RELIGION AND HEALTH: Public Health Research and Practice

Transcription

1 ?Annu. Rev. Public Health : Copyright c 2000 by Annual Reviews. All rights reserved religious involvement, spirituality Research examining the relationships between religion and the health RELIGION AND HEALTH: Public Health Research and Practice Linda M. Chatters Health Behavior and Health Education, School of Public Health, University of Michigan, Ann Arbor, Michigan ; chatters@umich.edu Key Words Abstract of individuals and populations has become increasingly visible in the social, behavioral, and health sciences. Systematic programs of research investigate religious phenomena within the context of coherent theoretical and conceptual frameworks that describe the causes and consequences of religious involvement for health outcomes. Recent research has validated the multidimensional aspects of religious involvement and investigated how religious factors operate through various biobehavioral and psychosocial constructs to affect health status through proposed mechanisms that link religion and health. Methodological and analytical advances in the field permit the development of more complex models of religion s effects, in keeping with proposed theoretical explanations. Investigations of religion and health have ethical and practical implications that should be addressed by the lay public, health professionals, the research community, and the clergy. Future research directions point to promising new areas of investigation that could bridge the constructs of religion and health. INTRODUCTION The relationship between religion and health has been of longstanding interest in the health, social, and behavioral sciences, spanning a period of >100 years (117, ). Despite recognized methodological and analytical issues (39), overall the findings indicate a consistent and salutary influence of religious factors on individual and population health (110, 122). In the past several years, systematic research on religious involvement and physical and mental health has begun to explore the functional mechanisms linking these constructs. This period has seen (a) development and expansion of systematic programs of research and inquiry, (b) improvement in the quality of research and scholarship around these issues, and (c) expansion of interest in this topic to diverse fields of research (e.g. psychoneuroimmunology) and clinical practice (16, 39, 106, 108, 110, 141). In recent years, several books have been published (69, 70, 105, 121, 124, 147, 151, 159) and major journals in public health and medicine have featured /00/ $

2 336 CHATTERS Annu. Rev. Public Health : Downloaded from empirical research, literature reviews, and special issues on these and related topics (16, 39, 100, 103, 106, 121, 124, 130, 132, 175, 180). Whereas religion and health were once considered marginal to serious scientific inquiry, they are currently? enjoying an unprecedented level of research interest and prominence. In the health and medical sciences, there is growing recognition that religious and spiritual concerns are important for understanding health-related behaviors, attitudes, and beliefs and are particularly salient for persons whose health is compromised (6, 71). In the area of clinical practice, several distinguished hospitals and health care centers have initiated programs incorporating religious/spiritual approaches and content as complements to standard treatment regimens, and a number of medical and nursing schools have begun to incorporate these issues into their curriculums (5, 112). Furthermore, in the fields of health education and health promotion, the concepts of spirituality and spiritual health are discussed and incorporated into notions of overall health (8, 28, 177). Finally, interest in the connections between religion and health is also reflected in the general culture, with the publication of numerous books and newspaper and magazine articles (179a) written for general audiences. Although these developments are noteworthy, it is important not to overstate the extent to which the research and practice communities endorse the proposition that religious involvement is significant for individual and population health. Skeptical resistance to this perspective is still evident (161) and reflects several factors. First, it is particularly challenging to understand the complex, multifactorial processes through which religion affects individual and population health. Furthermore, the study of religion and health involves multiple disciplines. Although this diversity enriches the study of religion and health, it is difficult to appreciate disciplinary differences in conceptual frameworks, methodological and analytical approaches, relevant contextual issues, and levels of inquiry (e.g. individuals vs populations). Second, longstanding scientific and professional perspectives have fostered stereotypes and misconceptions about issues of religion and/or openly antagonistic attitudes that preclude a consideration of these questions (16, 99, 105, 108, 110, 147, 184). Furthermore, researchers and practitioners in the behavioral, social, and health sciences may themselves be less religiously active than the general population (146, 184) and, therefore, may dismiss or deprecate the relevance of religion for human affairs. Finally, the absence of thoughtful and comprehensive discussions of the ethical considerations, practice and policy implications, and professional ramifications of the integration of religion and health has made researchers and practitioners wary of addressing these questions in their work (84, 110, 147, 160). The state-of-the-science in this research literature also contributes to ambiguity concerning the nature of religious effects on health outcomes. A number of conceptual and methodological limitations in this field hamper the development of a comprehensive assessment of religious effects on health (see section on measurement and methodological issues). Studies of religious effects often involve large sample sizes and multivariate analytical approaches, so that the estimation of religious effects on health outcomes occurs within the context of specific, and

3 ?RELIGION AND HEALTH 337 often different, sets of statistical controls. Accordingly, relevant information on the nature of the association between religion and health changes depending on the configuration of control variables that have been included in the analysis, and, across investigations, there is little agreement as to which factors are important and are consistently controlled. The literature is also characterized by studies on particular subgroups of the population (e.g. demographic and religious subgroups). Although they provide important information on relevant subgroups, comprehensive assessments of the magnitude of religious effects are hampered by these differences in samples. Finally, an assessment of the overall magnitude of religious effects on health is confounded by extant differences in the conceptual and methodological approaches used by investigators (see section on conceptual models that link religion and health). Clearly, this is not a trivial concern and is one that continues to pose a serious problem for advancements in the field. This diversity of conceptual and methodological approaches to the study of religion and health reflects both the richness of this area of research and its major challenge. Nonetheless, there are several programmatic efforts that investigate the connections between religion and health and that demonstrate sophisticated conceptual, methodological, and analytic approaches. In particular, current efforts reflect (a) greater clarity in defining the nature and boundaries of relevant content areas (e.g. conceptual definitions and multidimensionality); (b) the development, specification, and testing of explicit theoretical linkages connecting religion and health; (c) more careful consideration of the research methodologies and procedures (e.g. measurement, sampling, and study design) that are appropriate to the research question; and (d) the development of conceptually valid measures of religious involvement for use in religion and health research. The range of topics and issues that fall under the rubric of religion and health is extensive, in part because the constructs themselves are broad. Furthermore, the boundaries separating religion and spirituality from approaches and philosophies of health care (e.g. holistic care) within complementary and alternative medicine (CAM) are not always explicit (64), and there is often a tendency to aggregate these distinct research and practice traditions. [There are a number of excellent reviews of studies of religious and spiritual healing and the nature, functional properties, and efficacy of various healing modalities, e.g. CAM (64). Practice traditions and philosophies, e.g. holistic nursing and naturopathy, that encompass questions of religion and spirituality, e.g. Dossey s review (30) of holistic nursing, are not discussed here except as they address issues relevant to other topics of interest, e.g. professional perspectives and ethical considerations.] Given that religion and health encompass a broad substantive area, a selective approach to the literature was taken to identify the topics and issues to be addressed. First, except for cases involving particularly pivotal work, this review reflects published research from the 1990s. A number of excellent reviews of the literature on this topic that was published before the 1990s are noted for further reading. Chapter Parameters and Overview

4 338 CHATTERS Annu. Rev. Public Health : Downloaded from Second, I review several types of literature, including work that focuses on the development and testing of explicit theories and research questions pertinent to the mechanisms that link religion and health (e.g. psychosocial, biological, epidemiological, and physiological theories and research). Theory and research on concepts? that are pertinent for understanding the functional mechanisms for religion s health effects (e.g. social support) are also reviewed. Research that is primarily descriptive is placed within an appropriate theoretical/conceptual framework. Studies involving Christian denominations and sects currently dominate research into the effects of religion on health. Accordingly, this literature embodies a strong American Protestant character (49, 55, 56), as well as distinctive subgroup differences (e.g. views on race/ethnicity and denomination) from within American Protestantism. This has implications for the relevance of findings to other religions and faith traditions. Regarding terminology used in this review, religion and religious involvement are used to indicate these phenomena in a general sense, whereas specific designations refer to discrete forms of religious expression (e.g. prayer or service attendance). Finally, given the range of topics addressed in this literature, the chapter is representative but not exhaustive in its coverage. This literature review focuses on studies that examine discrete religious phenomena (e.g. affiliation, public and private behaviors, and beliefs), as well as spirituality. There are several arguments (130) for restricting the focus to explicitly religious factors (e.g. difficulties in conceptualizing and measuring spirituality). Spirituality, although distinct from religion, is nonetheless a related construct, a differentiation that is commonly acknowledged and validated by the lay public (171). The convergence of religion and spirituality is particularly evident for beliefs and experiences that specifically have God as their reference point. These types of transcendent, numinous experiences (e.g. feelings of closeness to God, peacefulness, and sacredness), which are frequently reported by respondents (104), emphasize a relationship to something greater than oneself (i.e. the sacred or God) and are invested with a sense of personal meaning and significance that may have important consequences for health and well-being (e.g. via behaviors, attitudes, and emotional states). However, other aspects of religious spirituality (i.e. worldand humanity-oriented spirituality) and other types of spirituality (i.e. nonreligious spirituality) (55) are not addressed. DEFINITIONS AND INDICATORS OF RELIGION AND RELIGIOUS INVOLVEMENT One of the most difficult and perplexing issues in the field concerns what is meant by religion and basic conceptual definitions of what constitutes religious involvement. One review of religious and social and behavioral science writings on the topic defines religion as... a process, the search for significance in ways related to the sacred (147, p. 32). This definition incorporates the themes of both the substantive content (e.g. beliefs, practices, and feelings directed toward God) and the functional aspects (e.g. a process focused on

5 ?RELIGION AND HEALTH 339 questions of ultimate meaning and concern) of religion. Systematic empirical work in the development of conceptual definitions of religious involvement indicate that it is a multidimensional construct (17, 35, 60, 91, 111, 147, 158, 182). Unfortunately, different disciplines and fields (e.g. psychiatry, social epidemiology, and clinical medicine) often emphasize particular dimensions or forms of religious expression (e.g. church attendance or denominational affiliation), further complicating efforts to define and measure religious involvement comprehensively. This problem is particularly worrisome because it fails to fully appreciate the multidimensional character of religious involvement, as well as the various mechanisms through which religious effects have impact on diverse health outcomes (39, 120, 122, 158, 182). One of the primary ways that religious involvement is conceptualized distinguishes between behavioral and subjective dimensions. The behavioral component pertains to individual characteristics and activities that reflect organizational or public religious expression (e.g. denominational affiliation and religious-service attendance), as well as private activities that may be performed outside (i.e. nonorganizational practices) of religious institutions (e.g. private prayer and devotional reading). Subjective dimensions of religious involvement include attitudes, beliefs, experiences, and self-perceptions and attributions that involve religious or spiritual content (e.g. religious identity and feelings of closeness to God) (120, 182). Despite sustained attention to these concerns, health research (i.e. epidemiological and medical research) is generally unfamiliar with extant developments in the conceptualization and measurement of religion involvement (130). Instead, this work often uses a restricted set of religious indicators that do not possess theoretically justifiable linkages to diverse health outcomes (39, 60, 105, 111, 158, 182). This state of affairs argues for conceptually based, empirically validated measures of the components of religion that are consequential for health and permit the investigation of the proximate causes and mechanisms (e.g. social support, coping resources, and personal control) that link religion and health (39, 44, 60, 147, 152, 182). Several conceptualization and measurement efforts are underway, including research and working groups supported by the National Institutes of Health (i.e. National Institute on Aging, National Institute of Mental Health, and National Institute on Alcohol Abuse and Alcoholism) and private organizations such as the Templeton Foundation and the Fetzer Institute (44). These efforts build on a tradition of research on religious involvement from the social and behavioral sciences and have as a guiding objective understanding religion s multifaceted relationship to health. RELIGIOUS EFFECTS ON HEALTH OUTCOMES Physical Health Outcomes Evidence from epidemiological and clinical studies and medical research supports the impact of religious affiliation and involvement on a diverse array of mental and physical health indicators and disease states. This literature encompasses studies

6 340 CHATTERS Annu. Rev. Public Health : Downloaded from Mental Health Outcomes and Well-Being of cancer, hypertension, stroke, other cardiovascular conditions, gastrointestinal diseases, overall and cause-specific mortality, indicators of physical disability, selfratings of health status, and reports of symptomatology (13, 25, 27, 31, 42, 58, 62, 65, 67, 76, 81, 83, 93, 94, 98, 100, 117, 134, 137, 144, 163, 185), encompassing numerous disease entities or types of rates (110). Several investigations indicate that religious involvement is associated with better outcomes for persons who are recovering from physical and mental illness (145, 156, 157). One recent study of immune system function in a sample of older adults (82) found a weak association between religious-service attendance and immune system status, independent of effects of depression and negative life events. Overall, better physical health status, as measured by a variety of indicators, is moderately associated with higher levels of religious involvement, even when defined by numerous indicators and examined within diverse groups (i.e. as defined by clinical disorder, gender, age cohort, denomination, race/ethnicity, and social class) within the population (39, 110, 130). Evidence concerning the impact of religion on indicators of mental health (100, 108, 110) indicates strong positive associations between religious involvement and mental health outcomes. Studies (primarily epidemiologic) indicate that religious factors have a salutary influence on a diverse set of outcomes, including depression, drug and alcohol use, delinquent behavior, suicide, psychological distress, and certain functional psychiatric diagnoses (24, 47, 61, 66, 76, 80, 81, 109, 156, 183). Investigations involving clinical populations include one that indicates the significance of religious factors for self-ratings of depression among a sample of men at 6-month follow-up (75). Intrinsic religiosity (but not service attendance or private religious activities) was associated with shorter time to remission of depression in a sample of older patients (77). Musick and associates (138) found that the effects of religious activity in lowering depressive symptoms were stronger for Black than Caucasian elderly patients with cancer. A study of denominational (i.e. mainline Protestants, conservative Protestants, and Pentecostals) and age cohort (i.e. baby boomers and older adults) differences in psychiatric disorders found high rates of psychiatric disorder among younger Pentecostals and of alcohol abuse/dependence among older Pentecostals (79). Furthermore, persons who attended services infrequently were at greatest risk. A large body of research indicates that religion is beneficial to a sense of personal well-being and overall adjustment (3, 10, 32, 34, 36, 43, 59, 89, 90, 108, 109, 111, 118, 135, 154). Research on older adults finds positive effects for religious involvement by using various mental health scales and indices (e.g. Center for Epidemiological Studies Depression Scale, General Well-Being Scale, Life Satisfaction Index A, Beck, and Affect Balance Scale), as well as single-item assessments of life satisfaction and happiness (121). These positive effects are observed net of the effects of factors that are associated with well-being (e.g. health, functional

7 ?RELIGION AND HEALTH 341 ability, and sociodemographic factors) and that occur across diverse samples of the population. A substantial body of literature has examined the associations between religious involvement and health and lifestyle behaviors (11, 124). Much of this work (54) focuses on the health/lifestyle behaviors of persons within identified religious groups and denominations (e.g. Mormons or Seventh Day Adventists), which may, in turn, result in lower levels of morbidity for various diseases. Other religious-involvement factors (e.g. service attendance and importance) have been studied in relation to alcohol, tobacco, and drug use (9, 47, 50, 53, 78, 86, 89, 178) and adolescent sexual behavior and contraceptive use (7, 26, 136, 155, 164, 174). Religious factors are associated with lower levels of behavioral risk such as alcohol use (2, 45) and higher levels of health-promoting behaviors (e.g. diet and exercise) among adolescents (178). Religious strategies may be particularly important for coping with mental and physical illness and disability. Persons who use religious coping appear to handle their conditions more effectively than those who do not (75, 145, 147, 149, 150, 152, 156, 183). Several studies indicate that religious coping is significant for mental and physical health outcomes for a variety of life circumstances, especially health problems (74, 145) and bereavement (90, 133). Religious coping also appears to reduce levels of depression and anxiety (72, 183) in connection with bereavement and other loss events (131). It is interesting that religious coping may moderate the relationships between physical illness and functional disability (59) and between disability and depression. A cross-sectional study of medically ill chronically institutionalized older adults (88) found that those who reporting using religion to cope had higher levels of social support and better cognitive functioning. Religious copers also had more severe medical illness than did noncopers. The cross-sectional nature of the study leaves open several possible relationships between religious coping and health (see section on analytic models). For example, illness severity may encourage religious coping (i.e. religious coping increases in response to the stress of illness). Related to this, medically ill older patients who used religious coping strategies were less likely than their counterparts to have cognitive (but not somatic) symptoms of depression (e.g. boredom, loss of interest, social withdrawal, and feelings of restlessness, failure, and hopelessness) (74). A comparison of various religious coping (positive vs negative) and nonreligious coping strategies indicates that positive religious coping was related to better mental health status, whereas negative religious coping and nonreligious coping were associated with poorer physical health, greater depression, and poorer quality of life (87). Religious resources (i.e. Health and Lifestyle Behaviors and Health Care Utilization Religious Coping

8 342 CHATTERS Annu. Rev. Public Health : Downloaded from prayer and service attendance) appeared to moderate the associations between discrimination and other stressors and psychological distress and depression among African Americans; the results indicated differential effects based on type of stressor and gender (MA Musick, DR Williams, JS Jackson, unpublished data). Finally,? a longitudinal study of older adults (93) found that religious coping factors offset the effects of stress (e.g. from neighborhood deterioration) on self-rated health status (but not functional disability). With few exceptions, the studies represented in the religious-coping literature are primarily based on cross-sectional data and are thus limited in that they do not address religious coping in a prospective manner (this issue is discussed later in the section on measurement and methodological issues). CONCEPTUAL MODELS LINKING RELIGION AND HEALTH Early systematic efforts (123) to describe the health effects of religious involvement outline several broad classes of explanations (e.g. hereditary, behavioral, and psychosocial effects and psychodynamics of belief, ritual, and faith) that link these constructs. Recent work, particularly based in frameworks from sociology (35) and psychology (147), delineates the specific mechanisms through which religious involvement influences health (37, 39, 44, 58, 68, 109). Despite differences in the origins and emphasis of these models, they share a number of common elements. First, these models acknowledge that religion and health are multidimensional constructs. Religion encompasses behavioral, attitudinal, public, and private activities, all of which potentially involve different antecedent factors and consequences for health outcomes. Similarly, the construct of health involves a diverse array of mental and physical indicators and disease states and relies on different types of information (subjective as well as objective data). Consequently, the impact of religion on physical and mental health involves a number of possible explanations and functional mechanisms (e.g. direct effects and mediated effects through other factors). Second, in these models, a given religious involvement factor (e.g. service attendance) is multifactorial and therefore may reflect the operation of several possible mechanisms (e.g. instrumental support, congregational climate, and positive self-perceptions). Third, in these models, religious involvement can be understood only within a context defined by shared group norms, expectations, and meanings (39). Consequently, the significance and relationship of a given religious factor to health outcomes will potentially vary across distinct social categories (e.g. race/ethnicity, denomination, age, social class, and region). This is true for comparisons within defined population groups (e.g. regional variations among African Americans), as well as for comparisons between groups (e.g. race/ethnicity and denominational differences). Finally, in discussions of these models, attention is

9 ?RELIGION AND HEALTH 343 paid to the possibility that various selection factors (e.g. extroversion and functional status) may be operative (and should be accounted for) in observed religion and health associations (1, 12). Several broad categories of factors have been proposed as linkages between religion and health (32, 39, 59, 107, 108, 123, 129). These categories include (a) specific lifestyle and health behaviors, including help seeking; (b) social resources; (c) coping resources and behaviors; (d) attitudes, beliefs, and emotional states and feelings; and (e) generalized beliefs about the world. Furthermore, within each category, both the potential positive and negative effects of religion on health are discussed. Lifestyle and Health Behaviors Explanations that focus on lifestyle and health behaviors suggest that religion is instrumental in shaping behaviors (e.g. risktaking and protective behaviors) that are consequential for physical and mental health. This includes directly and formally proscribing specific behaviors that are health risks (e.g. dietary restrictions and prohibitions against the use of alcohol and tobacco), as well as encouraging behaviors that are conducive to health (e.g. regular exercise). These distinctive patterns of lifestyle and health behaviors could result in lower rates of chronic and acute illnesses within identified religious groups. Additionally, religious adherents may have reduced risk for stressful life circumstances because religious teachings embody general guidelines for behavior (e.g. moderation and conformity) that discourage individual deviance and encourage interpersonal harmony (35). Ellison & Levin (39) outline several potential ways that religious involvement influences behavior, including self-monitoring via internalized norms for behavior, (fear of) direct external sanctions for violations of behavioral expectations, and positive emulation of valued community members. Religious institutions and clergy often function as gatekeepers for individual help-seeking behaviors and health care utilization, particularly among poor and disenfranchised populations (170). Although there is often a presumed beneficial effect, little is known about the role of clergy in health service delivery and the interface between faith communities and health and human service organizations (140). Social Resources Participation in religious groups confers a number of benefits in terms of enhanced social resources (12, 35, 38). These advantages include the size of one s social networks, frequency of interactions with network members, both actual and anticipated (subjective-support) exchanges of various types of informal and formal assistance (i.e. instrumental, socioemotional, and appraisal assistance), and positive perceptions of support relationships (e.g. satisfaction and anticipated help). Religious organizations, because they embody explicit norms Positive Aspects of Religious Involvement

10 344 CHATTERS Annu. Rev. Public Health : Downloaded from Coping Resources and Behaviors and beliefs regarding helping behaviors, constitute natural environments for the development of both formal (e.g. programmatic initiatives) and informal supportive relationships. A tradition of research documents the supportive roles that religious institutions have played in the development and maintenance of minority communities, particularly African Americans (46, 125, 165, 168, 172). The norms, characteristics, and functions of religious groups may facilitate the provision of support within these contexts and enhance perceptions of assistance. Religious organizations may be distinctive regarding information that they provide to members about their value and worth as individuals (e.g. sense of being cared for and loved by others) and thus contribute to positive perceptions of support (support satisfaction and anticipated assistance). Furthermore, because members share similar frames of reference and meaning, religious groups provide a unique context within which to interpret problematic life situations and proffer needed assistance (35, 39). Religious institutions actively shape attitudes and beliefs concerning significant life roles (e.g. parent, spouse, worker, friend, and neighbor) in ways that promote interpersonal harmony and effective role enactment and potentially reduce role stress (15, 57). Finally, personal investment in religious groups develops supportive ties and norms for support reciprocity (e.g. support credits ) that are maintained over time (165, 168). Interest in understanding the specific ways that individuals use religious resources to handle life problems has also taken the form of research on religious coping strategies and resources (6, 35, 40, 56, 68, 90, 129, 147, 148, 151, 152). Research on religious coping in its various forms provides a necessary link between general religious orientations and individual adjustment to negative life events (35, 147, 152). Religious-coping research includes a variety of approaches, from work that examines the associations between global religious expression and commitment (e.g. church attendance) and health outcomes to research that focuses on the more proximal, specific, and functionally oriented approaches to religion in response to personal difficulties (147, 148). Critiques of the religious coping literature (147, 148) describe several approaches in this research: (a) the indicators approach, in which global religious items are used as indicators of religious coping; (b) the overall approach, which assesses the overall degree of religious involvement in coping; (c) the general approach, which uses religious-coping items that are found in general scales of coping; (d) specific religious-coping methods that assess different methods of religious coping; and (e) patterns of religious coping. Broadly speaking, religiouscoping efforts encompass several types of explanations of religion s effects on health that were described previously (e.g. social support and control beliefs). In addition, specific research efforts describe a typology of religious-coping strategies, give more explicit attention to issues of both positive and negative outcomes of coping, and address issues of professional guidelines for practice (147).

11 ?RELIGION AND HEALTH 345 Programmatic work in the areas of religious coping within a medical context (85 88) and the psychology of religious coping (147) has made important contributions to the systematic investigation of the use of religion for coping with problematic life circumstances. Religious coping can involve different arenas (e.g. individual approaches and support from congregation members and clergy) and types of coping strategies [e.g. religious reframing involving cognitive reappraisals of stressful events and religious control orientations (129, 147)]. Investigations of the specific content of religious coping suggest that it is multidimensional, serves a variety of purposes (e.g. emotional comfort, meaning and purpose, and personal control), and involves a diverse array of strategies such as religious reappraisals and attempts to secure spiritual support (147, 152). Individuals may combine religiousand nonreligious-coping strategies to resolve problems (6, 150), or their strategies may be organized into positive vs negative religious patterns (152). Several critical areas for continuing research have been outlined (40), including how social location, problem type, and religious orientations potentially predict the use of specific religious-coping methods. Initial work indicates that religious coping is used and apparently most effective for specific population groups such as women, older persons, and African Americans, and for specific types of problems such as bereavement and serious health problems (40, 43, 131). Other research suggests that religious coping may also be effective for moderating stress related to poor neighborhood conditions (93) and unfair discrimination (MA Musick, DR Williams, JS Jackson, unpublished data). Finally, because religious-coping perspectives involve adaptation and change over time, prospective studies are needed to understand the causal linkages (e.g. coping mobilization processes) between problems, religious coping, and relevant health outcomes (152). Positive Attitudes, Beliefs, and Emotions Religious doctrines may support positive views of human nature and the self that engender attitudes and emotional states that are associated with better physical and mental health outcomes. Belief in the intrinsic value and uniqueness of each individual may promote feelings of self-esteem. Religious injunctions may shape interpersonal behaviors and attitudes toward others in ways that emphasize a variety of positive and prosocial goals (e.g. interpersonal warmth and friendliness, love, compassion, harmony, tolerance, and forgiveness) and that reduce the likelihood of noxious and stressful interpersonal interactions (33, 35). Religious ritual (e.g. prayer, communion, and confession) may be important for instilling positive emotional states (e.g. joy and ecstatic experiences) and/or generating more global beliefs and worldviews that are associated with better health (e.g. optimism and forgiveness). Finally, the use of religion to promote individual and community healing (i.e. restorative activities) has been associated with the experience of strong, positive emotions regarding the self, such as feelings of self-worth, competence, and connection with others (92, 129).

12 346 CHATTERS Annu. Rev. Public Health : Downloaded from Negative Aspects of Religious Involvement? Along with the presumed benefits of religious involvement for health, religion may also be associated with negative outcomes, such as poorer mental and physical health status, negative coping behaviors, and inappropriate use of health services (39, 106). Lifestyle and Health Behaviors Religious beliefs and commitments may also encourage general lifestyles and patterns of behavior that are harmful to health (40). Religious teachings may proscribe specific medical procedures and treatments and, in extreme forms, promote and reinforce social deviancy and aberrant behavior that is detrimental to health and well-being (e.g. cult membership). As outlined previously, there is little systematic information that focuses on the mechanisms through which religious groups and clergy influence the use of professional health services (140, 182). Religious groups may negatively influence patterns of informal and self-care, discourage professional help-seeking behaviors for health care, promote the inappropriate use of services (e.g. delays in the timing of service use), and encourage exclusive treatment by clergy (e.g. for mental and emotional problems). Social Resources Negative effects of religion on health could result from problematic social relations within religious institutions. Although participation in religious groups may confer benefits in terms of social network size and various types of social support, these relations can also be a source of distress (35, 39, 96, 97). The social processes that operate to shape appropriate behaviors and constrain inappropriate ones involve various forms of social pressure. Failure to conform to institutional expectations and other forms of deviance from group norms may be openly sanctioned by other church members and the clergy. Religious settings are suitable environments for support exchanges because of several organizational (e.g. expressed norms that advocate supportive interactions) and interpersonal characteristics (e.g. participant similarity regarding status characteristics and common language) that are important for support provision. Furthermore, social support exchanges are governed by norms of reciprocity and balance that are enacted over a period of time. However, interpersonal difficulties may arise when there are violations (actual or perceived) of these expectations or when helping results in paternalistic attitudes (129). Furthermore, problematic support exchanges may occur if support provider and recipient do not share similar views of support provision, beliefs about responsibility for problems and their resolution, and/or the role of religious vs secular coping. For example, support may be withheld because of a belief that the individual bears responsibility for the problem and/or its resolution. Similarly, the belief that God never gives us more than we can handle may effectively limit requests for support (both secular and religious) from those in need and/or it may constrain the efforts of potential helpers to provide assistance. For organizational exchanges, religious institutions

13 ?RELIGION AND HEALTH 347 often require members to invest significant amounts of effort and money (e.g. tithing) that may strain financial and other resources (35), and they may exhort members to participate in supportive relationships even when their capacity to provide assistance to others is limited. Emergent research investigates these types of negative social interactions (e.g. criticism and excessive demands) between members and/or church leaders and members within religious settings (96, 97) to document their occurrence and possible effects on health and well-being. Interpersonal conflicts and negative interactions may generate significant amounts of distress for individuals because they are unanticipated, relatively rare occurrences and are at odds with expectations and norms for behaviors within religious settings (96). Coping Resources and Behaviors Programmatic work (147, 152) also explores negative religious coping and identifies a number of situations in which the use of religious resources is detrimental to coping outcomes and adjustment. A study of patterns of positive and negative religious coping (152) found that persons who use negative coping tend to view the world as threatening, express a less secure relationship with God, and demonstrate a sense of spiritual struggle. Negative coping was associated with greater depression and psychological symptoms, poorer life quality, and less sociability. Negative Attitudes, Beliefs, and Emotions Religious doctrines may engender specific emotional states (e.g. guilt, shame, and anxiety) and attitudes and beliefs that reinforce negative views of human nature and the self (129). These emotional states, attitudes, and beliefs may, in turn, have negative physical and mental health consequences (117). The doctrine of original sin may function to undermine a sense of individual value and self-worth, whereas beliefs in human fallibility and imperfection may undermine a sense of personal mastery. Certain religious beliefs may emphasize the need for vengeance and retribution for wrongs, resulting in the harboring of anger, hostility, and vengeful feelings. Furthermore, religious doctrines and beliefs may instill and maintain global beliefs and worldviews that shape individual behaviors and interpersonal interactions. In particular, religious doctrines and beliefs may encourage pessimistic attitudes concerning the state of human affairs and a general distrust of others (particularly out-group members). RACE AND ETHNICITY IN RELIGION AND HEALTH RELATIONSHIPS Issues of race and ethnicity have occupied an important position in religion and health research. Similar to other substantive areas in which race and ethnicity issues are of interest, this research has often focused on documenting basic group differences in religion-related health phenomena, with little attention devoted to

14 348 CHATTERS Annu. Rev. Public Health : Downloaded from MEASUREMENT AND METHODOLOGICAL ISSUES understanding the processes that contribute to those differences, even though a considerable body of literature exists describing the historical importance of religion and religious institutions for minority groups in particular, African Americans (46, 125). This work describes various behavioral and psychosocial mechanisms through which religion and religious observances have influenced the health and well-being of African Americans, including social networks and supportive exchanges operating within churches, religious coping strategies, development and maintenance of various coping resources (e.g. positive perceptions of the self and group), and extensive discussions of the role of religious organizations in the development of broader community institutions (e.g. educational, political, civic, health, and social services). These early accounts often described religion s significance in terms of coping and adaptation to adverse life circumstances and disadvantage (i.e. racial and economic discrimination and prejudice) associated with minority status (i.e. stress and coping processes). The distinctive elements of African American religious expression revolve around the historical and contemporaneous position of religious institutions in addressing and ameliorating adverse life conditions and improving the social, emotional, psychological, and spiritual well-being of individuals and groups (46, 125, 172). Current research on religion and health has made important strides in understanding the role of race and ethnicity in these relationships. A number of research efforts have contributed to this literature, including, for example, programmatic work by Levin & Markides ( , 127) and Markides et al (116, 126) on older Mexican Americans and efforts by Chatters et al (17 23), Ellison & Taylor (40), Levin et al (111, ), Taylor et al ( ), and others (33, 34, 36, 63, 89, 137, 138, 140, 162, 173) on African Americans. This body of work describes the nature and functions of religion both within and between racial and ethnic groups, explores these issues within the historical and current life contexts of these populations, and uses theoretically derived analytical models to examine the relationships between religion and health. In doing so, this work has helped to enrich our understanding of religious involvement among diverse populations and its potential roles in advancing individual and group health. Systematic work on religious involvement and health has advanced the quality of research and scholarship in this area; for example, religion is now recognized as a multidimensional construct comprising attitudes, behaviors, beliefs, values, and experiences (111, 147). Unfortunately, research on religion and health continues to grapple with a number of measurement and methodological issues, despite sustained attention to these concerns (39, 60, 108, 123, 182). This section addresses related topics, including measurement and sampling concerns, the nature of group differences, and analytic and study design issues.

15 ?RELIGION AND HEALTH 349 Measurement Advancements in the area of measurement are limited by several factors. First, health researchers are often unfamiliar with measurement strategies from the social and behavioral sciences that have produced brief, reliable, and content-valid instruments for measuring religious involvement (39, 60, 147). Second, perhaps because of administration ease and established practices, there is a tendency to rely on objective and behavioral reports of religious involvement (e.g. church attendance and denomination) and to ignore other aspects of religion that reflect functional relationships with health outcomes. Behavioral reports of religious involvement are essentially proxies for a whole range of phenomena (e.g. lifestyle, attitudes, and social support) that are not directly assessed but potentially bear a stronger relationship to health and well-being (39, 123). The situation is exceedingly complex, because these multiple components of religious involvement demonstrate divergent relationships to health outcomes (82). However, the thorough investigation of the proximal factors and functional mechanisms linking religion and health will advance our understanding of how these constructs are related. Sampling Research findings supporting a religion-health connection, because they are based on data from different types of samples (e.g. specific religious denominations; college students; older adults; national, regional, and community surveys; and patient populations), suggest that religious effects on health are robust. However, this diversity also leads to a situation in which it is difficult to determine the specific relevant factors involved and the relative strength of religion-health relationships. For example, religious factors might be especially important predictors of health outcomes for particular groups (e.g. older persons); however, their importance might be overlooked if, within the particular sample used (e.g. college students), these factors have demonstrated little or no effect on health. In this case, a valid mechanism might be ignored, and the religion-health relationship would be ambiguous. The investigation of specific mechanisms for religion s positive health effects requires attention to the issue of sample comparability and what it might reveal about differential, as well as universal, processes and mechanisms. Inter- and Intragroup Comparisons Research that is focused on group comparisons of religion and health often reflects two related themes. One concerns the potential health consequences of membership in discrete religious groups (e.g. Mormons or Seventh Day Adventists) that exhibit distinctive lifestyles, behaviors, and attitudes, whereas a related interest involves comparisons across religious groups to establish potential differences in the types, correlates, and consequences of religious involvement. Both endeavors are useful and important for advancing our understanding of religion and health relationships. Unfortunately, in their most rudimentary forms, research of this type is based on erroneous assumptions about the nature of inter- and intragroup comparisons.

16 350 CHATTERS Annu. Rev. Public Health : Downloaded from Analytic Models First, intragroup comparisons often assume denominational homogeneity in the underlying lifestyle practices, attitudes, and beliefs that characterize religious groupings. Based on this assumption, comparisons between faith categories (e.g. Jews, Protestants, and Catholics) are made in which it is assumed that these discrete categories and their constituent groups (e.g. Reform and Orthodox Jews, Lutherans and Baptists) represent homogeneous entities regarding underlying religious factors (122). However, to take the example a step further, broad designations like Baptist and Methodist may also conceal important intragroup differences (e.g. Southern vs Northern Baptists). In essence, there is significant within-denomination variation about the centrality and meaning of religious phenomena and their patterns of relationships to health outcome(s). These cautions are especially important when considering broad religious categories such as Conservative Protestants, who are presumed to be distinctive and homogeneous regarding beliefs, practices, and attitudes (184). Similarly, intergroup comparisons are based on the assumption that comparisons between distinct groups are valid because the specific religious factors of interest not only have the same meaning, but also are associated with health outcomes in a similar manner. For example, behavioral indicators of religious involvement (e.g. service attendance) demonstrate wide denominational (and regional) variations in norms for and the experience and meaning of public religious participation. Consequently, they may be related to health outcomes in diverse ways across relevant comparison categories (39). Unfortunately, interdenominational comparisons (like race comparisons) often fail to progress beyond establishing differences to exploring the mechanisms and processes underlying those differences. This is partially attributable to the lack of appropriate measures of functionally relevant aspects of religious involvement in most existing data sets (39, 138). However, the advent of new data collection efforts that incorporate multiple, functional aspects of religious involvement and are sensitive to variation in the meaning of these factors will allow for a more detailed understanding of denominational differences. Finally, changes in underlying population parameters (e.g. age distributions and changes in denominational affiliation) may have consequences for understanding religion-health relationships because these factors are consequential for religious involvement and/or health factors (48, 184). This discussion clearly suggests that a thorough understanding of the nature of intragroup differences is necessary for interpreting the meaning of intergroup comparisons. Research on religion and health typically uses a strategy in which religion is but one of several independent variables examined in relation to health and well-being outcomes. Theoretical developments (particularly from the stress and coping literature) describing the interrelationships among religion, health, and various mediating factors have given rise to more complex analytic models representing these associations. Five basic models have been discussed (39, 108) representing several possible relationships among stress, religion, mediating factors, and physical and mental health outcomes. The suppressor model