have identified negative mental and physical health consequences of some aspects of religion. Schumaker (44) listed 10 ways in which religion can be deleterious to mental health—for example, by encouraging guilt or self-denigration. Negative religious coping has
Annals of Behavioral Medicine been defined as seeing a crisis as punishment from God or as calling into question God’s power or love (42). Such negative effects of religion on attitudes and affect, particularly guilt and shame, may be especially applicable to persons with HIV/AIDS. It has been documented that a minority of ministers, generally those older and less well educated, believe that AIDS is a punishment from God and that people with AIDS deserve their illness (12). Correspondingly, 17% of an AIDS population felt that the illness represented punishment from God (41). Thus, our finding of a negative association between religion-based judgmentalness and survival is in concert with other observations of potentially negative health effects of aspects of religion. A key question in this research was whether people who survived longer were more religious to begin with or whether they became more religious when they got sick (the chicken-or-the-egg question). Fortunately, we had a small group of people whom we were able to ask about their religiousness/spirituality before and after getting sick with a serious AIDS symptom. It is not likely that the major findings of higher religiousness/spirituality being associated with long survival was due to changes in these variables after sickness appeared because, in our sample, no factors changed (increased) after sickness (a Category C symptom) appeared. However, in view of the importance of this question, Gail Ironson and David Larson recontacted the LTS participants and the next 20 people in the COMP group being seen as part of a longitudinal study and asked them to fill out questionnaires assessing their religiousness and spirituality at different times in their lives, including the year before they found out they were HIV positive, the year after they became HIV positive, and the years before and after they were diagnosed with AIDS (on a scale of 1 to 10, where 1 represents not religious/spiritual at all and 10 represents very religious/spiritual). Our preliminary data (ns = 23 and 20 for the LTS and COMP groups, respectively) showed that the LTS group was more religious or spiritual than the COMP group in the year before contracting HIV (Ms = 5.8 and 3.7 for the LTS and COMP groups, respectively; t = –2.27, p = .03). In addition, only the LTS group increased significantly in religiousness in the year after becoming HIV positive (Ms = 7.2 and 4.3 for the LTS and COMP groups, respectively; t = –3.09, p = .01; increase for LTS, t = –3.30, p = .01; change for COMP, t = –.74, ns). No further change occurred in the LTS group after getting AIDS (M = 7.1, t = .83, ns). The two findings (prospectively showing there was not an increase in religiousness after an AIDS-defining Category C symptom and the retrospective finding just noted) together suggest the big increase in religiousness/spirituality may come in the year after finding out one is HIV positive rather than after the appearance of a Category C AIDS-defining symptom. So the preliminary answer is “both”; the long survivors were more religious to begin with, and they became more religious when they discovered they had a serious illness (HIV). Future research is certainly needed in this area, particularly because it may represent a time when interventions may be helpful to facilitate a turn to religion as a source of comfort. Another time when people may turn to spirituality and religion relevant to HIV is during bereavement. Folkman and col-
Volume 24, Number 1, 2002 leagues (45) found that spirituality increased in 77% of a caretaking cohort after the loss of a partner from AIDS. These participants’ spiritual beliefs and experiences provided emotional and cognitive resources for coping with the high levels of distress associated with the loss. Religious institutions might be a particularly good venue to engage low-SES people where they can be reached. Low SES has long been associated with poorer health (46). Confirming this, in this study, higher education was related to long survival. However, lower education was related to higher religiosity. Taken together, these findings suggest that religion is one factor that low-SES people may use that is protective of their health. Religious settings may be a place to reach low-SES people to have a positive protective impact on their health. Further Issues and Limitations There are a number of limitations to this study. First, we had to select a small number of 22 items from the longer Ironson–Woods SR Index (17), which has 89 items, to give to our original sample due to practical considerations (participant burden). Therefore, although the four subscales, which were derived, are correlated with both psychological and physical health outcomes, they may not cover the gamut of dimensions of the complex concept of religiousness/spirituality (see 9 for a more complete description). As we noted, further work with a subset of 21 items from the 89-item version having to do with compassion for others and view of others uncovered another dimension (Judgmental). When a large enough sample is obtained with the full 89 items, more dimensions may be discovered. Second, although we included the Hoge (18) and Koenig, Meador, and Parkerson (19) scales for cross-validation purposes, many other scales could not be included, partly due to practical considerations (i.e., participant burden or scale not available at time our study began in 1996). These include (but are not limited to) two that may be of particular interest in health psychology: Pargament’s scale measuring methods of religious coping (42) and the Daily Spiritual Experience Scale developed by a Fetzer–National Institute on Aging effort (47). The similarities and differences between our scale and these scales and concurrence in outcomes prediction could be a topic for future research. A third limitation of this study is that cross-sectional data (i.e., cross-group comparison of the LTS vs. COMP groups) do not lend themselves to statements about causality. Thus, psychosocial factors (e.g., optimism) that might influence survival could also shape the responses on the spirituality questionnaire. Those psychosocial variables could be the drivers rather than the results of the levels of SR. Similarly, a fourth limitation is that in an attempt to determine whether people become more spiritual or religious as a result of the illness or whether more spiritual or religious people are more likely to stay healthy, we used a combination of retrospective and prospective reporting. Retrospective reporting is subject to obvious possible biases as subsequent events may shape the participant’s memory. Prospective reporting is much sounder but harder to do. A fifth limitation of the study is that these results are limited to the HIV population. It is important to note that the questions
Ironson–Woods SR Index and HIV/AIDS Survival
47
for the instrument were developed from interviews with HIV, cancer, and cardiac patients (9,17) and, therefore, are seen as relevant to other health populations as well. However, whether our findings with the HIV population extend to other health populations remain for future studies to determine. Although there are important differences between the populations (e.g., in terms of religiousness/spirituality, the AIDS population is more alienated from traditional religion than cancer or cardiac patients [9]), we hypothesize that similar relations would be found for other health groups as well and are currently testing this in a cancer population. A final limitation is that we could not examine all possible explanations for a religion–health relation. Although individuals may hope divine intervention is possible, the effects of social support, dietary practices, and the possibility that healthier people are better able to attend religious services need to be ruled out as alternative explanations. Finally, it is noteworthy that dimensions of religiousness/spirituality are related not just to long survival but to a large number of beneficial psychosocial measures (distress, hope, optimism, less anxiety), beneficial behaviors (safe sex, less use of alcohol, less smoking, more willingness to tell partner of HIV status), and physiology (less cortisol). Future research should look at how and under what conditions inner transformation toward increased religiousness/spirituality occurs. Such research can also address further the question of how becoming more responsible (e.g., safe sex, less use of drugs or alcohol), being more compassionate and optimistic, and feeling more at peace may be associated with improved health. The occurrence of a major medical illness such as HIV may represent a time when people, thinking about their own mortality, may turn to religion for comfort. It may also represent a potential time for change to occur. In conclusion, both religiousness and spirituality, although rarely assessed, remain an important force in the lives of those with HIV, regardless of many of these people having been rejected by traditional religion. They still remain religious or spiritual, and this religiousness/spirituality is related to a number of beneficial outcomes including less affective distress, lower cortisol, and long survival.
REFERENCES (1) Kingsley D: Health, Healing, and Religion. A Cross-Cultural Perspective. Upper Saddle River, NJ: Prentice Hall, 1996. (2) Shealy CN: Sacred Healing: The Curing Power of Energy and Spirituality. Boston: Element, 1999. (3) Dossey L: [Dust jacket]. In Shealy CN, Sacred Healing: The Curing Power of Energy and Spirituality. Boston: Element, 1999. (4) Levin J, Schiller P: Is there a religious factor in health? Journal of Religion and Health. 1987, 26(1):9–36. (5) Levin J: Religion and health: Is there an association, is it valid, and is it casual? Social Science and Medicine. 1994, 38:1475–1482. (6) McCullough ME, Hoyt WT, Larson DB, Koenig HG, Koenig HG: Religious involvement and mortality: A meta-analytic review. Health Psychology. 2000, 19:211–222. (7) Koenig HG, McCullough ME, Larson DB (eds): Handbook of Religion and Health. New York: Oxford University Press, 2002.
48
Ironson et al.
(8) Larson DB, McCullough ME, Swyers J: Consensus Report on Spirituality & Health. Rockville, MD: National Institute for Healthcare Research, 1998. (9) Woods TE, Ironson GH: Religion and spirituality in the face of illness. Journal of Health Psychology. 1999, 4:393–412. (10) Holt JL, Houg BL, Romano JL: Spiritual wellness for clients with HIV/AIDS: Review of counseling issues. Journal of Counseling and Development. 1999, 77:379–399. (11) Wagner GJ, Serafini J, Rabkin J, et al.: Integrations of one’s religion and homosexuality: A weapon against internalized homophobia? Journal of Homosexuality. 1994, 26(4):91–110. (12) Crawford I, Allison KW, Robinson WL, et al.: Attitudes of African-American Baptist ministers towards AIDS. Journal of Community Psychology. 1992, 20:304–308. (13) Jenkins RA: Religion and HIV: Implications for research and intervention. Journal of Social Issues. 1995, 51(2):131–144. (14) Levin JS, Taylor RJ: Panel analyses of religious involvement and well-being in African Americans: Contemporaneous vs. longitudinal effects. Journal for the Scientific Study of Religion. 1999, 695–709. (15) Ironson G, Solomon G, Cruess D, Barroso J, Stivers M: Psychological factors related to long-term survival with HIV/AIDS. Clinical Psychology and Psychotherapy. 1995, 2(4):249–266. (16) Covey SR: Spiritual Roots of Human Relations. Salt Lake City, UT: Deseret, 1995. (17) Ironson G, Woods T: IWORSHIP. An Unpublished Assessment Instrument of Religious and Spiritual Dimensions. Coral Gables, FL: University of Miami, 1998. (18) Hoge DR: A validated Intrinsic religious motivation scale. Journal for Scientific Study of Religion. 1972, 11:369–376. (19) Koenig HG, Meador K, Parkerson G: Religion Index for Psychiatric Research: A 5-item measure for use in health outcome studies. American Journal of Psychiatry. 1997, 154:885–886. (20) Koenig HG, Futterman A: Religion and health outcomes: A review and synthesis of the literature. Proceedings of the Conference on Methodological Approaches to the Study of Religion, Aging, and Health, 1995. (21) Carver CS, Scheier MF, Weintraub JK: Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology. 1989, 56:267–283. (22) Cohen S, Kamarck T, Mermelstein R: A global measure of perceived stress. Journal of Health and Social Behavior. 1983, 24:385–396. (23) Cohen S, Williamson GM: Perceived stress in a probability sample of the United States. In Spacapan S, Oskamp S (eds), The Social Psychology of Health. Newbury Park, CA: Sage, 1988. (24) Spielberger et al.: STAI Manual for the State Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologist Press, 1970. (25) Beck AT: Depression: Causes and Treatment. Philadelphia: University of Pennsylvania Press, 1967. (26) Beckham EE, Leber WR (eds): Handbook of Depression: Treatment, Assessment, and Research (Appendix 3). Homewood, IL: Dorsey, 1985. (27) Beck AT, Weissman A, Lester D, Trexler L: The measurement of pessimism: The Hopelessness Scale. Journal of Consulting and Clinical Psychology. 1974, 42:861–865. (28) Scheier MF, Carver CS: Optimism, coping and health: Assessment and implications of generalized outcome experiences. Health Psychology. 1985, 4:219–247.
Annals of Behavioral Medicine (29) The ENRICHD Investigators: Enhancing Recovery in Coronary Heart Disease Patients (ENRICHD): Study design and methods. American Heart Journal. 2000, 139:1–9. (30) Cook WW, Medley DM: Proposed hostility and pharasaic virtue scales for the MMPI. Journal of Applied Social Psychology. 1954, 13:99–125. (31) Barefoot JC, Dahlstrom WG, Williams RB: Hostility, CHD, incidence, and total mortality: A 25-year follow-up study of 255 physicians. Psychosomatic Medicine. 1983, 45:59–63. (32) Chesney MA, Ickovics JR, Chambers DB, et al.: Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: The ACTG adherence instrument. AIDS Care. 2000, 12(3):255–266. (33) Baum A, Gatchel RJ, Schaeffer MA: Emotional, behavioral, and physiological effects of chronic stress at Three Mile Island. Journal of Consulting and Clinical Psychology. 1983, 51:565–572. (34) Webster’s II: New Riverside Dictionary. New York: Berkley, 1984. (35) Robinson JP, Shaver PR, Wrightsman LS: Criteria for scale selection and evaluation. In Robinson JP, Shaver PR, Wrightsman LS (eds), Measures of Personality and Social Psychological Attitudes. San Diego: Academic, 1991, 1–16. (36) Baron RM, Kenny DA: The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology. 1986, 51:1173–1182. (37) Markman P, Salahuddin S, Veren K, Orndorff S, Gallo R: Hydrocortisone and some other hormones enhance the expression of HTLV-III. International Journal of Cancer. 1986, 37:67–72. (38) Cupps T, Fauci A: Corticosteroid-mediated immunoregulation in man. Immunological Reviews. 1982, 65:133–155. (39) Munck A, Guyre PM: Glucocorticoids and immune function. In Ader R, Felten DL, Cohen N (eds), Psychoneuroimmunology (2nd Ed.). San Diego, CA: Academic, 1991, 283–310. (40) Oxman TE, Freeman DH, Manheimer ED: Lack of social participation or religious strength and comfort as risk factors for death after cardiac surgery in the elderly. Psychosomatic Medicine. 1995, 57:5–15. (41) Kaldjian LC, Jekel JF, Friedland G: End of life decisions in HIV positive patients: The role of spiritual beliefs. AIDS. 1998, 12(1):103–107. (42) Pargament KI, Koenig HG, Perez LM: The many methods of religious coping: Development and initial validation of the RCOPE. Journal of Clinical Psychology. 2000, 56:519–543. (43) Solomon GF, Ironson G, Balbin E, Fletcher, MA: Distressing life events, hostility, CD4+ cell decline, and viral load increase in HIV+ persons. Brain, Behavior, and Immunity. 2000, 14:131–132. (44) Schumaker J: Mental health consequences of irreligion. In Schumaker J (ed), Religion and Mental Health. Oxford, England: Oxford University Press, 1992, 54–69. (45) Richards TA, Acree M, Folkman S. Spiritual aspects of loss among partners of men with AIDS: Postbereavement follow-up. Death Studies. 1999, 23:105–127. (46) Adler NE, Boyce T, Chesney MA, et al.: Socioeconomic status and health: The challenge of the gradient. American Psychologist. 1994, 49:15–24. (47) Underwood LG: Daily Spiritual Experience Scale. Fetzer–National Institute on Aging Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research. Kalamzoo, MI: Fetzer.
Measuring Spiritual Well-Being in People With Cancer: The Functional Assessment of Chronic Illness Therapy—Spiritual Well-Being Scale (FACIT-Sp) Amy H. Peterman, Ph.D. Northwestern University
George Fitchett, D. Min. Rush-Presbyterian-St. Luke’s Medical Center
Marianne J. Brady, Ph.D. Northwestern University
Lesbia Hernandez, Pharm.D., M.P.H. University of Puerto Rico
David Cella, Ph.D. Northwestern University
amined this relationship in community samples (1–3), among medical and surgical patients (4,5), and among cancer patients (6–9). Religious beliefs and practices have been demonstrated to have positive effects upon illness prevention, recovery from surgery, mental illness, and coping with physical illness (10). As Larson, Swyers, and McCullough (11) noted, definitions of religion and spirituality have changed in the past few decades. Up until the 1960s and 1970s, religion was seen as a broad construct, encompassing individual and institutional elements as well as spirituality. More recently, religion has become more narrowly defined, and spirituality has become distinguished from religiousness, or the practice of religious behavior. Recent definitions of spirituality include dimensions such as a personal search for meaning and purpose in life, connection with a transcendent dimension of existence, and the experiences and feelings associated with that search and that connection (12,13). Religion is seen, in contrast, as participation in the institutionally sanctioned beliefs and activities of a particular faith group. Although there has been important progress in research on religion/spirituality and health, at least two important methodological challenges persist. First, most of the research has examined the relation between one or more dimensions of religion and health, whereas the relation between spirituality and health has received little attention. Second, many studies of religion and health, including many of the studies of the role of religion in living with cancer, employed measures of religion whose reliability and validity were never established (14–16). Two factors underscore the importance of studying the relation between spirituality and health. First, several observers have reported a change in approach to religion among many members of the baby boom generation (17,18). This change is marked by a defection from organized religion and worship and a more personal search for spiritual fulfillment. Roof (17) labeled this cohort “highly active seekers.” There is no definitive study of the proportion of the population who would identify with this pattern, but one study found that as many as 20% of their respondents identified themselves as spiritual but not religious (19).
ABSTRACT A significant relation between religion and better health has been demonstrated in a variety of healthy and patient populations. In the past several years, there has been a focus on the role of spirituality, as distinct from religion, in health promotion and coping with illness. Despite the growing interest, there remains a dearth of well-validated, psychometrically sound instruments to measure aspects of spirituality. In this article we report on the development and testing of a measure of spiritual well-being, the Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being (FACIT-Sp), within two samples of cancer patients. The instrument comprises two subscales—one measuring a sense of meaning and peace and the other assessing the role of faith in illness. A total score for spiritual well-being is also produced. Study 1 demonstrates good internal consistency reliability and a significant relation with quality of life in a large, multiethnic sample. Study 2 examines convergent validity with 5 other measures of religion and spirituality in a sample of individuals with mixed early stage and metastatic cancer diagnoses. Results of the two studies demonstrate that the FACIT-Sp is a psychometrically sound measure of spiritual well-being for people with cancer and other chronic illnesses. (Ann Behav Med
2002, 24(1):49–58)
INTRODUCTION The past decade has seen a growing body of research examining the relation between religion and health. Studies have exThis research was supported by Grant No. 5 R01 CA61679 from the U.S. Public Health Service (National Cancer Institute) and by an unrestricted educational grant from Ortho-Biotech, Inc. We thank Helen Albrecht Morrow and Marla Moss Avery for their invaluable administrative assistance with this article. Reprint Address: A. H. Peterman, Ph.D., Center on Outcomes, Research and Education, 1000 Central Street, Suite 101, Evanston, IL 60201. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
49
50
Peterman et al.
A second reason to focus on the relation between spirituality and health is the possibility that it will facilitate more inclusive studies. Measures of religion often reflect the beliefs and practices of a specific religious group. For example, the items in the Religious Well-Being subscale of the popular Spiritual Well-Being Scale (20) focus on a personal relationship with God, a belief that is central to Evangelical Protestantism but not equally significant in other branches of Christianity or other faith traditions. Measures of religion that focus on such specific beliefs cannot be employed in studies of religiously diverse populations without distortion. This concern becomes more significant in light of America’s changing religious landscape. As of 1996, those who reported affiliation with a major religion other than Christianity or Judaism (e.g., Islam, Hinduism, Buddhism) or with a new religious movement were a small percentage of our population (1.1% and .4%, respectively). However, that proportion has doubled in the past 23 years and is currently equal to the proportion of Orthodox Christians in America. If the current pattern of growth continues, in the next decade the number of adherents of Islam, Buddhism, and Hinduism combined will be larger than the number of adherents of Judaism (21). By examining spirituality rather than specific religious beliefs and practices, investigators may be able to be more inclusive of America’s growing religious diversity, to study and compare people with diverse religious traditions as well as those who identify themselves as spiritual but not religious. Commenting on several studies of religion, Thomason and Brody (22) argued that further research is certainly needed to develop and test the validity of scales that measure spirituality independent of religiosity or religious practice. Such instruments need to assess spiritual needs in patients in language and concepts that are inclusive of the spiritual lives of nonreligious persons, as well as those for whom religious faith is at the core of their spirituality. (p. 97)
Annals of Behavioral Medicine from faith in coping with illness. However, during the item reduction phase of FACT-G development, the two spiritual well-being items were dropped due to low factor loadings on the main identified factors. However, the initial interviews, subsequent patient reports, and previous research on spirituality and chronic/terminal illnesses suggested the significance of spiritual and faith issues in this population. Therefore, we undertook the development of a scale containing items about spirituality to be used when assessment of this additional dimension of QOL is desired. STUDY 1 The first study of the FACIT-Sp was undertaken to establish the factor structure, reliability and initial validity of the instrument. Data were collected in conjunction with a large-scale validation of the FACIT measurement system across languages (Spanish vs. English), cultures (Hispanic vs. Black non-Hispanic vs. White non-Hispanic), and literacy (high vs. low). Results of the Spanish language translation and validation have been reported elsewhere (25). From 1994 to 1996, participants were recruited from four sites in the mainland United States (Rush-Presbyterian–St. Luke’s Medical Center and Cook County Hospital in Chicago and Emory University Medical Center and Grady Hospital in Atlanta, GA) and three sites in Puerto Rico (San Juan Veterans Administration Medical Center, Rafael Lopez Nussa Hospital, and the I. Gonzalez Martinez Hospital). Method Sample The sample contained 1,617 subjects, of whom 53% were female and 47% were male. The median age was 54.6 years, and the median length of time since diagnosis with disease was 29 months. The majority (83.1%) of patients had cancer. Additional demographic and disease-related characteristics of the sample can be found in Table 1.
In this article we report on the development and psychometric properties of the Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being (FACIT-Sp) scale, an instrument that was designed to provide an inclusive measure of spirituality that could be employed in research with people with chronic and/or life-threatening illnesses. At least two studies utilizing the FACIT-Sp (6,23) have already been published. However, this will be the first published demonstration of the scale’s psychometric properties. The FACIT-Sp1 is part of the larger FACIT measurement system of which the Functional Assessment of Cancer Therapy –General (FACT-G) is the core instrument. The FACT-G measures health-related quality of life (QOL) and was developed using interview data from 135 cancer patients and 15 oncology specialists (24). In these interviews, both patients and specialists emphasized the relation between spiritual concerns and quality of life and discussed the importance of strength and comfort
Measures Functional Assessment of Cancer Therapy - General (FACT-G). The FACT-G (22) is a widely-used measure of QOL. The core of the FACIT scales, it comprises 27 questions that assess well-being in four domains: physical (PWB), functional (FWB), social/family (SFWB), and emotional (EWB). PWB comprises reports of physical symptoms; FWB assesses the degree to which the respondent can participate in and enjoy normal daily activities; the SFWB questions assess social support and communication; and the EWB measures mood and emotional response to illness. These individual domains are summed to create a total QOL score. Both the total score and the individual subscale scores have good internal consistency reliability (in this study, α = .72–.85), and the instrument has been well validated (24).
1In several early publications, the scale was referred to as the Rush
FACIT - Spiritual Well-Being Scale (FACIT-Sp). The FACITSp was developed with the input of cancer patients, psychotherapists, and religious/spiritual experts (e.g., hospital chaplains),
Spiritual Beliefs Module.
Volume 24, Number 1, 2002
Measuring Spiritual Well-Being
TABLE 1 Study 1: Sample Demographic and Disease Characteristics Characteristic
Mdn
Range
Age Number of months post cancer/HIV diagnosis Education (years)
54.6 29.0 11.1
18–90 0–446 0–28
n
%
764 853
47.2 52.8
503 718 396
31.1 44.4 24.5
558 508 380 147 21
34.6 31.5 23.5 9.1 1.3
956 661
59.1 41.9
534 258 316 236 273
33.0 16.0 19.5 14.6 16.9
Characteristic Sex Male Female Ethnicity African American Latino European American ECOG PSR 0 1 2 3 4 Language preference English Spanish Diagnosis Breast cancer Colon cancer Lung cancer Head and neck cancer HIV/AIDS
Note. N = 1,617. ECOG PSR = Eastern Cooperative Oncology Group Performance Status Rating.
who were asked to describe the aspects of spirituality and/or faith that contributed to QOL. The responses emphasized a sense of meaning in life, harmony, peacefulness, and a sense of strength and comfort from one’s faith (26). Items included in the scale were taken from the original FACT-G interviews, subsequent validation and translation interviews with over 200 patients, and interviews with several hospital chaplains. The 12-item FACIT-Sp scale can be found in the Appendix.
Demographic, disease, and treatment information. Basic information regarding demographic characteristics, type and stage of disease, current and previous treatments, and Eastern Cooperative Oncology Group Performance Status Rating (ECOG PSR [27]) were obtained from each participant. The ECOG PSR is a widely used measure of functional status. It is a single item rating of five activity levels: 0 = fully ambulatory without symptoms; 1 = fully ambulatory with symptoms; 2 = requiring rest for 1 to 49% of the waking day; 3 = requires rest 50 to 99% of the waking day; and 4 = requiring complete bedrest. Patients rated their own performance status. Research assistants verified the disease and treatment information with the participant’s medical record.
51
Marlowe-Crowne Social Desirability Scale (MCSDS). The 10-item short form (28) of the MCSDS (29) provides a measure of the degree to which participants endorse socially desirable characteristics. The reliability and comparability of the short form have been established (30). A validated Spanish version of the MCSDS (31) was completed by the Spanish-speaking participants in the study. As one would hope to see no correlation between social desirability and spiritual well-being, the MCSDS was administered to evaluate the discriminant validity of the FACIT-Sp. Profile of Mood States– Short Form (POMS-SF). The POMSSF (32) is a widely used scale measuring subjective mood states, such as anxiety/tension, vigor, and depression. It also produces a Total Mood Disturbance Score. The POMS-SF is a reliable and valid measure of affective states and is available in both English and Spanish (32). Convergent validity would be demonstrated by a moderate association between general distress, as measured by the POMS-SF, and more specific spiritual distress, indicated by the FACIT-Sp. Procedure Potential participants were identified from the daily record of office visits, treatment visits, and inpatient hospitalizations. Individuals who were over the age of 18, able to give informed consent, and had current or past diagnoses of cancer and/or HIV infection/AIDS were eligible for the study. Each potential participant was provided with a full explanation of the study, in accordance with the guidelines for treatment of human participants of each site’s Institutional Review Board. Once informed consent was obtained, participants completed the packet of questionnaires. Questionnaires were administered in either Spanish or English, depending on the individual’s score on a short acculturation scale evaluating preferred language: The psychometric equivalence of the Spanish translation of the original English FACIT scales has been documented (25). Results Factor Analysis A principal components analysis with varimax rotation was performed on the 12 items of the FACIT-Sp to evaluate the unidimensionality of the scale (33). Three factors emerged with eigenvalues over 1.0 (3.2, 3.2, 1.5). However, the third factor comprised only 2 items that were worded negatively (as opposed to 10 positively worded items). Because the direction of the item phrasing—not content—seemed to drive the separation of this third factor, a two-factor solution was examined and found to be more interpretable. The results of the two-factor analysis can be found in Table 2. One factor, labeled Meaning/Peace, contains 8 items and assesses a sense of meaning, peace, and purpose in life. The other factor, labeled Faith, contains 4 items and measures several aspects of the relation between illness and one’s faith and spiritual beliefs: The sense of strength and comfort, as well as the sense that “things will be okay,” might be considered to be the “fruits” of faith/spiritual beliefs. The correlation between the two subscales was .54 (p = .0001). In the remainder of this section we
52
Peterman et al.
Annals of Behavioral Medicine
TABLE 2 FACIT–Factor Analysis With Varimax Rotation
FACIT–Sp Item 1. I feel peaceful. 2. I have a reason for living. 3. My life has been productive. 4. I have trouble feeling peace of mind.a 5. I feel a sense of purpose in my life. 6. I am able to reach down deep into myself for comfort. 7. I feel a sense of harmony within myself. 8. My life lacks meaning and purpose.a 9. I find comfort in my faith or spiritual beliefs. 10. I find strength in my faith or spiritual beliefs. 11. My illness has strengthened my faith or spiritual beliefs. 12. I know that whatever happens with my illness, things will be okay.
TABLE 4 Spearman Correlations Between FACIT–Sp and the FACT–G, POMS, and MCSDS
Factor 1 Loading
Factor 2 Loading
.28 .29 .25 .01 .42 .52
.63 .59 .65 .56 .63 .59
.49 –.10 .90
.61 .56 .14
.91
.12
.82
.09
.69
.31
Note. All loadings above .40 are bold. FACIT–Sp = Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being scale. aItem is reverse scored.
report the results for these two subscales and for an aggregated (summed) total score. Descriptive statistics can be found in Table 3. The distributions of both subscales and the total score were somewhat positively skewed. Reliability The reliability of the scale and subscales was evaluated with internal consistency coefficients, which reflect the degree to which all items on a particular scale measure a single (unidimensional) concept. The alpha coefficients for the total scale and the two subscales were quite good (Cronbach’s α = .81–.88): see Table 3. Validity There were moderate to strong correlations between the total FACIT-Sp score and QOL, as measured by the total FACT-G score and its subscale scores (see Table 4). Further, both the Meaning/Peace and Faith subscales were positively associated with the FACT-G and its subscales, with notably stronger correlations for Meaning/Peace versus Faith. The FACIT-Sp and its
Scale FACT–G PWB EWB FWB SFW POMS Tension POMS Depression POMS Anxiety POMS Vigor POMS Fatigue POMS Confusion POMS TMDS MCSDS
Meaning/Peace
Faith
FACIT–Sp Total
.62 .31 .57 .54 .46 –.46 –.54 –.44 .46 –.39 –.53 –.60 .22
.34 .09 .35 .31 .28 –.21 –.26 –.24 .25 –.20 –.24 –.30 .26
.58 .25 .55 .51 .44 –.41 –.48 –.41 .42 –.36 –.47 –.54 .27
Note. Ns = 1561–1589. All ps < .001. FACIT–Sp = Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being scale; FACT–G = Functional Assessment of Cancer Therapy–General; MCSDS = Marlowe–Crowne Social Desirability Scale; PWB = Physical Well-Being; EWB = Emotional Well-Being; FWB = Functional Well-Being; SFWB = Social/Family Well-Being; POMS = Profile of Mood States; TMDS = Total Mood Disturbance Score.
subscales were also correlated with the POMS and its subscales, and again the pattern of stronger associations with the Meaning/Peace subscale is evident. An additional test of discriminant validity was conducted using the scores on the POMS Depression subscale. That is, previous research has established an inverse relation between depression and religion (34). We therefore predicted that more depressed respondents would have lower FACIT-Sp scores. The sample was divided into equal thirds using their scores on the POMS Depression subscale. Greater depression was associated with significantly lower FACIT-Sp total scores, F(2, 1586) = 186.98, p = .0001. Finally, there are weak (positive) associations between the Spiritual Well-Being subscales and total score and social desirability, as measured by the MCSDS. Relation Between Spiritual Well-Being and Demographic and Disease Variables The relation between spiritual well-being and the sample demographic and disease characteristics was evaluated using an analysis of variance within a general linear model. These results
TABLE 3 FACIT–Sp Descriptive Statistics: Study 1
FACIT–Sp total Meaning/Peace Faith Note.
M
SD
Possible Range
Actual Range
Cronbach’s α
38.5 25.2 13.3
8.1 5.6 3.6
0–48 0–32 0–16
1–48 1–32 0–16
.87 .81 .88
N = 1,617. FACIT–Sp = Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being scale.
TABLE 5 Relationship Between FACIT–Sp and Sample Characteristics Characteristic Age 1 = 18–45 2 = 46–55 3 = 56–65 4 = > 65 F Education 1 = Less than high school 2 = High school degree 3 = College 4 = Graduate school or degree F Marital status 1 = Single 2 = Married 3 = Separated 4 = Divorced 5 = Widow F Ethnicity 1 = Latino 2 = African American 3 = European American F Language English Spanish t Religious affiliation 1 = Catholic 2 = Protestant 3 = Jewish 4 = Other 5 = None F Diagnosis 1 = Breast cancer 2 = Colorectal cancer 3 = Head and neck cancer 4 = HIV+ 5 = Lung cancer F Performance status 0 1 2 3/4 F
Meaning/Peace
Faith
FACIT–Sp Total
24.3 25.5 25.3 25.8 5.74*** 1 < 2, 4
12.7 13.4 13.5 13.7 6.12*** 1 < 2, 3, 4
37.0 38.9 38.8 39.5 7.59*** 1 < 2, 3, 4
25.0 25.2 25.7 26.4 2.09
13.6 13.3 12.5 12.8 4.8** 1>3
38.6 38.4 38.2 39.2 .33
24.4 26.3 24.3 25.0 25.9 12.81*** 2 > 1, 3, 4; 5 > 1
13.1 13.7 13.8 13.1 14.2 7.87*** 2, 5 > 1
37.6 40.0 38.1 38.1 40.1 13.39*** 2 > 1, 3, 4; 5 > 1
25.0 25.7 24.8 3.56* 2>3
13.5 14.0 12.2 29.8*** All
38.6 39.7 37.0 12.64*** All
25.2 25.2 –0.12
13.1 13.6 –2.99**
38.3 38.9 –1.37
24.9 26.2 23.6 25.2 23.0 8.98*** 2 > 1, 5; 1, 4 > 5
13.2 14.0 8.6 13.9 10.2 38*** 2, 4 > 1, 5, 3; 1 > 3, 5
38.2 40.2 32.2 39.1 33.3 21.1*** 2 > 1, 5, 3; 1, 4 > 5, 3;
26.1 26.1 25.2 22.6 25.3 21.19*** 2, 1, 3, 5 > 4
13.8 14.0 13.6 11.6 13.2 23.18*** 2, 1, 3, 5 > 4; 2 > 5
39.9 40.2 38.8 34.1 38.5 28.38*** 2, 1, 3, 5 > 4
26.7 25.0 24.2 23.5 23.28*** 0 > 1, 2, 3/4; 1 > 3/4
13.6 13.3 13.1 13.2 1.43 None
40.2 38.2 37.2 36.7 14.43*** 0 > 1, 2, 3/4
Note. Results based on analysis of variance under a general linear model: All post hoc tests are Tukey tests. N = 1,617. FACIT–Sp = Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being scale; All = all differences are signficant; None = no significant differences. *p > .05. **p < .01. ***p < .001.
53
54
Peterman et al.
can be found in Table 5. Age had a weak positive association with the total score and subscale scores, such that older participants reported higher levels of spiritual well-being. Education had a weak, negative association with the Faith subscale only, such that participants with more years of education reported lower Faith scores. Women had higher scores than men on all three scales, and in general, married and widowed participants had the highest scores. Scores on the two subscales and the total Spiritual Well-Being scale differed among ethnic and religious groups. African Americans had higher scores than Latinos, who had higher scores than European Americans on Faith and total Spiritual Well-Being, whereas African Americans had higher scores than European Americans on Meaning/Peace. Latinos did not differ significantly from the other two groups on Meaning/Peace. In general, Protestant participants had the highest scores on all three scales, whereas Jewish participants and those who claimed no religious affiliation had the lowest scores. Specific group differences can be found in Table 5. Finally, type of disease was associated with the three scales, with participants with any type of cancer having higher scores than participants with HIV. Regarding performance status, participants who were fully ambulatory with no symptoms (ECOG PSR = 0) had higher scores on the FACIT-Sp and the Meaning/Peace subscale than all other groups. Scores on the Faith subscale did not differ by ECOG PSR. STUDY 2 Having established the reliability of the FACIT-Sp and a significant relation between Spiritual Well-Being and QOL in patients with chronic disease in Study 1, a second study was undertaken to further validate the FACIT-Sp by examining its relation to existing measures of religion and spirituality. The data were collected in the context of a larger study that investigated the longitudinal association between fatigue and QOL in a sample of patients beginning chemotherapy for any solid tumor or hematological malignancy. Only data collected during the baseline assessment are reported herein. Method Sample The demographic and disease characteristics of the 131 participants enrolled in this study can be found in Table 6. As shown, the sample was well educated and largely European American (87%), with good performance status. The majority of participants had breast (44.3%), colon (10.7%), or lung (10.7%) cancer. Sixty-five percent reported an affiliation with a specific church or synagogue. Measures Religion and spirituality. To permit analysis of convergent validity, five measures of religion/spirituality were employed in Study 2. These measures assessed six different dimensions of religion and spirituality, including organizational religious activity (ORA) and non-organizational religious activity (NORA),
Annals of Behavioral Medicine TABLE 6 Study 2: Sample Demographic and Disease Characteristics Characteristic
Mdn
Range
Age Education (years)
56 16
20–82 4–24
Characteristic
n
%
39 92
29.8 70.2
2 3 12 114
1.5 2.3 9.2 87.0
79 36 16 1
60.3 27.5 11.5 0.8
58 14 14 7 11 27
44.3 10.7 10.7 5.3 8.4 20.6
Sex Male Female Ethnicity Latino Asian American African American European American Performance status 0 1 2 3/4 Diagnosis Breast cancer Colon cancer Lung cancer Ovarian Lymphoma Other known Note.
N = 131.
spiritual beliefs and religious social support, coherence, and intrinsic religiosity. Several individual items examining satisfaction with religion, outlook on life, and a sense of peace were also administered. ORA was assessed with three items taken from Chatters et al. (35). The first item asked if the respondent was a member of a church or other religious institution. The other two items asked about the frequency of participation in public worship and other activities at one’s religious institution. In this study, the internal consistency reliability estimate for these items was .70. NORA was assessed with two items also taken from Chatters et al. (35). The first item asked about the frequency of private prayer, and the second item about the frequency of religious or devotional reading. The correlation between responses for these two items was .53. The 15-item version of the Spiritual Beliefs Inventory (SBI) (36) has two subscales. A 10-item subscale assesses spiritual and religious beliefs and devotional practices, and a 5-item subscale measures social support obtained from one’s religious colleagues and leaders. The alpha coefficient for the SBI and its subscales in this study were .96, .96, and .90, respectively. The Coherence subscale of Reker’s Life Attitude Profile–Revised (LAP–R) (37) was also included in Study 2. Reker defined coherence as “having a logically integrated and consistent analytical and intuitive understanding of self, others, and life in general. Implicit in coherence is a sense of order and rea-
Volume 24, Number 1, 2002 son for existence, a clear sense of personal identity, and greater social consciousness” (p. 15). We employed the six-item Coherence subscale from the 1991 version of Reker’s LAP–R. Sample items include “I have a framework that allows me to understand or make sense of my life” and “A period of personal hardship and suffering can help give a person a better understanding of the real meaning of life.” The internal consistency coefficient for the Coherence subscale was .80 in this study. Four single-item measures were also used. The first was “My whole approach to life is based on my religion,” which is a commonly employed measure of intrinsic religiosity, or the importance of religion to the respondent (38). Three individual items from the Cancer Patient Behavior Survey (CPBS) were also used: “satisfaction with religion,” “outlook on life,” and “sense of peace.” For these items and a variety of other activities and relations, the CPBS asks patients to compare how they are now with how they were before their illness. The items are scored on a 5-point scale from 1 (worse) to 3 (same) to 5 (better). Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being (FACIT-Sp). The 12-item version of the FACIT-Sp that was tested in Study 1 was also administered to all participants in Study 2. Demographic, disease and treatment information. Basic information regarding demographic characteristics, type and stage of disease, and current and previous treatments were obtained from each participant. In addition, research assistants verified the disease and treatment information with the participants’ medical records. Procedure Potential participants were identified from the daily record of patients beginning chemotherapy for solid tumors or hematological malignancies in the outpatient oncology clinics of two medical centers in the Midwest. Eligible individuals must have been receiving their first chemotherapy treatment for this episode of disease, and, for participants who had received prior chemotherapy, there must have been at least a 6-month chemotherapy-free period prior to study enrollment. Additional eligibility criteria were age over 18 years, ability to give informed consent, and absence of brain metastases. Each potential participant was provided with a full explanation of the study, in accordance with Institutional Review Board guidelines for the treatment of human participants. Once informed consent was obtained, a research assistant collected demographic and disease information from the participant by interview and allowed the participant to complete the self-report questionnaires. Results Descriptive Statistics and Scale Reliability The mean scores on the full scale (M = 36.8, SD = 8.3) and subscales (Meaning/Peace, M = 25.0, SD = 5.4; Faith, M = 11.8; SD = 4.3) were roughly comparable to those obtained in Study 1. The correlation between the Meaning/Peace and Faith
Measuring Spiritual Well-Being
55
subscales was somewhat lower, although still significant, in this study (r = .27, p < .01). The internal consistency estimates were also similar for the FACIT-Sp, Meaning/Peace, and Faith (Cronbach’s α = .86, .81, and .86, respectively). Validity The primary purpose of Study 2 was to examine the relation between the FACIT-Sp and other measures of spirituality and religion. This comparison helps establish the convergent validity of the FACIT-Sp as well as document the degree to which the FACIT-Sp appears to measure something that is not captured by existing scales. Hence, moderate correlations above .30 between the FACIT-Sp and the other measures were considered to be supportive of convergent validity; correlations over .80 (64% shared variance) were suggestive of possible duplication of an existing construct. As demonstrated in Table 7, the total FACIT-Sp was moderately correlated with all of the other measures (rs = .31 with NORA to .48 with the total SBI score, ps < .0005). The Faith subscale demonstrated moderate to strong correlations with all of the other measures (rs = .38 with the Reker Coherence Subscale to .75 with the SBI total score, ps < .005). None of the correlations with the Meaning/Peace subscale met our criteria for a significant degree of shared variability. Following from the Study 1 finding that participants who were Jewish or who reported no religious affiliation had the lowest FACIT-Sp scores, Study 2 revealed a similar pattern. The FACIT-Sp total scores for those who reported a specific congregational affiliation (n = 82) were significantly higher than for those who reported no such affiliation (n = 45), t(123) = 3.4, p = .0009. Among those who reported a congregational affiliation, there were no significant faith group differences in FACIT-Sp scores, F(3, 78) = .24, ns, and no faith group differences among those who reported no specific congregational affiliation, F(3, 39) = .46, ns. TABLE 7 Spearman Correlations Between FACIT–Sp and Other Measures of Religion and Spirituality
Measure Organizational religious activity Nonorganizational religious activity Intrinsic religiousness SBI Total SBI Beliefs SBI Support Reker Coherence subscale CPBS Satisfaction With Religion CPBS Outlook on Life CPBS Sense of Peace
Meaning/ Peace
Faith
FACIT– Sp Total
.13 .05
.15** .54**
.34*** .31***
.13 .13 .09 .20* .28** .08
.61** .75** .74** .64*** .38*** .39***
.41*** .48*** .45*** .46*** .38*** .25***
.25** .28**
.36*** .34***
.37*** .36***
Note. N = 131. SBI = Spiritual Beliefs Inventory; CPBS = Cancer Patient Behavior Survey. *p < .05. **p < .005. ***p < .0005.
56
Peterman et al. DISCUSSION
As Thoresen (39) and others (22) pointed out, an important next step in exploring the relation between spirituality and religion and health is the development of psychometrically sound measures of spirituality. Taken together, the results of Studies 1 and 2 demonstrate the validity and reliability of the FACIT-Sp, a new measure of spiritual well-being designed for use in research with people with chronic and life-threatening illnesses. The total scale and its two subscales show strong internal consistency reliability and a significant association with QOL. The total scale and Faith subscale demonstrate good concurrent validity with other measures of spirituality and religion. The Meaning/Peace subscale is correlated in the expected direction with several of the other spirituality measures, but the limited number of significant correlations and their relatively small size suggest that this subscale measures a unique concept not assessed by the other instruments used in Study 2. That is, some definitions of spirituality refer to the sense of meaning and purpose that spirituality provides, as well as a feeling of harmony and peace deriving from a connection to something larger than the self. The items in the Meaning/Peace subscale of the FACIT-Sp appear to be a good measure of these aspects of spirituality. This conclusion is based on the face validity of the items in the subscale. It is important to continue to examine the construct validity of the subscale as other validated measures of this aspect of spirituality become available. The FACIT-Sp has several important strengths that researchers might consider when evaluating the measure for use. First, this measure appears to be a good choice for assessing spirituality across a range of religious traditions and for respondents who identify themselves as spiritual but not religious. For example, the data from Study 2 suggest that among respondents with similar levels of religious commitment, the FACIT-Sp provides a measure of spirituality that is not biased for or against a particular religious group. In addition, the items in the scale make no reference to specific religious beliefs or practices, such as belief in God or use of prayer. Other validated measures of spirituality contain items that are more closely tied to religion in general (SBI–15) (36) or to the beliefs of a specific denomination (Spiritual Well-Being Scale [20]). A second strength of the FACIT-Sp is that one subscale (Faith) has a moderate to strong association with religion, whereas the other subscale (Meaning/Peace) is not significantly associated with existing religion measures. This suggests that the Faith subscale may measure a dimension of spirituality that overlaps with, or is enhanced by, religion, whereas the Meaning/Peace subscale measures a dimension that is more independent and is not assessed by existing instruments. There is significant ongoing controversy, both in the research literature and the lay press, over the degree to which religion is a component of spirituality. Therefore, it may be desirable to use such a measure that allows the assessment of both the attitudes or behaviors associated with religion or faith and a sense of meaningfulness in life that is independent of any religion or specific belief.
Annals of Behavioral Medicine A third strength of the FACIT-Sp is that the original validation took place in a large sample that was relatively diverse in ethnicity, religious affiliation, age, type of cancer, and stage of illness/prognosis. This provides initial assurance of the scale’s acceptability and utility among a range of samples, although as noted next, additional testing in other populations will be necessary. Fourth, the results of this study and two others (6,23) address a criticism that recently appeared in the literature (12). In our study, there were moderate correlations between FACIT-Sp scores and affective distress as derived from the POMS-SF. However, no more than 36% of the variance was shared between Meaning/Peace and any POMS-SF subscale. Brady et al. (6) reported a more comprehensive analysis of the unique relation between the FACIT-Sp and QOL in the same sample that was used in Study 1 here. Briefly, the FACIT-Sp total and subscale scores were still significantly associated with a single-item measure of contentment with QOL even after controlling for the effects of demographic and disease characteristics, mood as measured by the POMS, and social desirability. Cotton et al. (23) similarly demonstrated that the FACIT-Sp significantly predicted QOL among a sample of breast cancer patients, even after controlling for demographic characteristics, psychological adjustment to cancer, and another measure of spirituality. These findings do not support the conclusion, drawn by Koenig et al. (12), that the Meaning/Peace subscale is purely a measure of emotional well-being. Brady et al. (6) also examined the best predictors of contentment with QOL among three FACT-G subscales (PWB, EWB, SFWB), the Meaning/Peace subscale, and the Faith subscale. In a stepwise regression equation, Meaning/Peace was the best predictor of contentment with QOL and the Faith subscale entered the equation before SFWB. In further analysis, we found the magnitude of the correlations of FACIT-Sp total score and Meaning/Peace score with contentment with QOL (.48 and .49, respectively) were similar to the correlations of PWB and EWB with contentment with QOL (.47 for both). Finally, we found that when patients with high levels of fatigue or pain were examined, those with higher Meaning/Peace and Faith scores reported significantly greater global QOL compared to those with lower scores on the FACIT-Sp subscales. Three important limitations to the current version of the scale should be noted. First, as with other measures of religion and spirituality (40), there appear to be significant demographic differences in FACIT-Sp scores. Investigators should examine the need to control for gender, ethnicity, age, and/or marital status when using the FACIT-Sp. Also in keeping with findings for other measures of religion and spirituality, the FACIT-Sp total scale and subscale scores suffer from ceiling effects. This is especially evident in scores on the Faith subscale, where 46% and 30% of the subjects in Study 1 and 2, respectively, had the maximum score. Investigators should be aware of the potential impact of skewed distributions on some statistical analyses and might consider normalization of the distribution of the scores through standard statistical procedures. A third limitation is the possibility that the FACIT-Sp does not address a number of constructs that are important to a sense of spirituality for at least some people. These may
Volume 24, Number 1, 2002 include such aspects as forgiveness, generosity, and love. A 23-item version of the FACIT-Sp, the FACIT-Sp-Ex, includes items that assess these additional aspects of spirituality, and testing of the FACIT-Sp-Ex is currently underway. Additional psychometric testing of the FACIT-Sp is recommended. First, the performance of the scale should be examined in people of other religious traditions, particularly the Eastern religions and Islam. Relatedly, when employing measures of religion to establish the validity of measures of spirituality, it may be helpful to include an item that permits the stratification of the respondents according to the extent they identify themselves as spiritual and religious to meaningfully interpret the results. Second, as the FACIT-Sp was designed to be used in health-related research, it should be evaluated in samples of people with chronic or life-threatening conditions other than cancer and HIV/AIDS. Third, it would also be interesting to examine its utility and psychometric properties in a sample of healthy people: A nonillness version of the scale has been developed by altering the two scale items that refer to “my illness.” Fourth, the ability of the scale to prospectively predict future psychosocial/QOL outcomes, as well as morbidity and mortality, might be tested. Finally, sensitivity to change in response to an intervention that targets spirituality could also be assessed. In conclusion, evaluation of the FACIT-Sp indicates that it is a brief, reliable, valid measure of spirituality that may be especially useful in assessing the role of nonreligious spirituality in QOL and other health-related research. The scale is currently available in nine languages other than English: Dutch, French, German, Italian, Japanese, Norwegian, Portuguese, Spanish, and Swedish. REFERENCES (1) Hummer RA, Rogers RG, Man CB, Ellison CG: Religious involvement and US adult mortality. Demography. 1999, 36:273–285. (2) Musick MA, House JS, Williams DR: Attendance at religious services and mortality in a national sample. Meeting of the American Sociological Association, Chicago, 1999. (3) Strawbridge WJ, Cohen RD, Shema SJ, Kaplan GA: Frequent attendance at religious services and mortality over 28 years. American Journal of Public Health. 1997, 87:957–961. (4) Koenig HG, Pargament KI, Nielsen J: Religious coping and health status in medically ill hospitalized older adults. Journal of Nervous and Mental Disease. 1998, 186:513–521. (5) Oxman TE, Freeman Jr. DH, Manheimer ED: Lack of social participation or religious strength and comfort as risk factors for death after cardiac surgery in the elderly. Psychosomatic Medicine. 1995, 57:5–15. (6) Brady MJ, Peterman AH, Fitchett G, Mo M, Cella D: A case for including spirituality in quality of life measurement in oncology. Psycho-Oncology. 1999, 8:417–428. (7) Fehring RJ, Miller JF, Shaw C: Spiritual well-being, religiosity, hope, depression, and other mood states in elderly people coping with cancer. Oncology Nursing Forum. 1997, 24:663–671. (8) Holland JC, Passik S, Kash KM, et al.: The role of religious and spiritual belief in coping with malignant melanoma. Psycho-Oncology. 1999, 8:14–26.
Measuring Spiritual Well-Being
57
(9) Mickley JR, Soeken K, Belcher A: Spiritual well-being, religiousness and hope among women with breast cancer. Image: Journal of Nursing Scholarship. 1992, 24:267–272. (10) Matthews DA, McCullough ME, Larson DB, et al.: Religious commitment and health status: A review of the research and implications for family medicine. Archives of Family Medicine. 1998, 7:118–124. (11) Larson DB, Swyers JP, McCullough ME: Scientific research on spirituality and health: A consensus report. Rockville, MD: National Institute for Healthcare Research, 1998. (12) Koenig HG, McCullough ME, Larson DE: Handbook of Religion and Health. Oxford, England: Oxford University Press, 2001. (13) Zinnbauer BJ, Pargament KI, Scott AB: The emerging meanings of religiousness and spirituality: Problems and prospects. Journal of Personality. 1999, 67:889–919. (14) Acklin MW, Brown EC, Mauger PA: The role of religious values in coping with cancer. Journal of Religion and Health. 1983, 22:322–333. (15) Smith ED, Stefanek ME, Joseph MV, et al.: Spiritual awareness, personal perspective on death, and psychosocial distress among cancer patients: An initial investigation. Journal of Psychosocial Oncology. 1993, 11:89–103. (16) Yates JW, Chalmer BJ, St. James P, Follansbee B, McKegney FB: Religion in patients with advanced cancer. Medical and Pediatric Oncology. 1981, 9:121–128. (17) Roof WC: A generation of Seekers: The Spiritual Journeys of the Baby Boom Generation. San Francisco: Harper San Francisco, 1993. (18) Wuthnow R: After Heaven: Spirituality in America the 1950s. Berkeley: University of California Press, 1998. (19) Zinnbauer BJ, Pargament KI, Cole B, et al.: Religion and spirituality: Unfuzzying the fuzzy. Journal for the Scientific Study of Religion. 1997, 36:549–564. (20) Paloutzian RF, Ellison CW: Loneliness, spiritual well-being and quality of life. In Peplau LA, Perlman D (eds), Loneliness: A Sourcebook of Current Theory, Research and Therapy. New York: Wiley Interscience, 1982, 224–237. (21) Sherkat DE: Tracking the “other”: Dynamics and composition of “other religions” in the General Social Survey, 1973–1996. Journal for the Scientific Study of Religion. 1999, 38:551–560. (22) Thomason CL, Brody H: Inclusive spirituality. The Journal of Family Practice. 1999, 48:96–97. (23) Cotton SP, Levine EG, Fitzpatrick CM, Dold KH, Targ E: Exploring the relationships among spiritual well-being, quality of life, and psychological adjustment in women with breast cancer. Psycho-Oncology. 1999, 8:429–438. (24) Cella DF, Tulsky DS, Gray G, et al.: The Functional Assessment of Cancer Therapy (FACT) Scale: Development and validation of the general measure. Journal of Clinical Oncology. 1993, 11:570–579. (25) Cella D, Hernandez L, Bonomi AE, et al.: Spanish language translation and initial validation of the Functional Assessment of Cancer Therapy quality-of-life instrument. Medical Care. 1998, 36:1407–1418. (26) Fitchett G, Peterman AH, Cella D: Spiritual beliefs and quality of life in cancer and HIV patients. Presentation at World Congress of Psycho-Oncology. New York: October 1996. (27) Zubrod CG, Schneiderman M, Frie III E, et al.: Appraisal of methods for the study of chemotherapy of cancer in man: comparative therapeutic trial of nitrogen mustard and triethylene
58
(28)
(29)
(30)
(31)
(32)
(33)
Peterman et al.
Annals of Behavioral Medicine (34) McCullough ME, Larson DB: Religion and depression: A review of the literature. Twin Research. 1999, 2:126–136. (35) Chatters LM, Levin JS, Taylor RJ: Antecedents and dimensions of religious involvement among older Black Americans. Journal of Gerontology. 1992, 47(Suppl.):S269–S278. (36) Holland JC, Kash KM, Passik S, et al.: A Brief Spiritual Beliefs Inventory for use in quality of life research in life-threatening illness. Psycho-Oncology. 1998, 7:460–469. (37) Reker GT: The Life Attitude Profile–Revised (LAP–R) Procedures Manual: Research Edition. Peterborough, Ontario, Canada: Student Psychologists Press, 1992. (38) Gorsuch RL, McPherson SE: Intrinsic/Extrinsic measurement: I/E-revised and single-item scales. Journal for the Scientific Study of Religion. 1989, 28:348–354. (39) Thoresen CE: Spirituality and health: Is there a relationship? Journal of Health Psychology. 1999, 4:291–300. (40) Levin JS, Taylor RJ, Chatters LM: Race and gender differences in religiosity among older adults: Findings from four national surveys. Journals of Gerontology. 1994, 49(Suppl.):S137–S145.
thiophosphoramide. Journal of Chronic Diseases. 1960, 11:7–33. Strahan T, Gerbasi KC. Short homogeneous versions of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology. 1972, 28:191–193. Crowne DP, Marlowe D: A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology. 1960, 24:349–354. Fraboni M, Cooper D: Further validation of three short forms of the Marlowe-Crowne scale of social desirability. Psychological Reports. 1989, 65:595–600. Lara-Cantu MA, Suzan-Reed M: The Crowne and Marlowe Social Desirability Scale: A psychometric study [Spanish]. Salud Mental. 1988, 11:25–29. McNair DM, Lorr M, Droppleman LF: EdITS Manual for the Profile of Mood States. San Diego, CA: Educational and Industrial Testing Service, 1992. Peterman AH, Fitchett G, Cella D: Modeling the relationship between quality of life dimensions and an overall sense of well-being. Presentation at World Congress of Psycho-Oncology. New York: October 1996.
APPENDIX FACIT–Sp (Version 4) Below is a list of statements that other people with your illness have said are important. By circling one (1) number per line, please indicate how true each statement has been for you during the past 7 days.
Sp1 Sp2 Sp3 Sp4 Sp5 Sp6 Sp7 Sp8 Sp9 Sp10 Sp11 Sp12
I feel peaceful . . . . . . . . . . . . . . . . . . . . . . . . . . . . I have a reason for living . . . . . . . . . . . . . . . . . . . . . . My life has been productive . . . . . . . . . . . . . . . . . . . . . I have trouble feeling peace of mind . . . . . . . . . . . . . . . . I feel a sense of purpose in my life . . . . . . . . . . . . . . . . . I am able to reach down deep into myself for comfort . . . . . . . I feel a sense of harmony within myself. . . . . . . . . . . . . . . My life lacks meaning and purpose . . . . . . . . . . . . . . . . . I find comfort in my faith or spiritual beliefs . . . . . . . . . . . . I find strength in my faith or spiritual beliefs . . . . . . . . . . . . My illness has strengthened my faith or spiritual beliefs . . . . . . I know that whatever happens with my illness, things will be okay
. . . . . . . . . . . .
Not at all
A little bit
Somewhat
Quite a bit
Very much
0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4 4 4 4
Correlates of Self-Perceptions of Spirituality in American Adults Leila Shahabi, B.A. and Lynda H. Powell, Ph.D. Rush-Presbyterian-St. Luke’s Medical Center
Marc A. Musick, Ph.D. University of Texas-Austin
Kenneth I. Pargament, Ph.D. Bowling Green State University
Carl E. Thoresen, Ph.D. Stanford University
David Williams, Ph.D. University of Michigan
Lynn Underwood, Ph.D. Fetzer Institute
Marcia A. Ory, Ph.D. National Institute on Aging
tual and religious may be at particularly low risk for morbidity and mortality based on their good psychological status and ongoing restorative activities.
ABSTRACT To advance knowledge in the study of spirituality and physical health, we examined sociodemographic, behavioral, and attitudinal correlates of self-perceptions of spirituality. Participants were a nationally representative sample of 1,422 adult respondents to the 1998 General Social Survey. They were asked, among other things, to rate themselves on the depth of their spirituality and the depth of their religiousness. Results indicated that, after adjustment for religiousness, self-perceptions of spirituality were positively correlated with being female (r = .07, p < .01), having a higher education (r = .12, p < .001), and having no religion (r = .10, p < .001) and inversely correlated with age (r = –.06, p < .05) and being Catholic (r = –.08, p < .01). After adjustment for these sociodemographic factors, self-perceptions of spirituality were associated with high levels of religious or spiritual activities (range in correlations = .12–.38, all p < .001), low cynical mistrust, and low political conservatism (both r = –.08, p < .01). The population was divided into 4 groups based on their self-perceptions of degree of spirituality and degree of religiousness. The spiritual and religious group had a higher frequency of attending services, praying, meditating, reading the Bible, and daily spiritual experience than any of the other 3 groups (all differences p < .05) and had less distress and less mistrust than the religious-only group (p < .05 for both). However, they were also more intolerant than either of the nonreligious groups (p < .05 for both) and similar on intolerance to the religious-only group. We conclude that sociodemographic factors could confound any observed association between spirituality and health and should be controlled. Moreover, individuals who perceive themselves to be both spiri-
(Ann Behav Med
2002, 24(1):59–68)
INTRODUCTION The possibility that religiousness or spirituality is a protective factor that promotes good physical health has led to a proliferation of epidemiological and psychological research in this area. Two recent reviews of the literature on religiousness and spirituality and physical health (1,2) came to the same conclusion: The strongest and most consistent evidence for any relation is in the link between church or service attendance and all-cause mortality. Those who attend church or services once a week or more frequently have a 25% reduction in risk of death from any cause. This reduction is independent of demographic factors, physical health status, physical function, social support, depression, and healthy risk-factor profiles (2). Because this association cannot be accounted for by these potential confounders, the responsible factors remain unknown. Of particular interest is whether all churchgoers are equally protected or whether there is are subgroup members who, if they could be identified, are at particularly low risk. People attend church or services for a variety of reasons. Some may have intrinsic motivations and live their religion by absorbing it and finding ways to practice the moral precepts taught (3). Others may go for more extrinsic reasons and use their religion as a means of socialization, fulfilling an obligation, or gaining status within an organization (3). Those with intrinsic motivations may be a particularly low risk group if their moral development fosters high levels of psychosocial protective factors including, for example, altruism and forgiveness rather than hostility, joy rather than anger, hope and faith rather than hopelessness and depression, and restorative practices such
Reprint Address: L. H. Powell, Ph.D., Department of Preventive Medicine, Rush-Presbyterian-St. Luke’s Medical Center, Suite 470, 1700 West Van Buren, Chicago, IL 60612. © 2002 by The Society of Behavioral Medicine.
59
60
Shahabi et al.
Annals of Behavioral Medicine
as prayer and meditation rather than chronic sympathetic arousal. Efforts to identify those with a more intrinsic religiousness have resulted in such conceptualizations as private religious practices (e.g., reading the Bible, frequent prayer, listening to religious radio) and depth of religiousness (e.g., deriving strength and comfort from religion, self-perceptions of how deeply religious one is). However, there are, at best, weak links between these constructs and physical health (2), signaling the need for better measures of intrinsic religiousness. Spirituality is an appealing construct because it connotes an intrinsic characteristic that may or may not be induced by religiousness. Whereas religiousness tends to be defined by organizational beliefs, spirituality tends to be defined in personal, experiential terms (4). The personal nature of spirituality, however, creates problems for its scientific study. There is little agreement on what spirituality is, how to measure it, and the extent of its overlap with religiousness. Studies that have examined the health implications of spirituality have used a variety of measures that assess a variety of constructs, which, for the most part, have failed to find any links to physical health (2). Despite the lack of conceptual clarity in the research literature, individuals appear to have little trouble rating themselves on their own level of spirituality. Two studies using small samples of participants selected either because they were medically ill (5) or because they came from preselected religious backgrounds (6) described characteristics of individuals who perceived themselves to be spiritual or religious. The “spiritual” people, in contrast to the “religious” people, were of higher socioeconomic status (5,6), less often from a Catholic background (5), more likely to be independent and reject traditional organized religion (6), and more likely to have a horizontal belief system (7) where concerns were less with connectedness to God and more with the interconnectedness of all living things (5,6). Because these findings were from highly selected samples, their generalizability to more representative populations is not known. The purpose of this study was to identify sociodemographic, behavioral, and attitudinal correlates of self-perceptions of spirituality in a large, nationally representative sample of American adults. We identified these characteristics by comparing and contrasting individuals who considered themselves to be spiritual with those who considered themselves to be religious or to be neither spiritual nor religious. By so doing, we aimed to contribute to a refinement of the concept of spirituality and to identify potentially confounding factors that should be controlled in studies examining the role of spirituality on physical health. METHODS Sample Data for the study came from the General Social Survey (GSS) (8) distributed by the Roper Center for Public Opinion Research. The GSS is an independently drawn sample of persons 18 years old or older who are English-speaking and live in noninstitutionalized settings in the United States. The survey, which has been collected annually from 1972 to 1998, has been
used in numerous studies aimed at the investigation of social characteristics of the American adult public. Its main areas of interest are social in nature, including socioeconomic status, social mobility, social control, family, race relations, civil liberties, and morality. It has been shown to be both reliable and representative of the adult population of the United States. The GSS employs a split-ballot design to collect data. That is, some questions on the survey are asked of all respondents, whereas others are asked of a subset only. The choice of respondents to answer subsets is random; consequently, the smaller subsamples are representative of the larger sample and of the population as a whole. In the 1998 GSS survey, a special ballot on religiousness and spirituality was added. This ballot was designed by a group of experts in the field to represent the multiple dimensions included in the constructs of religion and spirituality (9). Preliminary psychometric data for each of these scales were presented earlier (9). The 1,422 individuals who responded to the special ballot formed the population under investigation. Variables The aim of this study was to create a sociodemographic, behavioral, and attitudinal profile of individuals who perceived themselves to be spiritual. Although the special ballot included questions about self-perceptions of spirituality and religiousness, public and private religiousness, religious denomination, daily spiritual experiences, religious history, religious social support, religious coping, religious beliefs and values, religious commitment, and forgiveness (9), we limited the religious variables chosen to keep the focus on the nonreligious descriptions. Thus, we selected only the behavioral factors (public and private religiousness), religious denomination, and the daily spiritual experiences items from the special ballot. Included, in addition, were attitudinal variables from the larger GSS data set. Table 1 presents a selected group of these variables; their scales; and, where appropriate, an alpha coefficient that estimates the internal consistency reliability of the scale. Self-perceptions of spirituality and religiousness. Respondents were asked two questions regarding their overall spirituality and religiousness: “To what extent to do you consider yourself a spiritual person?” and “To what extent do you consider yourself a religious person?” For each question, respondents rated themselves along a 4-point scale ranging from 1 (very spiritual or very religious) to 2 (moderately spiritual or moderately religious) to 3 (slightly spiritual or slightly religious) to 4 (not at all spiritual or not at all religious). A four-level typology was created by dichotomization of each of the two scales and then combination of the two scales into a 2 × 2 table. For example, the spiritual and religious group (n = 742, 52%) was composed of those respondents who answered very or moderately to both the question about spirituality and the question about religiousness. The spiritual-only group (n = 143, 10%) was composed of those respondents who answered very or moderately to the spiritual question and slightly or not at all to the religion question. The religious-only group (n = 128, 9%) was composed of those respondents who answered
Volume 24, Number 1, 2002
Correlates of Spirituality
61
TABLE 1 Variable Descriptions and Reliability Estimates for Scales Variable Religious/Spiritual activity Service attendance Private prayer Meditation Reads the Bible Daily Spiritual Experiences
Religious subscale Spiritual subscale Well-being Psychological distress
Cynical mistrust
Self-rated health Beliefs and values Conservatism
Fatalism
Justice
Nihilism
Question
Anchors
a
How often do you attend religious services? How often do you pray privately in places other than at church or synagogue? Within your religious or spiritual tradition, how often do you meditate? How often have you read the Bible in the last year? The following questions deal with possible daily spiritual experiences. To what extent can you say you experience the following: (a) I feel God’s presence (b) I find strength and comfort in my religion (c) I feel deep inner peace or harmony (d) I desire to be closer to or in union with God (e) I feel God’s love for me, directly or through others (f) I am spiritually touched by the beauty of creation Items a, b, d, and e Items c and f
0 (never) to 8 (several times a week) 1 (never) to 8 (more than once a day) 1 (never) to 8 (more than once a day) 1 (less than once a week) to 5 (several times a day) 1 (never or almost never) to 6 (many times a day)
— —
In the past 30 days how often did you feel: (a) So sad nothing could cheer you up, (b) nervous, (c) restless or fidgety, (d) hopeless, (e) that everything was an effort, and (f) worthless. Do you think most people would try to take advantage of you if they got a chance or would they try to be fair? Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people? Would you say that most of the time people try to be helpful or that they are mostly just looking out for themselves? Would you say your own health, in general, is excellent, good, fair, or poor?
1 (none of the time) to 5 (all of the time)
.85
1 (would try to be fair) or 2 (would take advantage of you)
.67
I’m going to show you a 7-point scale on which the political views that people might hold are arranged from extremely liberal to extremely conservative. Where would you place yourself on this scale? Do you agree or disagree? There is little that people can do to change the course of their lives. Some people say that people get ahead by their own hard work; others say that lucky breaks or help from other people are more important. Which do you think is the most important? Do you agree or disagree? In my opinion, life does not serve any purpose.
— — .91
.91 .70
1 (most people can be trusted) or 2 (can’t be too careful) 1 (try to be helpful) or 2 (just look out for themselves) 1 (poor) to 4 (excellent)
—
1 (extremely liberal) to 7 (extremely conservative)
—
1 (strongly disagree) to 4 (strongly agree)
—
1 (luck most important) to 3 (hard work most important)
1 (strongly disagree) to 4 (strongly agree)
—
(continued)
62
Shahabi et al.
Annals of Behavioral Medicine TABLE 1 (Continued)
Variable Intolerance (total score) Atheists subscale
Racists subscale
Militarists subscale
Communists subscale
Homosexuals subscale
Question
Anchors
There are always some people whose ideas are considered bad or dangerous by other people. For instance, somebody who is against all churches and religion … (a) If such a person wanted to make a speech in your (city/town/community) against churches and religion, should he be allowed to speak or not? (b) Should such a person be allowed to teach in a college or university, or not? (c) If some people in your community suggested that a book he wrote against churches and religion should be taken out of your public library, would you favor removing this book, or not? Or consider a person who believes that Blacks are genetically inferior. (same set of follow-up questions as Atheists subscale) Consider a person who advocates doing away with elections and letting the military run the country. (same set of follow-up questions as Atheists subscale) Now, I should like to ask you some questions about a man who admits he is a communist. (same set of follow-up questions as Atheists subscale) And what about a man who admits that he is a homosexual? (same set of follow-up questions as Atheists subscale)
slightly or not at all to the spiritual question and very or moderately to the religious question. The neither spiritual nor religious group (n = 409, 29%) was composed of those respondents who answered slightly or not at all to both the spiritual and religious questions. Sociodemographic variables. The demographic indicators included age, sex, ethnicity, city size, geographic location, and marital status. Indexes of socioeconomic status were education and family income. The coding for these variables is presented in Table 2. Religious denomination. Because respondents who identified themselves in different ways along the spiritual–religious continuum may have clustered along the lines of particular religious faiths, we included several dichotomous measures of religious affiliation. Respondents who stated that their religious preference was Catholic or no religion were coded as such. Respondents who stated that their preference was Protestant were then asked to indicate their denomination. With the classification scheme designed by Roof and McKinney (10), Protestant denominations were classified into three groups: conservative (e.g., Southern Baptists, Pentecostals), moderate (e.g., Luther-
a .87 .71
1 (yes, allowed to speak) or 2 (not allowed)
1 (yes, allowed to teach) or 2 (not allowed) 1 (not favor) or 2 (favor)
.71
(same anchors as Atheists subscale)
.74
(same anchors as Atheists subscale)
.79
(same anchors as Atheists subscale)
.73
(same anchors as Atheists subscale)
.79
ans, Methodists), and liberal (e.g., Episcopalians, Presbyterians). Religious or spiritual activity. Several variables that assessed common religious behaviors were included to assess extent of religious activity. Church or service attendance represented the extent of public religious activity. Religious and spiritual practices included private prayer, meditation, and Bible reading. The Daily Spiritual Experience Scale (9) is a six-item scale that measures the frequency with which the respondent has spiritual experiences on a day-to-day basis. This scale has items that represent spiritual experiences phrased in religious terminology (Religious subscale α = .91) and items that represent spiritual experiences that do not require religious terminology (Spiritual subscale α = .70). Well-being. We include three measures of well-being. The first, Psychological Distress, is a six-item index designed to assess depressive and anxious feelings over the month preceding the survey. The second measure is a three-item scale measuring cynical mistrust of the motives of others (11). The third measure is a standard item assessing self-rated health. Although self-rated health measures contain information about both phys-
Volume 24, Number 1, 2002
Correlates of Spirituality TABLE 2 (Continued)
TABLE 2 Description of Sociodemographic Factors and Religious Activity of the Special Spirituality/Religion Panel of the 1998 GSS Population Variable Sample size Sociodemographic factors Female Ethnicity White African American Other Agea City size City Suburb Unincorporated area Town or village Open area (< 2,499) Geographic location New England Middle Atlantic South Atlantic North Central South Central Mountain Pacific Marital status Married Widowed Divorced/Separated Never married Education Less than high school High school Associate/Junior college Bachelor’s degree Graduate degree Family income Less than $10,000 $10,000–$19,999 $20,000–$29,999 $30,000–$59,999 $60,000–$89,999 > $90,000 Religious beliefs and activity Religious affiliation Protestant Catholic Jewish Other None Religious attendance Several times a week Every week Nearly every week About once a month Several times a year
n
63
Variable
%
1,422 775
54.5
1,124 194 104 45.6
79.0 13.6 7.3 17
496 526 75 194 131
34.9 37.0 5.2 13.7 9.2
70 225 260 339 256 91 181
4.9 15.8 18.3 23.8 18.0 6.4 12.7
674 142 258 348
47.4 10.0 18.2 24.5
228 747 106 228 107
16.0 52.5 7.5 16.0 7.5
159 184 401 399 156 123
12.7 12.9 28.2 28.1 11.0 8.6
774 365 26 60 192
54.4 25.7 1.8 4.6 13.5
115 242 203 115 302
8.1 17.0 14.3 8.1 21.3 (continued)
Less than once a year Never Frequency of prayer Several times a day Once a day Several times a week Once a week Less than once a week Never Frequency of meditation More than once a day Once a day A few times a week Once a week A few times a month Once a month Less than once a month Never
n
%
151 274
10.6 19.3
349 421 201 108 308 24
24.5 29.6 14.1 7.6 21.7 1.7
123 187 129 68 60 46 94 654
8.6 13.2 9.1 4.8 4.2 3.2 6.6 46.0
Note. GSS = General Social Survey. aThese values are mean and standard deviation, respectively.
ical and mental health (12,13), they have been found to predict mortality (14–16) and coronary heart disease (17). Beliefs and values. This domain is measured with four single-item indicators. The first, Conservatism, is an indicator of the political leanings of the respondent, where high scores denote political conservativism and low scores denote political liberalism. The second, Fatalism, measures the degree to which respondents believe that they can control their destiny. The third, Justice, measures beliefs about success being a function of hard work or lucky breaks. The fourth, Nihilism, measures beliefs about whether life has any purpose. Intolerance. Because religious activity has been linked to intolerance toward certain groups (18), it is possible that respondents will vary in levels of intolerance based on their religious position. The GSS Intolerance Scale is composed of a total score and subscales scores that measure intolerance toward each of five different groups: atheists, racists, militarists, communists, and homosexuals. For each group, three questions were posed about whether a member of the specific subgroup should be allowed to make a speech, teach in a college or university, or have his or her books in a public library. Statistical Analyses The analytic goal of this study was to identify correlates of self-perceptions of spirituality. Toward this end, analyses followed several steps. First, to describe the population under study, frequency distributions for selected variables were determined. Second, to identify correlates of self-perceptions of spirituality and compare them to self-perceptions of religiousness, partial correlations between spirituality and religiousness and
64
Shahabi et al.
the full set of demographic, behavioral, and attitudinal variables were calculated. To ensure that these correlations were not confounded by extraneous factors, they were determined first after the other spiritual/religious variable was partialled out and then after the other spiritual–religious variable and the full set of sociodemographic correlates were partialled out. Next, we moved to an examination of the correlates of the four-level spiritual–religious typology in which information about both self-perceived spirituality and self-perceived religiousness was combined. To examine sociodemographic correlates of the four-level typology, we assumed that the direction of causality was that these factors predict placement on the typology rather than the reverse. Consequently, we modeled the association by regression of the typology (the dependent variable) on the sociodemograhic factors (the independent variables) using multinomial logistic regression. This procedure is akin to logistic regression in that it computes the likelihood of being in one category of the dependent variable versus another, based on levels of the independent variables. However, it differs from logistic regression in that it allows the dependent variable to be more than two categories. Because our greatest interest is in understanding spirituality, we calculated the odds for these analyses with the spiritual and religious group as the referent and compared it to the other three groups. In the final stage of the analyses, we posited that other factors, such as behaviors and attitudes, were associated with the spirituality–religiousness typology. Consequently, we modeled these associations by generating mean scores of the attitudinal–behavioral items based on levels of the typology. Given that levels of the typology were influenced by sociodemographic confounders, we adjusted first for these factors. All analyses were completed with the PROC CORR, PROC GLM, and PROC CATMOD procedures within the SAS software package. RESULTS Table 2 presents a description of the 1,422 respondents that were evaluated in the spirituality–religion panel of the 1998 GSS. The average age was 45.6 years, with 68% of participants falling within the approximate range of 28 to 62. The population was 54.5% female, 79% White, and 47.4% married, and 16% had less than a high school education. Approximately 34.9% lived in an urban area, and 36.3% lived in the South. These characteristics essentially mirrored those of the overall United States, where according to the 2000 census or the 2000 Current Population Survey (19), 50.9% were female, 75.1% were White, 52.9% were married, and 21.5% had less than a high school education. With regard to religious activities, 80% of the sample was either Protestant (54.5%) or Catholic (25.5%), and 13.5% reported that they had no religion. Approximately 31.6% attended church or services nearly every week or more, 54.1% prayed at least once a day, and 21.8% meditated at least once a day. This was comparable to religious activity in the U.S. population (20). Self-perceptions of spirituality and of religiousness were highly correlated (rxy = .63). Table 3 presents correlations between these self-perceptions and the entire set of independent
Annals of Behavioral Medicine variables. In general, there was a wide range of significant correlations (range = .07–.52), and those that were weak (e.g., .07) were nonetheless significant because of the large sample size. Among the demographic factors, self-perceptions of spirituality, after adjustment for religiousness, were associated with being female, younger, and better educated. In contrast, self-perceptions of religiousness, after adjustment for spirituality, were associated with being in an ethnic minority, older, and less well educated; living in a smaller town; and living in the South. Self-perceptions of spirituality were inversely correlated with being Catholic and positively correlated with having no religion, whereas self-perceptions of religiousness were correlated with being a conservative or moderate Protestant and inversely correlated with having no religion. After adjustment for these sociodemographic differences, self-perceptions of spirituality were correlated significantly with all of the religious activities. However, these correlations were weaker than the corresponding correlations with religiousness. The only exceptions to this were for meditation and the Spiritual subscale of the Daily Spiritual Experience Scale, which were both more strongly related to self-perceptions of spirituality (rxy = .28 and .38, respectively) than they were to self-perceptions of religiousness (rxy = .09 and .20, respectively). Among the indexes of well being, self-perceptions of spirituality were inversely associated with cynical mistrust (rxy = –.08), and self-perceptions of religiousness were associated with better self-rated health (rxy = .07). Among the beliefs and values, spirituality was inversely related to being conservative (rxy = –.08), in contrast to religiousness, which was positively related to being conservative (rxy = .14). Spirituality was unrelated to being intolerant, in contrast to religiousness, which was associated with the Intolerance total score (rxy = .11) and intolerance on the Atheist, Militarist, and Homosexual subscales (rxy = .12, .10, and .13, respectively). Our next step in the analyses was to examine the correlates of self-perceptions when information about spirituality and religiousness was combined to form the four-group typology. Table 4 focuses on the sociodemographic variables and presents the independent odds of having each of the variables, with the spiritual and religious group as the referent. Relative to this referent, the spiritual-only group had an 80% lower odds of being minority, a 46% lower odds of being from the South, and a 52% lower odds of being married. In addition, the spiritual-only group was younger and better educated. These demographic differences were less extreme when the religious-only group was compared to the referent and more extreme when the neither spiritual nor religious group was compared to the referent. The only exception to this was the educational level of the neither spiritual nor religious group, which was lower than the referent and considerably lower than the spiritual-only group. Table 5 presents the behavioral and attitudinal profiles of each of the spiritual–religious subgroups, after the sociodemographic confounders were controlled. The spiritual and religious group engaged in more religious or spiritual activity than any of the other three groups. This included more frequent service attendance, private prayer, meditation, Bible reading, and daily
Volume 24, Number 1, 2002
Correlates of Spirituality
65
TABLE 3 Correlates of Self-Perceptions of Spirituality and Religiousness Spirituality Variable Sociodemographic factors Female Minority Age City sizec South Married Education Family income Denomination Conservative Protestant Moderate Protestant Liberal Protestant Catholic No religion Religious/Spiritual activity Service attendance Private prayer Meditation Reads the Bible Daily spiritual experiences Religious subscale Spiritual subscale Well-being Psychological distress Self-rated health Cynical mistrust Beliefs and values Conservatism Fatalism Nihilism Justice Intolerance Total score Atheists subscale Racists subscale Militarists subscale Communists subscale Homosexuals subscale
Religiousness
Adjusted for Religiousness
Multiple Adjustmentsa
Adjusted for Spirituality
Multiple Adjustmentsb
.07** .03 –.06* .02 .02 –.02 .12*** –.02
— — — — — .02 — —
.05 .09*** .17*** –.07* .10*** .11*** –.11*** –.02
— — — — — .07* — —
–.03 .01 .00 –.08** .10***
— — — — —
.16*** .08** .03 .06 –.42***
— — — — —
.12*** .18*** .27*** .19*** .35*** .29*** .37***
.12*** .20*** .28*** .18*** .34*** .28*** .38***
–.02 .00 –.05
.00 –.03 –.08**
–.09*** –.03 –.06* .00
–.08** .00 –.02 .01
–.05 –.05 –.01 –.08* –.02 –.02
.01 .01 .03 –.03 .03 .04
.44*** .43*** .16*** .32*** .48*** .52*** .25*** .01 .00 .04
.34*** .30*** .09** .26*** .40*** .43*** .20*** –.01 .07* .05
.19*** .04 –.09** .01
.14*** –.03 –.09** –.02
.21*** .21*** .13*** .18*** .10** .19***
.11** .12*** .04 .10** .03 .13***
aAdjusted for religiousness, sex, age, race, education, income, city size, region, and religious affiliation. bAdjusted for spirituality, sex, age, race, education, income, city size, region, and religious affiliation. cLarger values indicate larger cities. *p < .05. **p < .01. ***p < .001.
spiritual experience. They experienced significantly less psychological distress and cynical mistrust than the religious-only group. They tended to be more politically conservative and less nihilistic (e.g., more inclined to believe that life has purpose) than the nonreligious groups. They were similar to the religious-only group on their level of intolerance, which was significantly higher than that observed in either of the two nonreligious groups. In contrast, the spiritual-only group was more politically liberal and more likely to claim to have no religion (34%) than
any other group, including the neither spiritual nor religious group (26%). They engaged in religious and spiritual activities but not as frequently as the spiritual and religious group. Compared to the religious-only group, they engaged in fewer religious activities (e.g., service attendance, prayer, Bible reading, Daily Spiritual Experience–Religious subscale) but more spiritual activities (e.g., meditation, Daily Spiritual Experience–Spiritual subscale; nonsignificant trend, p < .10). They were more inclined to be nihilistic (e.g., to believe that life has
66
Shahabi et al.
Annals of Behavioral Medicine
TABLE 4 Independent Sociodemographic Correlates of Spirituality/Religion Typology
Variable Female Minority Age City sized South Married Education Family incomee Model fit statistics χ2 df Pseudo R2 N
Spiritual Onlya 0.76 0.20*** 0.98*** 1.03 0.54** 0.48*** 1.09* 1.01
Religious Onlyb
Neither Spiritual Nor Religiousc
0.69 0.51* 1.00 1.01 0.63* 0.90 0.95 1.00
0.59*** 0.45*** 0.98*** 1.06* 0.57*** 0.61** 0.91*** 1.03*
166.78 24 .10 1,422
Note. Odds ratios are shown with the spiritual and religious group (n = 742) as the referent. an = 143. bn = 128. cn = 409. dLarger values indicate larger cities. eLarger values indicate higher incomes. *p < .05. **p < .01. ***p < .001.
no purpose) than either one of the two religious groups and were more tolerant than the religious and spiritual group only. The religious-only group (those who perceived themselves to be slightly spiritual or not at all spiritual) was distinctive in that they were higher on distress and cynical mistrust than the spiritual and religious group. Thus, within the religious groups, the self-perception of being very or moderately spiritual was associated with better psychological status.
DISCUSSION To understand better what people mean when they call themselves spiritual, we analyzed data from a nationally representative population of American adults. By comparing, contrasting, and integrating self-perceptions of spirituality with self-perceptions of religiousness, we developed a sociodemographic, attitudinal, and behavioral profile. From these analyses, we made several observations that can help to advance our understanding of self-perceptions of spirituality, a construct that has been referred to as “obscure” (21) and “fuzzy” (6). First, people who see themselves as spiritual have a different sociodemographic profile than those who see themselves as either religious or nonreligious. Self-perceptions of spirituality are associated with being younger, female, and highly educated. This is in contrast to self-perceptions of religiousness, which are more likely to be associated with being older, in an ethnic minority, less well educated, and living in the South. Furthermore, compared to those who see themselves as nonspiritual and nonreligious, those who perceive themselves to be both spiritual and religious differ on each of the sociodemographic factors studied. Because such factors as being younger, female, and highly edu-
cated are all related to good health, they could confound any observed relation between spirituality and health. That is, a finding that spirituality is associated with good health may be alternatively explained by the younger age, more frequent being female, or the higher socioeconomic status of those identified as spiritual. We believe that it is imperative to control for these sociodemographic factors before making inferences about any association between spirituality and health. Second, people who see themselves as spiritual are not a homogeneous group. When information about spirituality and religiousness was combined into a four-level typology, it was clear that spiritual people actually comprised two subgroups that differ on ethnicity, age, education, marital status, and geographic region of residence. The larger group, spiritual and religious, composed 52% of our sample and had a demographic profile that was characteristic of religious people. The smaller group, spiritual only, composed only 10% of our sample and had a demographic profile that was characteristic of religious independents (22). It was only when we studied the spiritual people as two distinct groups that we came on our most important finding: The spiritual and religious group appeared to be the most intrinsically religious. This group attended services, prayed, meditated, read the Bible, and had more daily spiritual experiences than any other group. In particular, they were higher than the spiritual-only group on their frequency of meditating, feeling deep inner peace or harmony, and being spiritually touched by the beauty of creation. Moreover, they were less distressed and less mistrusting than the religious-only group. They appeared to take the best from the spiritual side and the best from the religious side, an observation also made by others (5,6). The only characteristic on which they were not highly functioning was intolerance, which appeared to be a byproduct of religiousness, regardless of level of spirituality. The spiritual-only group also appeared to be developing their spirituality but not as frequently or to the same degree. They prayed, meditated, and had daily spiritual experiences less often than the spiritual and religious group but more often than the neither spiritual nor religious group. The spiritual-only group has been described as “New Age” (6), but this observation may have been biased by the large number of individuals from New Age groups that were involved in that study. Although New Age characteristics were not measured in our study, we were impressed with how mainstream this group actually appeared. The main factors that distinguished them were their political liberalism, tolerance of the expression of diverse points of view, and unwillingness to claim any religious denomination. They appear to be more like Roof’s (23) description of baby boomers who renounced traditional religion to pursue their own, perhaps more circuitous, path for spiritual development. One mechanism by which religion or spirituality may enhance physical health is via salutary psychosocial and psychophysiological status (24). As such, these data suggest the hypothesis that a subgroup of healthy individuals, or of healthy service attenders, who may be at particularly low risk for morbidity and mortality is the group who perceive themselves to be
Volume 24, Number 1, 2002
Correlates of Spirituality
67
TABLE 5 Comparison of Adjusted Mean Levels of Selected Variables by Spirituality and Religiousness Typology
Variable Sociodemographics (%) Married Denomination (%) Conservative Protestant Moderate Protestant Liberal Protestant Catholic No religion Religious/Spiritual activity Service attendance Private prayer Meditation Reads the Bible Daily spiritual experiences Religious subscale Spiritual subscale Well-being Psychological distress Self-rated health Cynical mistrust Beliefs and values Conservatism Fatalism Nihilism Justice Intolerance Total score Atheists subscale Racists subscale Militarists subscale Communists subscale Homosexuals subscale
Group 1: Spiritual and Religious
Group 2: Spiritual Only
Group 3: Religious Only
Group 4: Neither Spiritual Nor Religious
51a,c
39b
51c
41
29a,c 19 07 24b 04a,c
16 13 04 22b 34b,c
22 19 07 35c 02c
21 14 05 23 26
4.86a,b,c 6.60a,b,c 4.22a,b,c 2.79a,b,c 4.61a,b,c 4.63a,b,c 4.57a,b,c
2.26b 4.71b,c 3.55b,c 1.71b 3.46b,c 3.16b,c 4.03c
3.79c 5.67c 2.61 1.93c 3.85c 3.83c 3.88c
1.82 3.62 2.08 1.50 2.70 2.52 3.07
1.99b 3.10 1.97b
2.03 2.96 1.90
2.14c 3.11 2.17
1.96 3.02 2.05
4.17a,c 1.93 1.48a,c 2.53
3.60b,c 1.89 1.72b 2.59
4.13 1.83 1.49c 2.60
3.93 1.98 1.72 2.55
3.8a,c 3.5a,c 4.5c 4.0a,c 4.2a,c 2.7a,b,c
2.9 2.4b 3.9 3.9 3.7 1.7
3.5c 3.6c 3.8 3.8 4.2 1.9
2.9 2.3 3.3 3.3 3.7 1.7
Note. Means are adjusted for sex, age, race, education, income, city size, and region. Group differences indicate significant differences (p < .05) between the groups as indicated by the following superscripts: a = different than Group 2; b = different than Group 3; c = different than Group 4. Group 1, n = 742; Group 2, n = 143; Group 3, n = 128; Group 4, n = 409.
both spiritual and religious. They have good psychosocial status by virtue of having low levels of distress and cynical mistrust, two risk factors for mortality (25–27), and potentially good psychophysiological status by virtue of having high levels of potentially restorative activities (e.g., service attendance, prayer, meditation, and daily experiences of comfort, peace, and awe), which may promote ongoing shifts from sympathetic arousal to parasympathetic relaxation or lower levels of cortisol secretion (see Ironson et al., this issue). The strength of this study is its large, representative sample size. However this large sample is limited because even weak associations become statistically significant and more in-depth description of attitudes and behaviors is not possible. Moreover, a large number of statistical tests were conducted, and working at .05 level of significance, approxi-
mately 5% of the tests would be significant due to chance alone. However, we place validity in our findings because of their consistency with those of two similar studies conducted on highly select samples (5,6). We believe these studies complement ours and should be referred for more in-depth descriptions of attitudes and behaviors of people who perceive themselves to be spiritual. In summary, the inherently personal nature of spirituality is perhaps its greatest strength but also its greatest weakness, at least as far as scientific study is concerned. Its strength lies in the possibility that it reflects the internalization of virtues advocated across numerous religious traditions and in the promise such a conceptualization offers for establishing a relation to physical health. Its weakness is that the wide variety of conceptualizations and assessment options limit comparisons across studies
68
Shahabi et al.
and leave unanswered questions about consistency in research findings. We took the approach of going to the general public and asking them to define for themselves where they believed they fell along the spirituality continuum rather than relying on our own preconceived notions about what spirituality is. We gained trust in the validity of these self-perceptions because they were consistent with religious or spiritual restorative behaviors and salutary psychological status. The combination of being both spiritual and religious is an intriguing concept because it may suggest that efforts toward spiritual development are maximized when pursued within the context of a supportive religious environment. That is, going each week to a serene place that encourages reflection, listening to sermons that advocate love and service to others, and experiencing the joy of joining in song in praise of faith may provide direct, ongoing, and reinforcing experiences of the virtues enjoyed by highly spiritual people. Conversely, it is not inconceivable that people who are already highly spiritual would seek out religious contexts and, in the process, develop their religiousness. We look forward to future studies that test whether individuals who perceive themselves to be both religious and spiritual are at particularly low risk for death and disease.
Annals of Behavioral Medicine
(10)
(11) (12)
(13)
(14)
(15)
(16)
(17)
(18)
REFERENCES (1) McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoresen CE: Religious involvement and mortality: A meta-analytic review. Health Psychology. 2000, 19:211–222. (2) Powell LH, Shahabi L, Thoresen CE: Religion and spirituality: Linkages to physical health. Society for Epidemiologic Research Annual Meeting. Toronto, Ontario, Canada: 2001. (3) Allport GW, Ross JM: Personal religious orientation and prejudice. Journal of Personality and Social Psychology. 1967, 5:432–443. (4) Spilka B, McIntosh DN: Religion and spirituality: The known and the unknown. American Psychological Association Annual Meeting. Toronto, Ontario, Canada: 1996. (5) Woods TE, Ironson GH: Religion and spirituality in the face of illness. Journal of Health Psychology. 1999, 4:393–412. (6) Zinnbauer BJ, Pargament KI, Cole B, et al.: Religion and spirituality: Unfuzzying the fuzzy. Journal for the Scientific Study of Religion. 1997, 36:549–564. (7) Davidson J: Glock’s model of religious commitment: Assessing some different approaches and results. Review of Religious Research. 1975, 16:83–93. (8) Davis JA, Smith TW: General Social Surveys, 1972–1991. Chicago: National Opinion Research Center, 1991. (9) Fetzer Institute/National Institute on Aging Working Group: Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research: A Report of the Fetzer Institute/Na-
(19) (20) (21)
(22)
(23) (24)
(25)
(26)
(27)
tional Institute of Aging Working Group. Kalamazoo, MI: Fetzer Institute, 1999. Roof WC, McKinney W: American Mainline Religion: Its Changing Shape and Future. New Brunswick, NJ: Rutgers University Press, 1987. Demaris A, Yang R: Race, alienation, and interpersonal mistrust. Sociological Spectrum. 1994, 14:327–349. Fylkesnes K, Forde OH: Determinants and dimensions involved in self-evaluation of health. Social Science and Medicine. 1992, 35:271–279. Manderbacka K, Lundberg O, Martikainen P: Do risk factors and health behaviors contribute to self-ratings of health? Social Science and Medicine. 1999, 48:1713–1720. Idler EL: Self-assessed health and mortality: A review of studies. In Maes S, Leventhal H, Johnston M (eds), International Review of Health Psychology. New York: Wiley, 1992, 33–54. McCallum J, Shadbolt M, Wang D: Self-rated health and survival: A 7-year follow-up study of Australian elderly. American Journal of Public Health. 1994, 84:1100–1105. Dasbach EJ, Klein R, Klein BEK, Moss SE: Self-rated health and mortality in people with diabetes. American Journal of Public Health. 1994, 84:1775–1779. Moller L, Kristensen TS, Hollnagel H: Self-rated health as a predictor of coronary heart disease in Copenhagen, Denmark. Journal of Epidemiology and Community Health. 1996, 50:423–428. Ellison CG, Musick MA: Southern intolerance: A Fundamentalist effect? Social Forces. 1993, 72:379–398. United States Department of Commerce: U.S. 1990 Census data [online]. Available from http://www.census.gov Gallup Organization: 2000 Gallup poll [online]. Available from http://www.gallup.com Hood RW, Spilka B, Hunsberger B, Gorsuch R: The Psychology of Religion: An Empirical Approach. New York: Guilford, 1996. Hayes BC: Religious independents within Western industrialized nations: A socio-demographic profile. Sociology of Religion. 2000, 61:191–207. Roof WC: A Generation of Seekers: The Spiritual Journeys of the Baby Boom Generation. San Francisco: Harper, 1993. Levin J: How religion influences morbidity and health: Reflections on natural history, salutogenesis and host resistance. Social Science and Medicine. 1996, 43:849–864. Miller TQ, Smith TW, Turner CW, et al.: A meta-analytic review of research on hostility and physical health. Psychological Bulletin. 1996, 119:322–348. Allison TG, Williams DE, Miller TD, et al.: Medical and economic costs of psychologic distress in patients with coronary artery disease. Mayo Clinic Proceedings. 1995, 70:734–742. Denollet J, Brutsaert DL: Personality, disease severity, and the risk of long-term cardiac events in patients with a decreased ejection fraction after myocardial infarction. Circulation. 1998, 97:167–173.
Should Clinicians Incorporate Positive Spirituality Into Their Practices? What Does the Evidence Say? Walter L. Larimore, M.D. Focus on the Family
Michael Parker, D.S.W. Department of Social Work The University of Alabama
Martha Crowther, Ph.D., M.P.H. Department of Psychology The University of Alabama
Until the 20th century, religion and medicine were closely linked (1). The famous Johns Hopkins physician, Sir William Osler, wrote in 1910 in the first edition of the British Medical Journal, “Nothing in life is more wonderful than faith—the one great moving force which we can neither weigh in the balance nor test in the crucible” (2). The earliest practice linkages of social work were religious. Religious organizations were the first sponsors of social service programs in America, and most of the first social workers in the Charity Organization Society and settlement house movements shared a common spiritual mission. Despite these historical roots, the social work profession gradually went through secularization and professionalization processes that emulated psychiatry and the medical model (3). During the 20th century, medicine, psychiatry, psychology, and social work replaced religion and spirituality with naturalism, empiricism, secular humanism, and libertarian morality as the primary sources of ethics and values (4). Yet until the last few decades of the 20th century, the medical sciences had not begun to study the relation between measures of religion and spirituality and mental and physical health. During the 20th century, religion and science were considered by the academic, scientific, and medical communities to be separate realms of thought whose presentation in the same context leads to misunderstanding of both (5). Religiousness was labeled “equivalent to irrational thinking and emotional disturbance” (6, p. 637). It has now been demonstrated that such notions were not based on any scientific evidence but rather on non-evidence-based clinical impressions. Research indicated that religious and spiritual beliefs and practices were widespread among the American population and that these beliefs and practices had clinical relevance. Therefore, toward the end of the 20th century, professional organizations increasingly called for greater sensitivity and better training of clinicians concerning the management of religious and spiritual issues in the assessment and treatment of patients; these organizations included the American Psychiatric Association in 1989, the American Psychological Association in 1992, the Accreditation Council for Graduate Medical Education in 1994, the Council on Social Work Education in 1995, the Joint Commission on the Accreditation of Healthcare Organizations in 1996, the Ameri-
ABSTRACT Most of the rhetoric decrying the incorporation of basic and positive spiritual care into clinical practice is not based on reliable evidence. We briefly review the current evidence, which demonstrates that (a) there is frequently a positive association between positive spirituality and mental and physical health and well being, (b) most patients desire to be offered basic spiritual care by their clinicians, (c) most patients censure our professions for ignoring their spiritual needs, (d) most clinicians believe that spiritual interventions would help their patients but have little training in providing basic spiritual assessment or care, (e) professional associations and educational institutions are beginning to provide learners and clinicians information on how to incorporate spirituality and practice, and (f) anecdotal evidence indicates that clinicians having received such training find it immediately helpful and do apply it to their practice. We point out the reasons that much more research is needed, especially outcome-based, clinical research on the effects of these spiritual interventions by clinicians. We conclude that the evidence to date demonstrates trained or experienced clinicians should encourage positive spirituality with their patients and that there is no evidence that such therapy is, in general, harmful. Further, unless or until there is evidence of harm from a clinician’s provision of either basic spiritual care or a spiritually sensitive practice, interested clinicians and systems should learn to assess their patients’ spiritual health and to provide indicated and desired spiritual intervention. Clinicians and health care systems should not, without compelling data to the contrary, deprive their patients of the spiritual support and comfort on which their hope, health, and well-being may hinge. (Ann Behav Med
2002, 24(1):69–73)
Although the views expressed in this article reflect our exclusive opinions, we gratefully acknowledge the assistance of the John A. Hartford Foundation’s Geriatric Faculty Scholars Program. Reprint Address: M. Crowther, Ph.D., M.P.H., The University of Alabama, Department of Psychology, Box 870348, Tuscaloosa, AL 35487–0348. E-mail:
[email protected] © 2002 by The Society of Behavioral Medicine.
69
70
Larimore et al.
can Academy of Family Physicians (AAFP) in 1997, the American College of Physicians in 1998, and the Association of American Medical Colleges in 1998. A random survey of almost 300 physicians at the 1996 meeting of the AAFP revealed that 99% believed that spiritual well-being can promote health and healing. Seventy-five percent believed that others’ prayers could promote healing (7). Another survey reported that the majority of family physicians believed spiritual well-being is an important component in health (8). Despite this belief, most of these physicians reported infrequent discussions of spiritual issues with patients and infrequent referrals of hospitalized patients to chaplains. Why? The physicians reported not that they lacked the interest, but they lacked appropriate training. For example, Ellis, Vinson, and Ewigman (8) reported that 59% of physicians feel uncomfortable taking a family history, whereas 56% reported uncertainty about how to determine which patients desired spiritual discussion, and 49% reported uncertainty about how to handle spiritual matters. Medicine is not the only profession that has not prepared its clinicians to provide basic spiritual services to patients. Surveys of social work faculty have suggested that almost 90% report that religious or spiritual issues were never or rarely presented in their graduate social work studies, and yet over 80% of those being surveyed favored the development of specialized elective courses to address spirituality and practice (9). In parallel to other professions, calls for spiritually sensitive strategies for research and evaluation of practice are increasingly being made by social work researchers and educators (10). For the past several years, Walter L. Larimore has taught a Continuing Medical Education course, sponsored by the Christian Medical Association, to more than 6,000 health care clinicians. The course emphasizes how clinicians could incorporate research-based positive spirituality into their clinical practices. Over 99% of the attendees, in precourse surveys, reported interest in the ethical and practical how-to’s of incorporating basic spiritual skills (how to utilize a spiritual assessment, how and when to provide a spiritual consult or referral, and how and when to pray with a patient or family) into their practices. In postcourse surveys, over 97% of attendees reported satisfaction with the training, and over 95% predicted that they could use one or more of these spiritual interventions in their practice. When attendees were surveyed 6 or 12 months after the course, over 90% reported that they were able to incorporate and had continued to incorporate this training into their practice. However, there are those who question whether the use of such spiritual interventions is either wise or ethical. For example, two recent commentaries (11,12) attempted to minimize the ability of and to question the ethics of clinicians that desire to assess and address their patients’ spiritual needs. In the first commentary, two PhDs and several theologians and chaplains from New York City wrote, “It is not clear that physicians should engage in religious discussions with patients as a way of providing comfort ” (11). Another group of academicians suggested that “it is a general mandate of modern devel-
Annals of Behavioral Medicine oped societies to keep professional roles separate … [as] distinct spheres of activity … [to] ensure competence and boundaries” (12). They asserted that clinicians “might need to explain to patients why [spiritual] activities usually fall better under the purview of competent pastoral care” (12). Unfortunately, although these assertions may appear to be evidence based, they are unaccompanied by any outcome-based research. Our view of the evidence is significantly different. We believe that there are sufficient, research-based reasons for clinicians to provide basic spiritual interventions, albeit cautiously, for example, with their patient’s permission, and with respect and sensitivity to the multiple ethical issues such interventions entail. Over 35 systematic reviews have all concluded that in the vast majority of patients, the apparent benefits of intrinsic religious belief and practice outweigh the risks (13,14). Furthermore, surveys indicate that a sizeable majority of patients want their physician to address religious and spiritual issues in the context of a clinical visit (1,13). Fears of religious abuse and claims of possible negative effects of religion on health, although deserving of discussion, prevention, and investigation, are highly speculative and have no basis in population-based systematic reviews. In fact, the vast majority of the cross-sectional and prospective cohort studies have shown that religious beliefs and practices are consistently associated with better mental and physical health outcomes (1,15). Some critics have asserted that the magnitude of these effects is weak and inconsistent; others have claimed these effects do not reflect risk. The reader should be aware that these criticisms are the distinct minority of opinions among the 1,600 publications in this area (1). We believe that objective observers will conclude that the apparent health benefits of positive spirituality are not established beyond doubt and that better research is needed. Further, we would acknowledge the absence of a unifying theoretical framework that would foster interdisciplinary thinking about spiritual interventions by clinicians. We also join with our critics in recognizing that outcome-based, clinical research on the effects of spiritual interventions is almost nonexistent (16). However, to claim that there is no evidence to support either the training of clinicians in basic spiritual intervention or the practice of the same by experienced or interested clinicians is, in our view and in the view of others (1,15), uninformed. One group wrote that “the absence of compelling empiric evidence and the substantial ethical concerns (we raise) suggest that, at the very least, it is premature to recommend making religious and spiritual activities adjunctive medical treatments” (17). We simply and strongly disagree. Furthermore, we believe the current evidence speaks against such admonitions. First and foremost, if clinicians were to wait for controlled data to be available before utilizing interventions, many aspects of mental and physical health care would screech to a halt. That said, unwise would be the practitioner who would utilize therapy, without controlled data, that has a high probability of being harmful. However, therapies that are inexpensive, easy to apply, desired by the patient, and appear to be helpful (based on uncon-
Volume 24, Number 1, 2002 trolled data) with minimal risk of harm not only seem reasonable to clinicians, who must after all, live and practice in an imperfect world. Second, most who publish on the intersection of positive spirituality and patient care agree that clinicians should become comfortable addressing the basic spiritual and religious needs of their patients, including taking a religious history, supporting healthy religious beliefs, ensuring access to religious resources (e.g., religious reading materials, a chapel or prayer room, contact information for local clergy), providing spiritual referral or consultation, and viewing the clinical pastoral professional (clergy or chaplain) as an integral part of the health care team (1,18). One systematic review conducted by Matthews and colleagues concluded that practitioners who make some small changes in how patients’ religious commitments are broached in clinical practice might improve health care outcomes (13). Third, a growing number of clinical educators seem to disagree with our critics. In 2000, at least 65 of 126 U.S. medical schools and a growing number of residencies offered courses on the incorporation of religion or spirituality into clinical practice (1). In one allied field, survey research of accredited graduate schools of social work identified 17 schools in 1995 that offered courses with a spiritual or religious focus. By 2000, the number of schools offering courses related to spirituality had grown to over 50 (19). Studies have begun to describe the result of such courses. However, if such basic spiritual interventions were to be harmful, as claimed by some critics, one would expect such reports would be widespread. We are not aware of any published reports or systematic studies about clinicians having caused harm by addressing patients’ religious or spiritual needs. We concur that there is less agreement about some spiritual interventions, such as praying with patients or providing religious counsel. Those supporting these interventions are in nearly universal agreement that they should be patient centered, not practitioner centered (1,18). Furthermore, patients have a right to expect that religious counseling, like other forms of counseling, will be performed only by clinicians trained or experienced in such therapy. Medical ethicists are right in insisting the practitioner must honor the patient’s autonomy, follow the patients’ lead and needs, and utilize permission, respect, wisdom, and sensitivity (1,13,14,18). Current data indicate that a practitioner’s religious beliefs will influence whether and to what extent he or she addresses these issues (1). Nevertheless, almost 70% of primary care doctors agree that physicians should address at least some religious issues with patients. Between 46 and 78% of patients indicate that they would like their physician to pray with them. One third of primary-care physicians and two thirds of religiously devout physicians report doing so (1). So, why do some experts so vociferously argue that this patient-perceived need should not be met or that clinicians should discontinue their current practice of basic spiritual intervention until more research is available? In our view, a major part of the problem with the incorporation of basic spiritual interventions into health care has been
Positive Spirituality
71
the confusion associated with the terms faith, spirituality, religiosity, and religion. Religious variables in most early research were limited to religious affiliation. The current science of spiritual assessment suggests that the measurement of religiousness and spirituality must be multidimensional. Because there is a multiplicity of definitions for each of these terms (faith, spirituality, religiosity, and religion), we prefer to use the term positive spirituality. Positive spirituality, a term attributed to Parker, Fuller, Koenig, Bellis, and Vaitkus (20) and Crowther, Parker, Koenig, Larimore, and Achenbaum (21), is distinctive from broader terms in that positive spirituality involves a developing and internalized personal relationship with the sacred or transcendent. This relationship is not bound by race, ethnicity, economics, or class and promotes the wellness and welfare of others and self. We join those who assert that certain religious beliefs and activities can adversely affect both mental and physical health (1). Spirituality or religion can be restraining rather than freeing and life enhancing (22). Religion has been used to justify hypocrisy, self-righteousness, hatred, murder, torture, and prejudice. The aspects of spirituality or religiousness (e.g., hypocrisy, self-righteousness) that separate people from the community and family, that encourage unquestioning devotion and obedience to a single charismatic leader, or that promote religion or spiritual traditions as a healing practice to the total exclusion of research-based medical care are likely to adversely affect health over time. We have theorized that religious or spiritual beliefs and activities that encourage honesty, self-control, love, joy, peace, hope, patience, generosity, forgiveness, thankfulness, kindness, gentleness, goodness, faithfulness, understanding, and compassion and that provide hope and foster creative problem solving under difficult circumstances are more likely to be associated with mental and physical health benefits. The evidence to date seems to indicate that dependence on the transcendent helps an individual acknowledge his own self-limitations without despairing of his or her circumstances (23). Research has shown that when people become ill, many rely heavily on religious beliefs and practices to relieve stress, retain a sense of control, and maintain hope and a sense of meaning and purpose to life (24). To encourage clinicians to ignore such needs seems to us senseless and uncaring. Western religious traditions emphasize an intimate relationship with a transcendent force, place high value on personal relationships, and stress respect and value for the self, while placing an emphasis on self-sacrificing service and humility. The resulting emphasis on relationship (relationships to a transcendent force, to others, and to self) may have important mental health consequences, especially in regard to coping with the difficult life circumstances that accompany poor health and chronic disability (24). Positive spirituality may reduce the sense of loss of control and helplessness that accompany physical illness. Positive spiritual beliefs may also provide a cognitive framework that could reduce stress and increase purpose and meaning in the face of
72
Larimore et al.
illness (25). Spiritual activities such as prayer may reduce the sense of isolation and increase the patient’s sense of control over the illness. Public religious behaviors that improve coping during times of physical illness include but are not limited to participating in worship services, praying with others (and having others pray for one’s health), and visits from religious leaders such as a chaplain, pastor, priest, monk, or rabbi either at home or in the hospital. For the reader desiring to learn more about including spiritual assessment into their practice, we would recommend the following: 1. The work of the Fetzer Institute, which in collaboration with the National Institute on Aging, compiled 12 reviews reflective of different domains of religiousness and spirituality and a series of brief multidimensional measures for clinical use (which may be obtained by calling 616–375–2000). 2. A self-study module by the National Institute of Healthcare Research (available by calling 301–984–7162). 3. For Christian clinicians, the Christian Medical Association has developed a small-group video series for study (available by calling 888–230–2637). Each of these works point out the critical distinction between religiousness (specific behavioral, social, doctrinal, and denominational characteristics that involve a system of worship and doctrine shared within a group) and spirituality (individualistic, transcendent, ultimate meaning of life). In summary, this evidence points overwhelmingly to a positive association between what we call positive spirituality and mental and physical health and well-being. Most patients desire basic spiritual interventions by their care providers and decry that the profession is ignoring their spiritual needs. Most clinicians believe that spiritual interventions would help their patients but have little training in providing basic spiritual assessment or care. Professional schools and associations are encouraging and, in many cases, providing such training. Anecdotal evidence indicates that learners or clinicians seeking such training find it immediately helpful and apply it to their practice. Nevertheless, much more research is needed, especially outcome-based, clinical research on the effects of these spiritual interventions by clinicians. The evidence to date tells us that it is clear that clinicians should encourage positive spirituality with their patients. Until more evidence is available, we would encourage interested mental and behavioral health care providers and systems to learn to assess their patients’ spiritual health and to provide indicated and desired spiritual intervention. Clinicians should not, without compelling data to the contrary, “deprive their patients of the spiritual support and comfort upon which their hope, health and well being may hinge” (1).
REFERENCES (1) Koenig HG, McCullogh ME, Larson DB: Handbook of Religion and Health. New York: Oxford University Press, 2001.
Annals of Behavioral Medicine (2) Osler W: The faith that heals [Editorial]. British Medical Journal. 1910, 1:470–472. (3) Holland TP: Values, faith and professional practice. Social Thought. 1989, 15:28–40. (4) Koenig HG: The Healing Power of Faith. New York: Simon & Schuster, 1999. (5) National Academy of Sciences Committee on Science and Creationism: Science and Creationism. Washington, DC: National Academy Press, 1984. (6) Ellis A: Psychotherapy and atheistic values: A response to A. Bergin’s “Psychotherapy and Religious Values.” Journal of Consulting and Clinical Psychology. 1980, 48:635–639. (7) Waring N. Can prayer heal? Hippocrates. 2000, 14:22–24. (8) Ellis MR, Vinson DC, Ewigman B: Addressing spiritual concerns of patients: Family physicians’ attitudes and practices. Journal of Family Practice. 1999, 48:105–109. (9) Sheridan MJ, Wilmer CM, Atcheson L: Inclusion of content on religion and spirituality in the social work curriculum: A study of faculty views. Journal of Social Work Education. 1994, 30:363–376. (10) Canda ER: Conceptualizing spirituality for social work: Insights from diverse perspectives. Social Thought. 1988, 14:30–46. (11) Sloan RP, Bagiella E, VandeCreek L, et al.: Should physicians prescribe religious activities? New England Journal of Medicine. 2000, 342:1913–1916. (12) Post SG, Puchalski CM, Larson DB: Physicians and patient spirituality: Professional boundaries, competency, and ethics. Annals of Internal Medicine. 2000, 132:578–583. (13) Matthews DA, McCullough ME, Larson DB, et al.: Religious commitment and health status: A review of the research and implications for family medicine. Archives of Family Medicine. 1998, 7:118–124. (14) Matthews DA, McCollough ME, Swyers JP, et al.: Religious commitment and health status. Archives of Family Medicine. 1999, 8:476. (15) Levin, JS: Religion and health: Is there an association, is it valid, and is it causal? Social Science and Medicine. 1994, 38:1475–1482. (16) Sloan RP, Bagiella E: Data without a prayer. Archives of Internal Medicine. 2000, 160:1870. (17) Sloan RP, Bagiella E: Spirituality and medical practice: A look at the evidence. American Family Physician. 2001, 63:33–35. (18) Larimore WL: Providing basic spiritual care for patients: Should it be the exclusive domain of pastoral professional? [Medicine and Society]. American Family Physician. 2001, 63:36–40. (19) Russel R: Directors’ notes. Spirituality and Social Work Forum. 2001, 8:9. (20) Parker MW, Fuller G, Koenig H, Bellis JM, Vaitkus M: Soldier and family wellness across the life course: A developmental model of successful aging, spirituality, and health promotion: Part II. Military Medicine. 2001, 166:561–570. (21) Crowther M, Parker M, Koenig H, Larimore WL, Achenbaum WA: Successful Aging and Positive Spirituality. Rowe and Kahn’s Model of Successful Aging Revisited. Manuscript submitted for publication, 2002. (22) Pruyser PW: The seamy side of current religious beliefs. Bulletin of the Menninger Clinic. 1977, 41:329–348. (23) Taylor RJ, Chatters LM: Church members as a source of informal social support. Review of Religious Research. 1988, 30:193–202.
Volume 24, Number 1, 2002 (24) Bearon LB, Koenig HG: Religious cognition and use of prayer in health and illness. Gerontologist. 1990, 30:249–253.
Positive Spirituality
73
(25) Koenig HG, Pargament KI, Nielsen J: Religious coping and health status in medically ill hospitalized older adults. Journal of Nervous and Mental Disease. 1998, 186:513–521.
The Witches’ Brew of Spirituality and Medicine Raymond J. Lawrence, D.Min. New York-Presbyterian Hospital Columbia Presbyterian Center
concept is pulled in two quite different directions. For some, it connotes the inner life force or vitality, in which case any living being can be said to possess some kind of spirituality. For others, spirituality connotes a connection with an imagined or real power source outside of the natural world. In their treatment of patients, physicians would do well to keep an arm’s length from spirituality in both senses of the word (3). Hardly anyone would dispute the contention that spirituality, problematic though the concept’s definition is, probably has some bearing on health. Certainly a rare person would contend that the state of one’s mind (or spirit or soul) has no effect on one’s physical health. The simple matter of the will to live undoubtedly bears on one’s physical health. However, the claim that physicians should involve themselves in their patients’ spirituality is wrong-headed on several counts. First, such an assignment would mean a stunning inflation of the physician’s job description, and second, the notion that physicians are competent to assess their patients’ religious and spiritual condition trivializes all forms of religion and spirituality (4,5). The proposal that overworked physicians should expand their task into the arena of spiritual histories, updating them annually, is supremely unrealistic. The universal cry of physicians in this country at present concerns the negative impact of the clock on their patient care. This pressure put on them by managed care to cut to the chase with patients is not likely to go away anytime soon. It is pure fantasy to think that physicians will undertake the major new time-consuming task of examining patients’ spirituality with no way to bill for the time involved. If they do undertake this task, what important aspects of clinical practice will they exclude to make time for it? Furthermore, no one seems to have given much thought to problems and difficulties of taking spiritual histories and where a physician would learn to hone such a skill. Some suggest that a medical school elective would suffice (6,7). Others suggest that four simple questions that require yes or no answers will glean such a history, the ultimate trivialization. Even the best ministers and chaplains among us, after years of academic and clinical training, find the taking of spiritual histories a complex one. The world of spirituality, religion, and the imagination is often impenetrable, even by the most skilled and sensitive inquirer. Not every religious professional explores that world very well. It is shaped by a variety of traditions, some quite arcane, and it is a time-consuming undertaking. People who disclose the quality of their urine stream with reluctance and deception do not readily disclose their deeply held values and beliefs, many of which are ineffable, idiosyncratic, and often transparently irra-
ABSTRACT Recent proposals to join spirituality and medicine are facile and ill defined. The notion that physicians have the time or training to make assessments and recommendations about spirituality is misguided. Whenever a physician demonstrates personal caring for a patient, the healing process is likely enhanced, and in that sense, physicians often promote the spirituality of the patient. However, recent proposals to extend the physician’s task to that of assessing religion and directing the patient toward approved forms of spirituality are inappropriate. The languages of religion and science are radically different. The cultural body–mind split will not be solved by such simplistic solutions as having physicians endorse spirituality, which will result only in denigration of both medicine and religion. Physicians are encouraged to rely on clinically trained ministers for assistance in understanding the patient’s state of mind or spirit and its possible effects on the course of illness and health. (Ann Behav Med
2002, 24(1):74–76)
The recent widespread and frenzied promotion of a marriage of spirituality and medicine is a juggernaut that needs a more critical analysis than one finds today in the literature. The facile manner in which even the most prestigious medical centers—Harvard and Duke, for example—are promoting a union of spirituality and medicine is simply astonishing. Rather suddenly, in less than a decade, spirituality has become something of the darling of the medical world, and physicians, strangely enough, seem to be doing most of the talking. They are encouraging their fellow physicians to incorporate spirituality in their clinical practices, arguing that this will enhance their compassion and further a treatment of the patient as a whole person. They argue that spirituality benefits patients in terms of better health and longevity as well. They encourage physicians to endorse religious and spiritual practices for patients. They advocate that physicians take spiritual histories of their patients and update them annually, and they encourage physicians to pray with patients (1,2). The first, tangential problem, and an intractable one, is that of definition. The task of grasping precisely what any one individual means by spirituality is a daunting one. Spirituality as a Reprint Address: R. J. Lawrence, D.Min., New York-Presbyterian Hospital, Columbia Presbyterian Center, 622 West 168th Street, New York, NY 10032. E-mail:
[email protected]. © 2002 by The Society of Behavioral Medicine.
74
Volume 24, Number 1, 2002 tional. A clinically trained minister, chaplain, or pastoral psychotherapist would attempt to get a spiritual history of a patient only with the luxury of considerable time to invest and a clear willingness on the part of the patient to divulge very personal information, which by tradition is considered private. Physicians do not stand a chance. What would a physician do, for example, with the patient with metastatic lung cancer who acknowledges that her decades of heavy smoking has brought her to this predicament and who believes that she is being punished and expects to die for her sins? This is, at least in part, a religious or spiritual problem, not lending itself to a quick fix or a formulaic response. Anyone who proposes to make him- or herself available to this patient, whether chaplain, physician, social worker, or whoever, should be well informed in counseling and theology and should have the luxury of time to give. A typical physician has neither the time nor the training to offer either pastoral counseling or psychotherapy. A good many illnesses would seem similarly related to such factors as patient choices, values, history, religion, and patterns of living and, therefore, invite serious personal reflection in the context of what we call pastoral counseling. Few would dispute the contention that physicians generally would do well to become skillful in the art of interpersonal relations, or what used to be tritely called “bedside manner.” Patients typically seek in a physician something more than simply a scientist. They want to believe that the physician actually cares about them as persons. They want to feel the compassion of the professional. Without a personal bond between the two, the healing process may be jeopardized. Physicians do not have a sterling reputation generally in the arena of interpersonal relations. For better or worse, science has taken over the territory. Physicians could be more effective if they acted like they recognize that illness and health are overdetermined. Mental attitude, religious beliefs, family relationships, and other similar factors that impinge on health do not lend themselves readily to scientific scrutiny. Physicians would do well to be open to input from such nonmedical sources that might impact health. The patient treated for an ulcerated stomach, whose wife is continuing a sexual relationship with a family friend, is likely not to recover satisfactorily if his social life remains a source of trouble to him. Although this is a random and inflammatory illustration, probably the course of most illness is shaped in part by social, mental, and religious factors. Physicians would do well to enlist the aid of clinically trained chaplains and others similarly trained to assist in that arena. Whenever a physician demonstrates compassion or understanding in relation to a patient, it could be said that spirituality is being added to medicine. After very successful spinal surgery, my primary care physician paid me a visit in my hospital room. He sat for 30 min and listened to my saga. I was greatly elevated in spirit because of his expression of care for me. However, he was making me a special case, and he does not have hours enough in a day to give all his patients such attention. Furthermore, I was not dying or otherwise in distress, as many of his patients are. He took no spiritual history. He barely touched the tip of the iceberg as regards my total “spiritual condition,” but if he
75 finds ways similarly to hearten all his patients by communicating a genuine personal concern for them, he undoubtedly furthers their good health. It could even be said that he contributes to their spiritual well-being. However, such a modest goal, valuable though it is, is far removed from the grandiose, inflated objectives of the new spirituality and medicine advocates. The current literature promoting the union of spirituality and medicine is neither precise nor discriminating. Many writers take great comfort in the report that 70 medical schools are now teaching courses on spirituality and medicine, but the substance of such courses is subject to critical examination. To the extent to which such courses focus on the values, commitments, and vocation of medical students or on basic training in interpersonal skills, they might enhance medical practice. However, when such courses offer instant protocols for taking patients’ spiritual histories, as some suggest, they are greatly oversimplifying a subtle and complex task. Those who wish to mix religion and medicine do not understand that the language of science and the language of religion are radically different. Religion and spirituality exist in the realm of poetry and the imagination, not the realm of science where medicine dwells. The rules of discourse are different. To mix the languages can be highly misleading and disruptive. Physicians who are scientists cannot legitimately incorporate “relationships with transcendent beings” into their discourse as if it were another source of data like some laboratory result. To do so is schizophenogenic. The widespread claim in much of the literature is that spiritual and religious activity should be promoted by physicians as beneficial to health. Implicit in most of the recent spirituality literature, on the one hand, is the assumption that anything spiritual is beneficial (5). Hence, the frequent injunction to physicians to encourage patients who are attendees of religious worship services to continue in such activities. On the other hand, many of the new spirituality advocates have no hesitancy about rejecting certain spiritual beliefs that they consider ill advised or harmful, as for example, the belief that illness is punishment from God. They recommend steering patients away from such beliefs. We have, therefore, a revolutionary new claim abroad in the land, that physicians have the task of screening out unhealthy from healthy religious beliefs and practices. Not only is this a full-time job, but it is one that even a well-trained religious professional would find odious. Much worse, it would place physicians in the new role of arbitrating between healthy and unhealthy religious and spiritual practices, a role no physician in his or her right mind would wish to assume. Even more bizarre and irrational is the often-heard injunction that agnostic patients not be steered toward religious and spiritual practices, presumably out of respect for their agnosticism. Thus, the agnostics will be deprived of what in the spirituality literature is viewed as the best medicine, available only to the spiritually minded. Physicians caught up in this mess of contradiction and inflated job descriptions would ultimately envy even B’rer Rabbit. (For those physicians who actually do want to engage Tar Baby, I suggest as a random introductory exercise that they assess the differential therapeutic impact of the Evan-
76 gelical revival meeting, the Catholic mass, the sexual rites of Tantric Hinduism, Appalachian snake handling rites, and the whirling dervishes.) One can read in a wide array of sources these days the hapless advice that physicians pray with patients, especially if requested to do so, or at least stand in silence while the patient prays. Physicians should do neither. The response of choice should be a referral of the patient to a chaplain or to an appropriate religious authority. Prayer is by no means the simple, innocuous, and invariably benign activity that it is typically assumed to be, but it is a complex value-laden undertaking rooted deeply in the imagination and tradition of particular persons. To suggest that a physician is competent to decode what well-trained ministers often find to be difficult and impenetrable semiotics is once again to trivialize religion and spirituality. The problematic dichotomy between body and soul in modern medicine, as in the culture at large, is serious and debilitating. However, the notion that physicians will bridge that rift by assuming the role of religious and spiritual authorities is naive in the extreme. It may be that in some future more advanced culture than our own, one recovering from Cartesian dualism of mind and body, that physicians will combine science with the content of religion, including imagination, symbols, myth, and poetry in their treatment. In such a world, the physician would become something like a physician–priest/pastor/guru. The current spirituality and medicine movement seems to be lobbying for such a world. Such a monumental change, if it ever does occur, will not come cheaply or quickly. Scientific education will then be supplemented by significant religious education. The current advocates of the union of spirituality and medicine are trying to create a new world on the cheap, without appropriate preparation, and seemingly oblivious to the radical nature of their vision and its implications.
Annals of Behavioral Medicine Stirring spirituality and religion into the practice of clinical medicine in a facile manner, as promoted in much of the current discourse, will result only in a witches’ brew that will embarrass medicine and trivialize religion. The best we can hope for in the spirituality–religion and medicine arena at present is for physicians to rely more on clinically trained ministers for input into the general state of mind or soul of a patient. Sometimes patients will disclose to a minister what they will not disclose to others. Sometimes nonmedical aspects of a person’s life seem to have some bearing on the state of health, a thesis that should receive almost universal acceptance. In such conversation between physician and minister, we may bring some beneficial impact on the course of some person’ s health. To think that we can accomplish much more than this in the present environment is unrealistic.
REFERENCES (1) Koenig, H: Religion, spirituality, and medicine: Application to clinical practice. Journal of the American Medical Association. 2000, 284:1708. (2) Astrow AB, Puchalski CM, Sulmasy DP: Religion, spirituality, and health care: Social, ethical, and practical considerations. Journal of American Medicine. 2001, 110:283–287. (3) Lawrence RJ: The trouble with spirituality. Contra Mundum. 2001, 1:91. (4) Sloan RP, Bagiella E, VanderCreek L, et al.: Should physicians prescribe religious activities? New England Journal of Medicine. 2000, 432:1913–1916. (5) Sloan RP, Bagiella E, Powell T: Religion, spirituality, and medicine. Lancet. 1999, 353:664–667. (6) Puchalski CM: Spirituality and health: The art of compassionate medicine. Hospital Physician. 2001, Mar:30–36. (7) Koenig, HG: The Healing Power of Faith: Science Explores Medicine’s Last Great Frontier. New York: Simon & Schuster, 1999, 297–298.