Preferred Citation: Jamner, Margaret Schneider, and Daniel Stokols, editors. Promoting Human Wellness: New Frontiers for Research, Practice, and Policy. Berkeley:  University of California Press,  c2000 2000.




Edited by

Margaret Schneider Jamner

Daniel Stockols

University of California Press
Berkeley Los Angeles London

Preferred Citation: Jamner, Margaret Schneider, and Daniel Stokols, editors. Promoting Human Wellness: New Frontiers for Research, Practice, and Policy. Berkeley:  University of California Press,  c2000 2000.



Cornelius L. Hopper and Irene Bronston
  Acknowledgments xii
  Introduction: New Frontiers for Research, Practice,
and Policy
Margaret Schneider Jamner
  Part 1
New Directions in Human Wellness Promotion
1 The Social Ecological Paradigm of Wellness Promotion
Daniel Stokols
2 The Societal Context of Disease Prevention and
Wellness Promotion
Lester Breslow
3 Promoting Wellness: Biomedical versus Outcomes Models
Robert M. Kaplan


Community Participation, Empowerment, and Health:
Development of a Wellness Guide for California
S. Leonard Syme
5 Genetic Determinism as a Failing Paradigm in Biology
and Medicine: Implications for Health and Wellness
Richard C. Strohman
  Part 2
Wellness Promotion Research: Innovative Strategies
and Perspectives
6 Creating Health-Promotive Environments: Implications
for Theory and Research
Daniel Stokols
7 Theory-Based Evaluation: Investigating the How and
Why of Wellness Promotion Programs
Johanna Birckmayer and Carol Hirschon Weiss
8 Pregnancy Prevention Opportunities Focusing
on the Younger Sisters of Childbearing Teens
Patricia L. East
9 Immigrants May Hold Clues to Protecting Health during
Pregnancy: Exploring a Paradox
Sylvia Guendelman
10 Race and Health: Implications for Health Care Delivery
and Wellness Promotion
Mack Roach III
11 Health Promotion in Ethnic Minority Families:
The Impact of Exposure to Violence
Kathy Sanders-Phillips
12 Valuing Future Health in Social Policy and Human
Health Behavior
Theodore G. Ganiats and William J. Sieber


Part 3
Wellness Promotion Practice: Toward More
Comprehensive Approaches
13 Health Promotion at the Dawn of the 21st Century:
Challenges and Dilemmas
Meredith Minkler
14 Bridging the Clinical and Public Health Approaches to
Smoking Cessation: California Smokers' Helpline
Shu-Hong Zhu and Christopher M. Anderson
15 Disease Prevention versus Health Promotion: Pitfalls
of Preventive Care in the Geriatric Population
Andrew Duxbury
16 Preventing Disability in Older Americans:
The Challenge of the 21st Century
John C. Beck
17 An Educational Approach to Engage Health Care
Professionals in Wellness Promotion
Stuart J. Slavin and Michael S. Wilkes
18 University-Community Partnerships to Promote Wellness
in Children, Youth, and Families
Philip R. Nader
  Part 4
Wellness Promotion Policy: Toward a More Explicit
Consideration of the Political Context
19 Strategies for Reducing Youth Violence: Media,
Community, and Policy
Lawrence Wallack
20 Adolescent Sexuality and Health Care Reform
Adele Dellenbaugh Hofmann


Improving Health and Safety in the
Agricultural Workplace
Marc B. Schenker
22 Enhancing Women's Health: Current Status and
Directions in Research and Practice
Annette L. Stanton, Sharon Danoff-Burg, and
Sheryle J. Gallant
23 Cardiovascular Disease in Women: Exploring Myths
and Controversies
Amparo C. Villablanca
24 HIV/AIDS Prevention: Successes and Challenges
Craig R. Waldo and Thomas J. Coates
Sheldon R. Waldo and Joyce C. Lashof
  Notes on Contributors 701
  Index 713



On behalf of the University of California, we are proud to offer this collection of essays and Wellness Papers in the new Promoting Human Wellness text. This volume represents the fruits of a collaborative vision shared by The California Wellness Foundation (TCWF) and the University of California. Since 1991, Wellness Award Lectures have been published annually as policy analyses of varied elements of health promotion and disease prevention, emanating from a selection of the most innovative work of University of California faculty. In 1997, we accepted an invitation from the University of California Press to expand that vision into a text on human wellness, augmented by invited contributions from national experts who have written original papers on the state of women's health, AIDS prevention developments, and wellness promotion evaluation strategies.

The original Wellness Award topics and authors were chosen as a result of an annual university-wide call for abstracts and blinded peer review of submissions. The abstracts were screened for scholarly excellence and applicability to a wide array of health promotion and disease prevention issues. A distinguished university-wide multidisciplinary committee of deans, faculty, and other academic personnel reviewed the submissions and selected each year's awardees. With the addition of invited

essays, we have created a collection that brings into focus the major human health issues facing our nation today.

Our partnership with TCWF, which began in 1995, was intended to further faculty scholarship in response to critical public health problems facing families and communities in California. Among the concerns that the lectures have addressed have been the growth of violence in society, workplace hazards, teen pregnancy, socioeconomic disparities, aging and health promotion, environmental health planning, and other problems that might respond to thoughtful policy realignment and a community health approach.

Free public lectures based on these papers were delivered at University of California campuses each October. In 1997, the papers included proposed solutions to or analyses of different aspects of the problem of violence in our society, particularly gang proliferation and domestic violence recognition; a review that raises significant policy questions for occupational health in the agricultural workplace; an evaluation of a multiyear TCWF Violence Prevention Initiative and its impact on communitybased gun ordinances throughout California; and an analysis of The Wellness Guide Program, also supported by a grant from TCWF.

An important theme has emerged in several of these papers, namely, that empowerment of communities and individuals is a core prerequisite of good physical and mental health. The related concepts of resilience and self-determination have long been known to contribute to the individual's ability to thrive and grow. As a society, we need to revisit this premise and strive to imbue our community organizations and health and social service programs with the necessary resources and tools that empower individuals and families to recognize and build on their own strengths.

The goal of our Wellness Lectures Program has been to inform the ongoing health policy reform process and to contribute to a more humane future for California and the nation. We believe that the Wellness Award Lectures papers meet that challenge. Our intent is to raise questions in a thoughtful and provocative way among a new generation of health professionals, bioethicists, and social scientists—those who will help shape the humanity and texture of American society in the coming millennium. We hope that these papers will find their way into the classrooms of schools of medicine, nursing, public health, social work, and bioethics as well as college courses in women's studies, ethnic studies, psychology, sociology, economics, and social ecology. We hope that the book will

be of interest to health professionals, policy makers, researchers, and students in the field of wellness promotion.

Cornelius L. Hopper, M.D.
Vice President—Health Affairs

Irene Bronston, M.P.H.
Coordinator, Wellness Lectures Program

University of California Office of the President
January 2000



The editors would like to thank Roger Greaves, former CEO, Health Net, Gary L. Yates, President and CEO of The California Wellness Foundation (TCWF), and the University of California Office of Health Affairs for support of the UC Wellness Lecture Series and this volume. We would also like to thank the members of the editorial review board, including Lester Breslow, M.D., M.P.H.; Joyce Lashof, M.D.; Sheldon Margen, M.D., M.P.H.; and Mary Walshok, Ph.D., for their valuable contributions to the selection and review of the chapters to be included in this volume. A number of individuals deserve our gratitude for their efforts both on the Lecture Series and with the book project. These include Irene Bronston, M.P.H., Coordinator, TCWF/UC Wellness Lectures Program; Adele Amodeo, M.P.H., Coordinator, UC Health Policy and Legislation; Cornelius Hopper, M.D., UC Vice President of Health Affairs; and Don Prial, Public Relations Counsel. Finally, we would like to thank the following individuals for their comments and suggestions on draft manuscripts: Virginia Alhuesen, Ph.D.; Elaine Alpert, M.D.; Lisa Berkman, Ph.D.; Claire Brindis, M.D.; Ross Conner, Ph.D.; Jeannie Gazzaniga, Ph.D.; Joyce C. Lashof, M.D.; Shari MacMahan, Ph.D.; Lauri Pasch, Ph.D.; Arthur Rubel, Ph.D.; Norman Schneider, Ph.D.; Sora Park Tanjasiri, Ph.D.; Tammy Tengs, Ph.D.; Dawn Upchurch, Ph.D.; and Stewart Wolf, M.D.



New Frontiers for Research,
Practice, and Policy

Margaret Schneider Jamner

Wellness promotion, as this volume demonstrates, is at a critical juncture as we enter the new millennium. We have gained an appreciation for the complexity of the task and are beginning to develop methods for identifying the most effective strategies for improving the health-related quality of life among Americans. Moreover, we have expanded our sphere of influence to encompass not only the immediate causes of morbidity and mortality but also the more fundamental determinants that reside in the political, social, and physical environments. This volume illustrates the potential for promoting human wellness that has been generated by these developments. Future success in realizing this potential relies on recognizing that elements that are often encountered as barriers to health promotion (e.g., political agendas, idiosyncratic populations) can and must be embraced and incorporated into the methods that guide wellness research and practice. Only by employing these elements to serve the ends of wellness promotion will we sustain the current momentum toward creating a nation that supports and facilitates optimal health.

The chapters in this volume offer compelling evidence for the complex web of interrelated influences that operate dynamically to determine health and wellness. Regardless of the specific disease or disability being examined, it is clear that one must consider the likelihood that health status may be affected by variables at many levels, including (but certainly not limited to) the human genome, individual health behavior and psychological attributes, medical care, and the physical and social

environments. Moreover, each level has the potential to interact with factors from other dimensions. As described by Leonard Duhl (1996), “It is as if there were a ball of interconnected strands that could be picked up at any point, and a relationship to all other issues, institutions, people, and places would exist” (p. 259).

The multidimensional model that could be constructed to depict any specific health problem threatens to be overwhelming in its complexity. Nevertheless, conceptualizing wellness promotion from a systems perspective may turn out to be a requirement for effective intervention (Wandersman et al., 1996). It is important, therefore, to note that work presented in these pages also provides testimony that adopting a systems approach to wellness does not preclude elegant solutions to health problems and, in fact, may simplify matters by identifying optimally effective leverage points for intervention. In order to realize the potential within the systems approach for identifying parsimonious pathways to promoting health, it is crucial that health researchers, health practitioners, and policy makers maintain an exceptionally broad vision of the range of activities and targets that may fall within the health promotion mandate. It is equally important that wellness professionals address explicitly the nonscientific forces bearing on the translation of wellness knowledge into effective action.

Public health, the parent discipline to wellness promotion, has been said to permeate “through all the social, environmental, and other activities of populations” (Holland, 1997, p. 1645). Likewise, although the promotion of human wellness is often identified with orchestrating a change in lifestyle, such individual modifications “usually require some combination of educational, organizational, economic, and environmental interventions in support of change in both behavior and conditions of living” (Green et al., 1997, p. 125). Appropriate targets for change in the pursuit of enhanced health and wellness for a population therefore include elements within the individual, the social milieu, the physical environment, the medical care system, the economy, and the political arena. This point is vividly illustrated by the chapters in this volume, several of which present impressive evidence of the powerful force for change that results from directly addressing contextual factors.


A number of the authors featured in this volume argue that the influence of the political context on human wellness deserves greater attention. In

particular, the politicization of health-related issues often results in a markedly skewed allocation of resources with respect to research. Strohman (chapter 5), for example, opines that the share of research funds devoted to mapping the human genome is grossly out of proportion to the health benefits that this project is likely to deliver. His work suggests that far greater salutary outcomes might be expected to accrue to the population if sufficient funds were directed to mapping out the ways in which genes interact with their immediate (i.e., organismic) and distal (i.e., extraorganismic) environments to determine phenotypic expression. It is in the interest of the goals of health promotion that scientists make a concerted effort toward educating political decision makers regarding the connection that proposed health research initiatives have to the objective of improving the health of the population.

This advice should not be interpreted as a condemnation of basic research, whose relationship to the human condition may at times be obscure or difficult to discern. Basic research is and will continue to be of great importance since expanding our understanding of the mechanics of our world can serve us in many unforeseen and significant ways. Nevertheless, in a society characterized by limited resources for research, the way in which these resources are distributed should be continually reassessed in order to determine whether adjustments in the allocation are likely to result in greater health returns.

Another instance of value-driven political agendas leading to inequitable resource allocation is the greater emphasis placed on men, as compared to women, in health research. As discussed by Stanton et al. (chapter 22), the historical view of the female as the lesser “deviation” from the male norm has contributed to the disproportionate attention paid to men in health research. The nominal representation of women in the sciences and politics in the past also has helped maintain the illusion that important research questions could be adequately addressed through research on men only. Villablanca (chapter 23) reaffirms this pattern in the case of heart disease. Although coronary heart disease is the leading cause of death for both men and women, the latter have been largely excluded until recently from most heart disease prevention trials. In the last decade, a shift toward greater recognition of the health needs of women has occurred and has led to attempts at establishing greater gender equity in research. This new movement has fueled the Women's Health Initiative, which will yield a wealth of data concerning the factors that influence the health and wellness of women. Wellness professionals would do well to note this apparently successful culmination to years of

campaigning for gender equity in research. Results of the current wave of health research directed at women's issues will provide valuable scientific information useful in promoting women's health. In order to continue the momentum toward a more equitable health research agenda at the national level, wellness professionals must find effective ways to supply their expertise to the political decision-making process.

Values-laden political priorities also play a large role in determining the allocation of resources among interventions designed to improve or enhance health. Hofmann (chapter 20) eloquently lays out the argument for providing teens with complete information concerning contraception, yet recent federal legislation provides funds for school-based sex education that teaches abstinence-only pregnancy prevention. Given the strength of the evidence against the utility of the abstinence-only approach, it appears that this legislation is based not on scientific knowledge but rather on the values of individuals, lobbyists, and organized voter groups expressed as political will. Similarly, Waldo and Coates (chapter 24) describe the failure of HIV prevention programs and attribute this lack of success to a political climate that, for example, blocks widespread use of needle exchange despite ample evidence that allowing drug users to receive sterile syringes in exchange for used needles reduces HIV transmission without increasing drug use. A resolution issued in 1998 by the Presidential Advisory Council on HIV/AIDS (American Public Health Association, 1998) rebuked the president and the secretary of health and human services for failing to remove the ban on using federal funds for needle exchange programs and stated that “tragically, we must conclude that it is a lack of political will, not scientific evidence, that is creating this failure to act.”

The persistence of the “agrarian myth” (a pervasive belief in the salutary conditions of agricultural occupations) in the face of data concerning the health problems of agricultural workers represents another case of an area in which policy and legislation have lagged behind available scientific information. As explained by Schenker (chapter 21), current health and safety legislation designed to protect farmers and farm laborers in the United States is notably insufficient. For example, Schenker notes that rollover protectors for tractors have been legislated in Europe but not in the United States, even though they essentially eliminate rollover fatalities. These examples demonstrate that successful wellness promotion requires engaging the political process in a data-based evaluation of funding and legislative priorities and pushing for policies that have the

greatest likelihood for improving national health status by addressing the actual needs of the nation's constituent populations.

The theorem that the political context plays a large role as a force affecting how public health knowledge is translated into preventive action is generally acknowledged in the field of public health. In the first chapter of the 1997 edition of the Oxford Textbook of Public Health, Detels and Breslow state, “What can be done will be determined by the scientific knowledge and resources available. What is done will be determined by the social and political commitments existing at the particular time and place” (p. 3). One model of this process, proposed by Richmond and Kotelchuck (1983), posits three factors that contribute to the shaping of health policy: knowledge base, political will, and a social strategy. According to Richmond and Kotelchuck, the knowledge base refers to “the scientific and administrative data base upon which to make decisions.” Thus, epidemiologic research, needs assessments, clinical trials, and other forms of intervention evaluations all contribute to this knowledge base that may be used to inform health policy decisions. How or even whether this information is used, however, depends greatly on the political climate.

In discussing the Richmond and Kotelchuck model, Atwood et al. (1997) address the example of preventive priorities in the United States with respect to tobacco control. They point out that, because of a lack of political will, the proportion of resources currently allocated to preventing tobacco use does not correspond to the magnitude of the toll that tobacco takes on human health. These authors suggest that public health researchers should pay greater attention to how their work may be used to shape health policy and should consider this issue as integral to the research planning process. Planning a research agenda with findings that will be useful in shaping health policy is certainly one way to increase the social validity of public health research (Geller, 1991), yet it would be unnecessarily restrictive to confine the spectrum of public health research to one that speaks directly to policy issues.

More to the point, the authors in this volume demonstrate that wellness professionals need not remain detached from the political arena; rather, they may be able to dramatically impact community health by mobilizing political will. In fact, it has been suggested that “one of health education's major supportive functions is to enhance selfconfidence and provide the variety of skills needed by individuals and their communities to influence the policy-making process” (Tones, 1997,

p. 785). The enormous potential for enhancing the impact of wellness promotion activities through shaping political will is illustrated in the interventions described by Minkler and Wallack in this volume. At the neighborhood level, Minkler (chapter 13) demonstrates that residents of a high-crime neighborhood can be successfully mobilized to lobby for increased police protection and consequently create an environment that facilitates improved health behavior. Although the connection between community crime levels and individual health habits has not been clearly established, there is an intuitive link between, for example, a fear of walking in one's neighborhood and the likelihood of walking for exercise. Moreover, as the chapter by Sanders-Phillips (chapter 11) shows, there is evidence suggesting that being exposed to violence in one's community may induce negative psychological states (e.g., depression, hopelessness, ennui) that act as barriers to the establishment of a healthful lifestyle. The Violence Prevention Initiative (VPI) detailed by Wallack (chapter 19) offers a model of what can be accomplished via advocacy of public policy solutions to public health problems. The tools employed by the VPI toward the goal of reducing the widespread and easy availability of handguns to youth have included savvy use of a scientific and applied database, mobilization of a broad range of constituencies, and strategic use of the mass media. Coordination of these elements has resulted in a number of tangible results, including the passage of numerous local gun-control ordinances and statewide legislation (approved by the California State Legislature but later defeated by governor's veto) banning the sale and distribution of Saturday Night Specials.

The programs described by Minkler and Wallack are unique not only because they appear to succeed but also because they squarely address community-level issues that interfere with the “response-ability” of individuals to remain healthy. It is interesting to note that both of these programs are focused on violence. Whereas the VPI (Wallack) targets primarily potential perpetrators and victims of gun violence, the Tenderloin project (Minkler) addresses the indirect effects of living within a climate characterized by the threat of violence. Both programs, however, address the problem of violence through political influence exerted by members of the community. Together, these programs demonstrate that the future success of health promotion relies on a willingness both to tackle social problems that may in the past have been considered outside the domain of public health and to enlist political strategies in the process.

The focus on violence may reflect the growing concern of the American public with the problem of violent crime, a concern that has led to

such legislative developments as the “three strikes” law in California. It may be, therefore, that part of the success of these efforts should be attributed to a preexisting political climate that was hospitable to antiviolence innovations. In this way, then, these programs exemplify how programmatic outcomes may be enhanced when the political climate is not hostile to the intent of the intervention. Recent California gun-related legislative action in the wake of shooting incidents in school and day-care settings further demonstrates that when the political will is galvanized by immediate events, health-promoting legislation may be enacted quite rapidly.


In addition to mobilizing political will, another strategy with great potential for enhancing the translation of scientific expertise into wellness promotion is a greater reliance on nontraditional methods in both community and clinical settings. This point is made explicitly by Syme (chapter 4), who contrasts the success of The Wellness Guide, an intervention tool heavily influenced by qualitative research methods, with large-scale interventions such as MRFIT that were developed using a “top-down” methodology (i.e., expert driven). The idea that purveyors of health information should become familiar with the beliefs, attitudes, knowledge, and perceived needs of the populations they seek to reach is embodied in the tenets of social marketing (Novelli, 1990). Syme extends this approach to incorporate a consumer-driven perspective to selecting not only the method of intervention delivery but also the intervention content. Thus, the Guide that was eventually developed to meet the community's needs actually contained relatively little “health” information. Evaluation of the Guide suggests that it was used by the recipients and resulted in significant cognitive and behavioral changes with implications for health. Similarly, the program described by Minkler (chapter 13) evolved as it did because elderly residents of an inner-city neighborhood were given the opportunity to shape the program. As a result, the investigator's resources were directed toward assisting the residents in their efforts to reduce the threat of crime in their neighborhood. Although these programs do not conform to the traditional view of a health promotion intervention, they succeeded in the sense that they were embraced by the target communities, resulted in tangible improvements, and

facilitated beneficial behavior changes. These success stories offer considerable fuel to the imperative for wellness promotion professionals to step outside the boundaries of traditional public health paradigms and engage in greater attempts to obtain relevant information from the members of the communities that they seek to serve and to do so quite early during program development.

Innovative methods also can play an important role in program evaluation. Strong program evaluations are critical to the growth of community-based health promotion because they can both identify programs that work and provide clues about the reasons that some programs fail. Birckmayer and Weiss (chapter 7) provide a number of examples in their discussion of theory-based evaluation (TBE), in which the use of process evaluation contributes substantially to the interpretation of program evaluation results. Unlike traditional outcomes-only evaluations, documenting program activities in a process evaluation can permit evaluators to distinguish cases in which an intervention fails because of inadequate theory from cases in which an intervention fails because of inadequate program implementation. Since program evaluation remains a linchpin of wellness promotion, it is paramount that future interventions include this type of approach in order to facilitate effective program development.

Ganiats and Sieber (chapter 12), in their discussion of the complexity involved in attaching monetary values to future health outcomes, offer additional evidence in support of using nontraditional methodology in program evaluation. Policy analysts typically assign a numeric discounting rate to both dollars and health in conducting a standard cost-effectiveness analysis. Unfortunately, since most health promotion programs expend dollars in the present for health outcomes in the future, programs tend to fare poorly in these analyses. As Ganiats and Sieber point out, the value that an individual might place on a future health outcome can vary considerably, depending on personal characteristics and circumstances. Only through careful and population-specific studies is it possible to obtain useful estimates concerning how future health outcomes should be valued for a particular program. These authors suggest, therefore, that the qualitative dimension of time preferences with respect to future health outcomes needs to be better understood in order to permit useful comparisons of cost-effectiveness across programs.

Wellness promotion in the clinical setting also stands to gain from incorporating methods that transcend the traditional medical model. Slavin and Wilkes (chapter 17) address this topic from the perspective

of the training provided to physicians. In describing their innovative doctoring curriculum, they emphasize the value of a person-centered diagnostic approach in which physicians consider social facets of patients' health problems. They give the example of detecting domestic violence through patient-centered interviewing techniques and mobilizing community resources to address not only the immediate injury but also the potential for future injury. A person-centered (rather than a diseasecentered) approach to medical treatment decisions also is encouraged by Duxbury (chapter 15), who reviews the pros and cons of screening for prostate cancer. The high probability for false positives in prostate cancer screening and the likelihood that quality of life will diminish following surgical intervention combine to argue against the ultimate benefit of screening to the patient. This conclusion rests, however, on the qualitative assessment of alternative treatment outcomes. In a broader discussion of health promotion strategies for the elderly, Beck (chapter 16) also favors a person-centered approach. Specifically, he describes a comprehensive preventive assessment for the elderly that takes into account physical, social, and medical resources and yields a prioritized set of recommendations. These recommendations go far beyond the typical physicians' advice and may include suggestions such as reducing or eliminating a medication, enrolling in a class at a community college, or installing shower rails for the handicapped.

Unlike the traditional biomedical model, which tends to reduce patients to a disease entity and focuses on isolating and eliminating the disease, the patient-centered approach put forth by these authors seeks to optimize functioning and well-being. In order to successfully achieve this goal, clinicians must include qualitative assessment methods in their diagnostic procedures and, similarly, must consider the impact of treatment on the patient as a whole being. This method of evaluating alternatives for patient treatment has been formalized in the General Health Policy Model (GHPM) described by Kaplan (chapter 3). The GHPM uses a standard metric, quality-adjusted life years (QALYs), to compare intervention strategies. Quality-adjusted life years are based on “the current life expectancy adjusted for diminished quality of life associated with dysfunctional states and the duration of stay in each state” (p. 50). A key component of the GHPM is generated by individuals drawn from the general population who rate various health outcomes according to their relative importance. In other words, laypersons assign relative values to a series of health outcomes in order to quantify their feelings about various states of disability. Consequently, treatment decisions

based on the GHPM are informed by the qualitative dimension of alternative outcomes. All the innovative approaches described by Slavin and Wilkes, Beck, Duxbury, and Kaplan consider the impact of medical intervention on the patient as a functioning person (rather than as a host to a disease) as a factor in treatment decisions. This humanistic framework offers great potential for maximizing patients' health in the face of disease. The future of wellness promotion as an aspect of medical care relies on the institutionalization of these types of innovative approaches in clinical settings.


The various interventions reported by Syme, Minkler, and Wallack, described previously, illustrate what may be accomplished when program planners are receptive to input from the target community and are willing to engage in activities outside the parameters of what has traditionally been considered health promotion. Similarly, The Smokers' Helpline, a smoking cessation program detailed by Zhu and Anderson (chapter 14), represents a novel intervention that is based on extensive qualitative study of the target population. In addition, the Smokers' Helpline has achieved a unique hybridization of the clinical approach with the public health approach to give rise to a highly efficacious—and effective—program.

The Helpline is unusual in several ways. The menu of services offered to smokers who call the Helpline is one key element to the marriage between a population-based reach (the public health approach) and intensive one-on-one treatment (the clinical approach). Also critical is the proactive strategy in which Helpline counselors follow up on initial calls rather than waiting for smokers to call back to begin counseling. A third component that distinguishes the Helpline is the scheduling of counseling sessions such that more sessions occur during the time with the greatest probability of relapse. The availability of counseling in several languages contributes as well to the Helpline's success. Ultimately, of course, dissemination of the Helpline's toll-free numbers is critical and has been accomplished by including the numbers in ads funded by the Tobacco Control Section of the California Department of Health and by establishing partnerships with primary care physicians.

Although not explicitly, the Smokers' Helpline also acts as a model of a health promotion intervention that works within the political context

to further its effectiveness. As explained by Zhu and Anderson, the Helpline was funded through the monies generated by Proposition 99, the California Tobacco Tax Initiative. This proposition was an expression of Californians' support for interventions that would reduce the health threat of tobacco use. Unlike the effort to restrict public smoking, however, the Helpline has not encountered organized political resistance. Very likely, this tacit acceptance stems from the congruence between the conservative point of view that individuals should take responsibility for their own health and behavior and the Helpline's focus on assisting smokers in their own attempts to quit. Even the tobacco industry would be hard-pressed to muster an argument against providing smokers with assistance toward quitting. Consequently, the Helpline has enjoyed a sort of political immunity that has provided room for its growth and success.


A number of the chapters presented in this volume exemplify an approach to promoting human wellness based on the strategy of identifying highimpact leverage points for intervention. In this approach to health promotion, the goal is to identify and make use of “certain behaviors, social roles, and situational conditions [that] can exert a disproportionate influence on personal and collective well-being” (Stokols, 1996, p. 291). East (chapter 8) identifies one such high-impact leverage point in her discussion of the increased risk for pregnancy among younger sisters of childbearing teens. Specifically, she explains how both the indirect effect generated by sisters' shared family and community environment and the direct effect on younger sisters of having an older sister with a child (e.g., orientation to child-rearing, witnessing of social status attributed to the older sister) combine to place younger sisters of childbearing teens at elevated risk for teen pregnancy. East's research points to the potential impact on teen pregnancy rates that might be brought about through interventions targeting these younger sisters of childbearing teens.

In a somewhat different approach to identifying potential leverage points for health promotion, Guendelman (chapter 9) identifies attributes that may bring about desirable birth outcomes among immigrant mothers. She first describes a paradox wherein immigrants from Mexico and Southeast Asia experience more favorable birth outcomes

than would be expected on the basis of their socioeconomic status. She posits that these surprising findings may reflect health-promoting cultural factors, including salutary dietary habits, strong social cohesion, and relatively low substance abuse. Guendelman's approach suggests that wellness promotion should focus not only on intervening to reduce risk factors for disease but also on encouraging attributes associated with more positive health outcomes.

Like Guendelman, Roach (chapter 10) employs an analysis of epidemiologic data to explore an issue of inequity in health outcomes. In this case, the focus is on disentangling the influence of race from that of environmental and behavioral factors associated with race. Through his examination of differential cancer mortality in Blacks versus Whites, Roach calls into question the medical community's adherence to certain disease categorization systems that may mask differences in the extent of disease on diagnosis by a physician. Roach's scrutiny of the data suggests that, by using a grouping scheme that does not differentiate finely enough between stages of cancers, researchers and clinicians may be overlooking the underlying reasons for greater cancer mortality among Blacks. The implication of Roach's argument is that identification of effective leverage points for ameliorating the cancer epidemic will require altering current clinical diagnostic categories and looking beyond race to modifiable variables that may increase the risk of dying from cancer. The more general lesson to be taken from Roach's argument is that clinical diagnostic categories may artificially obscure important differences between patients and that these differences may provide clues for primary or secondary disease prevention.

Another technique for promoting wellness through identifying critical leverage points is to specify a particular context within which individuals may be affected by an intervention and then design a program for that context. Nader offers an example of this strategy in his description of the school-based CATCH intervention (chapter 18). Schools often have been put forth as locales within which children may be reached and effectively influenced to enhance their health. Nader adds to this traditional perspective the proposition that partnerships between universities and schools yield mutually beneficial avenues for promoting children's health. Such partnerships can offer valuable field experience to university students, supply meaningful data to university-based researchers, and bring about positive changes in behavior among schoolchildren. Thus, the partnership between the university and schools may be a potent leverage point both for improving community health directly

and for enhancing the education of individuals who will be in a position to influence community health at the conclusion of their training.


Whether addressing the question of who should be targeted, what to target, or where an intervention should be delivered, identifying optimal leverage points should be a priority in wellness promotion. It has been suggested, for example, that the primary task for those interested in promoting human wellness is to “set up the conditions that optimize health … [including] such naively elementary ideas as abolishing war, meeting basic needs (not wants) and redistributing the wealth of the planet” (Duhl, 1996, p. 261). More specifically, Syme (chapter 4) states that the most critical factors related to health are problems of inequity and that, therefore, all available energy should be devoted toward minimizing the unequal distributions of resources within the society of the United States. This goal is perhaps so daunting to most individuals that it may induce a sense of helplessness. In fact, however, there are many specific instances of inequity in American society that appear more vulnerable to influence when viewed independently. Access to health care services is one issue frequently mentioned in discussions of inequity and health. Many factors contribute to the fact that some people have greater and easier access to health care than others. Lack of insurance, transportation, child care, or language skills are but a few of the barriers that can prevent individuals from receiving health care. These impediments are very real and quite prevalent. Nevertheless, one example in this volume—the Smokers' Helpline—demonstrates a program in which all these problems have been minimized. The smoking cessation program is funded through cigarette tax dollars, thus eliminating the need for health insurance. The program is administered via telephone, which does away with the need for transportation or child care. Finally, the service is available in several languages to accommodate non-English speakers and allow some tailoring to specific cross-cultural concerns, such as issues of confidentiality. It would be naive to suggest that similarly elegant solutions could be found for the great host of health care services that are available to some and not to others. Still, it is encouraging to note that in at least one case technology, legislation, and innovative programming have been synthesized into a service that is available free of charge to the great majority of California residents.


The present volume aims to stimulate progressive action in the wellness field that speaks to the several themes outlined in this introduction (i.e., consideration of the political context and mobilization of political will, deployment of nontraditional assessment and intervention strategies, and identification and appropriate exploitation of high-impact leverage points). There are a number of specific topic areas both within and outside the sphere of traditional public health that are not covered in this collection and that have clear implications for health. For example, there are dramatic contrasts in the quality of schooling provided to the nation's youth. Since education level is quite strongly predictive of various health behaviors as well as of overall health status, it would be appropriate for those interested in improving the health of Americans to turn attention toward equalizing educational opportunities. Other areas that are not featured in this volume include work-site health promotion, unintentional injuries (e.g., vehicular accidents), and substance abuse. The reader should bear in mind that this volume is not intended to be a comprehensive review of the wellness field; rather, it attempts to motivate innovation within the field by drawing attention to topics that have received inadequate attention by wellness researchers and practitioners until now.

This introduction has attempted to highlight some of these previously underemphasized themes. The necessity of adopting a systems perspective toward wellness promotion, the value added by innovative assessment strategies, and the utility of considering political influences as a force in health promotion are acknowledged in these pages by physicians, sociologists, psychologists, biomedical researchers, and those whose training is grounded in the field of public health. That so many disparate disciplines find common ground with regard to wellness promotion is further evidence that the complex web that defines the field reaches into multiple academic domains and requires an interdisciplinary effort to understand the relevant issues and devise appropriately elegant solutions. It should therefore not be surprising that the varied works presented here do in fact share a great deal of overlap in terms of the underlying messages regarding using a multifactorial approach to wellness, examining wellness issues in context, and targeting intervention points that are maximally effective. With these tenets as guidelines, the future of human wellness promotion promises to be replete with innovative solutions to old problems, with rapid responses to new problems, and with greater synergy between existing strategies for coping with ongoing problems. The current trajectory of the field thus holds great potential for improving the health of our nation and its constituent populations.



“American Public Health Association” . (1998). AIDS panel chastises administration for inaction on needle exchange. The Nation's Health, April, 1.

Atwood, K., Colditz, G. A., and Kawachi, I. (1997). “From public health science to prevention policy: Placing science in its social and political contexts.” American Journal of Public Health87, 1603–1606.

Detels, R., and Breslow, L. (1997). “Current scope and concerns in public health.” In R. Detels, W. Holland, J. McEwen, and G. S. Omenn, eds., Oxford Textbook of Public Health.Pp. 3–17. New York: Oxford University Press.

Duhl, L. J. (1996). “An ecohistory of health: The role of “Healthy Cities”.” American Journal of Health Promotion10, 258–261.

Geller, E. S. (1991). “Where's the validity in social validity?” Journal of Applied Behavior Analysis24, 189–204.

Green, L., Simons-Morton, D., and Potvin, L. (1997). “Education and life-style determinants of health and disease.” In R. Detels, W. Holland, J. McEwen, and G. S. Omenn, eds., Oxford Textbook of Public Health.Pp. 125–137. New York: Oxford University Press.

Holland, W. W. (1997). “Overview of politics and strategies.” In R. Detels, W. Holland, J. McEwen, and G. S. Omenn, eds., Oxford Textbook of Public Health.Pp. 239–243. New York: Oxford University Press.

Novelli, W. D. (1990). “Applying social marketing to health promotion and disease prevention.” In K. Glanz, F. M. Lewis, and B. Rimer, eds., Health Behavior and Health Education: Theory, Research and Practice.Pp. 342–369. San Francisco: Jossey-Bass.

Richmond, J. B., and Kotelchuck, M. (1983). “Political influences: Rethinking national health policy.” In C. McGuire, R. Foley, A. Gorr, et al., eds., Handbook of Health Professions Education.Pp. 386–404. San Francisco: Jossey-Bass.

Stokols, D. (1996). “Translating social ecological theory into guidelines for community health promotion.” American Journal of Health Promotion10, 282–298.

Tones, K. (1997). “Health education, behavior change, and the public health.” In R. Detels, W. Holland, J. McEwen, and G. S. Omenn, eds., Oxford Textbook of Public Health.Pp. 783–814. New York: Oxford University Press.

Wandersman, A., Valois, R., Ochs, L., et al. (1996). “Toward a social ecology of community coalitions.” American Journal of Health Promotion10, 299–307.




The chapters in part 1 introduce a trend toward innovation in the field of human wellness and present both general paradigms and specific examples of the multidimensional and person-centered perspective on wellness promotion. Stokols (chapter 1) describes an emerging paradigm of wellness promotion, traces the societal developments that have prompted greater interest in wellness promotion since the 1970s, and highlights the challenge and value of developing broadergauged and more comprehensive wellness promotion strategies for the future. Breslow (chapter 2) continues this line of argument and focuses specifically on the need to incorporate the social context as a factor in wellness promotion planning and intervention. Strohman (chapter 5) expands the discussion in an unusual direction with his examination of the role of genetics in human wellness. He argues for a greater appreciation of the complex inter-relationships between the genome and both micro- and macroenvironments in the determination of phenotypic expressions related to health. Finally, Kaplan (chapter 3) and Syme (chapter 4) provide concrete examples of the types of programs that can emerge from a more integrative paradigm of wellness. Kaplan describes an alternative model for evaluating medical interventions based on “quality of life years,” that is, treatment outcome in terms of both extending the duration of life and improving the quality of life. Kaplan's General Health Policy Model yields a different order of priorities for medical procedures than the traditional approach of valuing interventions solely in terms of

effect on longevity. This alternative system of prioritizing reflects the effect of medical intervention on quality of life and suggests that population health status might be enhanced if resources were shifted away from procedure-based reimbursement toward primary prevention. Finally, Syme focuses on community-based wellness interventions and follows a synopsis of noted failures in this area with more encouraging information about several successes employing nontraditional approaches. In particular, he presents initial findings from a theory-based evaluation of The Wellness Guide, a printed community resource guide distributed to mothers on the Women, Infants, and Children (WIC) program. These and other data provide fuel for his advice that wellness professionals “should learn from the mistakes of the past … to be creative and inventive enough to become experts in the role of not being experts.” Thus, as a whole, part 1 functions as a rallying call to the wellness field to step outside the lines of traditional wellness programming and think creatively about new means for promoting human wellness at both the individual and the community level.



Daniel Stokols


Prior to the 1970s, efforts to improve individual and population health focused almost entirely on the medical treatment of disease. The concepts of health promotion and wellness were little known, and financial investment in preventive health care was quite limited, accounting for only 2.5% of the nation's annual health care expenditures during the early 1970s (Brennan, 1982).

Less than three decades later, societal commitment to disease prevention strategies has increased dramatically. Local, state, and federal governments have enacted public policies designed to curtail individuals' use of tobacco products, protect workers' safety, and reduce driving-related injuries and fatalities (Breslow and Johnson, 1993; Pertschuk and Shopland, 1989; Wells et al., 1997; Williams et al., 1983). Businesses and managed care organizations have invested substantial funds toward the development and implementation of disease prevention programs (Fielding and Piserchia, 1989; Satcher and Hull, 1995; U.S. Public Health Service, 1992, 1993). In 1996, the U.S. Food and Drug Administration successfully imposed regulatory constraints on the sale and distribution of cigarettes and smokeless tobacco products to minors (U.S. Food and Drug Administration, 1996). It is likely that even more stringent regulatory constraints on the production and sale of tobacco products will be enacted at state and federal levels during the coming years (Newsweek, 1998).

The burgeoning interest in disease prevention and health promotion

over the past 30 years has been prompted by several societal developments and concerns. During the 1960s, the U.S. surgeon general's reports on smoking and health and the health consequences of smoking (U.S. Public Health Service, 1964, 1967) gave official, widespread recognition to the fact that a behavioral factor, cigarette smoking, is a cause of cancer and other serious diseases. Additional evidence for the links between smoking behavior and lung cancer was presented in the Surgeon General's Report on Health Promotion and Disease Prevention (U.S. Public Health Service, 1979), which declared that “cigarette smoking is the single most important preventable cause of death” (p. 7). In recent years, other developments—including the exponential rise in national health expenditures, growing concerns about the deficiencies of the medical care system, heated debates about health care reform, evidence for the health and financial benefits of disease prevention programs, and growing use of unconventional or “alternative” medical therapies among both physicians and the lay public—brought the concepts of wellness and health promotion programming to the forefront of the national agenda (Breslow and Johnson, 1993; Eisenberg et al., 1993; Fielding and Halfon, 1994; Fries et al., 1993; Satcher and Hull, 1995; Schauffler, 1993; Weiss, 1991).

Several scientific studies have documented the substantial health benefits and financial savings associated with disease prevention and health promotion programs (DeJoy and Wilson, 1995; O'Donnell and Harris, 1994; Pelletier, 1996; Stokols et al., 1995). Effective wellness promotion strategies include employee health risk appraisal, counseling, and lifestyle change programs (Erfurt et al., 1991; Fries et al., 1994), cultural change strategies within organizational settings (Allen and Allen, 1986; Bellingham, 1990), and the provision of clinical preventive services to enhance maternal and child health (Thompson et al., 1995; U.S. Preventive Services Task Force, 1989). A mid-decade review of progress toward meeting the Healthy People 2000 goals in the United States found substantial reductions in adult use of tobacco products and in alcoholrelated automobile deaths and moderate gains in the proportion of adults exercising regularly and eating less fatty diets (McGinnis and Lee, 1995). Also, the proportion of workplaces providing health promotion programs for their employees increased significantly between the mid-1980s and early 1990s (U.S. Department of Health and Human Services, 1992).

Although many health promotion programs have been effective, others have failed or achieved only limited success. For example, although

employers have made substantial efforts to bring their workplaces into compliance with state and federal regulations aimed at reducing occupational injuries and illnesses, the adverse health and economic impacts of work-related conditions continue to be enormous. In 1995, occupational injuries accounted for $121 billion in lost wages and administrative and health care costs (Rosenstock, Olenac, and Wagner, 1998). Moreover, even the best-designed work-site health promotion programs reach only a small proportion of the total workforce. Participants in these programs tend to be healthier, better paid, more educated, and more motivated to change their health habits than nonparticipants (DeJoy and Wilson, 1995; O'Donnell and Harris, 1994). Further, lifestyle-change programs that focus narrowly on modifying specific health behaviors often neglect the contextual circumstances that lead to high relapse and attrition rates once the interventions have ended (Marlatt and Gordon, 1985; Prochaska et al., 1992). In addition, certain health risks such as exposure to community violence, obesity, teen pregnancy, substance abuse, financial barriers to medical and preventive services, and lack of adequate health insurance remain “segmented in pockets of heightened prevalence” (Fisher, 1995), particularly among low-income and minority groups in the population (Adler et al., 1994; Fontanarosa, 1995; Margen and Lashof, this volume's afterword; McGinnis and Lee, 1995; Satcher and Hull, 1995).

To improve the health of vulnerable populations and reduce the selfselection biases and attrition rates associated with many intervention programs, broader-gauged strategies of health promotion will be required that combine behavioral, organizational, environmental, regulatory, and political initiatives to alleviate community sources of illness and injury (Atwood et al., 1997; Montes et al., 1995; Winkleby, 1994; Winett et al., 1989). The limitations of earlier disease prevention and health promotion programs, noted previously, highlight the need for a major paradigm shift away from narrowly focused interventions aimed primarily at changing individuals' health behavior toward more comprehensive formulations that address the interdependencies among socioeconomic, cultural, political, environmental, organizational, psychological, and biological determinants of health and illness (Richmond and Kotelchuck, 1991; Schneider Jamner, this volume's introduction; Stokols, 1996; Winett et al., 1989).

Researchers in the field of public health have recognized for many years that patterns of health and illness are closely linked to a variety of sociocultural, political, and physical-environmental conditions within

communities (Breslow, 1996; Cassel, 1964; Catalano, 1979; Detels and Breslow, 1997; Durkheim, 1951; McKinlay, 1975; Syme, chapter 4). The “new public health” outlined in the Ottawa Charter gave explicit emphasis to social causes of illness, above and beyond the physicalenvironmental health threats that exist in certain communities (Kickbush, 1989; Ottawa Charter for Health Promotion, 1986). The social ecological paradigm for health promotion extends these earlier notions by providing a set of conceptual and methodological principles, drawn largely from systems theory, for organizing comprehensive, communitybased health promotion programs (Emery and Trist, 1972; Green and Ottoson, 1994; McLeroy et al., 1988; Miller, 1978; Moos, 1979).

The next section of the chapter outlines some of the developments, both in the field of public health and in society as a whole, that have led to the development of a broader, more integrative approach to wellness promotion—the social ecological model—which emphasizes the joint influence of behavior and environment on wellness rather than focusing exclusively on either category of health-determining factors.


We live in an era in which the dominant causes of morbidity and mortality are strongly linked to human behavior at the individual, community, and government levels (Detels et al., 1997). A crucial challenge for the 21st century is to develop programs and policies that will establish health-promotive environments at local, regional, and global levels—those that minimize individuals' exposure to health-threatening conditions and support their efforts to promote personal and collective wellness. The enormity of this task stems from the complex web of health-determining factors that impinge on individuals, organizations, and whole communities. Although these complexities make the task of promoting human wellness seem rather daunting, the “small wins” approach to social problems (Weick, 1984) suggests that as incremental health promotion and environmental protection strategies are adopted in local communities, they can exert a positive, albeit gradual, influence on population health.

Essential prerequisites for developing effective environmental design and public policy programs to create healthful surroundings are sound theoretical analyses of key concepts such as “health,” “wellness,” “wellness

promotion,” and “healthy environments.” A review of the relevant research literature on topics such as health promotion, environmental stress, and environmental risk assessment, however, reveals important gaps in our understanding of these issues. For example, health is often defined in individualistic and physicalistic terms with explicit emphasis on “soundness of body or mind and freedom from disease or ailment” (Webster's, 1989). Analyses that define health simply as the absence of personal illness or injury, however, give little or no consideration to issues of collective well-being (e.g., social cohesion and sense of community; cf. Sarason, 1974) and optimal states of wellness (e.g., strong feelings of personal commitment to one's social and physical milieu; cf. Pelletier, 1994). The term “wellness” is used in this volume to refer to the broader conception of health, which encompasses not only the absence of illness symptoms but also very positive states of well-being.

The terms “disease prevention” and “health protection” have been used to describe various medical and public health strategies aimed at preventing the onset of physical and mental illness (e.g., inoculation against infectious diseases, enhanced community sanitation services, reduction of workplace hazards, and governmental regulation of food and drug safety). The concepts of health or wellness promotion, however, differ from the disease prevention orientation in that they place greater emphasis on the role of individuals, groups, and organizations as active agents in shaping health practices and policies to optimize both individual and collective well-being (e.g., U.S. Public Health Service, 1979, 1991; Williams, 1982; Winett et al., 1989; World Health Organization, 1984).

The majority of health promotion programs implemented in corporate and community settings have focused on changing individuals rather than their environments, organizations, or institutions. That is, they have been designed to modify individuals' health habits and lifestyles (e.g., exercise and dietary regimens) rather than to provide environmental resources and interventions that promote enhanced well-being (e.g., installation of improved ventilation systems within buildings to enhance indoor air quality, design of safe stairways to reduce falls and injuries, modification of agricultural machinery to reduce occupational injuries, and provision of insurance coverage for preventive risk-factor screenings among the elderly). Much recent research, however, suggests the potential value of environmental and institutional interventions as an adjunct to behaviorally oriented health promotion programs (e.g., Archea, 1985; Archea and Connell, 1986; Beck, chapter 16; Green and Kreuter, 1990;

Greenberg, 1986; Hedge, 1989; Karasek and Theorell, 1990; Lawrence, 1990; Mendell and Smith, 1990; Robertson, 1986; Schenker, chapter 21; Syme, 1990; Williams, 1982; Winett et al., 1989).


The chapters in this volume reflect recent efforts to develop broader and more comprehensive strategies for promoting human wellness. These strategies recognize the multiplicity of factors that influence personal and collective well-being and emphasize the multidisciplinary foundations of scientific research, professional practice, and health policy analysis in the field of wellness promotion. Taken together, these chapters reflect several themes that are intrinsic to the social ecological paradigm as a basis for understanding the community and environmental origins of public health problems and for organizing disease prevention and wellness promotion programs that can effectively ameliorate those problems. Specifically, the chapters in this volume highlight the following:

  1. The advantages of implementing disease prevention and wellness promotion programs that target multiple health risks and illnesses (e.g., cancer, cardiovascular disease, AIDS and HIV, unintentional injuries, and community violence) rather than focusing narrowly on singular disease categories
  2. The strategic value of identifying and reducing threats to public health at several community levels and in the context of multiple settings (e.g., residential settings, schools, workplaces, and health care facilities)
  3. The well-documented links among poverty, minority status, and susceptibility to disease and the importance of targeting vulnerable groups in the population for preventive services and wellness programs
  4. The advantages of combining multiple strategies of disease prevention and wellness promotion (e.g., lifestyle change, health education, medical practice, environmental enhancement, media campaigns, and regulatory initiatives) within comprehensive health promotion programs
  5. The challenge of rigorously evaluating the health benefits, costeffectiveness, and sustainability of alternative disease prevention and wellness promotion programs

A major goal of this chapter is to delineate the core theoretical and programmatic assumptions that underlie the social ecological approach to

wellness promotion and the ways in which these assumptions are relevant to the research and intervention programs described in several chapters of this volume.


The term “ecology” pertains broadly to the interrelations between organisms and their environments (Hawley, 1950). From its early roots in biology, the ecological paradigm has evolved within several disciplines (e.g., sociology, psychology, economics, and public health) to provide a general framework for understanding the nature of people's transactions with their physical and sociocultural surroundings (e.g., Barker, 1968; Cassel, 1964; Catalano, 1979; Park and Burgess, 1925; Rogers-Warren and Warren, 1977). The field of social ecology gives greater attention to the social, institutional, and cultural contexts of people-environment relations than did earlier versions of human ecology, which were more closely oriented to biological processes and the geographic environment (e.g., Alihan, 1964; Binder et al., 1975; Michelson, 1970). The social ecological perspective encompasses certain core assumptions about the dynamics of human health and the development of effective strategies to promote personal and collective well-being. These assumptions are outlined here.

First, the healthfulness of a situation and the well-being of its participants are assumed to be influenced by multiple facets of both the physical environment (e.g., geography, architecture, and technology) and the social environment (e.g., culture, economics, and politics). Moreover, the health status of individuals and groups is influenced not only by environmental factors but also by a variety of personal attributes, including genetic heritage, psychological dispositions, and behavioral patterns. Thus, efforts to promote human well-being should be based on an understanding of the dynamic interplay among diverse environmental and personal factors rather than on analyses that focus exclusively on environmental, biological, or behavioral factors (cf. Moos, 1979; Stokols, 1996). For example, chapter 5 by Strohman highlights the joint influence of individuals' genetic heritage and the social environment on their susceptibility to various illnesses. Also, Minkler's research demonstrates the value of community organizing techniques aimed at ameliorating health-threatening conditions in a residential area (e.g., the threat of

street crime and violence) as a basis for enhancing the health-promotive behavior of elderly neighborhood members. And Margen and Lashof emphasize the ways in which poverty, inequality of income distribution, and minority status jointly undermine the health of individuals and vulnerable subgroups in the population.

Second, analyses of health and health promotion should address the multidimensional and complex nature of human environments. As noted previously, environments can be described in terms of their physical and social components, but they also can be characterized in terms of their objective (actual) or subjective (perceived) qualities and their scale or immediacy to individuals and groups (proximal vs. distal). Furthermore, environments can be described as an array of independent attributes (e.g., lighting, temperature, noise, space arrangement, and group size) or in terms of the composite relationships among several features, as exemplified by constructs such as behavior settings, person-environment fit, and social climate (Stokols, 1987). The highly variegated nature of human environments has direct implications for the design and evaluation of health promotion programs, as illustrated by several of the chapters in this volume. For example, chapter 9 by Guendelman highlights the diversity of environmental factors that affect pregnancy outcomes in immigrant and nonimmigrant populations, including the level of social support available to individuals, the nutritional content of their diet, and the quality and availability of prenatal care. Schenker's chapter similarly illustrates the multiple environmental factors that influence agricultural workers' health, including their levels of exposure to sunlight, pesticides, and unsafe equipment. And the chapters by Stanton, Danoff-Burg, and Gallant; Roach; and Villablanca reveal the subtle ways in which societal and cultural values influence the direction and funding of health research programs and sometimes lead to racially, culturally, and gender-biased interpretations of research findings.

Third, just as environments can be described in terms of their relative scale and complexity, the participants in those environments can be studied at varying levels ranging from individuals, small groups, and organizations to larger communities and populations. Rather than focusing solely on individuals or aggregates, the social ecological perspective incorporates multiple levels of analysis and diverse methodologies (e.g., medical exams, questionnaires, behavioral observations, environmental recordings, and epidemiologic analyses) for assessing the healthfulness of settings and the well-being of individuals and groups. For

example, chapter 7 by Birckmayer and Weiss shows that the evaluation of wellness promotion programs can employ many types of measures at different levels of intervention, from indices of individual attitude and behavior change to measures of the health impacts of organizational policy changes. The chapters by Syme and Schneider Jamner also note that evaluations of community-based wellness promotion programs, ideally, should incorporate qualitative as well as quantitative measures to provide a more complete assessment of intervention outcomes.

Moreover, the social ecological perspective assumes that the effectiveness of health promotion programs can be enhanced through the coordination of individuals and groups acting at different levels, for example, family members who make efforts to improve their health practices, corporate managers who shape organizational health policies, and public health officials who supervise community health services (e.g., Green and Kreuter, 1990; Pelletier, 1984; Winett et al., 1989). Accordingly, chapter 24 by Waldo and Coates suggests that HIV and AIDS prevention efforts should engage multiple social units to promote health-enhancing behavioral changes, including individuals, small groups, organizations, and community decision-making groups. The chapter by East highlights the influential role of childbearing teens in shaping the sexual practices and likelihood of unwanted pregnancy among their younger sisters, and that by Nader underscores the value of collaborative partnerships linking universities with other community groups as a basis for promoting wellness among children and their families.

Fourth, the social ecological perspective incorporates a variety of concepts derived from systems theory (e.g., interdependence, homeostasis, negative feedback, and deviation amplification; see Cannon, 1932; Emery and Trist, 1972; Katz and Kahn, 1966; Maruyama, 1963) to understand the dynamic interrelations between people and their environments. Thus, people-environment transactions are characterized by cycles of mutual influence whereby the physical and social features of settings directly influence their occupants' health, and, concurrently, the participants in settings modify the healthfulness of their surroundings through their individual and collective actions. These cycles of mutual influence between individuals and their environments are evident in chapter 11 by Sanders-Phillips, showing that individuals' routine exposure to community violence leads to feelings of disempowerment, helplessness, and depression, which in turn preclude their efforts to adopt health promotive practices related to dietary improvement and physical activity. Similarly,

Duxbury's chapter shows that wellness among the elderly depends on the degree of fit between individuals' functional abilities on the one hand and environmental constraints on their daily activities on the other. Either or both of these elements can be modified to promote higher levels of person-environment fit (cf. Lawton, 1999).

Moreover, the social ecological model views human environments as complex systems in which local settings and organizations are nested within more complex and remote regions. Accordingly, efforts to promote human well-being must take into account the interdependencies that exist among immediate and more distant environments. For example, the occupational health and safety of community work settings are directly influenced by state and national ordinances aimed at protecting environmental quality and public health (cf. Schenker, chapter 21). Similarly, research conducted as part of the Violence Prevention Initiative in California (Wallack, chapter 19; Wintemute, 1992) demonstrates the substantial impact of public policies (especially those regulating the production and sale of firearms) in lowering rates of community violence, injury, and homicide. Additionally, Hofmann (chapter 20) contends that managed care policy reforms guaranteeing age-appropriate health care to teenagers would be one of the most effective strategies for reducing unwanted teen pregnancies and abortions.

Finally, owing to the complexity of human environments and an explicit emphasis on multilevel and multimethod analyses of behavior, the social ecological perspective is inherently interdisciplinary in its approach to health research and the development of health promotion programs. The ecological perspective draws on the fields of medicine and public health, as well as the behavioral and social sciences, in the study and promotion of human well-being. The chapter authors represented in this volume, for example, bring a wide range of disciplinary training and perspectives to the field of wellness promotion, including molecular genetics, medicine, epidemiology, psychology, sociology, anthropology, and economics. Moreover, social ecological approaches to wellness promotion link the community-wide, preventive strategies and epidemiologic orientation of public health with the individual-level, therapeutic and curative strategies of medicine. For example, the California Smokers' Helpline described in chapter 14 by Zhu and Anderson epitomizes a health-promotive intervention that combines the broad scope and “reach” of a statewide public health program with the intensity of a more personalized clinical intervention. Similarly, the medical education programs described in chapter 17 by Slavin and Wilkes introduce physicians

to a variety of public health and behavioral medicine concepts and train them to be more effective change agents for promoting wellness among their patients.

The social ecological perspective also incorporates the behavioral and social science emphases on (1) the active role played by individuals and groups in modifying their own health behavior and well-being, (2) the development and testing of theoretical models describing people environment transactions, and (3) the importance of conducting evaluative studies to assess the cost-effectiveness and social impact of wellness promotion programs (e.g., Cassel, 1964; Engel, 1976; Evans, 1988; Ganiats and Sieber, chapter 12; Henderson and Scutchfield, 1989; Kaplan, chapter 3; Schwartz, 1982; Williams, 1982; Winett et al., 1989). For example, the statewide Wellness Guide described in chapter 4 by Syme exemplifies a community-level intervention whose development and evaluation were guided by psychological and sociological theories of the links between personal and community empowerment on the one hand and the likelihood and efficacy of individuals' health-promotive behavior on the other.

The core assumptions and themes inherent in ecological approaches to wellness promotion are elaborated on in subsequent chapters of the volume. For example, the multiple dimensions of environments and the ways in which they are related to individual and collective well-being are examined in the chapter by Stokols in part 1. The interactions among several categories of environmental and personal determinants of health also are discussed in that chapter. Moreover, social ecological perspectives on health suggest that the effectiveness of wellness promotion efforts can be enhanced through multilevel intervention “packages” (Geller, 1987; Weiss, 1991; Williams, 1982; Winett et al., 1989) combining both behavioral and environmental modification strategies. Thus, chapter 16 by Beck describes a multifaceted intervention for promoting wellness among the elderly that includes a social, medical, and environmental assessment, followed by the provision of specific recommendations for individual behavior change and environmental modifications (e.g., installation of handrails in bathrooms to prevent falls).

Taken together, the chapters in this volume highlight the growing importance of developing more integrative and broader-gauged strategies for promoting human wellness. The principles of social ecology outlined here and elaborated on in subsequent chapters provide a valuable foundation for establishing more integrative and effective approaches to wellness promotion research, practice, and policy during the 21st century.



The author thanks Dr. Margaret Schneider Jamner for her helpful comments on an earlier version of the chapter.


Adler, N. E., Boyce, T., Chesney, M. A., et al. (1994). “Socioeconomic status and health: The challenge of the gradient.” American Psychologist49, 15–24.

Alihan, M. A. (1964). “Social ecology: A critical analysis.” New York: Cooper Square Publishers.

Allen, J., and Allen, R. F. (1986). “Achieving health promotion objectives through cultural change systems.” American Journal of Health Promotion1, 42–49.

Archea, J. C. (1985). “Environmental factors associated with stair accidents by the elderly.” Clinics in Geriatric Medicine1, 555–569.

Archea, J., and Connell, B. R. (1986). “Architecture as an instrument of public health: Mandating practice prior to the conduct of systematic inquiry.” In H. Wineman, R. Barnes, and C. Zimring, eds., Proceedings of the seventeenth annual conference of the Environmental Design Research Association.Pp. 305–309. Washington, D.C.: Environmental Design Research Association.

Atwood, K., Colditz, G. A., and Kawachi, I. (1997). “From public health science to prevention policy: Placing science in its social and political contexts.” American Journal of Public Health87, 1603–1606.

Barker, R. G. (1968). Ecological psychology: Concepts and methods for studying the environment of human behavior.Stanford, Calif.: Stanford University Press.

Bellingham, R. (1990). “Debunking the myth of individual health promotion.” In M. E. Scofield, ed., Occupational medicine: Worksite health promotion.665–675. Philadelphia: Hanley and Belfus.

Binder, A., Stokols, D., and Catalano, R. (1975). “Social ecology: An emerging multidiscipline.” Journal of Environmental Education7, 32–43.

Brennan, A. J. (1982). “Health promotion: What's in it for business and industry?” Health Education Quarterly9, 9–19.

Breslow, L. (1996). “Social ecological strategies for promoting healthy lifestyles.” American Journal of Health Promotion10, 253–257.

Breslow, L., and Johnsen, M. (1993). “California's Proposition 99 on tobacco and its impact.” Annual Review of Public Health14, 585–604.

Cannon, W. B. (1932). The wisdom of the body.New York: W. W. Norton.

Cassel, J. (1964). “Social science theory as a source of hypotheses in epidemio logical research.” American Journal of Public Health54, 1482–1487.

Catalano, R. (1979). Health, behavior, and the community: An ecological perspective.Elmsford, N.Y.: Pergamon Press.

DeJoy, D. M., and Wilson, M. G., eds (1995). Critical issues in worksite health promotion.Boston: Allyn and Bacon.


Detels, R., Holland, W. W., McEwen, J., and Omenn, G. S., eds. (1997). Oxford textbook of public health:Volumes 1, 2, 3. 3rd ed. New York: Oxford University Press.

Detels, R., and Breslow, L. (1997). “Current scope and concerns in public health.” In R. Detels, W. W. Holland, J. McEwen, and G. Omenn, eds., Oxford textbook of public health, Volume 1: The scope of public health. 3rd ed. Pp. 3–17. New York: Oxford University Press.

Durkheim, E. (1951). Suicide: A study in sociology.New York: The Free Press.

Eisenberg, D. M., Kessler, R. C., Foster, C., et al. (1993). “Unconventional medicine in the United States.” New England Journal of Medicine328, 246–252.

Emery, F. E., and Trist, E. L. (1972). Towards a social ecology: Contextual appreciations of the future in the present.London: Plenum Press.

Engel, G. L. (1976). “The need for a new medical model.” Science196, 129–136.

Erfurt, J. C., Foote, A., and Heirich, M. A. (1991). “The cost-effectiveness of worksite wellness programs for hypertension control, weight loss, and smoking cessation.” Journal of Occupational Medicine33, 962–970.

Evans, R. I. (1988). “Health promotion—Science or ideology?” Health Psychology7, 203–219.

Fielding, J. E., and Halfon, N. (1994). “Where is the health in health system reform?” Journal of the American Medical Association271, 1292–1296.

Fielding J. E., and Piserchia, P. V. (1989). “Frequency of worksite health promotion activities.” American Journal of Public Health73, 538–542.

Fisher, E. B., Jr. (1995). “Editorial: The results of the COMMIT Trial.” American Journal of Public Health85, 159–169.

Fontanarosa, P. (1995). “The unrelenting epidemic of violence in America: Truths and consequences.” Journal of the American Medical Association273, 01792–1793.

Fries, J. F., Harrington, H., Edwards, R., et al. (1994). “Randomized controlled trial of cost reductions from a health education program: The California Public Employees' Retirement System (PERS) Study.” American Journal of Health Promotion8, 216–223.

Fries, J. F., Koop, C. E., Beadle, C. E., et al. (1993). “Reducing health care costs by reducing the need and demand for medical services.” New England Journal of Medicine329, 321–325.

Geller, E. S. (1987). “Applied behavior analysis and environmental psychology: From strange bedfellows to a productive marriage.” In D. Stokols and I. Altman, eds., Handbook of environmental psychology:Volume 1. New York: John Wiley and Sons, 361–388.

Green, L. W., and Kreuter, M. W. (1990). “Health promotion as a public health strategy for the 1990s.” Annual Review of Public Health11, 319–334.

Green, L. W., and Ottoson, J. M. (1994). Community health. 7th ed. St. Louis: Mosby.

Greenberg, M. R. (1986). “Indoor air quality: Protecting public health through design, planning, and research.” Journal of Architectural and Planning Research3, 253–261.

Hawley, A. H. (1950). Human ecology: A theory of community structure.New York: Ronald Press.


Hedge, A. (1989). “Environmental conditions and health in offices.” International Review of Ergonomics2, 87–110.

Henderson, D. A., and Scutchfield, F. D. (1989). “Point-counterpoint: The public health versus medical model of prevention.” American Journal of Preventive Medicine5, 113–119.

Karasek, R., and Theorell, T., eds. (1990). Healthy work: Stress, productivity, and the reconstruction of working life.New York: Basic Books.

Katz, D., and Kahn, R. L. (1966). The social psychology of organizations.New York: John Wiley and Sons.

Kickbush, I. (1989). “Approaches to an ecological base for public health.” Health Promotion4, 265–268.

Lawrence, R. J., ed. (1990). “Housing, health, and well-being.” Special issue of the Journal of Sociology and Social Welfare17.

Lawton, M. P. (1999). “Environmental taxonomy: Generalizations from research with older adults.” In S. L. Friedman and T. D. Wachs, eds., Measuring environment across the lifespan.Pp. 91–124. Washington, D.C.: American Psychological Association.

Marlatt, G. A., and Gordon, J. R. (1985). Relapse prevention: A self-control strategy for the maintenance of behavior change.New York: Guilford Press.

Maruyama, M. (1963). “The second cybernetics: Deviation-amplifying mutual causal processes.” American Scientist51, 164–179.

McGinnis, J. M., and Lee, P. R. (1995). “Healthy People 2000 at mid decade.” Journal of the American Medical Association273, 1123–1129.

McKinlay, J. B. (1975). “A case for refocusing upstream: The political economy of illness.” In A. J. Enelow and J. B. Henderson, eds., Applying behavioral science to cardiovascular risk.Washington, D.C.: American Heart Association.

McLeroy, K. R., Bibeau, D., Steckler, A., and Glanz, K. (1988). “An ecological perspective on health promotion programs.” Health Education Quarterly15, 351–378.

Mendell, M. J., and Smith, A. H. (1990). “Consistent pattern of elevated symptoms in air-conditioned office buildings: A reanalysis of epidemiologic studies.” American Journal of Public Health80, 1193–1199.

Michelson, W. H. (1970). Man and his urban environment: A sociological approach.Reading, Mass.: Addison-Wesley.

Miller, J. G. (1978). The theory of living systems.New York: McGraw-Hill.

Montes, J. H., Eng, E., and Braithwaite, R. L. (1995). “A commentary on minority health as a paradigm shift in the United States.” American Journal of Health Promotion9, 247–250.

Moos, R. H. (1979). “Social ecological perspectives on health.” In G. C. Stone, F. Cohen, and N. E. Adler, eds,. Health psychology: A handbook.Pp. 523–547. San Francisco: Jossey-Bass.

Newsweek. (1998). Making tobacco say “Aaaaah.” March 16, 33.

O'Donnell, M. P., and Harris, J. S., eds. (1994). Health promotion in the workplace. 2nd ed. Albany, N.Y.: Delmar Publishers.

“Ottawa Charter for Health Promotion.” (1986). Health Promotion, 1, iii–iv.

Park, R., and Burgess, E., eds. (1925). The city.Chicago: University of Chicago Press


Pelletier, K. R. (1984). Healthy people in unhealthy places: Stress and fitness at work.New York: Dell Publishing.

Pelletier, K. R. (1994). Sound mind, sound body: A new model for lifelong health.New York: Simon and Schuster.

Pelletier, K. R. (1996). “A review and analysis of the health and cost-effective outcome studies of comprehensive health promotion and disease prevention programs at the worksite:” 1993–1995 update. American Journal of Health Promotion10, 380–388.

Pertschuk, M., and Shopland, D. (1989, September). Major local smoking ordinances in the United States. National Institutes of Health Publication 90479. Washington, D.C.: U.S. Government Printing Office.

Prochaska, J. O., DiClemente, C. C., and Norcross, J. C. (1992). In “search of how people change: Applications to addictive behaviors.” “American Psychologist” 47, 1102–1114.

Richmond, J. B., and Kotelchuck, M. (1991). “Coordination and development of strategies and policy for public health promotion in the United States.” In W. W. Holland, R. Detels, and G. Knox, eds., Oxford textbook of public health.Pp. 441–454. Oxford: Oxford Medical Publications.

Robertson, L. S. (1986). “Behavioral and environmental interventions for reducing motor vehicle trauma.” In L. Breslow, J. E. Fielding, and L. B. Lave, eds., Annual Review of Public Health7, 13–34. Palo Alto, Calif.: Annual Reviews Inc.

Rogers-Warren, A., and Warren, S. F., eds.. (1977). Ecological perspectives in behavior analysis.Baltimore: University Park Press.

Rosenstock, L., Olenac, C., and Wagner, G. R. (1998). “The National Occupational Research Agenda: A model of broad stakeholder input into priority setting.” American Journal of Public Health88, 353–356.

Sarason, S. B. (1974). The psychological sense of community.San Francisco: Jossey-Bass.

Satcher, D., and Hull, F. (1995). “The weight of an ounce.” Journal of the American Medical Association273, 1149–1150.

Schauffler, H. H. (1993). Health promotion and disease prevention in health care reform. Contract report to The California Wellness Foundation. Berkeley: School of Public Health, University of California.

Schwartz, G. E. (1982). “Testing the biopsychosocial model: The ultimate challenge facing behavioral medicine.” Journal of Consulting and Clinical Psychology50, 1041–1053.

Stokols, D. (1987). “Conceptual strategies of environmental psychology.” In D. Stokols and I. Altman, eds., Handbook of environmental psychology, Volume 1. New York: John Wiley and Sons, 41–70.

Stokols, D. (1996). “Translating social ecological theory into guidelines for community health promotion.” American Journal of Health Promotion10, 282–298.

Stokols, D., Pelletier, K. R., and Fielding, J. E. (1995). “Integration of medical care and worksite health promotion.” Journal of the American Medical Association273, 1136–1142.

Syme, S. L. (1990). “Health promotion: Old approaches, new choices, future imperatives.” Presented at the Conference on “The New Public Health: 1990,” Los Angeles, April.


Thompson, R. S., Taplin, S. H., McAfee, T. A., et al. (1995). “Primary and secondary prevention services in clinical practice: Twenty years' experience in development, implementation, and evaluation.” Journal of the American Medical Association273, 1130–1135.

“U.S. Department of Health and Human Services.” (1992). 1992 National Survey of Worksite Health Promotion Activities. Final Report PB93-500023. Washington, D.C.: U.S. Government Printing Office.

“U.S. Food and Drug Administration.” (1996). Executive summary: The regulations restricting the sale and distribution of cigarettes and smokeless tobacco to protect children and adolescents.

“U.S. Preventive Service Task Force.” (1989). Guide to clinical preventive services: An assessment of the effectiveness of 169 interventions.Baltimore: Williams and Wilkins.

“U.S. Public Health Service.” (1964). Smoking and health: Report of the Advisory Committee to the Surgeon General of the Public Health Service. PHS Publication 1103. Washington, D.C.: U.S. Government Printing Office.

“U.S. Public Health Service.” (1967). The health consequences of smoking: A Public Health Service review. PHS Publication 1696. Washington, D.C.: U.S. Government Printing Office.

“U.S. Public Health Service.” (1979). Healthy people: The Surgeon General's report on health promotion and disease prevention. DHEW Publication (PHS) 79-55071. Washington, D.C.: U.S. Government Printing Office.

“U.S. Public Health Service.” (1991). Healthy People 2000: National health promotion and disease prevention objectives. PHHS Publication (PHS) 91-50212. Washington, D.C.: U.S. Government Printing Office.

“U.S. Public Health Service.” (1992). Business responds to AIDS.Washington, D.C.: U.S. Government Printing Office.

Webster's Encyclopedic Unabridged Dictionary of the English Language. (1989). New York: Portland House.

Weick, K. E. (1984). “Small wins: Redefining the scale of social problems.” American Psychologist39, 40–49.

Weiss, S. M. (1991). “Health at work.” In S. M. Weiss, J. E. Fielding, and A. Baum, eds., Perspectives in behavioral medicine: Health at work.Pp. 1–10. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Wells M., Stokols D., McMahan S., and Clitheroe, C. (1997). “Evaluation of a worksite injury and illness prevention program: Do the effects of the REACH OUT Training Program reach the employees?” Journal of Occupational Health Psychology2, 25–34.

Williams, A. F. (1982). “Passive and active measures for controlling disease and injury: The role of health psychologists.” Health Psychology1, 399–409.

Williams, A. F., Karpf, R. S., and Zador, P. F. (1983). “Variations in minimum licensing age and fatal motor vehicle crashes.” American Journal of Public Health73, 1401–1403.

Winett, R. A., King, A. C., and Altman, D. G. (1989). Health psychology and public health: An integrative approach.New York: Pergamon Press.


Winkleby, M. A. (1994). “The future of community-based cardiovascular disease intervention studies.” American Journal of Public Health84, 1369–1372.

Wintemute, G. J. (1992). “From research to public policy: The prevention of motor vehicle injuries, childhood drownings, and firearm violence.” American Journal of Health Promotion6, 451–464.

World Health Organization. (1984). “Health promotion: A discussion document on the concept and principles.” Health Promotion1, 73–76.



Lester Breslow

From time immemorial, people have been concerned, collectively as well as individually, about avoiding health impairment and prolonging life. In society they have developed two increasingly specialized arms for advancing their health interests: (1) sets of healers, now long dominated by physicians, for dealing with individuals' problems, and (2) community health protection, to which public health personnel are devoted. The first set focuses on people one at a time and the second on the population as a whole. Physicians have tended to emphasize cures, whereas public health has focused attention on prevention, though considerable overlap in these emphases has occurred. Both are supported by society to varying degrees in different times and situations, generally corresponding to the value that society places on them for safeguarding health.


Most health problems arise in a societal context. Thus, tuberculosis, influenza, pneumonia, and other respiratory-borne and acute intestinal diseases dominated the health scene during the early days of the industrial revolution and even into the 20th century. Exhausting factory work, crowded housing and work space, polluted water, and inadequate nutrition largely determined that disease pattern. The latter has steadily been yielding both to social reform of poor living conditions and to social

implementation of scientific-technical advances in microbiology and immunity.


Progress against the communicable diseases has led to longer lives. Steady improvement in technology and living conditions, however, created opportunity for increasing segments of the population—at first for the more affluent and subsequently most other people—to luxuriate in ways that induced other kinds of health damage. Physical inactivity; excessive calorie consumption, especially in the form of fats and sweets; and access to new forms of tobacco eventuated in the 20th-century epidemics of cardiovascular disease, lung cancer, diabetes, and other chronic diseases.

In the struggle against this new set of health problems, treatment dominated in the beginning, but prevention is now emerging as the major approach. The 1957 report of the Commission on Chronic Illness, which brought together the American Medical Association, American Public Health Association, American Hospital Association and American Public Welfare Association, introduced and delineated the key concepts for chronic disease prevention: primary and secondary prevention.1 The former was defined as “averting the occurrence of disease” and the second as “halting the progression of disease from its early unrecognized stage to a more severe one and preventing complications or sequelae of disease.”2 Continued development of knowledge concerning the causation of chronic disease, especially behavioral risk factors, has enhanced prospects for primary prevention. Similarly, continuing technological advances in screening for early signs of disease have led to improvements in secondary prevention. As the 20th century comes to a close, social support for primary and secondary prevention of chronic disease is growing rapidly.


Meanwhile, something beyond prevention of disease as a health goal is gaining recognition. The World Health Organization (WHO) has defined health as “a state of complete physical, mental and social wellbeing, not merely the absence of disease and infirmity.” 3 Though denigrated for years as wooly-headed and not useful in scientific work, the

concept recently appears to be achieving greater credibility. Documents such as the Lalonde Report 4 and the Ottawa Charter for Health Promotion,5 as well as the growing use of the term “health promotion” (often in connection with disease prevention), indicate a fundamental change in thinking about health.4, 5 That change reflects socially determined progress against disease to the point that it is possible to envisage a higher state of health than the absence of disease.

Gradual acceptance of that idea appears due to continuing advances not only against the communicable diseases but also against the noncommunicable chronic diseases. The prescience of the first WHO leaders is shown in the fact that mortality from major forms of chronic disease such as heart disease and cancer was still rising at the mid-century point; since then, however, both have passed their epidemic peaks and are declining, thus opening the path to a new view of health.

Overcoming the chronic diseases is resulting both from their declining incidence and from lowering their fatality. Social support of primary and secondary prevention has played a substantial role in that development.


An important consequence of delaying mortality among those afflicted with chronic disease has been the extension of life expectancy within the population as a whole. Aging of the population, with its accompanying accumulation of multiple pathologies from life's stresses, is stimulating concern about functional impairment, with its social as well as individual implications.

One sign of that concern may be seen in efforts to measure health impairments. One of the first to emerge was the set of activities of daily living (ADLs) proposed by Katz, Downs, Cash, and Grotz in 1970.6 They listed six activities necessary for living, such as eating and toileting, and began measuring the extent to which individuals could perform these functions without aid from others. This process permitted delineating the need for care, with treatment as the primary aim. Since the ADLs earmarked only those persons requiring extensive care, a next step was to define instrumental activities of daily living (IADLs).7 These included such items as ability to perform household chores or to go shopping. Subsequently, advanced activities of daily living (AADLs) have been proposed 8 and include such behaviors as participation in regular exercise.

Noteworthy in this evolution of ADLs, IADLs, and AADLs has been

the shift in focus from disease to impairments. That shift reflects a growing social concern in the health arena with loss of capacities for living, not merely with the traditional “causes” of disease and death. The focus on these functional criteria for health status also reflects the greater numbers of people living with chronic disease and their associated disabilities due to life-extending treatments and medications.


Wellness as a health goal is still a relatively crude concept. It has not yet become sufficiently explicit and standardized to permit clear delineation in a person or measurement in a population. In dealing with health, we are still strongly inclined to diagnose and count diseases and to determine the number of ADLs, IADLs, and AADLs that people can perform.

The term “wellness promotion” implies a step beyond prevention of disease and impairment. All the items that might be listed under the rubric of wellness apply to relative states of health rather than to gradations of disease or impairment. Promotion of wellness may similarly be viewed as actions to achieve higher levels of health above and beyond the reduction of impairment.

Disease and impairment have been quite well delineated; wellness has not. It is therefore necessary to make wellness as a state of health more explicit in order to advance understanding of how to promote it. This constitutes an important, current challenge to health science.9


Despite the elementary state of the wellness concept, it already appears that many of the same things that must be attacked in order to prevent disease also need attention in achieving wellness. For example, avoiding smoking both prevents lung disease and enhances pulmonary function; physical exercise both prevents cardiac disease and strengthens musculoskeletal function. Thus, in many respects disease prevention and wellness promotion constitute the two aspects of a positive health strategy.

Several levels of action are desirable in this strategy: individual, interpersonal, and social. In behavior affecting one's own health (and that is an increasingly important factor), the individual makes the final decision; he/she ultimately determines what and how much to eat, how much alcohol

to drink, and the like. Because health determinants nowadays consist largely of access rather than exposure to risk factors, opportunity and inducements for individual health-related choice must be the intervention end target. One does not exercise such choice in a vacuum. Rather, that is done in the context of family, friends, peers, professional advisers, and the whole social milieu. Therefore, simple appeals to individuals may not be very effective since they constitute only one influence on the choice to be made; also, they may easily take the form of “blaming the victim.” Interpersonal intervention, on the other hand—for example, by peers or physicians—may carry considerable weight with the individual because of the substantial tendency to follow such leads.

The third, and probably most effective, level of intervention for behavior that protects health is the social environment. Since behavior adverse to health arises so largely from societal influences—for example, inducing youngsters to smoke cigarettes, tolerating excessive use of alcoholic beverages, and failing to provide adequate opportunity for physical exercise in the inner city—such influences must be dealt with on the social level. In an earlier day, health advance necessitated creating a social milieu that favored hand washing and not spitting; now it has become necessary to establish new social hygiene standards: not smoking, drinking modestly if at all, exercising, and eating sensibly.

Exemplifying the societal approach to disease prevention and wellness promotion, and the struggles involved, is the California tobacco control program.10 Activists against the number one cause of death in the United States—in major voluntary health agencies, public health and medicine, and elsewhere—had not prevailed in the legislature over tobacco industry interests. Therefore, taking advantage of the state's legislative initiative process, they succeeded in getting Proposition 99 passed in a general election. The latter increased the tax on cigarettes 25 cents a pack and allocated 20 percent of the funds to prevention efforts, the rest to medical, hospital, research and other purposes. The broad-scale program implemented with the prevention funds by the California Department of Health Services stimulates tobacco control activity in local health departments, schools, and a wide variety of community agencies throughout the state. These local efforts, and subsequent state action, stopped smoking in public buildings, workplaces, and restaurants; stepped up enforcement of laws forbidding tobacco sales to minors; removed vending machines from many places; and otherwise severely discouraged smoking. Vigorous communication via television, radio, and other media did likewise. Schools introduced materials and in other

ways encouraged students to combat pro-smoking influences. The social milieu in the whole country was, of course, tilting strongly against smoking. The program's specific effectiveness, however, is indicated by the fact that the rate of smoking decline in California doubled through the next six years, bringing the prevalence down below all other states except Utah.

The tobacco industry fought back intensively with political contributions and lobbying, formation of phony “public” organizations to oppose local control activities, and heavy promotion of products. The governor and the legislature then joined in substantially disrupting the program by diverting prevention monies to medical services, until the antitobacco forces obtained court action to stop such diversion of funds. Now efforts are under way to restore the program.

The California tobacco control experience has indicated how sharply social trends and conflicts can affect health-related behavior and thus disease prevention and wellness promotion. Other elements of such efforts are likewise embedded in social context, which must be considered in order to achieve effective action.


1. “Commission on Chronic Illness.” 1957. Chronic Illness in the United States.Cambridge, Mass.: Harvard University Press.

2. “Commission on Chronic Illness.” 1957. Chronic Illness in the United States. Volume I: Prevention of Chronic Illness.Cambridge, Mass.: Harvard University Press.

3. “World Health Organization.” 1984. Health promotion: A discussion document on the concept and principles. Health Promotion1, 73–76.

4. “Government of Canada.” 1974. A new perspective on the health of Canadians (Lalonde Report). Department of National Health and Welfare, Ottawa.

5. “Ottawa Charter for Health Promotion.” 1986. Canadian Journal of Health Promotion77, 425–436.

6. Katz, S., Downs, T. D., Cash, H. R., and Grotz, R. C.1970. “Progress in development of the index of ADL.” Gerontologist10, 20–30.

7. Avlund, K., Schultz-Larsen, K., and Kreiber, S.1993. “The measurement of instrumental ADL: Content validity and construct validity.” Aging5, 371–383.

8. Reuben, D. B., Laliberte, L., Hiris, J., and Mor, V.1990. “A hierarchical exercise scale to measure function at the Advanced Activities of Daily Living (AADL) level.” Journal of the American Geriatrics Society38, 855–861.

9. Breslow, L.1999. “From disease prevention to health promotion.” Journal of the American Medical Association281, 1030–1033.

10. Breslow, L., and Johnsen, M.1993. “California's Proposition 99 on tobacco and its impact.” Annual Review of Public Health14, 585–604.



Biomedical versus Outcomes Models

Robert M. Kaplan

This chapter grew out of my experience of being named as the Wellness Lecturer at the University of California, San Diego, in 1991. It is remarkable to consider what changes have occurred over these last few years. Yet, it is also remarkable that we still face many of the same problems. In 1991, managed care had captured only a small percentage of the health care market in the United States. Today, managed care is a dominant force in most areas of the country. In 1991, few observers challenged the autonomy of physicians as sole decision makers. Today, practice guidelines for the individual health care providers are an accepted part of practice. Finally, the use of public health and preventive health care approaches, although recognized as important in 1991, were not common parts of practice. Today, the reorganization of health care has provided new opportunities for incorporation of prevention paradigms. These changes have stimulated rethinking of the basic foundations of health care organization and delivery, and this chapter addresses some of the related issues. We will begin by considering the framework proposed in 1991.


The major problems in the American health care system might be described as the three A's: affordability, access, and accountability (Kaplan, 1993). Similar problems were identified by Relman (1989).



The affordability problem results from our inability to pay for all the health care that is desired. Health care costs in the United States grew remarkably between 1940 and the mid-1990s. In 1940, approximately four billion dollars per year were spent on health care in the United States. That amount tripled by 1950 and continued to escalate at an exponential rate through the early 1990s. Health expenditures were over a trillion dollars in 1996. It now takes just over a day to match the yearly expenditure from 1940.

The high costs of American health care cause very serious problems for U.S. products in the world marketplace. This is because the costs of health care are represented in every product that the United States exports. Since our per capita expenditures on health care are twice what they are in countries such as Great Britain or Japan, health care costs contribute proportionally more to the expense of American exports. We pay for expensive health care in many different ways. In some cases patients pay more for services. However, we also pay for high health care costs when we purchase consumer products. Part of the price of each product is the health insurance paid on behalf of the workers. More important, workers are doubly affected by increased health care costs. When their employers pay more for health insurance, workers get lower wages and retirement benefits (Center for Health Economics Research, 1994). If spending on health benefits rises, other aspects of compensation may be held constant or may decline.


The United States remains the only industrialized country that does not provide universal access to health care. Part of the problem is that health care is usually unaffordable without insurance. The exact number of uninsured people is difficult to determine. Current estimates suggest that over 40 million Americans have no regular source of health care. The only group that has universal coverage is the elderly since virtually all Americans older than age 65 are covered by the Medicare program. The uninsured are not necessarily the unemployed. In fact, the majority of those without health insurance are working or dependents of workers. However, many employers provide either no health insurance or inadequate coverage.



Health care may be the only major American industry that is not held accountable for what it produces. Although we produce more health care services per capita than any country in the world, it is not clear that Americans are in any way healthier than residents of other developed countries, where considerably less is spent on health care. Patient satisfaction surveys suggest that consumers are significantly more satisfied with health care services in countries that spend considerably less. One analysis compared satisfaction and expenditures in 10 countries (the United States, Canada, France, Germany, Sweden, Australia, the Netherlands, Italy, Japan, and the United Kingdom). Among these countries, the United States spends more per capita and is significantly lower in the percentage of consumers satisfied with the services they receive (Blendon et al., 1990).

In summary, the U.S. health care system is in serious trouble. Solutions to these problems require that we consider all three dimensions. In addition, we must challenge some of the most basic models of health care. The accountability piece of the puzzle is perhaps the most challenging. In order to address accountability, we must address central ideas about the purpose of health care. Most of this chapter reviews methods for accounting for health care benefits.


Health care has been dominated by a traditional biomedical model. According to this model, human pain and suffering are caused by disease processes. Disease activity is measured by judgments of trained physicians and by physiological measures, including blood chemistry or radiographic evidence of pathology. The traditional medical model recognizes behavioral factors as predictors of these outcomes. Behavioral risk factors might be cigarette smoking, high-risk behaviors, or the consumption of a high-fat diet (Kaplan, 1984). In addition, the traditional biomedical model suggests that disease process is determined by genetic predispositions, environmental exposures, and the aging process itself. The disease process is also affected by medical care and the use of regular medical tests (Wilson and Cleary, 1995).

According to the traditional biomedical model, the purpose of medicine is to find disease pathology and to fix it. We sometimes refer to this

as “find it–fix it medicine.” For problems such as high blood pressure, for example, the physician's task is to diagnose the problem and to administer a medicine that will make blood pressure normal. The measure of success is a blood pressure reading that falls within a defined range of normality. Unfortunately, many medical procedures may affect biological processes, but may not affect life expectancy or life quality. It has been estimated that 30% to 50% of all medical procedures have little effect on long-term outcomes (Brook and Lohr, 1987). Further, some procedures may have a negative effect on survival and quality of life.

An alternative model for health care, known as the outcomes model, is similar to the traditional biomedical model. However, the ultimate outcome is not a measure of disease process. The goals of health care are to extend the duration of life and/or to improve the quality of life. Disease processes are of interest because pathology may either shorten life expectancy or make life less desirable. The same variables that predict disease process may also predict life expectancy or quality of life. However, in contrast to the traditional biomedical model, behaviors or biological events may affect life expectancy independently of disease process. Further, the measures of success in the outcomes model are different than those in the traditional biomedical model. The outcomes model emphasizes quality of life and life duration instead of clinical measures of disease process. As similar as these two models appear, they lead to substantially different approaches to organizing, financing, and delivering health care (Kaplan, 1990). These distinctions are addressed in the following sections.

Valuing Health Services

The traditional biomedical model uses procedures to fix biological problems. The greater use of procedures in the United States than in other countries resulted in the American system becoming more expensive than the systems in other countries. By 1990, it was clear that cost control would dominate the health policy agenda throughout the decade. Often, cost reduction is considered the major objective of health care reform. Pauly (1995), for example, argues that cost should be the central consideration in policy analysis. However, too much attention to cost may neglect the primary mission of health care. For example, if cost is the only criterion, the development of guidelines for appropriate care may exclude expensive services. In order to choose between alternative health programs, it is best to evaluate not only the costs but also the

benefits (Sturm and Wells, 1995). Such an evaluation recognizes all financial and health outcomes as either a cost or a benefit. Financial outcomes are easily understood, but clinical outcome measures are often poorly understood, especially from a patient perspective. For example, a change in an arterial blood gas value is not an ideal health outcome measure because it may not mean much to a patient or to a public policy maker. On the other hand, restoration or preservation of the ability to perform activities of daily living is the goal of many therapies. Because patient-centered outcomes are measurable and meaningful, a paradigm shift in medicine is beginning to embrace patient-centered reports.

Despite the improvements in measuring patient outcomes, determining the value of health services has been particularly difficult. In contrast to cost-benefit analysis, which focuses on the dollar returns for investing in particular programs, or consumer willingness to pay for service, cost-utility analysis used in some health services evaluations considers the health outcomes of a program, weighted by patient preferences for outcomes, in relation to the financial costs of the program. In economics, the value of a product is related to the willingness of consumers to pay for it. For example, the value of a Mercedes-Benz automobile is set by the price that consumers are willing to pay for the car. If the price is too high, few cars are sold. Health services are difficult to value in this manner because consumers rarely pay for them directly. Instead, the charges are paid by third parties. Third-party payment leaves consumers out of the loop and makes it difficult to establish whether the services are valuable to patients.


In order to understand health outcomes, it is necessary to build a comprehensive theoretical model of health status. The major aspects of the model include mortality (death) and morbidity (health-related quality of life). In several papers, we have suggested that diseases and disabilities are important for two reasons. First, illness may cause the life expectancy to be shortened. Second, illness may make life less desirable at times prior to death (diminished health-related quality of life) (Kaplan and Anderson, 1988; Kaplan et al., 1993).

Over the last two decades, a group of investigators at the University of California, San Diego, has developed the General Health Policy Model (GHPM). Central to the model is a general conceptualization of

quality of life. The model separates aspects of health status and life quality into distinct components. These are life expectancy (mortality), functioning and symptoms (morbidity), preference for observed health states (utility), and duration of stay in health states (prognosis). Each component is described here in more detail.

Mortality. A model of health outcomes necessarily includes a component for mortality. Indeed, many public health statistics focus exclusively on mortality through estimations of crude mortality rates, age-adjusted mortality rates, and infant mortality rates.

Morbidity. Health-related quality of life is also an important outcome. Most public health indicators are relatively insensitive to variations toward the well end of the death-wellness continuum. Measures of mortality, to give an extreme example, ignore all variations of morbidity: a person in a coma is considered equivalent to an asymptomatic person at full function. Both, after all, are alive. In addition, disability measures often ignore those who are relatively healthy. For example, the RAND Health Insurance Study reported that about 80% of the general population have no dysfunction. Thus, they would estimate that 80% of the population is well. But in studies that assess symptoms and function, only about 12% of the general population report no symptoms on a particular day (Kaplan et al., 1976). In other words, health symptoms or problems are a very common aspect of the human experience. Some might argue that symptoms are unimportant because they are subjective and unobservable. However, symptoms are highly correlated with the demand for medical services, expenditures on health care, and motivations to alter lifestyles. Further symptoms lower quality of life even if a disease cannot be detected. Thus, we feel that the quantification of symptoms is very important when assessing morbidity. The GHPM, using the Quality of Well-Being (QWB) scale described later, considers functioning in three areas (mobility, physical activity, and social activity) and symptoms.

Utility (Relative Importance). Not all outcomes are equally important. For example, a treatment that prevents nausea is not equivalent to one that prevents death. Given that mortality and the various components of morbidity can be tabulated, it is important to consider their relative importance. A key component of the GHPM attempts to scale the various health outcomes according to their relative importance. This exercise adds the “quality” dimensions

to health status. In the preceding example, the relative importance of dying would be weighted more than developing nausea. The weighting is accomplished by rating all states on a quality continuum ranging from 0 (for dead) to 1.0 (for optimum, asymptomatic functioning). These ratings are typically provided by independent judges who are representative of the general population. Using this system it is possible to express the relative importance of states in relation to the life-death continuum. A point halfway on the scale (0.5) is regarded as halfway between optimum function and death. The quality-of-life weighting system for the QWB has been described in several different publications (Kaplan et al., 1976, 1978, 1979).

Prognosis. In the GHPM, the term “prognosis” refers to the probability of transition among health states over the course of time and includes consideration of duration of problems. A headache that lasts one hour is not equivalent to a headache that lasts one month. A cough that lasts three days is not equivalent to a cough that lasts three years. In considering the severity of illness, duration of the problem is central. As basic as this concept is, most contemporary models of health outcome measurement completely disregard the duration component. The GHPM considers the point at which the problem begins. A person may have no symptoms or dysfunctions currently but may have a high probability of health problems in the future. The prognosis component of the model takes these transitions into consideration. A discount rate is used for future outcomes if the utility of a future outcome is not the same as that of a present outcome. For example, a daylong headache that will begin a year from now may be less of a concern than a daylong headache that will start immediately.

The components of the model can be integrated to express outcomes in terms of quality-adjusted life years (QALYs). A QALY is defined as the equivalent of a completely well year of life, or a year of life free of any symptoms, problems, or health-related disabilities. A principal advantage of the QALY is that it provides a common metric that allows different programs to be directly compared. The quality-adjusted life expectancy is the current life expectancy adjusted for diminished quality of life associated with dysfunctional states and the duration of stay in each state.

Consider, for example, a person with a rare lung disease and in a state of functioning and symptoms that is rated by community peers as 0.5 on

the 0-to-1.0 utility scale described previously. If the person remains in that state for one year, he or she would have lost the equivalent of onehalf of one year of life. However, a person who has the flu may also be rated as 0.50. In this case, the illness might last only three days, and the total loss in QALYs might be 3/365 0.50, which is equal to 0.004 QALYs. By itself, it is clear that the flu does not produce as significant a health outcome as the lung disease. But suppose that 5,000 people in a community get the flu. The QALYs lost would then be 5,000 0.004, which is equal to 20 years, or greater than the one person with the rare lung disease. This indicates that the flu may be a greater health policy problem than the rare disease.

Now suppose that a vaccination becomes available and that the threat of the flu can be eliminated by vaccinating the 35,000 people in the community. The cost of the vaccine is $5 per person, or $175,000. The average cost/utility of the program would be as follows:


Using the concept of the QALY, the net cost/utility ratio of two alternative programs can be calculated as follows:


It is important to consider the marginal, or incremental, cost/utility of programs. Although we could compare the simple costs and effects of different programs, we most often have to decide how much we are willing to pay in order to add an additional benefit. For example, behavior modification has been shown to be valuable for helping people kick the smoking habit. However, the benefits might be enhanced if nicotine replacement therapy is added to behavior modification. Analysis might show that both behavior modification and behavior modification plus nicotine replacement reduce tobacco use. However, we must decide if we are willing to pay the extra expense for adding nicotine replacement. Usually, we must choose between several programs.

Another way to evaluate outcomes is within “policy space.” Various approaches to cost/benefit and cost/utility analysis occasionally produce different results. The output for cost/benefit analysis is in monetary terms—a program that produces cost savings. Cost/utility analysis


Figure 3.1. Two-dimensional policy space from Anderson et al. (1986). Tobacco excise tax may be one of the few policy examples that is in the upperright quadrant.

focuses on the cost to produce a QALY. Anderson et al. (1986) integrated the concepts of QALYs and net dollars returned within a common framework. This was accomplished by creating a two-dimensional policy space as illustrated in Figure 3.1. The x -axis in the figure represents net dollars returned per person. Returns are defined as benefits minus costs in dollar units. The y -axis displays well years lost or gained through a particular treatment program, clinical intervention, or policy change.

The right half of the plane would be used to represent programs in which benefits exceed costs, while the left half would display situations in which costs exceed benefits. The upper half of the figure displays outcomes that have positive health effects in terms of QALYs. Those in the bottom half of the figure would be used to represent negative health outcomes in well-year units.

The two-dimensional space yields four quadrants. One quadrant, the lower left, represents unsuitable alternatives. In these cases, dollars are being spent, and negative health consequences occur. Administration of a uniformly toxic treatment might be represented by this quadrant. The upper-right quadrant represents the most attractive alternatives. Here, QALY health benefits are gained, and there are also economic benefits. Increasing tobacco excise tax may be one of the few true examples of such a program. The upper-left quadrant shows QALY gains, but with more significant costs associated with these improvements. Transplantation surgery for the elderly might be described by this quadrant. Here

there are significant health benefits, but the recipients may not return to the productive economic sector.

The lower-right quadrant represents another level of economic tradeoff. Here, society may be willing to sacrifice some health benefits in exchange for cost savings. Anderson and colleagues suggested that these trade-offs may be common in studies involving nuclear power; pollution control; occupational, environmental, and consumer product safety; highway speed limits; and so on.


Although the traditional biomedical model and the outcomes model are similar in many ways, they lead to different decisions about the use of resources for prevention. In the following sections, several examples are reviewed.

Diagnosis versus Outcomes

The traditional biomedical model is centered on medical diagnoses. Diagnosis defines the problems that have been found and gives direction about what needs to be fixed. The traditional system pays providers for using diagnostic tests to find problems and for using therapeutic interventions to fix the problems. Despite the importance of diagnosis, it often obscures or confuses the importance of some health problems. There are at least three reasons why focusing on diagnosis may have led us in some wrong directions. First, diagnoses do not always lead to better health outcomes. Often, people are placed in categories, but identification of a condition does not necessarily mean that an effective treatment can be applied. Second, diagnoses are not always correct, and in some cases, individuals will be treated for conditions they do not have or will fail to be effectively treated because the correct diagnosis was overlooked. Third, in many cases, poor health outcomes result from risky behavior or from exposure to risk factors. Community health outcomes can be enhanced by removing the risk factor or by modifying behavior. The identification of a disease on the pathway between the risk factor and the outcome is interesting but not essential.

There are approximately 2,150,000 deaths in the United States each

year. Deaths are tallied according to major and underlying cause. The traditional biomedical model emphasized disease-specific causes of death, and therefore pathways to prevention typically considered risk factors for particular diseases. For example, cigarette smoking is associated with deaths from cancer of the lung. Thus, efforts to reduce lung cancer concentrate on smoking cessation. However, most of the major causes of death are associated with a variety of different risk factors. Further, many risk factors are associated with death from a variety of different causes. For example, tobacco use causes not only lung cancer but also a wide variety of other malignancies, heart disease, stroke, and birth complications (Kaplan et al., 1995). By concentrating on diagnoses, the traditional model often misses the relationship between behaviors and outcomes.

Major nongenetic contributors to mortality were examined in an important analysis by McGinnis and Foege (1993). When these external factors are considered independent of the disease model, clear priorities for prevention emerge. A summary of the estimates for actual causes of death in the United States is presented in Table 3.1. Tobacco use is associated with more than 400,000 deaths each year, while diet and activity patterns account for an additional 300,000. These dwarf the number of deaths associated with problems that the public is generally concerned about, such as illicit drug use. The McGinnis and Foege analysis challenged us to think differently about the way we track health indicators in the United States. Only a small fraction of the trillion dollars the United States spends annually on health care is devoted to the control of the major factors that cause premature mortality in the United States. Estimates suggest that less than 5% of the total annual health care budget is devoted to prevention efforts (Rothenberg et al., 1987). If the focus of attention shifts from finding and fixing diseases to producing QALYs, it becomes clear that preventive efforts to reduce tobacco, drug, and alcohol use and to promote exercise deserve greater attention.

It is commonly argued that traditional fee-for-service medicine provides few incentives to offer preventive services. Indeed, the higher the rates of service utilization, the greater the revenue. One attractive feature of managed care is that there are substantial incentives to prevent illness and to reduce health care utilization. From a public health perspective, managed care organizations have responsibility for a defined population. If they can keep this population healthy by investing in prevention, they may ultimately profit by having reduced costs and higher consumer satisfaction.

Factor Deaths Percentage
SOURCE: McGinnis and Foege (1993).
Tobacco 400,000 19
Diet-activity patterns 300,000 14
Alcohol 100,000 5
Microbial agents 90,000 4
Toxic agents 60,000 3
Firearms 35,000 <2
Sexual behavior 30,000 1
Motor vehicles 25,000 1
Illicit use of drugs 20,000 <1
Total 1,060,000 50

The traditional biomedical model has dominated thinking about prevention. Most of the 3% to 5% of the health care dollar used for prevention is devoted to clinical preventive services offered by physicians. For example, the great majority of expenditures on prevention relate to screening for diseases, such as breast cancer, cervical cancer, and prostate cancer. The purpose of the prevention service is to detect a disease that already exists and intervene medically so that progression is retarded. The screening tests have become profitable for the providers who offer them, and there is growing concern about abuses or profiteering by those who administer tests to people who do not need them.

Guidelines for preventive services might limit reimbursement to those services recommended in the Guide to Clinical Preventive Services. The U.S. Preventive Services Taskforce has prepared this excellent document that describes the appropriate use of a wide variety of medical preventive services. The strong emphasis in the Guide is on the appropriate use of medical screening tests by physicians. The latest version of the guide includes 70 chapters. There are 11 chapters devoted specifically to counseling (tobacco, physical activity, diet, motor vehicle injuries, household and recreational injuries, youth violence, low-back pain, dental/periodontal disease, HIV and STD, unintended pregnancy, and gynecological cancers). However, most of the guidelines advise physicians on testing and the medical implications for those who have been found to have a disease. For example, 47 of the 60 sets of guidelines in the first edition of the Guide concern the application of medical screening tests. Another five consider immunizations and chemoprophylaxis. The remaining eight guidelines are on counseling.


The guidelines for preventive services arise from the traditional find it–fix it model. The emphasis is on using medical screening tests that identify diseases that already exist. These are the services that have traditionally generated fees for providers. The limited yield from this approach is reviewed in the next section. The current environment favors an outcomes-oriented approach. According to the outcomes model, diagnosis and treatment are important only if they make life longer and/or better.

Medical Geography

The traditional biomedical model of health care rests on several important assumptions. One assumption is that medical technology enhances health outcomes and that there is a pool of untreated disease on which these technologies might be successfully deployed. If this is true, then more applications of medical technologies should result in better health outcomes. A second assumption is that any well-trained doctor presented with the same problem will come to the same diagnosis. However, it has been known for some time that there is substantial variation in the rate at which problems are diagnosed and treated in different communities. Thus, even though the distribution of a disease may be the same in different communities, the rate at which diseases are diagnosed varies substantially. Wennberg and colleagues have devoted the past quarter of a century to a description of this problem (Wennberg, 1996).

Wennberg argues that a major factor in the use of medical services is “supplier-induced demand.” This implies that providers create demand for services by diagnosing illness. Most surgical subspecialists would agree there is a need to perform surgery on some well-defined cases. These might include amputation of a toe with gangrene, removal of some well-defined tumors, or intervention to repair a compound fracture. However, there is also substantial discretion in the use of some medical and surgical procedures. This is well illustrated by the comparison of procedures in two communities: Boston, Massachusetts, and New Haven, Connecticut.

Boston and New Haven are similar in a variety of ways. Both are traditional New England cities that have multiethnic populations. The two cities have approximately the same climate, and both cities are home to prestigious Ivy League universities. Since the cities are near each other, we would expect that the costs of medical care should be approximately the same. Using data from the mid-1970s, Wennberg and

colleagues (1990) demonstrated that, in fact, medical care costs in Boston were nearly twice as high as they were in New Haven.

Figure 3.2 shows the distribution of costs in Connecticut service areas and in Massachusetts services areas in the 1970s. In 1975, Medicare was paying $324 per recipient per month in Boston and only $155 per month for residents of New Haven. The situation has not changed much. In 1989, per capita hospital expenditures for acute care were $1,524 for residents of Boston and $777 for those living in New Haven.

Further study by Wennberg and his colleagues showed that Boston has more hospital capacity than does New Haven. In Boston, there are

4.3 hospital beds for every 1,000 residents, while in New Haven, there are fewer than 2.3 beds per 1,000 residents. Residents of Boston are more likely to be hospitalized for a wide variety of acute medical conditions than are residents of New Haven. For many different medical conditions, such as pneumonia or congestive heart failure, Bostonians are more likely to be cared for as hospital inpatients, while residents of New Haven are treated outside the hospital.

Some of the differences between Boston and New Haven might be attributed to the greater development of hospital facilities. New Haven has only one major medical school (Yale), while Boston has three medical schools. The Harvard Medical School is associated with several different teaching hospitals. Further, Boston has four hospitals associated with different religious establishments, while there is only one religious affiliated hospital in New Haven.

The Boston–New Haven comparison is particularly interesting from a public policy perspective. Medicare is a federal program that hopes to provide equal benefit to all its recipients. Yet, on average, Medicare spends twice as much per capita in Boston as it does in New Haven (Wennberg et al., 1996) for the same care. Are New Haven residents getting a bad deal? Since the government is spending less on New Haven residents, it might be argued that their health should suffer. However, outcomes evidence does not show that residents of Boston are any healthier than residents of New Haven. In fact, some evidence implies that Boston residents may be worse off. For example, people in Boston are more likely to be rehospitalized for the same condition than are people in New Haven. Residents of Boston appear to have more complications from medical treatment. More may not necessarily be better. Indeed, there is some evidence that more may be worse (Fisher et al., 1994). In the next section, we consider this question. Under a traditional model that focuses on diagnosis, more facilities will lead to more diagnoses.


Figure 3.2. Distribution of costs in Connecticut service areas and in Massachusetts service areas in the 1970s. Each dot represents one of the 11 most populated market areas in Connecticut or Massachusetts. Per capita expenditures for hospitals are generally lower in Connecticut, but there is a twofold range of variation. The circled dots represent the New Haven and Boston markets, where the majority of hospitalizations occur in teaching hospitals. Adapted from Wennberg (1990).


The outcomes model regards diagnosis as important only if it leads to patient betterment.

A War on Cancer

Numerous studies raise questions about the association between volume of medical care and community health status. Two recent examples come from studies on the treatment of cancer and the treatment of cardiovascular disease. In 1971 Congress passed the National Cancer Act, which was also described as President Nixon's War on Cancer (National Cancer Act, PL 99-158, 1971). The purpose of the National Cancer Act was to deploy significant resources toward the eradication of cancer. Most of those resources have been directed toward treatment, with relatively few resources devoted to cancer cause and prevention. Progress in the War on Cancer was recently evaluated by Bailar and Gornik (1997). Figure 3.3 summarizes recent trends in cancer mortality in the United States between 1996 and 1994. Mortality from cancer appeared to peak in about 1991 and has gone down slightly since then. Overall, there have been slight increases in cancer mortality since the War on Cancer began in 1971. However, changes in cancer death rates have been relatively modest. The American Cancer Society provides data on cancer mortality trends over the past 60 years for men (Figure 3.4) and women (Figure 3.5). For both men and women, there have been significant declines in cancers of the stomach and significant increases in cancers of the lung. For women, there have also been significant declines in cancers of the uterus and small declines in cancers of the colon and rectum. However, for most sites, the proportions of people dying of cancers have been relatively unaffected by major changes in medical care. The rapid increase in deaths from cancers of the lung can be attributed almost exclusively to the use of cigarettes. It is encouraging that deaths from lung cancer appear to have peaked for males by 1990and are now declining as cigarette use has decreased. Rates of lung cancer for women, however, are continuing to increase.

One example of the differences between the traditional biomedical and the outcomes models concerns screening and treatment for prostate cancer. The War on Cancer followed a traditional find it–fix it biomedical model. The identification of cancer dictates treatment, which in turn is evaluated by changes in biological process or disease activity. In the case of prostate cancer, a digital rectal exam may identify an asymmetric prostate, leading to a biopsy and the identification of prostate


Figure 3.3. Trends in cancer mortality in the United States, 1986 and 1994. Data from Bailar and Gornik (1997).

cancer. Diagnosis of cancer often leads to a radical prostectomy (surgical removal of the prostate gland). The success of the surgery would be confirmed by eradication of the tumor, reduced prostate-specific antigen (PSA), and patient survival.

In contrast to the traditional biomedical model, an outcomes perspective embraces public health notions of benefit. Instead of focusing on disease process, benefit is defined in terms of life duration and quality of life. Studies have demonstrated that serum PSA is elevated in men with clinically diagnosed prostate cancer (Hudson et al., 1989) and that PSA levels above 4.0ng /ml have positive predictive value for prostate cancer. Despite the promise of PSA screening, there are also significant controversies. Prostate cancer is common for men age 70 years and older (Lu-Yao et al., 1994). Averaging data across eight autopsy studies, Coley et al. (1997) estimated the prevalence of prostate cancer to be 39% in 70- to 79-year-old men. The treatment of this disease varies dramatically from country to country and within regions of the United States. For example,


Figure 3.4. Trends in cancer mortality by site for men. Source: American Cancer Society (1997).

radical prostectomy is done nearly twice as often in the Pacific Northwest as it is in New England, yet survival rates and deaths from prostate cancer are no different in the two regions (Fleming et al., 1993). PSA screening finds many cases. However, in the great majority of cases, the men would have died of another cause long before developing their first symptom of prostate cancer. When the disease is found, it is often “fixed” with surgical treatment. However, the fix has consequences, often leaving the man incontinent and/or impotent. The outcome model considers the benefits of screening and treatment from the patient's perspective. Often, using information provided by patients, it is concluded that quality-adjusted life expectancy is optimized without screening and treatment (Kaplan, 1997).

In considering changes in mortalities since 1970, Bailar and Gornik


Figure 3.5. Trends in cancer mortality by site for women. Uterine cancer death rates are for cervix and corpus combined. Source: American Cancer Society (1997).

(1997) concluded that cancer has not been defeated. The find it–fix it model has found and treated significantly more cancer, but the increased treatment has not produced clear public health benefits. In fact, they argue that it is time to reevaluate the dominant strategy of the past 40years, which placed most emphasis on improving treatments and little emphasis on prevention. The major increases in cancer have been associated with cigarette smoking, yet few of the resources have been devoted to the eradication of tobacco use. The outcomes model clearly leads toward an emphasis on factors that alter health outcomes. Bailar and Gornik concluded, “A national commitment to prevention of cancer, largely replacing reliance on hopes for universal cures, is now the way to go” (1997,p. 1574).


Looking into Arteries

Other examples come from the treatment of cardiovascular disease. Acute myocardial infarction is the most common cause of morbidity and mortality in both the United States and Canada. However, the two countries approach the treatment of cardiovascular disease differently. Invasive cardiac procedures, such as coronary angiography, are performed considerably more often in the United States than in Canada. Some years ago, we noted that about eight of every ten well-insured patients in San Diego received angiography following a heart attack, if they were treated in private hospitals (Nicod et al., 1991). However, only 40% patients at the San Diego Veterans Affairs Health Center received the procedure following a heart attack. In Vancouver, only 20% of post–myocardial infarction patients got angiography, and only 10% of patients in Sweden received the procedure. This variation would be acceptable if we knew that more care led to better health. However, there was little evidence that more aggressive care produced better results. Controlling for the seriousness of the heart attack (measured by the ejection fraction), the probability of surviving a heart attack in San Diego, Vancouver, and Sweden was comparable.

More recently, the use of invasive cardiac procedures in the United States and Canada was evaluated for 224,258 elderly Medicare recipients in the United States and 9,444 older patients in Ontario, Canada. Each of these patients had been the victim of heart attack after 1991. Among American patients, 34.9% underwent coronary angiography, while only 6.7% of the Canadian patients received this procedure. Having coronary angiography increases the likelihood that other invasive procedures will be performed. Among the American patients, 11.7% underwent transluminal coronary angioplasty in comparison to 1.5% of the Canadian patients. Further, 10.6% of the American patients versus only 1.4% of the Canadian patients underwent coronary artery bypass surgery. Figure 3.6 summarizes both the procedure rates and the mortality rates for these patients. It might be presumed that American patients are better off because they are more likely to obtain the latest procedures. However, mortality rates 30 days following the attack were comparable in the two countries (21.4% vs. 22.3%). Further, the mortality rates one year later were virtually identical (34.3% in the United States vs. 34.4% in Canada). These data suggest that the use of high-technology medical procedures is much more likely in the American system than in


Figure 3.6. Rates of coronary angiography (section a), cardiovascular procedures (section b), and heart disease mortality (section c) in the United states and Canada. Adapted from Tu et al.(1977).

the Canadian health care system. However, there is no clear evidence that patients benefit, at least in terms of survival (Tu et al., 1997).

These findings suggest that the find it–fix it approach to established coronary heart disease may have limited benefits. The procedures are expensive but may not extend life. An alternative might be to invest in programs that attempt to enhance outcomes by promoting health in entire communities. For heart disease, this might be accomplished by changing behaviors to reduce cholesterol. Programs to lower cholesterol might have only a small benefit for individuals but might have a substantial benefit for communities. As many as 40% of men and 20% of women have serum cholesterol levels less than 240 mg/dl. One analysis considered the benefits of population-wide heart disease prevention programs in California and Finland. Some of these programs have been criticized because they reduce serum cholesterol by only 1% to 4%. However, these slight reductions in average serum cholesterol may have contributed to as much as one-third in the decline in coronary heart disease in the United States since the mid-1960s. The education programs in Finland and California use media campaigns and face-to-face instruction. The programs cost about $4.95 per person per year and, on average, produce a reduction of about 2% in serum cholesterol. The programs produce a qualityadjusted year of life at about $3,200 for individuals at risk for coronary

heart disease. A more intensive program that reduces serum cholesterol by about 3% might cost $16.55 for the first year and $8.28 per year thereafter and could produce a year of life at about $6,100. Even though advances in medical care have cut mortality from coronary heart disease (Hunink et al., 1997), the evidence suggests that population-based efforts to reduce serum cholesterol should become part of U.S. health policy (Tosteson et al., 1997).


There have been a few attempts to apply outcomes thinking to health policy. One heroic attempt to apply an outcomes model to public policy was considered by the State of Oregon. In the late 1980s, Oregon was faced with the fact that costs of health care were expanding much more rapidly than the budgets for Medicaid. Collectively, Oregon citizens spent approximately $6 billion on health care in 1989, which is about three times what they spent in state income taxes. Oregon also recognized that American health care is not a two-tiered system but rather a three-tiered system. The three-tiered system includes people who have regular insurance and can pay for their care, people enrolled in Medicaid, and a growing third tier of people who had no health insurance at all. By 1993, this third tier represented about 450,000 citizens, or about one-fifth the population of the state. In addition, another 230,000 were underinsured, and the trend indicated that the number of uninsured and underinsured was steadily increasing. In order to address the growing crisis of funding for health care, the only alternatives were to (1) change eligibility criteria and remove some individuals from the Medicaid rolls, (2) continue to provide care to a large number of people but limit the care they receive, or (3) increase revenues (taxes) to pay for the current system.

Led by a grassroots citizens group known as Oregon Health Decisions, it was argued that Oregon, like other states and countries, was already rationing health care. The problem was that rationing was implicit and not open to public scrutiny. In fact, people were being rationed rather than services. In other words, many individuals in need of care received none because they were in the wrong category. Pregnant women, for example, were covered. A young woman employed as an hourly worker may be ineligible for health care, but an unemployed woman would be eligible if she were to become pregnant. Thus, the system created incentives to become pregnant in order to have a regular source of health care. The system allowed health care under Medicaid for

poor families with young children but disallowed coverage for poor families with older children. Oregon, like many other states, defined Medicaid eligibility for the Aid for Families with Dependent Children (AFDC) as 50% of the poverty line. That policy set the criterion income at about $5,700 per year for a family of three. A hard-working independent carpenter earning $11,000 annually might be completely excluded by the system even though he was at high risk for injury.

These arguments caught the attention of John Kitzhaber, M.D., the president of the state senate and now the governor of Oregon. Under Kitzhaber's leadership, Oregon passed three pieces of legislation to attack this problem (Kitzhaber, 1990). In this chapter we focus most specifically on Senate Bill 27, which mandated that health services be prioritized from most to least important. The purpose of the prioritization was to eliminate services that did not provide benefit. The process of creating the prioritized list was an extremely difficult one. The commission began by attempting to create a prioritized list of all health services. However, it soon became apparent that this was a nearly impossible task. Thus, the commissioners began searching for a combination of conditions and treatments that could be lumped together. They refer to these as condition-treatment pairs. Examples of these condition treatment pairs are shown in Table 3.2. For example, the condition of rectal prolapse is paired with the treatment partial colectomy, while osteoporosis is paired with medical therapy.

A Health Services Commission was created in order to develop the prioritized list. This commission obtained several sources of information. First, it held public hearings to learn about preferences for medical care in the Oregon communities. These meetings helped clarify how citizens viewed medical services. Various approaches to care were rated and discussed. On the basis of 48 town meetings that were attended by more than 1,000 people, 13 community values emerged. These values included prevention, cost-effectiveness, quality of life, ability to function, length of life, and so on. The major lesson from the community meetings was that citizens wanted preventive services. Further, the people consistently stated that the state should forgo expensive heroic treatments for individuals or small groups in order to offer basic services for everyone. In order to pay for these basic services, it was necessary to reduce spending elsewhere, and it was therefore important to rank services according to their desirability as determined by studies of community values. The commission chose to evaluate services using the QWB scale and a modified GHPM.

Condition Treatment
Rectal prolapse Partial colectomy
Osteoporosis Medical therapy
Ophthalmic injury Closure
Obesity Nutritional and lifestyle counseling

Two factors influenced the commission's decision to modify the standard GHPM approach. First, the commission could not conduct clinical trials for each of the many condition-treatment pairs. Further, estimation of treatment benefit using the QWB cannot be left to laymen. Thus, the commission formed a medical committee that had expertise in essentially all specialty areas and nearly all the major provider groups in the state. The Oregon citizens were concerned about using weights from California in assigning priorities in their state. Thus, 1,001 Oregon citizens participated in a separate weighting experiment. The weights were obtained in a telephone survey that was conducted by Oregon State University.

Multiple Lists

The Oregon Health Services Commission created multiple versions of the prioritized list. The first list was created in 1990. This list rankordered 1,692 condition-treatment pairs. The Health Services Commission used rough estimates of QWB changes and cost to create this list. Although this list has been the target of extensive criticism, the Health Services Commission probably intended it to be a working document rather than a final list. While they were working on the list, they completed a first draft that was discussed at a public meeting.

The list was put together in a hasty manner. As a result, there were some obvious problems. For example, 314 different medical therapies were all valued at $98.51, and an additional 177 were all estimated to cost $2,560.65 (Tengs et al., 1996). The last 193 pairs on the 1,692-item list were ranked alphabetically according to diagnosis. In addition, there were many counterintuitive orderings on the list. For example, treatment for thumb sucking was ranked higher than treatment for AIDS. Although the 1990 list was clearly flawed, it is also important to emphasize that it was never officially released. There are many reasons why the list may not have been dependable. Perhaps most important was that the

overburdened staff was required to evaluate many condition-treatment pairs in a very short period of time. In most cases, systematic data were not available to guide their evaluations.

Perhaps the most discussed list was the one released in 1991. This list included 709 items. The list was based on the QWB scale to estimate benefit and included estimates of costs. In the 1991 list, the system was reorganized according to three basic categories of care: essential, very important, and valuable to certain individuals (Table 3.3). Within these major groupings there were 17 subcategories. Nine of the 17 subcategories were classified as essential. The commission decided to place greatest emphasis on problems that were acute, fatal, and reversible. In these cases, treatment prevents death, and there is full recovery, for example, appendectomy for appendicitis and nonsurgical treatment for whooping cough. Other subcategories classified as essential included maternity care, treatment for conditions that prevent death but do not allow full recovery, and preventive care for children.

Listed as very important were treatments for nonfatal conditions that would return the individual to a previous state of health. Also included in this category were acute, nonfatal, one-time treatments that might improve quality of life. These might be hip replacements and cornea transplants. At the bottom of the list were treatments for fatal or nonfatal conditions that did not improve quality of life or extend life. These might be progressive treatments for the end stages of diseases such as cancer and AIDS or care for conditions for which the treatments were known not to be effective. In the revised approach, the commission decided to ignore cost information and to allow their own subjective judgments to influence the rankings on the list. Table 3.4 summarizes the conditions selected from the top of the list, the middle of the list, and the bottom of the list. Unfortunately, the final exercise in Oregon resulted in many deviations from the GHPM. However, the exercise demonstrates an attempt to resolve the healthcare crisis on the basis of health outcome.

One of the most important criticisms of the Oregon plan has been offered by Tengs and her associates (Tengs, 1996; Tengs et al., 1996). These investigators used rank-order correlational methods to compare the Oregon lists with cost-effectiveness analyses that have been published in the literature. They report that the 1990 list showed essentially no correlation with published cost-effectiveness analyses. The 1991 list was correlated 0.39 with published studies, while the correlation of the 1992 list was 0.25. The correlation with the 1993 list was 0.24. The Tengs analyses are important, and they have received widespread attention as criticisms

NOTE: Every person is entitled to services necessary for a diagnosis. Each health service on the list is presumed to include necessary ancillary services, such as hospital care, prescription drugs, and medical equipment and supplies necessary for successful treatment. SOURCE: Oregon Health Services Commission.
1. Acute fatal—treatment prevents death and allows full recovery: appendectomy
for appendicitis; nonsurgical treatment for whooping cough; repair of
deep, open wound in neck; nonsurgical treatment for infection of the heart muscle
2. Maternity care, including most newborn disorders: obstetrical care for pregancy;
care of the newborn
3. Acute fatal—treatment prevents death but does not allow full recovery:
nonsurgical treatment for stroke; all treatment for burns; treatment for severe head
4. Preventive care for children: immunizations and well-child exams
5. Chronic fatal—treatment improves life span and quality of life: nonsurgical
treatment for insulin-dependent diabetes; medical and surgical treatment for treatable
cancer of the uterus; medical treatment for asthma; drug therapy for HIV disease
6. Reproductive services—excludes maternity and infertility services: birth
control and sterilization
7. Comfort care: pain management and hospice care for the end stages of disases
such as cancer and AIDS
8. Preventive dental care—adults and children: exams; cleaning and fluoride
9. Proven effective preventive care for adults: mammograms; blood pressure
screening; Pap smears
Very Important
10. Acute nonfatal—treatment causes return to previous health: nonsurgical
treatment for acute thyroiditis; medical treatment for vaginitis; fillings for cavities
11. Chronic nonfatal—one-time treatment improves quality of life: hip replacement;
corneal transplants for cataracts; rheumatic fever
12. Acute nonfatal—treatment without return to previous health: relocation of
dislocated elbow; repair of cut to cornea
13. Chronic nonfatal—repetitive treatment improves quality of life: nonsurgical
treatment for rheumatoid arthritis; gout; migraine headaches
Valuable to Certain Individuals
14. Acute nonfatal—treatment speeds recovery: medical treatment for viral sore
throat; diaper rash
15. Infertility services: medical treatment for infertility; in vitro fertilization;
artificial insemination
16. Less effective preventive care for adults: routine screening for those people
not otherwise at risk, such as diabetes screening if the person is under 40 years old
and not pregnant
17. Fatal or nonfatal—treatment causes minimal or no improvement in quality
of life: aggressive treatment for end stages of diseases such as cancer and AIDS; medical
treatment for nongenital warts.

Top 10
1. Medical treatment for bacterial pneumonia
2. Medical treatment of tuberculosis
3. Medical or surgical treatment for peritonitis
4. Removal for foreign body from pharynx, larynx, trachea, bronchus,
and esophagus
5. Appendectomy
6. Repair of ruptured intestine
7. Repair of hernia with obstruction and/or gangrene
8. Medical therapy for croup syndrome
9. Medical therapy for acute orbital cellulitis
10. Surgery for ectopic pregnancy
Middle 10
350. Repair of open wounds
351. Drainage and medical therapy for abscessed cysts of Bartholin's gland
352. Medical therapy for pilonidal cyst with abscess
353. Medical therapy for acute thyroiditis
354. Medical therapy for acute otitis media
355. Drainage tubes or tonsil-adenoidectomy for chronic otitis media
356. Surgical treatment for cholesteatoma
357. Medical therapy for sinusitis
358. Medical therapy for acute conjunctivitis
359. Medical therapy for spina bifida without hydrocephalus
Bottom 10
700. Mastopexy for gynecomastia
701. Medical and surgical therapy for cyst of the kidney
702. Medical therapy for end-stage HIV disease (comfort care excluded—
it is high on list)
703. Surgery for chronic pancreatitis
704. Medical therapy for superficial wounds without infection
705. Medical therapy for constitutional aplastic anemia
706. Surgical treatment for prolapsed urethral mucosa
707. Paracentesis of aqueous for central retinal artery occlusion
708. Life support for extremely low birth weight (<500 g) and under 23-week
709. Life support for anencephalous
of the Oregon process and the GHPM. However, the Tengs et al. analyses did not apply the GHPM and do not consider health-related quality of life. In fact, all of the analyses used in their evaluation were studies that valued the cost per year of life saved, and studies considering cost per quality-adjusted life year were systematically eliminated. Since one of the most important aspects of medical care is to improve
functioning and the quality of life, this could be a serious problem. The rationale for disregarding health-related quality of life was that too few studies have evaluated these outcomes. On the other hand, it is not clear that the correlation between rigorous cost-effectiveness analysis and the Oregon process would have been improved by inclusion of quality of life data. As Tengs (1996) notes, problems with the 1991 Oregon list included poor measures of cost, failure to discount future costs, and considering health outcomes for only the first five years after treatment.

It is of interest to note that the 1991 list, the one closest to using the GHPM, showed the strongest correlation with the published literature. The correlation of .39 is statistically significant and substantial. It is important to emphasize that the Oregon project and the previously published analyses were conducted using very different methods. Further, the Tengs's analysis makes the assumptions that the rank ordering of previously published studies is correct and that failure to replicate this rank ordering must be incorrect. It is known that there is substantial variability in the estimates of cost-effectiveness across published studies.

Perhaps the most misunderstood aspect of the Oregon experience is the belief that Oregon rationed health care on the basis of cost-effectiveness. Tengs (1996) argued that Oregon intended to allocate resources using a systematic approach. However, in response to political pressures and political realities, the state abandoned the use of systematic decision analysis. They first eliminated cost for consideration, and then they eliminated quality of life from the analysis and focused on subjective estimates of treatment effectiveness over a short time interval of five years. Not only did the system applied in Oregon differ from systematic cost-effectiveness analysis, but in fact the priorities were only weakly related to results from systematic analysis.

Update on the Oregon Plan

Despite all the criticisms of the Oregon plan, there is some evidence that the outcomes approach produced benefits. The Medicaid portion of the Oregon Health Plan enrolled 120,000 new members during its first year. This was equal to the number projected for the five-year life span of the demonstration project. The initial evaluation showed that emergency room visits declined 5.3% in 1994 and that urgent care visits declined by 1%. A 1995 evaluation suggested that emergency room visits dropped an additional 2.1% overall and 6.2% in Oregon's rural areas. One of the advantages of the program is that insuring more people reduces bad debt

and the need for cost shifting. In 1994, charity care declined 18.7%, and bad debts 10.6%, in Oregon. In the Portland metropolitan area, these reductions were 23.8% and 15.7%, respectively. In 1995, charity care was reduced by over 30% relative to the 1994 levels.

One of the interesting results of the plan was that the number of Oregon families receiving AFDC declined. Since it was no longer necessary to be in AFDC in order to get health care, the incentives for being on welfare were removed (Conviser, 1997).


This chapter has outlined a new paradigm for thinking about alternatives in health care. In short, it is suggested that limited health care resources be used to maximize life expectancy and health-related quality of life. Services that do not achieve their objectives should not be funded, and the savings should be used to extend basic health care benefits to people currently uninsured. The proposed model is consistent with the thinking of several groups, including scholars in the United States (Office of Technology Assessment, 1979; Russell, 1986; Weinstein and Stason, 1976, 1977), the United Kingdom (Drummond et al., 1987; Maynard, 1991; Williams, 1988), Canada (Torrance, 1986, 1987), and Australia (Richardson, 1991). Although the exact methodologies proposed by these different research groups vary slightly, the theory is nearly identical. Patrick and Erickson (1993) offered a detailed account of the methodological steps required to implement the system. Methods are now available to begin guiding policy decisions. However, our information base for the implementation of the model is still incomplete.

The attractiveness of the outcomes model arises from several of its qualities. For example, it makes all choices explicit, so that anyone can evaluate the assumptions used to make decisions. The model requires a large amount of data on health outcomes, much more than what we have available for most health care programs. Little research has been completed that measures outcomes in a manner that would be useful in applying the model. This requires either that we obtain the data directly (through clinical trials or other means) or that the data be estimated. Though there is extensive experience in using estimated data (e.g., Naglie and Detsky, 1992; Weinstein and Stason, 1976), there is evidence that the methods used in this approach may be faulty (Fryback et al., 1993). A second drawback is the effort required to complete an evaluation.

Work is needed on methods to simplify the evaluation without detracting from the strengths of the model.

In conclusion, problems in health care might be characterized by the three A's: affordability, access, and accountability. These three problems are interrelated. Health care became expensive because a traditional biomedical model rewarded providers for doing procedures on the basis of diagnoses. The excessive expense of care made the costs prohibitive, and many people lost access to the system. However, the expensive system has been unable to demonstrate that it provides benefits to patients. An alternative to the biomedical model, known as the outcomes model, emphasizes that health services must be valued in terms of their impact on life expectancy and their effects on patient-reported quality of life. The outcomes model suggests that population health status might be enhanced if resources are shifted away from procedure-based reimbursement and toward primary prevention.


Supported in part by a scholars' grant from the American Cancer Society.


“American Cancer Society.” (1997). Cancer Facts and Figures—1997.Atlanta: American Cancer Society.

“American College of Physicians.” (1997). “Screening for prostate cancer.” Annals of Internal Medicine126, 480–484.

Anderson, J. P., Bush, J. W., Chen, M. M., and Dolenc, D. C.(1986). “Policy space areas and properties of benefit-cost/utility analysis.” Journal of the American Medical Association255 (6), 794–795.

Bailar, J. C., and Gornik, H. L.(1997). “Cancer undefeated.” New England Journal of Medicine336, 1569–1574.

Blendon, R. J., Leitman, R., Morrison, I., and Donelan, K.(1990). “Satisfaction with health systems in ten nations.” Health Affairs9, 185–192.

Brook, R. H., and Lohr, K. N.(1987). “Monitoring quality of care in the Medicare program: Two proposed systems.” Journal of the American Medical Associations258, 3138–3141.

“Center for Health Economics Research.” (1994). The Nation's Healthcare Bill: Who Bears the Burden?Waltham, Mass.: Center for Health Economics Research.

Coley, C. M., Barry, M. J., Fleming, C., et al. (1997). “Early detection of prostate cancer. Part II: Estimating the risks, benefits, and costs.” Annals of Internal Medicine126, 468–479.


Conviser, R.(1997). “A brief history of the Oregon Health Plan and its features.” Unpublished paper, Oregon Health Services Commission, Salem, Oregon.

Drummond, M., Stoddart, G., Labelle, R., and Cushman, R.(1987). “Health economics: An introduction for clinicians.” Annals of Internal Medicine107, 88–92.

Fisher, E. S., Wennberg, J. E., Stukel, T. A., and Sharp, S. M.(1994). “Hospital readmission rates for cohorts of Medicare beneficiaries in Boston and New Haven.” New England Journal of Medicine331, 989–995.

Fleming, C., Wasson, J. H., Albertsen, P. C., et al. (1993). “A decision analysis of alternative treatment strategies for clinically localized prostate cancer.” Journal of the American Medical Association269, 2650–2658.

Fryback, D. G., Dasbach, E. J., Klein, R., et al. (1993). “The Beaver Dam Health Outcomes Study: Initial catalog of health-state quality factors.” Medical Decision Making13 (2), 89–102.

Hudson, M. A., Bahnson, R. R., and Catalona, W. J.(1989). “Clinical use of prostate specific antigen in patients with prostate cancer.” Journal of Urology142, 1011–1017.

Hunink, M. G., Goldman, L., Tosteson, A. N., et al. (1997). “The recent decline in mortality from coronary heart disease, 1980–1990: The effect of secular trends in risk factors and treatment.” Journal of the American Medical Association277 (7), 535–542.

Hunink, M. G., Goldman, L., Tosteson, A. N., et al. (1993b). “Application of a general health policy model in the American healthcare crisis.” Journal of the Royal Society of Medicine86, 277–281.

Hunink, M. G., Goldman, L., Tosteson, A. N., et al. (1997). “Decisions about prostate cancer screening in managed care.” Current Opinion in Oncology9, 480–486.

Kaplan, R. M.(1984). “The connection between clinical health promotion and health status: A critical overview.” American Psychologist39 (7), 755–765.

Kaplan, R. M.(1990). “Behavior as the central outcome in health care.” American Psychologist45, 1211–1220.

Kaplan, R. M., and Anderson, J. P.(1988). “A general health policy model: Update and applications.” Health Services Research23, 203–235.

Kaplan, R. M., Anderson, J. P., and Ganiats, T. G.(1993). “The quality of wellbeing scale: Rationale for a single quality of life index.” In S. R. Walker and R. Rosser, eds., Quality of Life Assessment and Applications.Pp. 65–94. London: MTP Press.

Kaplan, R. M., Bush, J. W., and Berry, C. C.(1976). “Health status: Types of validity and the index of well-being.” Health Services Research11 (4), 478–507.

Kaplan, R. M., Bush, J. W., and Berry, C. C.(1978). “The reliability, stability, and generaliz ability of a health status index.” Proceedings of the Social Status Section, American Statistical Association, 704–709.

Kaplan, R. M., Bush, J. W., and Berry, C. C.(1979). “Health status index: Category rating versus magnitude estimation for measuring levels of well-being.” Medical Care17 (5), 501–525.

Kaplan, R. M., Orleans, C. T., Perkins, K. A., et al. (1995). “Marshaling the evidence for greater regulation and control of tobacco products: A call for action.” Annals of Behavioral Medicine17, 3–14.

Kitzhaber, J.(1990). “The Oregon Basic Health Services Act.” Oregon State Senate, Salem, Oregon.


Lu-Yao, G. L., Baron, J. A., Barrett, J. A., and Fisher, E. S.(1994). “Treatment and survival among elderly Americans with hip fractures: A population based study.” American Journal of Public Health84, 1287–1291.

Maynard, A.(1991). “Economic issues in HIV management.” In A. Maynard, ed., Economic Aspects of HIV Management.Pp. 6–12. London: Colwood House Medical Publications.

McGinnis, J. M., and Foege, W. H.(1993). “Actual causes of death in the United States.” Journal of the American Medical Association270, 2207–2212.

Naglie, I. G., and Detsky, A. S.(1992). “Treatment of chronic nonvalvular atrial fibrillation in the elderly: A decision analysis.” Medical Decision Making12, 239–249.

Nicod, P., Gilpin, E. A., Dittrich, H., et al. (1991). “Trends in use of coronary angiography in subacute phase of myocardial infarction.” Circulation84 (3), 1004–1015.

“Office of Technology Assessment, U.S. Congress.” (1979). A Review of Selected Federal Vaccine and Immunization Policies: Based on Case Studies of Pneumococcal Vaccine.Washington, D.C.: U.S. Government Printing Office.

Patrick, D. L., and Erickson, P.(1993). Health Status and Health Policy: Allocating Resources to Healthcare.Cambridge, Mass.: Cambridge University Press.

Patrick, D. L., and Erickson, P.(1995). “When does curbing health costs really help the economy?” Health Affairs (Millwood) 14, 68–82.

Relman, A.(1989). “Confronting the crisis in health care.” Technology Review, July, 31–40.

Richardson, J.(1991). “Economic assessment in healthcare: Theory and practice.” Melbourne, Australia: Monash University National Centre for Health Program Evaluation.

Rothenberg, R., Masca, P., Mikl, J., et al. (1987). “Cancer.” American Journal of Preventive Medicine3 (Suppl.), 30–42.

Russell, L.(1986). Is Prevention Better Than Cure?Washington, D.C.: The Brookings Institution.

Sturm, R., and Wells, K. B.(1995). “How can care for depression become more cost-effective?” Journal of the American Medical Association273, 51–58.

Tengs, T. O.(1996). “An evaluation of Oregon's Medicaid rationing algorithms.” Health Economics5 (3), 171–181.

Tengs, T. O., Meyer, G., Siegel, J. E., et al. (1996). “Oregon's Medicaid ranking and cost-effectiveness: Is there any relationship?” Medical Decision Making16, 99–107.

Torrance, G. W.(1986). “Measurement of health state utilities for economic appraisal: A review.” Journal of Health Economics5, 1.

Torrance, G. W.(1987). “Utility approach to measuring health-related quality of life.” Journal of Chronic Diseases40, 593–600.

Tosteson, A. N., Weinstein, M. C., Hunink, M. G., et al. (1997). “Cost-effective-ness of populationwide educational approaches to reduce serum cholesterol levels.” Circulation95, 24–30.

Tu, J. V., Pashos, C. L., Naylor, C. D., et al. (1997). “Use of cardiac procedures and outcomes in elderly patients with myocardial infarction in the United States and Canada.” New England Journal of Medicine336, 1500–1505.


Weinstein, M. C., and Stason, W. B.(1976). Hypertension: A Policy Perspective.Cambridge, Mass.: Harvard University Press.

Weinstein, M. C., and Stason, W. B.(1977). “Foundations of cost-effectiveness analysis for health and medical practice.” New England Journal of Medicine296, 716.

Wennberg, J. E.(1990). “Small area analysis and the medical care problem.” In L. Sechrest, E. Perrin, and J. Bunker, eds., Research Methodology: Strengthening Causal Interpretations of Nonexperimental Data.Pp. 177–206. USPHS/UDHHS Publication (PHS) 90-3454. Washington, D.C.: U.S. Government Printing Office.

Wennberg, J. E.(1996). “On the appropriateness of small-area analysis for cost containment.” Health Affairs (Millwood) 15, 164–167.

Williams, A.(1988). “The importance of quality of life in policy decisions.” In S. Walker and R. Rosser, eds., Quality of Life: Assessment and Application.Pp. 279–280. London: MTP Press.

Wilson, I. B., and Cleary, P. D.(1995). “Linking clinical variables with health related quality of life: A conceptual model of patient outcomes.” Journal of the American Medical Association273, 59–65.



Development of a Wellness
Guide for California

S. Leonard Syme


Interventions are the key element in public health practice. Every discipline in public health attempts, directly or indirectly, to prevent disease or promote health through interventions. Many of these have successfully improved public health through the implementation of government regulations for water, milk, and food quality as well as for sewage control, housing standards, highway and automobile safety, and occupational hazards. However, in comparison to such structural changes, when interventions have required that people learn the facts and then change their own high-risk behavior, results have not been as impressive.

We need to review this failure and think about alternative ways of attacking this problem.


It is instructive to begin with the classic failure, that of the Multiple Risk Factor Intervention Trial (MRFIT).1 That trial, which began in 1971 and ended in 1981, was intended to reduce the death rate from coronary heart disease in the United States by getting men in the top 10% risk group to lower their risk through behavioral change. A massive screening of 361,662 men in 20 cities throughout the country identified those at highest risk, using high serum cholesterol, cigarette smoking, and high blood pressure as determinants. After three intensive screening exams,

6,428 men were selected and randomized to intervention in MRFIT clinics while another 6,438 men, chosen at random, were informed of their risk and sent back to their own doctors.

Prior to randomization, all these men were informed about the requirements of the trial. They had to agree ahead of time to be randomly assigned either to the clinic or to their own doctors. Those randomized to the clinic had to be willing to stop smoking, change their diet, take medication for high blood pressure, agree to come to the clinic once or twice a week during the early phases of the trial, bring family members when requested, and continue their participation for six to eight years. Participants were urged not to enter the trial if they had any doubts or reservations about these terms. In addition, some volunteers were rejected by staff psychologists because it was thought that they might not be good participants. As a result, the final group of men was carefully selected, highly motivated, and highly informed.

The MRFIT cost $180 million. It was probably the most intensive and expensive clinical trial ever developed to educate people and get them to change their behavior. The trial failed.2 Despite the fact that each clinic in the trial had a large staff of specially trained counselors who worked very closely with each participant for the entire period, after six years 65% of the participants were still smoking, only half the men with hypertension had their blood pressure under control, and there were so few dietary changes that they are not even worth mentioning.

There are several explanations for the failure of MRFIT. One, of course, is that the men in the special intervention group did not change their behavior enough; another is that men in the control group changed too much.

During posttrial debriefing with the men in the control group, it was learned that when they were randomized to the control group, many simply changed behaviors on their own. This is an important phenomenon because it suggests that providing people with information about their risk is not a waste of time, and it also points out the importance of how that information is transmitted. In MRFIT, the control group members were provided with personalized and relevant information about threats to their health, but they were forced to find their own solutions. The men in the special intervention group with whom the clinics worked so intensively did not have better results than those men relegated to the control group.

The results from MRFIT were, of course, disappointing. Even more discouraging is the fact that even when people do successfully change

high-risk behaviors, new people continue to enter the “at-risk” population to take their place. For example, every time a man in MRFIT stopped smoking, one or two children in a school yard somewhere were probably taking their first tentative puffs on a cigarette. In trials like MRFIT, nothing is done to change the distribution of disease in the population because such programs do not address the forces in society that caused the problem in the first place.


The MRFIT is just the most dramatic in a long list of failures we have experienced over the years in attempting to get people to change their behavior and maintain those changes over time. As a result of the MRFIT failure and others like it, some of us in public health began to explore new and different intervention alternatives. One of the most interesting of these was to develop community intervention programs rather than continuing to focus on individuals. If we could not get individuals to change their behavior, perhaps we could succeed by changing the communities in which people live.

In 1984, I began a collaboration with Enid Hunkeler of the Kaiser Health Plan Division of Research to design and execute a communitybased smoking intervention program in Richmond, California. At the beginning of this project, we assessed smoking behavior, knowledge, and attitudes in a random sample of residents living in a targeted area of Richmond. At the end of five years, we did a second survey with another random sample in the area to assess changes in behavior, knowledge, and attitudes. In the interval between the two surveys, we implemented an ambitious and intensive antismoking program. We worked with business and government leaders and identified a local captain on every block of the research area. We helped local children make an antismoking video featuring their own dancers and musicians performing their own music. (This video was shown for the first time in a large auditorium to a standing-room-only audience.) Data from an extensive process evaluation study directed by Professor Troy Duster of the Institute for Social Change at the University of California, Berkeley, confirmed that the project was implemented very effectively. Unfortunately, after five years of work, the smoking cessation results in Richmond were no better than they were in our two comparison cities (San Francisco and Oakland).

It is not entirely clear why this carefully designed project failed, but

we can speculate on some possible reasons. For years, Richmond has been battling unemployment, crime, drugs, severe air pollution from several nearby oil refineries, and an absence of community services. Of all the problems experienced by residents of this community, cigarette smoking is very low on their list. Our research group did not take adequate account of this, instead descending on the residents with a set of plans and priorities that reflected the assumption that our problems were their problems. The children's video, a project developed on the initiative of the residents, may have been the most successful part of the intervention; unfortunately, since our project, sponsored by the National Institutes of Health, focused on adults and not children, funds to evaluate the video project were not available.

The nationwide Community Intervention Trial for Smoking Cessation (COMMIT) 3 followed basically the same design as the Richmond project. This trial involved more than 10,000 heavy smokers in 11 intervention cities. A matched group of another 11 cities served as the control group. In the intervention cities, the goal was to create a social climate that did not support tobacco use. Efforts were made to implement smokefree policies at work sites and elsewhere in the community, provide a newsletter and other information, and train people to become smoking counselors. At the end of this massive and ambitious nationwide trial, no difference was observed in quit rate among heavy smokers between the intervention and control communities. Approximately 18% quit in both communities. Among light to moderate smokers, 31% quit in the intervention cities while 28% quit in the control communities. It is important to note that this study was expertly done by some of the best smoking cessation interventionists in the country, and it, too, failed.

During the time of the MRFIT, Richmond, and COMMIT studies, three major community intervention programs targeting coronary heart disease were also under way. These three interventions were the Stanford Five-City Coronary Heart Disease Project,4 the Minnesota Heart Health Program,5 and the Pawtucket (Rhode Island) Community Heart Health Program.6 These were long-term trials involving intensive educational campaigns focused on coronary heart disease risk factors. Each of these projects was well designed, and all were implemented with energy and enthusiasm. None have yet reported their morbidity and mortality data, but they have published their risk-factor results. No differences were found in Minnesota and Rhode Island between the communities targeted for intervention and those that served as control communities; some modest changes were seen in the Stanford study for two risk factors.

Given the monumental effort involved in the studies, these results can be seen only as disappointing.


Notwithstanding the failures of intervention efforts that relied primarily on the dissemination of information, it is important to recognize the successes that have taken place. Death rates from coronary heart disease in the United States have plummeted since 1968 despite the failure of various clinical trials designed to do just that. A conference, sponsored by the National Heart, Lung, and Blood Institute, was organized to explain this remarkable phenomenon, but it could not account for it in terms of less fat in the American diet, better control of hypertension, reduction in cigarette smoking, increases in physical activity, new surgical techniques, or better community resuscitation programs. Although the reduction in death rates is real and something is happening to explain it, we do not know what that something is.

Cigarette smoking rates have plummeted as well. Compared to overall prevalence rates of around 45% several years ago, current rates are now in the mid-20% range. Although more than 90% of this reduction has been achieved by individuals working on their own, it does appear that this success is due, in part, to information provided to them on the dangers of smoking.7

More discouraging is the fact that despite our best efforts to inform people about the risks of smoking, the rate among many minority populations is unchanged, and the rate among women continues to rise. And, most troubling, the smoking rate among American high school students has risen from 27% in 1991 to 35% in 1995. Among young Black males, the rate of smoking has doubled during that time (28%), while among young White males and females, the prevalence of smoking is now 37% and 40%, respectively. Among 8th-, 10th-, and 12th-graders, daily smoking rates increased by about 50% between 1991 and 1996. One of every five 12th-graders now smokes daily.8 This is very sad news.

More people in the United States are physically active than ever before, but this phenomenon occurs mostly among the upper-middle class.9 In the general population, the prevalence of obesity is actually increasing, with one-third of Americans now overweight.10 And pregnant women drank more alcohol in 1995 than they did four years earlier (3.5% of pregnant women had seven or more drinks per week, a 400% increase from 1991).11


Clearly, there have been successes and failures, but we do not always know why, and we have not always been able to stop the harmful things from happening. Thus, the current situation is confusing. Providing information is certainly of value, but we very often find that it is not enough.


The field of public health faces some difficult challenges. Suggestions have been made to improve this situation, but while they are reasonable, they are not, in my opinion, enough to solve the problem. One of these suggestions is that we develop and use more innovative and rigorous research designs. Part of that process would require us to improve our definitions as well as our measurement and assessment methods. The way in which we define disease, for example, has actually caused us much harm. We typically use a disease classification scheme that is firmly rooted in the clinical model of disease. Thus, we study such chronic problems as coronary heart disease, cancer, and arthritis as separate entities. This type of classification may be useful in the individual diagnosis and treatment of disease, but it is not helpful if our goal is prevention. It would make more sense to develop a classification scheme based on the risk factors that clinical conditions have in common.

Historically, this is the approach that was used so successfully in the study of infectious diseases. In that work, the key public health concepts were the study of diseases that were water, food, vector, and air borne. In a similar vein, in the study of diseases in the developed world today, it probably would be more useful to study smoking diseases, poverty diseases, nutritional deficiency or excess diseases, and sexually transmitted diseases. By studying only clinically defined diseases, we make it virtually impossible to see underlying commonalities, to see what forces in the community are causing problems, and to see how better to target in-terventions.12 This risk-factor focus would be of great value in improving assessment and measurement methods because it would be more likely to influence researchers to measure the right things.

Another suggestion is that we use qualitative methods more frequently in our work. There is often a reluctance to utilize such methods in favor of more rigorous quantitative assessment, even when quantitative approaches may be missing important issues. For example, we all know that in studying older people, quality of life may be more important than

a tabulation of illnesses, but we are not very practiced in ways to measure this. Similarly, we all know that early childhood development is important for adult health and well-being, but it is rarely studied in public health because children do not have enough disease to fit the usual epidemiologic model. Clearly, we are not well served by focusing solely on the quantitative assessment of the usual issues. In addition, we tend to rely on linear multivariate statistical methods to analyze our data when we might be better served by such alternative approaches as “grade of membership” or other nonlinear methods.

While these suggestions have merit and need to be considered, there is an even more important problem that must be addressed if we hope to deal effectively with the intervention challenge we face. We need to understand better the community and environmental forces that cause disease and that might be successfully targeted for intervention. It is, of course, not easy to select important community and environmental risk factors, but there is one factor about which there can be no argument: social class.


Until recently, social class had not been widely studied in public health despite the fact that the lower people are in terms of social class—however that term is defined—the higher are the rates of virtually all diseases and conditions. Although this phenomenon is universally recognized, almost nothing is known about why it occurs. While the list of possible causes is long and well recognized—including such factors as poverty, substandard housing, unemployment, poor nutrition, inadequate medical care, and low education—the relative importance of these various factors is not clear because social class is so rarely studied as the focus of research. In fact, social class is of such overwhelming power that it is typically “held constant” in research so that other things can be studied. If that were not done, social class would swamp all other factors, and it would be impossible to see the role of any other issues. Consequently, virtually nothing is known about the various subcomponents associated with social class.

One might justifiably argue: Why do we need to study the relative importance of this or that factor associated with social class? Is it not

enough to recognize that people who are lower down the social ladder have higher rates of disease and then take some kind of action on the basis of that information? Do we need to fuss about the statistical details? Unfortunately, if we simply say that people in lower social class positions are in trouble and should be helped, we often end up being criticized for trying to change the world and are told that it is too difficult, too revolutionary, and too impractical. Then we withdraw, and instead we try to help people stop smoking cigarettes. If it were possible to isolate and identify a few important issues about social class, we could at least begin to think about interventions in a more practical and concrete way.

The Study of British Civil Servants

An example of how we might study social class is provided by the work of Michael Marmot and his associates. In their study of British civil servants, these researchers showed that those at the very bottom of the civil service hierarchy had heart disease rates that were four times higher than those at the top.13 After adjusting for blood pressure, serum cholesterol, smoking, fibrinogen, social support, and other well-known heart disease risk factors, the difference between these groups was still threefold. But Marmot's study revealed something else as well: Those civil servants one step down from the top of the hierarchy—people who were professionals and executives, such as doctors and lawyers—had heart disease rates that were twice as high as those at the very top. Those at the very top were upper-class directors of agencies, all of whom had been educated at Oxford and Cambridge and most of whose careers were destined to end in knighthood.

It is not surprising that individuals at the bottom had higher rates of disease than those at the top, but it is surprising that doctors and lawyers, one step down from the top, also had higher rates. Doctors and lawyers are not poor, they do not live in run-down houses or suffer from inadequate medical care, and they are not hampered by poor education or poor nutrition. Therefore, it is not just that those at the bottom who have the highest rates of heart disease; there is a gradient of disease from the top of the British civil service hierarchy to the bottom. At first it was thought that this phenomenon might for some reason be unique to the British civil service. It is not. A similar gradient has been found almost everywhere in the industrialized world and for virtually every disease that has been studied.14 How can this gradient be explained?



One hypothesis that seems to make sense in explaining this phenomenon is that as one moves down the social class hierarchy, one has less opportunity to influence the events that affect one's life. If this idea about control of one's destiny turns out to be useful, there are things that can be done to intervene, such as providing information to people so that they can take greater control of their lives. The term empowerment is much used these days, and even though it has lately taken on a faddish connotation, it may nevertheless be a central and crucial concept.15 Indeed, this idea in one form or another has been used by scholars for many years under the headings of mastery, self-efficacy, locus of control, learned helplessness, controllability, predictability, desire for control, sense of control, powerlessness, hardiness, competence, and sense of coherence.

The concept of control has recently been found to be of great value in studying job stress and health. Early efforts to study the impact of job stress by focusing exclusively on job demands failed to identify any health consequences. However, when Robert Karasek and Tores Theorell added the concepts of discretion, latitude, and control to the notion of job demands, they were able to show very important effects on health.16 Interestingly, Marmot and his associates recently published a paper that finally solved the gradient-of-disease problem among the British civil servants: When the concept of control was added to their analysis, the gradient of disease was virtually eliminated.17

A Successful Intervention

One very impressive and successful community intervention began in 1971 in North Karelia, Finland. The North Karelia project was instituted as a result of demands from the community and began when a local newspaper reported that North Karelia had the world's highest death rate from coronary heart disease. This article caused great consternation and prompted community leaders to demand that national public health authorities do something to help them. In other words, the priorities for intervention were established by the citizens themselves, not by public health professionals. When the government team arrived, it was warmly greeted by an enthusiastic population. The intervention itself focused on structural changes rather than just information and education; for example, local meat and dairy producers agreed to lower the fat content of

their foods, and, in the evenings, schools were used as community discussion centers.18

Between 1970 and 1997, the death rate from coronary heart disease in North Karelia dropped more than 60%. This is a faster rate of decline than was recorded elsewhere in Finland or in any other European country.19 This decline is not simply due to improved medical care, because case-fatality rates were unchanged; the big improvement was a reduction in new and recurrent cases of coronary heart disease, a change due primarily to improved primary prevention. While declines in mortality rates and risk factors have also been observed during this same period of time in other parts of Finland, the changes in North Karelia have been by far the most impressive.20 One major difference between the highly successful North Karelia study and the other major community-level coronary heart disease intervention projects that failed was the active involvement of community workers in initiating and facilitating the project.


Drawing on the concepts of social class and control of destiny and the strategies of community-initiated interventions (as in North Karelia), a group of us in the Berkeley School of Public Health have spent approximately 10 years developing The Wellness Guide for California. Our goal was to inform Californians about the determinants of a healthy lifestyle and to communicate that there are choices that they can make and community resources that are available to help them. The Guide was developed with the active and intensive involvement of more than 400 community members who selected the topics to be covered, dictated much of the content, and guided the writing. It was specifically designed to empower people in lower social class positions because these are the people who bear the largest burden of disease and possess the fewest resources for dealing with these problems.

Interestingly enough, community members were initially involved only reluctantly. The first effort consisted of an 80-page guide, written entirely by university professionals, to help people deal with life problems from conception and birth to old age and death. We thought that we had done a good job in preparing this document. However, when we pretested it among the people for whom it was intended, we got a rude shock. Most of them found our efforts amusing but not very helpful or relevant. One man said, “You folks up at the university have done it again. If you want to know what problems we face, why don't you just ask us?”


We spent the next full year learning from our intended audience what they saw as their major life problems and what they knew about ways to solve those problems. This was an eye opener. For example, we had not realized that one of the biggest problems faced by 8- to 12-year-old girls was “being stuck with a bad reputation.” Nor were we aware that many people did not know precisely what a budget was or how to make one.

In the end, children helped with the section on children, people with mental illness helped with that section, old people worked on the aging section, and so on. We also received significant input from members of the Hispanic community in developing and adapting the Guide into its Spanish-language version: La Guía del Bienestar. Following a difficult year of listening and learning, we received major funding from The California Wellness Foundation to design, print, and distribute the Guide to 100,000 mothers in the Women, Infants and Children (WIC) Program, a nutrition program intended for low-income pregnant women and mothers of small children. Before distributing the Guide, we held 12 regional workshops throughout the state to train WIC staff in how to use the Guide with their clients most effectively.

Once our work was under way, The California Wellness Foundation selected and funded an independent group of researchers to evaluate our project. Three outcomes were selected for study: Did the mothers who received the Guide, in comparison to those who did not, (1) have more sense of control and confidence in solving life problems, (2) know more about ways to solve those problems, and (3) change their behavior?

A complex, multistage, stratified random sampling design was used to select WIC mothers for study. In the end, 1,189 mothers were chosen for the baseline survey from 36 clinics in the state. As shown in Table 4.1, 816 mothers received the Guide in English or Spanish, while 373 did not. Approximately four months after the baseline survey, another random sample was selected to see whether mothers who had received the Guide had improved regarding their attitudes, knowledge, confidence, sense of control, and behavior. This sample consisted of 1,889 mothers, 994 of whom had received the Guide and a random sample of 895 mothers who had not. Finally, eight months after the baseline survey, a third independent sample was chosen, this time consisting of 672 mothers. Of this group, 362 had received the Guide, and 310 had not. Only a few mothers ended up in more than one sample; the three samples were independent of one another.

During all three time periods, mothers who received the Guide were

  Mothers Who
Received the Guide
Mothers Who Did
Not Receive the Guide
At baseline 816 373
4 months later 994 895
8 months later 362 310
Total mothers 2,172 1,578
  (24 WIC clinics) (12 WIC clinics)
  8 Months after
Receiving the Guide (%)
Had read the Guide 86
Still had the Guide 74
Changed behavior 26
Called telephone numbers 16
Intended to use the Guide in the future 65
almost identical to those in the control group in terms of age, level of education, literacy, and breast-feeding status. The only difference was that more mothers who received the Guide preferred to speak and read Spanish.

The first encouraging finding that the evaluators noted in this survey was that eight months after receiving the Guide, 86% of the mothers had read it, and 74% still had their copies. As shown in Table 4.2, 16% had called one or more of the telephone numbers in the Guide, 26% had actually made behavioral changes, and 65% intended to use the Guide in the future.

The second encouraging finding was that mothers who received the Guide were more knowledgeable and confident about solving life problems, as shown in Tables 4.3 and 4.4. The data in Tables 4.3 and 4.4 show the percentages of improvement of mothers who received the Guide versus those who did not in dealing with the following problems: What would you do if (1) you needed free medical care for your children, (2) a friend told you that he or she had a serious drug problem, (3) for the past few months your family had been spending more money than it made, and (4) a friend thought that he or she was being denied government benefits he or she deserved? These four problems were actually discussed in the Guide, but a fifth

  Responses 4 Months
after Baseline (%)
Responses 8 Months
after Baseline (%)
Problem Recipients Non-recipients Recipients Non-recipients

*Chi-squarc test of difference between recipients and nonrecipients, p <.01.

Getting medical care
for children
14 2 [*] 11 −10 [*]
Helping friend with
drug problem
18 9 [*] 24 −1 [*]
Spending too much
31 14 [*] 25 −3 [*]
Helping friend
who is being denied
government benefits
30 −4 [*] 29 0 [*]
Helping friend who
is owed a tax refund
19 6 [*] 31 −1 [*]

problem was not. The purpose of asking about the fifth problem was to see whether mothers who read the Guide would be empowered by that reading to deal more effectively with a problem that was not referred to in the Guide. The fifth question was, “Suppose a friend thinks the government owes her a refund on her tax return? What would you do?”

As can be seen in Table 4.3, mothers who received the Guide were significantly more knowledgeable about finding information to solve problems than mothers in the control group. This is true for each of these problems at four months, and these differences are even greater at eight months. In my view, the extent of improvement in the mothers who received the Guide is remarkable. It is also noteworthy that the mothers who did not receive the Guide actually showed a decrease in knowledge over time.

As can be seen in Table 4.4, mothers who received the Guide were also significantly more confident at four months and at eight months in solving problems than mothers who had not received the Guide. This is important because confidence is crucial in dealing with life challenges. If people know ahead of time that they will not be able to meet such challenges, little effort will be made even to try. If people lack confidence

  Responses 4 Months
after Baseline (%)
Responses 8 Months
after Baseline (%)
Problem Recipients Non-recipients Recipients Non-recipients

**Chi-square test of difference between recipients and nonrecipients, p <.01.

*Chi-square rest of difference between recipients and nonrecipients, p <.05.

Getting medical care
for children
17 −2 [**] 22 0 [**]
Helping friend with
drug problem
7 −3 [**] 10 0 [*]
Spending too much
1 −1 [*] 9 −1
Helping friend
who is being denied
government benefits
7 4 14 13
Helping friend who
is owed a tax refund
5 5 11 10 [**]
in their ability to solve problems, it is not likely that they will be able to control their destiny.

Mothers who received the Guide were also asked whether they had actually done anything different as a consequence of reading it. This was an open-ended question, and a content analysis of the results was carried out. We were pleased to note that 20% of mothers had made a change after four months, and a surprising 26% of them made a change after eight months. Some of the most frequent responses are shown in Table 4.5. This type of behavioral change based on a printed document may be unique in the health education literature.

One key function of the Guide is its role as a directory of community resources. Every page has a listing titled “For More Help.” This listing refers people to the Community Services section of their local phone book. Initially we found that the major telephone directory publishers each had a different way of classifying community services, but eventually we were able to get them all to standardize their listings and adopt the taxonomy used in the Guide. As a result, Californians in virtually every community can use the Guide in conjunction with their own phone directory

Found a counselor Decided to breast-feed
Used housing assistance Visited an employment agency
Gained a better understanding
of how to talk to kids
Located Alcoholics Anonymous
group for friend
Started school Learned how to choose a pediatrician
to find local services. GTE Directories has agreed to use our system nationwide; we will soon work with the other major telephone companies to encourage them to do the same. Once this is accomplished, a national version of the Guide can be published.

While successes like this are rare and it is tempting to be excited about these results, it is important to realize that we have not yet been able to show that these improvements in knowledge, confidence, problemsolving skills, and behavioral changes do finally result in better health. Nevertheless, the epidemiologic evidence suggests that when people have greater control over their lives, when they are better able to influence the events that impinge on them, their health is better. There is growing evidence from the field of psychoneuroimmunology that this type of psychosocial factor can have important impacts on immunologic function,21 and so we are optimistic and will continue to pursue this work. We were gratified that the WIC Program agreed with our assessment. To date, WIC has purchased and distributed one million copies of the Guide to mothers around the state.


While The Wellness Guide was developed with extensive community input, it nonetheless is focused on individuals and not on the social, economic, organizational, or political situations that are at the root of most problems. In fact, it might seem cruel and ineffective to suggest that people simply learn to adjust to and cope with an unfair and damaging world. This is another very important challenge that we must face.

To some extent, this issue has been addressed by a group of us at the School of Public Health in a study dealing with hypertension among San Francisco bus drivers. On the basis of a review of their preemployment medical records, we determined that the blood pressure of these bus drivers was relatively normal before they began driving city buses. The

longer they drove a bus, the higher their blood pressure became, even after adjusting for increasing age. After many years of research on this problem, we believe that the cause of this situation is not to be found in the individual drivers but in the job itself. An unrealistic and brutal bus schedule not only causes enormous stress in drivers but encourages poor nutrition and little opportunity for physical exercise as well. The schedule also has a damaging effect on family life because drivers, to recover from their stressful days, often go to local taverns where they drink too much and then arrive home late. Over time, many of these drivers end up suffering from severe fatigue and depression.

Proposals to the San Francisco bus company about changing the bus schedule originally met with great resistance. The company much preferred a program that taught drivers better coping skills so that they could deal with stress more effectively. The company was also willing to instruct the drivers about nutrition and exercise and encourage them to seek medical help for high blood pressure. However, once we began working with individual drivers, two important things happened. First, the bus company became aware of the fact that the drivers had high rates not only of hypertension but also of musculoskeletal problems, gastrointestinal difficulties, and respiratory complaints—all of which were causing high rates of absenteeism, accidents, and early retirement. If we were to offer individual recommendations to drivers that focused only on blood pressure, it would not help their other problems. Further, even if some current drivers were helped, the new drivers who followed them would continue to be at risk. The company began to see that the individual approach to hypertension was not going to solve the larger problem, and they began to discuss the possibility of structural and organizational changes.

The drivers also began to appreciate the limits of an intervention that focused on them as individuals. They knew that even if they succeeded in making changes in their diet, exercise, and coping patterns, the continued heavy pressure from their jobs would eventually break them. As a consequence, the drivers, through their union, began pressing for structural changes to the job itself.


It is often easier to begin intervention programs with individuals because their problems are obvious and salient. It is not as easy to recognize the

often subtle or indirect relevance of social, economic, organizational, and political forces. With The Wellness Guide, we are training people to seek help from community resources. When those resources are found to be ineffective or absent, people begin complaining and banding together for help. Several community agencies, in fact, have used these complaints as ammunition to obtain increased funding. In a sense, then, the Guide is serving as a community mobilization tool.

The Guide is also being used to teach parents and children what they can and should expect from their schools. These “health promoters” are then encouraged to recruit other parents and students to address schoolwide issues such as safety and improved parent-child communication. We are hopeful that this will provide the type of community support the schools need to increase their effectiveness and resources. In fact, schools may be the most important venue for such interventions.

It might be argued that The Wellness Guide approach is not ultimately the best way to deal with the effects of economic inequality in our society and that such an approach distracts us from focusing on fundamental societal reforms and makes things easier for those who remain committed to the status quo. However, by insisting on only fundamental and revolutionary social change, we may be doomed to programs that will not take effect for generations. Moral outrage about inequality is appropriate, but if we really want to change the world, we may have to begin by avoiding this type of artificial dichotomy. It seems clear to me that empowering individuals is a crucial and necessary first step in the movement toward societal change. One cannot stand without the other.

In my view, a good beginning toward reconciling the individual-versus-community argument can be made by empowering people very early in life. A wonderful example of this type of skills training is provided by the Perry Preschool Program in Ypsilanti, Michigan, in the 1960s.22 Poor, Black three- and four-year-old children were invited to attend a new preschool program (a model that eventually led to the development of Head Start). The sponsors of the program were overwhelmed by the number of applicants. To be fair, they accepted a group at random for admission. The children who were accepted had one or two years of this preschool experience and no further special intervention.

These children were then interviewed in a follow-up study when they were 19 years of age. The results were astonishing. Almost 100% of them were located, and their life circumstances were compared to a random group who had not had the preschool experience. In comparison to that control group, the children from the preschool had twice the

rates of high school graduation and college admission and one-half the rates of arrest, unemployment, and welfare dependency; the girls had one-half the rate of teenage pregnancy. A more recent follow-up of these children at age 27 reveals similarly powerful life changes.

It is not clear which elements of this program are responsible for these results. One feature of the program is that children who enroll in the class are asked what they would like to do. If they do not know, they are assigned to work with children who have declared an interest. When the new children eventually indicate areas of interest, all the resources of the school are marshaled to help them. If, for example, they are interested in airplanes, they learn, with help, to make a paper airplane and fly it. When it crashes, they are helped to redesign it and fly it again. When it crashes again, they are helped a third time. And so on, all day.

This approach encourages children to develop their own interests and to learn different ways to solve problems in the face of difficulties or failure. In short, they learn how to succeed. When these children go on to kindergarten and grade school, they do better than other poor, Black children, and this improvement tends to be cumulative over time. It is this type of empowerment that changes the lives of children and their families, and it is empowered individuals that are necessary to the initiation and maintenance of effective societal change.


The key lesson we health professionals should learn from the mistakes of the past is to be creative and inventive enough to become experts in the role of not being an expert.23 An excellent article titled “Sustaining Interventions in Community Systems: On the Relationship between Researchers and Communities” reveals the problem we face in its last sentence: “Those psychologists willing to apply their expertise to community health will face extraordinary challenges in translating their expertise to the needs of communities.” 24 It has been suggested by Professor Meredith Minkler that this sentence should read, “Those psychologists [and others, of course] willing to acknowledge the expertise of communities will face extraordinary challenges in letting go of power so that communities can use their expertise (aided by outside resources) to build on their strengths and address what they define as the needs of their communities.”

John McKnight eloquently expresses the problem this way:

The dilemma we face is lack of familiarity with the real community. We have great professional skills in managing and working within our systems, but

our skills are much less developed once we leave the system's space and cross over the frontier into the community. Indeed, many professionals are confused and frustrated when they attempt to work in community space, which seems very complex, disordered, unstructured, and uncontrollable. And many health professionals begin to discover that their powerful tools and techniques seem weaker, less effective, and even inappropriate in the community.25

Most educational interventions, either individual or community based, have thus far not proven to be effective. Most people do not change high-risk behaviors, and those who do seem to do so for reasons unrelated to our special efforts. It is important to learn what we can from the successes that we have seen. In my view, the common element in these successes is that people have found ways to influence the events that impinge on them and to change behaviors that do not support a healthy lifestyle. To do this, of course, they need information that they can shape to fit their life and social circumstances. This is a major challenge to those of us in public health. We have not paid sufficient attention to this problem in our training, research, or intervention programs. We need to do better.


Prepared for the 1997 California Wellness Foundation/University of California Wellness Lecture Series under a grant from The California Wellness Foundation.


1. “Multiple Risk Factor Intervention Trial Research Group.” 1981. The Multiple Risk Factor Intervention Trial.Preventive Medicine10, 387–553.

2. “Multiple Risk Factor Intervention Trial Research Group.” 1982. Multiple Risk Factor Intervention Trial: Risk factor changes and mortality results. Journal of the American Medical Association248, 1465–1477.

3. “Community Intervention Trial for Smoking Cessation (COMMIT): II.” 1995. Changes in adult cigarette smoking prevalence. American Journal of Public Health85, 193–200.

4. Farquar, J. W., Fortmann, S. P., Flora, J. A., et al. 1990. Effects of community wide education on cardiovascular disease risk factors. Journal of the American Medical Association264, 359–365.

5. Luepker, R. V., Murray, D. M., Jacobs, D. R., et al. 1994. Community education for cardiovascular disease prevention: Risk factor changes in the Minnesota Heart Health Program. American Journal of Public Health84, 1383–1393.


6. Lefebvre, R. C., Lasater, T. M., Carleton, R. A., and Peterson, G.1987. “Theory and delivery of health programming in the community: The Pawtucket Heart Health Program.” Preventive Medicine16, 80–95.

7. Fiore, M. C., Novotny, T. E., Pierce, J. P., et al. 1990. “Methods used to quit smoking in the United States: Do cessation programs help?” Journal of the American Medical Association263, 2760–2765.

8. Novotny, T. E.1996. “Smoking among Black and white youth: Differences that matter.” Annals of Epidemiology6, 474–475.

9. Caspersen, C. J., Christenson, G. M., and Pollard, R. A.1986. “Status of the 1990 physical fitness and exercise objective: Evidence from NHIS 1985.” Public Health Reports101, 587–592.

10. 1997. “Update: Prevalence of overweight among children, adolescents, and adults—United States” , 1988–1994. Morbidity and Mortality Weekly Report6 (9), 199–202.

11. “U.S. Department of Health and Human Services.” 1997. The Ninth Special Report to the U.S. Congress on Alcohol and Health.Washington, D.C.: U.S. Department of Health and Human Services, Public Health Service, National Institute of Alcohol and Alcohol Abuse.

12. Syme, S. L.1996. “Rethinking disease: Where do we go from here?” Annals of Epidemiology6, 463–468.

13. Marmot, M. G., Rose, G., Shipley, M., and Hamilton, P. J. S.1978. “Employment grade and coronary heart disease in British civil servants.” Journal of Epidemiology and Community Health3, 244–249.

14. Adler, N. E., Boyce, W. T., Chesney, M. A., et al. 1994. “Socioeconomic status and health: The challenge of the gradient.” American Psychologist49, 15–24.

15. Syme, S. L.1990. “Control and health: An epidemiological perspective.” In Rodin, J., Schooler, C., and Schaie, K. W., eds., Self-Directedness: Cause and Effects throughout the Life Course.Pp. 213–219. Hillsdale, N.J.: Lawrence Erlbaum Associates.

16. Karasek, R., and Theorell, T.1990. Healthy Work: Stress, Productivity, and the Reconstruction of Working Life.New York: Basic Books.

17. Marmot, M. G., Bosma, H., Hemingway, E., et al. 1997. “Contribution of job control and other risk factors to social variations in coronary heart disease incidence.” The Lancet350, 235–239.

18. Puska, P., Nissinen, A., Tuomilehto, J., et al. 1985. “The community based strategy to prevent coronary heart disease: Conclusions from the ten years of the North Karelia Project.” Annual Review of Public Health6, 147–193.

19. Salomaa, V., Miettinen, H., Kuulasma, K., et al. 1996. “Decline of CHD mortality in Finland during 1983–1992: Roles of incidence, recurrence, and case-fatality.” Circulation94, 3130–3137.

20. Vartiainen, E., Puska, P., Jousilahti, P., et al. 1994. “Twenty-year trends in coronary risk factors in North Karelia and in other areas of Finland.” International Journal of Epidemiology23, 495–504.

21. Ader, R., ed. 1981. Psychoneuroimmunology.New York: Academic Press.

22. Berrueta-Clement, J. R., Schweinhart, L. J., Barnett, W. S., et al. 1984. Changed Lives: The Effects of the Perry Pre-School Program on Youths through Age 19.Ypsilanti, Mich.: High/Scope Press.


23. Minkler, M., ed. 1997. Community Organizing and Community Building for Health.New Brunswick, N.J.: Rutgers University Press.

24. Altman, D.1995. “Sustaining interventions in community systems: On the relationship between researchers and communities.” Health Psychologist14 (6), 526–536.

25. McKnight, J. L.1994. “Two tools for well-being: Health systems and communities.” American Journal of Preventive Medicine10 (3), 23–25.



Implications for Health and Wellness

Richard C. Strohman


The trouble with the extended theory of the gene is that genetic elements, while critical, are only one aspect of biological regulation. They cannot, in themselves, specify details of organismal phenotype, including complex diseases like sporadic cancer and cardiovascular diseases. To be sure, there are cases in which genes may be said to “cause” attributes of an organism, but these are rare; in the realm of human diseases they account for about 2% of our total disease load.1 For the most part, complex attributes … phenotypes of organisms … are not caused by genes even though genes are the ultimate agents used to create phenotypes. But if genes don't determine us, if our disease causality cannot be located in genetic agents alone, if developmental processes characterized by high fidelity adherence to species form cannot be reduced to genetic programs, if the source of evolutionary change is not traced solely to random genetic mutation, then what does determine us? Where is disease causality located, where and what is the nature of programmed growth and development in living organisms, and what is the creative source of new morphology and function acting as substrates for natural selection? In short, if the program for life is not in the genes … and organisms are clearly programmed …, then where is the program?

The short answer is that the program is in no one place; it is distributed at many levels of the organism, and all levels are open to environmental signals. Controls may be found distributed in gene circuits, in metabolic networks, in cytoskeletal structures, in membrane units, in extracellular

matrix elements, and finally in the cell as a whole and in networks of cells at the various levels of organization above the cell. These levels of control each have their own rules, and all levels are interactive with one another and, in the case of cells and organisms, with the world around them. The major new idea here is that these levels of control are not reducibly connected; it is not possible, for example, to reduce common cancer to rules that govern DNA,2 just as it is not possible to reduce intelligence simply to the laws governing ion fluxes in brain neurons. DNA is involved in the phenotype “cancer” or “intelligence,” but the cause of both lies elsewhere at higher levels of organization, including the level of the cell as a whole and the level of cell-cell networking.

This short answer is already extremely complex compared to the idea of reducibility, that ultimate control is in the gene. Part of the current maturing of biology is the surrender of simple “storybook” explanations for how life works and the acceptance that life is beginning to appear more like a complex adaptive system than like a gene machine. It is not the purpose of this chapter to treat in detail the various levels of complexity in cells and organisms, and in any case that would be beyond the reach of the author. What is attempted here is an effort to define the deficiencies evident in genetic reductionism and the problems presented by these deficiencies to medical thinking and to concepts of wellness.

Holism and Epigenesis:
Alternatives to Reductionism in Biology

Reductionism is being questioned at many levels.3,4 We are hearing of concerns about the lack of relationship between genomic and morphological complexity of cells and organisms. We hear questions concerning whether genomic databases provide the information necessary to define function at higher levels.5,6 We are becoming aware of theories of development that do not rely so heavily on genetic mutation as the source of new morphology and action but that instead emphasize the presence of robust generic processes of cells and organisms that generate new phenotypes.7 And we are learning of theories of evolution distinctly different from standard brand neo-Darwinism.8 These are not special creation theories but scientific theories that truly aim to incorporate developmental processes into a new and more complex theory of evolution. Finally, we are beginning to see a view of complex human disease that is not reductionist in nature and does not rely on causal explanations rooted in gene mutation but rather sees disease as a function of

organismal or at least organ- and tissue-level dysfunction.9,10 In short, we are bearing witness to the reemergence of the organism as a legitimate focus of research in biology and of the genome as an embodied, experience-contingent entity. What used to be referred to as the book of life written in the concrete of DNA is now being referred to as the flexible genome. Genes alone are vitally important; they are necessary but not sufficient to determine function or dysfunction in cells and organisms (the exceptions are the rare monogenic diseases discussed here).

If we are seeing a shift away from formal reductionism that identifies genes and genetic programs as the causes of complex diseases, it is also a shift toward an emphasis on higher levels of analysis and in many ways involves a turn toward physiological levels analysis informed perhaps by complex adaptive systems theory.7,8 While this shift is just beginning to be appreciated in the basic research community, where it and the accompanying revolution will certainly take some time to complete itself, the implications for medicine and for the further evolution of our concepts of health and wellness require immediate attention.11–13 This is so for many reasons, not the least of which is that medical research continues to be dominated by molecular/genetic analysis and by a reductionist program that resists any tendency toward hierarchical analysis in which the gene appears not as sole causal agent but merely as an important part of the overall complex biological system. Genetic analysis does contribute uniquely to rare monogenic diseases (see the following discussion) but cannot extend the notion of unique genetic cause to complex diseases (common cancer and heart diseases for example) or to regulatory levels above the gene where the issues of health and wellness and their relationship to the world are most likely to be joined. This is a major problem (discussed in the following) since the stated goal of one of our most ambitious and expensive scientific projects, the Human Genome Project (HGP), is to map all complex human diseases to “Mendelian” genes.

Biological Complexity and Concepts of Wellness

Why is it that molecular/genetic reductionism does not address issues of wellness and health? First, molecular and genetic focus is almost exclusively on the diagnosis and cure of disease symptoms. This focus provides a valuable contribution because some diseases are truly genetic in the strict sense and because the design of many drugs is seen to depend more and more on a molecular understanding. However, the strategies (causal pathways) for health and wellness are profoundly different from

those of genetic diseases, as will be explained here. Second, real genetic diseases are rare and account for less than 2% of the disease load in the economically advanced sectors of the postindustrial world.1,14 Common diseases like most cancer and cardiovascular diseases that account for over 70% of premature morbidity and mortality are not genetic in the strict Mendelian sense. Nevertheless, the vast majority of our research budget is assigned to genetic-related problems.15 This 70% represents multifactorial diseases involving many genes whose interactions with one another and with their encoded proteins define an open network sensitive to environmental signals.1,15 The problem here is that, while the HGP will be able to provide a detailed genetic map for complex polygenic diseases, it cannot provide the instructions for reading these maps. Therefore, insights into the vast majority of complex human diseases and into their prevention are not to be expected from the HGP as such. Third, therefore, multifactorial diseases and states of health and wellness are to be seen as emergent features of these interactive informational networks. They are not reducible solely to the actions of single or even multiple genetic agents or to the actions of their encoded proteins. Fourth, while the economics of managed care forces an emphasis on disease prevention and on the superficial aspects of wellness, there is no theoretical insight into the concepts of wellness and health from fundamental research in experimental biology centered in a reductionistic genetics. Concepts of health and wellness are characteristics of whole organisms and of processes that are time and place dependent—dynamic processes open to environmental signals and contextualized by an individual's life experience. This is simply to say that in the matter of complex disease, it is necessary to go beyond the (genetic) information given, just as in the matter of a mental process, it is necessary to go beyond information in the neural cells and structures associated with that process. In complex adaptive systems like organisms and brains, even a complete description of initial conditions (even if that were possible) cannot provide a sufficient basis for predicting the outcome of the system.8

For all the previously stated reasons, we now find ourselves in a most critical situation in health care. Driven by economic requirements of managed care, we are coming to recognize the importance of disease prevention combined with health promotion/maintenance, and that is a giant step forward. The direction here is to extend the period of healthy life without necessarily increasing life expectancy. That is, there is increasing evidence that populations in economically advanced countries are rapidly approaching a maximum life expectancy and that further

investments in this direction will be expensive without necessarily being productive. As reviewed in this chapter, even the elimination of cancer and heart disease is expected to provide only a marginal increase in life expectancy in the population as a whole. Real improvement, both in quality of life and in lowering the cost of health care, is seen increasingly to come from the effort to bring more of our population into the realm of a maximum life expectancy now enjoyed mostly by the more affluent sectors. There is overwhelming evidence that increases in life expectancy have come in the past through holistic measures and not from applications of medical technology (see the following discussion). At the same time, the community of fundamental biological researchers, as exemplified by current directions within the National Institutes of Health (NIH) and the HGP, continue to emphasize a molecular/genetic approach to disease, an approach that theoretically could marginally extend the lives of the affluent few and of the fewer with rare monogenic diseases but that says very little about substantially extending the lives of the many who, because of socioeconomic reasons, do not possess the environments necessary for a long and relatively inexpensive-to-maintain healthy life. One of the arguments made here is that our national research portfolio needs to be balanced with a larger effort dedicated to understanding the processes by which the organism integrates its world of experience into a phenotype of health and well-being. Knowledge of this integration process at fundamental levels of cells and organisms cannot fail to support the further development of health and wellness concepts now seeking recognition and resources in our health care system.

This is not in any way meant to discredit or undermine research in gene/molecular-based biology, which continues to be essential. Indeed molecular genetic research as defined by newer epigenetic approaches is essential to our understanding of the pathways from environment to genome and to the changes in patterns of gene expression that take place during exposure to disease-related stress. The argument here is for balancing our national research effort with a commitment to inquiry into the more complex issues of living organisms and of their interactions with the world in which they live.


The Medical-Epidemiological Background

Substantial evidence from diverse studies now points to the possibility that most human diseases in the Western world are manageable and that

we are reaching a limiting plateau in our attempts to extend life. In addition, there is mounting evidence that disease management and longer life expectancy are more related to the presence of an environment appropriate to the conserved human genome than they are to the total medical intervention effort. Life span is a species constant,16,17 and in the United States we appear now to be rapidly approaching a maximum life expectancy of age 85.18 Even the elimination of the most serious premature killers—cancer and cardiovascular diseases—is predicted to provide a mere two to three years of additional life for the population at large.19,20 Increases in life expectancy coming from molecular genetic approaches are not expected since monogenic diseases remain stable at only 2% of total diseases and since the afflictions of older people are seen to be multifactorial, polygenic, and therefore ultimately beyond the reach of applied molecular genetics. That is to say, progeroid syndromes have a genetic basis fundamentally different from the simple monogenic diseases afflicting mostly younger people. In younger people, but not in the older, the power of modern molecular biology is seen as sufficient to provide, in theory, a successful genetic analysis and even therapy based on a linear (single gene → single disease) format. Attempts by gene cloners, armed with advanced statistical devices, to redefine common polygenic diseases in terms of genetic tendency 21 and attempts by behavioral scholars of various backgrounds to apply monogenic “software” to the reality of polygenic human traits 22 all appear to discount the warnings coming from cell and molecular embryological studies 23,24 that genetic approaches alone are not sufficient to yield a satisfactory picture of complex phenotypes. These, as well as other studies discussed later in more detail, include examples of nongenetic but nevertheless cellular responses to developmental environments in which the genotype is constrained by local circumstances.

The Biomedical Paradigm
and the Problem of Informational Redundancy

The major assumption of modern biomedical research is that unique genes have unique effects. This assumption is essential in the following areas:

Medical genetics, which seeks isomorphic mapping of human diseases to Mendelian genes 25

Molecular biology, which seeks to identify unique, genetically based mechanisms driving cellular processes 26


Developmental biology, which presupposes (1) the presence of genetic programs, (2) additivity of gene effects, and (3) the ability to map complex developmental stages to additive programmatic sequences in DNA 27

These assumptions and presuppositions, now experiencing major problems, are also the major features of the HGP. The HGP has become the centerpiece of the biomedical paradigm and has distilled a simplistic guide for future research and application. This guide is summarized as follows:

  1. All major noninfectious diseases are caused by defective genes.
  2. Diagnosis and therapy are available through genetic analysis alone.
  3. Aging and other complex human behavior is genetic, and all may be mapped to Mendelian factors.

As Brenner 28 and Wilkins 27 have pointed out, however, the uniqueness assumption of genetic determinism,

Unique Genes → Unique Effects,

is undermined by an emerging body of evidence showing functional informational redundancy in cell regulation. Here the focus is on redundant genes that more than one gene may specify any given function.29 In this case the reductionistic plan to associate genetic causality with complex phenotype is brought into question since the major research approach, saturation mutagenesis, depends completely on the uniqueness equation. This approach to understanding disease will generate a map or network of factors that interact to provide a useful background for a complex phenotype. However, as argued here, ultimate behavior is encoded not in DNA but rather in the environmentally interactive cellular epigenetic network, which includes the genome.

Levels of Biological Regulation

It is important here to distinguish three modes of gene activity that are operative in determining complex phenotype in organisms. The first is monogenic, which specifies a one gene → one trait pathway. This path


Figure 5.1. Genetic and epigenetic regulation.

way is often influenced by environment or by other genes, but in some cases, when mutation involves a specific DNA sequence, environment is seen to be irrelevant. Diseases like sickle cell anemia and Duchenne muscular dystrophy come to mind as prime examples of monogenic diseases. The second pathway is polygenic, which refers to the fact that phenotype is determined by many genes acting together.

The third path is epigenetic, which may involve both single-gene and multigene interaction. Epigenesis implies a level of complexity beyond gene-gene interaction and extends to interaction between genes, between genes and gene products (proteins), and between all of these and environmental signals, including, of course, the individual organismal experience. But in addition, epigenetic pathways are usually thought by developmental biologists to involve progressive states of organization, each succeeding state depending on the prior state. Epigenetic pathway therefore implies great complexity of interaction as well as the production of entire states of organization arising from that interaction (see Figure 5.1). Finally, an epigenetic change in a cell, in a strict sense, is heritable; initial cellular responses not restricted to genomic alterations, usually called phenotypic or physiological adaptations, may persist over time and become stable so that change is transmitted to daughter cells during mitosis.

The heritable aspect of epigenetic change is an obvious aspect of differentiation where many different cell types, all with identical genomic sequences, maintain their differences over many generations. Of course, secondary changes in DNA may also contribute to the stabilization of cellular change, but these changes are not programmed by genes; they are rather programmed into DNA by regulatory events about which we now know quite a bit. Changes in methylation pattern, in DNA-binding proteins, or chromatin structure are examples of inherited secondary changes in DNA. These epigenetic changes result in altered transcriptional patterns and therefore in altered patterns of behavior at all levels

of the organism, from cellular to integrated psychophysiological action.30 Epigenesis has been given a modern definition as follows:

Classical genetics has revealed the mechanisms for the transmission of genes from generation to generation, but the strategy of the genes in unfolding the developmental programme remains obscure. Epigenetics comprises the study of the mechanisms that impart temporal and spatial control on the activity of all those genes required for the development of a complex organism from the zygote to the adult.31

As such, the definition establishes the basis for a level of organizational control above the genome, a level that is now well established in fact, but it is a level of complexity that continues to evade decisive theoretical insight. That is, epigenetic regulation is already extending and stretching the limits of our ability to draw the limits of interactional networks that are at work in governing a major phenotype like a complex disease. For example, the mechanisms of DNA marking (e.g., methylation) may be elucidated, but what is missing is any understanding of the question, “Where and how are these mechanisms deployed in cells … what are the rules, the boundary conditions for such deployment?” These questions are being addressed,10 but currently we have no consensus in biology that is necessary for a major new direction to be implemented. Courage and vision may be required on the part of our research leadership if we are to progress. Meanwhile we expect that a full description of a genetic network will come complete with a set of rules for its operation as an open system. But the rules do not come with the network diagram; they have to be discovered by human ingenuity. The differences between a genetic and an epigenetic informational system are depicted in Figure 5.1.

We have wrongly extended the theory of the gene to another area altogether; we have been lulled into reasoning that if the gene theory works at one level, from DNA to protein, it must work at all higher levels as well. We have thus extended the theory of the gene to the realm of gene management. But gene management is an entirely different process involving interactive cellular processes that display a complexity that may be described only as transcalculational, a mathematical term for “mind-boggling.” This interactive complexity is epigenetic in nature; it involves open networks of genes, proteins, and environmental signals that may turn out to be coextensive with the cell itself. It is as if the cell has interposed between its genome and its behavior a second informational

system able to integrate environmental and genetic information into its dynamical process and able to generate from this integration responses that are functional, or adaptive.

Genetic pathways specify organismal function only in rare cases, as in monogenic diseases like sickle cell anemia or muscular dystrophy, where mutation produces dysfunction in a protein of crucial importance. In these cases the cell (mostly but not always) has no compensatory mechanism, and environmental influences are nil; redundant information at either the genetic or the epigenetic level appears to be absent, and the mutant gene becomes the disease. But this rare event has such a powerful effect in making real the critical issues of disease and health that it has commanded our attention in other areas of our lives. Common diseases like cancer and cardiovascular problems that account for over 70% of premature morbidity and mortality are not the effects of single genes.

Epigenetic networks have been described as cellular neural networks and, given their great complexity and openness to environmental signals, most probably utilize a (nonlinear) logic and set of rules quite different from the comparatively linear rules needed for completing the genetic sequence of events. This comparison also emphasizes feedback from epigenetic networks to the genome, feedback that includes changing the patterns of gene expression. This change in pattern of gene expression is accomplished by enzymatic changes in chromosome structure and by “marking” sections of DNA chemically without changing the genetic code in any way. What is changed is the accessibility of genes to expression pathways. But the decisions to mark or not to mark are in the epigenetic and not the genetic pathway. These details of epigenetic biology, as defined by Jablonka and Holliday,30,31 are well known and are thoroughly covered in the literature. We can see at once that failure to include epigenetic processes and their rules in predicting outcomes and basing outcome analysis only on information in DNA will lead to the anomalies that are now being seen. Thus, information for cellular integration and response is encoded not only in DNA, and there are no genetic programs for this process; rather integration and response come out of the dynamics of the interactive system itself. The system response includes the genome but is not reducible to it. The cell is starting to look more like a complex adaptive system rather than a factory floor of robotic gene machines, and that is well and good.

In what follows, whenever I refer to polygenic traits or diseases, I assume,

along with mainstream biology, strong environmental interaction.

For my purposes, therefore, polygenic and epigenetic are synonymous. The basic assumption is that complex disease states, at a cellular level, involve heritable changes that may include gene mutation but that also include persistent cytoplasmic changes. In addition, it must be clear what classical developmental biologists mean when they discuss complex phenotypes in terms of genotypes. What is usually meant is that all complex traits (e.g., intelligence, aggressiveness, and cancer) have some genetic basis. But this basis is so polygenic (interactive and epigenetic)—it may extend to the entire genome—that there is little in the way of practical meaning given to “genetic basis.” For example, there is a genetic basis for speaking French, but the meaning of this does not go beyond the idea that there is a genetic basis for being human. In order to speak any language, we need to have something called a human genome (of which there are as many different kinds as there are humans) consisting of about 100,000 genes. But while these genes are necessary for speaking French, they are not sufficient. We also need the appropriate environment, the appropriate body, and the appropriate experience, all of which provide information not contained in the genome. Unfortunately, most behavioral and medical geneticists continue to believe that even the most complex human behavior can be reduced to genetic circuits. We now turn to examples where predictions and diagnoses based on genetic analysis alone have generated conflict and anomalous results.


The foundations of applied molecular genetics are twofold. The first is found in the statistical approaches designed by Fisher and Wright 32 to describe efficiency of selection in producing desired traits in agricultural populations. The second is found in the singular successful attempt in 1908 by A. Garrod to map a metabolic disease, alkaptonuria, to a Mendelian pattern of inheritance.33 Garrod would later offer the concept of “inborn errors of metabolism” to describe a range of metabolic disorders, leading to the general emphasis on genetic disease. The wide acceptance of the concept of genetic diseases, and the confusion of rare monogenic diseases of the Garrod type with the more common

polygenic human diseases, is seen as the single most important historical development underlying the widespread belief in the paradigm in question.34

The tension between agricultural genetics and medical genetics has been described and analyzed most recently by Wahlsten.35 In brief, the argument is that the major statistical tool, analysis of variance (ANOVA), as developed by Fisher, is insensitive to the heredity-environment interaction. This insensitivity is minimized in the agricultural breeding experiments for which ANOVA was designed because large sample size is normally the rule. In medical genetic studies (extended families) or in behavior genetics (twin studies), the sample sizes are small, so that error is large in detecting lack of interaction between heredity and environment. As Wahlsten points out, a newer statistical approach, multiple regression, is replacing ANOVA, but for the kinds of studies being discussed here, the two procedures are essentially equivalent. Experts in agricultural genetics generally accept significant interaction between genes and environment and are extremely cautious in applying heritability coefficients or in assigning any significant numeric value to genetic cause when dealing with complex traits. Their position is that if gene effects are interactive (not additive) with environmental effects, it is incorrect to use ANOVA for assessing genetic contribution to a particular phenotype across a range of environments. Medical geneticists, however, using the same ANOVA but with significantly smaller sample size, not surprisingly do not find evidence for interaction and therefore assume that heredity and environment are additive. They then assign great significance to heritability coefficients and are confident that these numbers describe quantitatively the contribution of separate heredity and environment to any particular phenotype. We have a medical literature, then, that asserts with great confidence, but with serious theoretical reservations from sectors of population genetics, that this or that complex behavior or disease, while having an environmental component, also has a separate genetic component that can be discovered and utilized in pursuit of some hypothetical treatment strategy. It is beyond the scope of this review to enter this controversy fully. It is enough to state the minimum conclusion that medical/behavioral genetics, with a linear view of gene-disease causality, finds itself in serious debate with a significant segment of its parent science, population genetics, which sees complex traits, including disease, as highly interactive and impossible to reduce to genetic elements alone (Figure 5.1).



Since the work of Garrod in 1908, a large number of monogenic diseases have been discovered, and there is a general misconception that all diseases are open to monogenic logic and to solution through gene therapy of some kind. In fact, the total percentage of monogenic diseases has remained constant at less than 2%. While rare monogenic diseases are legitimate targets of the new technology, most of the rest of the 98% of human diseases, including cancer and heart diseases, are not. The latter are polygenic, multifactorial diseases for which genes may be necessary but not sufficient.1

Diseases may be distributed according to whether they are determined before or after fertilization.36,37 Those determined before fertilization (2%) are, of course, genetic and are mostly not preventable. Of those determined after fertilization (98%), there may be multiple causality, including early developmental effects, but in theory at least these are all preventable.

There is a second level at which the biomedical paradigm is in conflict with actual disease distribution. The problem for medical genetic theory is that the common diseases of cancer and of the circulatory system appear to be new; they were not significant causes of death and disability in the early part of the 20th century.37 They are now the major cause of premature death and suffering in the industrial world. Clearly, this sudden shift in causality cannot be based on genetic change. Evolutionary theory and molecular biology agree completely that genetic adaptation due to mutation would take thousands of years and that change due to genetic recombination would also require much more time than the mere 50 to 100 years involved. The reasoning of medical genetics, however, is that these new diseases attack people mostly in older (post-60) age-groups. As such, the responsible genes would be beyond the reach of natural selection, which operates effectively at younger prereproductive ages. This being the case, it is argued that heart and cancer diseases are “old” entities, have always been with us (as have their genes), but show up significantly now because it is only recently that our population has aged sufficiently for them to become a problem. If this is true—so goes the argument—then these are genetic diseases, pure and simple, and may be attacked as such.

But the natural history of our complex diseases shows that, in all probability, these are not genetic diseases but are diseases of civilization.

Of course, they have some genetic basis, but this basis is so broad as to be trivial with regard to providing precise genetic answers. Evidence that diseases of civilization are not simply genetic includes the following. First, twin studies show extremely low concordance for most cancers and heart disease. Second, these same diseases show remarkable variation in identical populations over time and over geographic and migratory patterns. These variations disclose, for example, that diseases tend to be place (environment) specific and that when people migrate, they tend to have those diseases that are common to their host population, not those that are common to the genes they brought with them, that is, not common to their native population. These variations are reversible. Finally, these diseases are rare in populations that have not come under Western habits. Natural history studies all indicate that our major noninfectious diseases are not genetic in any straightforward causal sense; they are diseases associated with changes in environment. That is the message from the past and the present. That message, extended into the future, is that new noninfectious diseases, their prevention and therapy, will also be associated with environmental change.36,37


Hypertension, Myocardial
Infarction, and the ACE Mutation

Restriction fragment length polymorphisms (RFLP) are being used to generate maps of genes and gene products that interact to produce a disease phenotype. The general idea here is that unique DNA sequences (mutations) can be linked to inheritance of phenotype and then mapped to specific chromosomes. Ultimately this analysis may lead to identifying mutated genes of known function and, theoretically, to gene or gene product replacement therapy. While this approach is applicable to singlegene diseases, it is highly suspect when applied to polygenic, multifactorial diseases.

The starting point for much of RFLP work was the analysis by Lander and Botstein 38 applied to the hypertensive rat.39 This work revealed linkage of hypertension to a mutation at the ACE (angiotensin converting enzyme) locus, a gene responsible for converting angiotensin I to angiotensin II, a protein crucial to blood pressure regulation. Subsequent work, however, showed that ACE mutation was not linked with hypertension

in humans.40,41 More recent studies provide a strong suggestion that, even for the hypertensive rat, early developmental changes will neutralize ACE mutation and provide for near normal phenotype. Thus, if young rats are taken from genetic mothers bearing the ACE mutation and nursed by normal mothers of a related strain, the pups show decreased levels of hypertension.42

Myocardial infarction in humans has also been linked to ACE mutation.41 However, in this study many individuals were identified with the identical mutation who had no heart disease. Clearly, other factors are involved. How many other genes or other factors might there be? In studies like this, the question is rarely asked. But the physiology of heart function clearly reveals that ACE-related diseases will most likely be multifactorial, polygenic entities. If so, then one expects that each of the many genes will have a small effect,43 redundancy will be present, and any one gene or even several functionally related genes may be necessary but not sufficient to precipitate a heart disease. In other words, one anticipates that in this situation genetic diagnosis will not be a robust predictor of phenotype. The environment and individual natural history will be major determining factors. In the case of angiotensin-related function, it is clear that redundant epigenetic regulation will dominate a single genetic defect. Why? It is well known that in the normal or diseased human ventricle, ACE is a minor source of angiotensin II. There are many other (gene coded) serine proteases that provide for 90% of ventricular angiotensin II levels.44

We conclude that ACE mutation will predict neither hypertension nor myocardial infarction in humans. While an ACE mutation might have some effect, at the wider physiological–nervous system level, there will be further interactional complexity and phenotypic adaptation, including central nervous system override of renin production. These and other elements of the hypertensive control network will confound simple genetic determinism. Examples include complex cortical and medullary regulation of heart and blood pressure rhythms that are exquisitely sensitive to environmental input and personal experience.45 While the use of ACE inhibitors may be a useful therapy for hypertension, an ACE screen for heart disease is not predicted to be efficient. Yet the biomedical community persists in calling for the use of an ACE gene screen to predict tendency for heart diseases and to emphasize genetic models of hypertension in general.46,47 Why?

Molecular biologists are compelled to find as much detail as possible in gene-based networks like the one for hypertension, and RFLP approaches

do provide the appropriate tool. In time, a gene map of extremely high density for this network will become available. But such maps, for each multifactorial trait or disease, will include perhaps hundreds of genes and interactive gene products all with input from the environment. The complexity of such a system can be described only as indeterminate since, when only six genes are involved in shaping a trait, the total number of phenotypes possible is over 4,000. There will be little of predictive value in the individual bits of genetic information defining this system. Rather, hypertension or other disease phenotype will be defined by the system as a whole and by the responses the system makes to the appropriate internal and external signaling pattern. The search for a screen that depends on a single or even a multiple number of genetic variants is an understandably oversimplified approach born of an optimism that all controls reside in genetic elements. But, as has been pointed out here, such optimism, realistic when applied to the rare monogenic diseases, is misplaced when applied to complex multifactorial systems such as hypertension.


This is a model of multifactorial diseases. It is the major cause of death in North America and in a number of European countries.48 Current research focus is on a hypothesis in which “response to injury” offers the most promise for understanding and perhaps control.49 The hypothesis involves a complex etiology of atherosclerosis that includes disorders of lipid metabolism, clotting, blood pressure regulation, and carbohydrate regulation. The events proximal to the disease include lipid infiltration of blood vessel walls and loss of control of intimal cell proliferation that is postulated by some to involve, in the case of restenosis following angioplasty, tumor suppressor gene (p53) inactivation.50,51 Thus, assuming that the hypothesis is correct, it remains highly unlikely that diagnosis or prediction of atherosclerosis will emerge from gene mutation analysis. Why? Mutation in p53 or related genes is identified as an end product in a long line of causal factors. Epidemiological studies reveal over 200 risk factors at work,52,53 and molecular studies suggest that hundreds of genes may be involved; as many as 200 different genes are probably active in lipid metabolism alone. Using new techniques of linkage disequilibrium,10 one may be able to detect the influences of perhaps 500 genes affecting atherosclerosis. What is the predicted benefit of measuring small effects of 500 genes? Nil. Why? Because each small effect will

display strong environmental shaping, will be open to redundancy at the gene level, and will be interactive with other genes so that cooperative and compensating (epigenetic) effects can take place, all of which makes early diagnosis or prediction of final outcome based on genetic information alone extremely difficult if not impossible.

There is, however, a role for genetic analysis in the diagnosis of complex polygenic diseases like atherosclerosis. New approaches attempt to link genomic variation with specific environmental and physiological states to predict disease phenotype. They recognize that genetic and environmental signals are strongly interactive and that few signals of either kind will exert independent effects on the determination of disease susceptibility. These approaches assume “that interactive effects are translated through quantitative variation of intermediate biological and physiological agents that link discrete genome type variations and variation in risk of disease.” 52 For example, Sing and his collaborators 54 have proposed a nonparametric statistical strategy for selecting combinations of genotypes, intermediate risk factor traits (physiological states), and environmental agents that can be associated with subsets of individuals showing a disease phenotype. One early result of this has been the strong association (odds ratio greater than 2) of high body mass index combined with unique apolipoprotein E genotypes with coronary artery disease. This strategy may be extended to many genomic, physiological, and environmental interactors and may reveal the role of genome type variation in nonlinear relationships with a great variety of other interactors.55 Approaches such as this represent an upper limit of using genetic analysis coupled with other signals in a dynamic epigenetic framework to predict disease outcome. Other examples of nonlinear approaches to predict disease susceptibility rely on dynamical measures of physiological states alone and will be discussed here.

Cancer and Mutation in Regulatory Genes

The purpose of this section is, once again, to focus on an epigenetic perspective as an alternative way of thinking about disease causality. It is not the intent to dismiss mutations as an important aspect of tumor formation. However, a significant and persistent criticism of the mutational theory of cancer remains in evidence, and it behooves us to be reminded of it as a possible missing piece of our mainstream approach to cancer detection and diagnosis. This criticism should be kept in mind as we

review examples here of conflicts in this area. Basically this criticism, drawn from different sources,56–60 states the following:

  1. The mutation of single genes of “major” importance, in itself, is insufficient to cause cancer at least in the early stages. Tumor suppressor genes or oncogenes are examples of major genes.
  2. Early stages are often reversible and display tissuewide cellular changes inconsistent with single-gene mutation causality.
  3. Early-stage changes are seen as epigenetic adaptations to environmental signals. These changes progress through intermediate states to end-stage tumors that do show many mutations that may preclude any spontaneous remission. In what follows here, discussion is restricted to polygenic cancers where there is no Mendelian segregation associated with the phenotype.

Cancer, in its multiple forms, has often been described as one of our most multifactorial and enigmatic diseases. While strong evidence exists for a genetic background, for many cancers much of this evidence is potentially confounded by congenital and familial effects; we forget that many things are inherited in addition to genes. In addition, we know that many forms of cancer have strong environmental determinants. Current research emphasis, however, is on mutation in tumor suppressor genes, which, while they will play some role in cancer, may also prove to be constrained by other factors. Here I analyze several cases and conclude that an epigenetic basis for polygenic cancer is an attractive but missing research component.

The p53 and rb Genes as Tumor Suppressors The most recent trend has been to associate unique cancers with mutation in growth control or tumor suppression genes such as p53 or retinoblastoma (rb).61,62 These genes code for DNA-binding proteins that delay or inhibit cell replication. Mutation in both alleles would then produce defective regulation of growth and tumor formation. If one of these genes is defective at birth, then one inherits a tendency or susceptibility for cancer; the disease itself is then predicted to occur when, through somatic mutation, the second allele is also defective. But it is now clear that some form of redundancy for p53 and rb is present in cells, making it difficult or impossible to use mutational analysis alone for predicting cancer. For example, a mouse has been constructed with both p53 alleles absent (homozygous

knockout).63 In this case, it was expected that growth control in all cells would be defective with dire effects for all affected individuals. However, the affected animals were all normal at birth, and early development and growth was normal. It was only after adulthood was reached that some, but not all, of these individuals developed tumors in excess of that found in control populations. Clearly, the early phases of development that depend on stringent growth controls remain independent of p53 input or have redundant pathways around p53. The same must be said of the p53 mutant adult individuals that did not show any tumor formation. Finally, when a normal p53 gene is inserted into cancer cells, it may or may not restore normal growth regulation.64

More recent evidence shows that p53 protein may form heterodimers with many other cellular proteins,65 including replication protein A, which is involved in the initial stage of DNA replication.66 Thus, p53 regulation is a prime candidate for epigenetic control in which the final effect is modulated by a complex interaction of many bits of genetic and environmental information. Much is learned about DNA replication in p53 studies, but the emerging picture shows not single-gene control of cancer but complex interactive regulation. Epigenetic interactions of p53 protein with other gene products form a basis for explaining the varied effects observed when p53 is mutated in different genetic backgrounds 63 or when wild-type p53 fails to restore normal growth regulation to p53-defective cells.62

A similar story may be told for the retinoblastoma (rb) mutation (for a review, see reference 67), which arises either spontaneously or via heredity associated with a deletion in or absence of chromosome 13 in 20% to 30% of affected cases.68,69 But 20% to 30% is not 100%, so clearly other factors are involved. Many individual rb tumors do not show mutated rb genes.70 In addition, while expression of wild-type rb in some rb-defective cells will restore normal growth,71 such transfection and expression fails to produce normal growth when these cells are transplanted to the eye of nude mice.72 Homozygous rb knockout in the mouse is lethal but only late in development after lineage determination is complete and after millions of cell replications have been completed.62 This gene, therefore, while it plays an important role in cell replication, is not essential during early development and is not sufficient to cause cancer. We must assume that epigenetic control of rb is taking place.

Finally, one of the most outstanding characteristics of neuroblastoma is spontaneous remission at 10- to 100-fold greater than that seen for

any other human cancer.73 Reversibility of tumor growth is normally thought to be inconsistent with mutation causality, especially when one eliminates from consideration the possibility of immune surveillance as the cause of remission.

A newer explanation for remission is apoptosis, but one is left with an epigenetic regulation because some aspect of cellular behavior must be presumed to signal cell death or to engage the so-called apoptosis program. Here again we become aware of the facile nature of molecular thinking on the issue of genetic programming and apoptosis.74 For example, in a recent review we read, “Within a few months of birth the programme is activated: cells then die successfully by apoptosis and the (rb) tumor shrinks.” 73 When one asks the question, “What activates the program?,” the answer is usually in terms of other genes for growth factors or DNA-binding proteins. But cells reacting to stressful signals such as x-irradiation or to a loss of normal tissue environments when explanted into cell culture will display a variety of epigenetic adaptations that might easily trigger cell death.75

Genes for Breast Cancer Human breast cancer work is heavily invested in gene mutation causality even though a large population study tells us that less than 2.5% of breast cancer is associated with genetic determinism.76 Many mutations are found in later stages of a variety of tumors. While it is likely that these play some role, it remains uncertain whether mutations are the cause or the effect of earlier, nongenetic lesions that, if reversed soon enough, would have deflected the tumors altogether.56,77

The latest focus in breast cancer has been on a familial study where linkage has been established for “cancer tendency” to a locus on chromosome 17q21.78 The lod score (see the glossary at the end of this chapter) for linkage was 5.98, well in the range to ensure a high probability of association between the cancer and the genetic anomaly. While the technology of this linkage study may be assumed to be state of the art, we must also be aware of its problems. For example, we do not know what the frequency of this mutation might be in the general population, nor do we know the extent to which other mutations might be present in the suspect or other chromosomes of the affected women. Nor do we know the pleiotropic and epistatic effects that other genes might have in altering the penetrance of the suspected mutation. These are all questions of fundamental importance in elementary genetics.43 We also have

no knowledge of what environmental influences might be required in order for tumor formation to occur in the presence of the mutation. The mutation may be necessary but not sufficient; it may require specific environments or specific other genetic background. Finally, as mentioned at the outset of this discussion, we also know that while there is some relationship of family history to breast cancer, only 2.5% of breast tumors are genetic in origin.76

This breast cancer gene, BRCA1—a putative tumor suppressor gene—has now been isolated.79 BRCA1 germ line mutations are linked to breast and ovarian cancer in a number of small families having multiple cases of these diseases. Women carrying mutant BRCA1 alleles have a significant increase of breast and ovarian cancer so this finding may prove to be extremely useful. The search for the breast cancer gene had been accompanied by an unnecessary hyperbolic publicity in both mass media and scientific press where an anxious expectation was developed that once the gene was found, we would be provided with important new clues for all breast cancer and perhaps for cancer in general. High expectations like this have proved mostly unfounded in the history of cancer research and serve only to frustrate public opinion and undermine confidence in research. It was sobering for many, therefore, to find that in nonfamilial breast and ovarian cancers, constituting more than 95% of the cases, BRCA1 mutations were not involved.80 Thus, this mutation “appears to play no role in common, nonhereditary forms of breast cancer that strike about 173,000 women in the U.S. each year—a finding that undermines some long-held assumptions about how the gene works.” 81

There has also been much attention paid to another tumor suppressor gene, CDKN2, that codes for the cell cycle regulatory protein cyclindependent kinase-4 inhibitor (p16).82 This (mutant) gene has been linked to a variety of tumors and has been a prime suspect for breast cancer. However, a recent report has now examined human breast carcinomas for mutations of this gene with negative results. Evidently, p16 is not involved in the formation of primary breast carcinoma.83 In addition, it has been found that p16 mutations are found in cell lines derived from many tumors but not in primary tumors within the patients, making it clear that so-called carcinogenic mutations may be a pure artifact of cell culture.83 The studies on p16 mutation are consistent with a hypothesis of cancer where early neoplastic change is an epigenetic one that includes mutation as an event after the fact of initiation.



Imaging Techniques

Computed tomography and magnetic resonance imaging have become widespread and extremely expensive additions to diagnosis. The problems inherent in imaging techniques have been recently analyzed 84 and are discussed briefly here as prologue to similar problems turning up with molecular measurements that, while extremely sensitive, are also without proven meaning when applied to disease manifestation.

Imaging techniques, because they are so sensitive, often measure not disease itself but early changes in tissue that are taken as evidence that disease will develop. Early changes may, however, be extremely misleading since they often reflect reversible processes or those with extremely long lag time to any clinical manifestation. As our machines are able to detect the most incipient stages, we experience several problems.84 First, as exemplified by thyroid cancer, is the problem of defining diseases as cellular changes that always progress to serious morbidity. In this case, clinical cancer (tumor size greater than 2 centimeters) is seen in only 0.1% of adults between the ages of 50 to 70 years. However, autopsy using increasingly thin sections of the gland could reveal at least one papillary carcinoma in 36% of adults. It was calculated that as sections became thinner, autopsy would show verifiable papillary cancer in 100% of cases. These “tumors” discovered at earliest stage represent an enormous reservoir of detectable but subclinical disease. Under these circumstances and for a variety of diseases, the patient may never experience clinical symptoms but, under aggressive medical management, may become involved in unnecessary and expensive medical procedures that are predicted to have little positive effect.84

The second major problem that arises has to do with the effect on reported disease frequency where frequency increases as the degree of measurement sensitivity increases. However, without any manifestation, early stage diagnosis makes it appear as if we are experiencing large increases in the disease itself. The third problem is the effect of statistical evaluations of various therapies for a disease. As the time between diagnosis and manifestation increases, it is made to appear as if various therapies are working even when nothing in the way of treatment need be involved in the statistical analysis.84


Antigen and Nucleic Acid Sequence Measurement

Nowhere in medical technology have we greater sensitivity of measurement than in antigen and nucleic acid chemistry. The possibility exists, however, that these measurements are often without predictive value for the diseases for which their measurement was designed. Increased levels of scrutiny can, for example, explain recent reported increased prevalence in breast, prostate, and thyroid cancer.84 Prostate antigen testing, together with other evaluation, may prove useful. However, these tests, used alone, can provide for an enormous increase in reported prevalence and an increased apparent time of survival and, unless carefully applied, could lead to unnecessary treatment.

Polymerase chain reaction (PCR) is a technology used extensively to report “viral loads” in patients and has replaced measurements of actual infective units of pathogen. What PCR does is to amplify by thousands of times sequences of DNA that are present in body fluids at extremely dilute concentration. For example, it has proved to be crucial for diagnosis of HIV-related AIDS, where it is used to amplify vanishingly small amounts of parts of HIV sequences. But such partial sequences tell us nothing about the presence of “live” virus and could be reporting on DNA fragments resulting from a variety of sources (e.g., viral degradation). In addition, PCR is notoriously difficult to quantitate, and recent publications fail to publish standards that might show actual viral numbers. As Nobelest Kary Mullis, the inventor of PCR, himself states in a paper that was refused publication, “The vice of the PCR is that it can find the biochemical equivalent of the needle in the haystack. Viral fragments that are present only in minute quantities can be amplified and identified, but this tells us nothing about whether replicating virus is present in sufficient quantities to do harm.” 85 The HIV-AIDS hypothesis remains plagued by the fact that most AIDS patients, until end-stage disease, rarely show HIV viremia (classically defined by actual virus replication), and diagnosis continues to rely on PCR and antibody measurements. In HIV and other infectious diseases, we may have abandoned, at great scientific cost, traditional rules for establishing disease causality.86 For example, HIV, as many contend, may be caused by agents other than or in addition to HIV. But reliance on PCR tends to conceal the absence of viral units and to conceal the possibility that other causes may be involved.



The conflicts described here arise from attempts to apply a strict geneticreduction analysis to the problem of complex phenotypes. Why has it been so difficult to use single-gene markers to predict polygenic cancer in individuals even when a mutation is present in an important tumor suppressor gene (rb or p53)? Why are there so many false positives and false negatives in predicting heart disease when there is a measurable defect in an important gene (ACE) involved in blood pressure regulation? Finally, why is it that the elimination (homozygous knockout) of both copies of a gene like p53—known to be important for controlled growth of cultured cells—fails to provoke any measurable defects in normal development? On the basis of a linear genetic analysis, these failed predictions should have been viewed as evidence of a failed cancer theory. However, the response from biotechnology has been to rescue the genetic theory in the face of these and many other exceptions to the rule that unique genes have predictable unique effects. Why this response should persistently come up has been treated elsewhere as part of a paradigm shift 87 in biology and medicine.88

The short answers, to the questions of lack of predictive power of gene analysis and of why we have thrown out the facts rather than the theory, are not too difficult. The explanation formulated here is that polygenic disease and growth regulation are not linear processes and cannot therefore be fully analyzed by a linear logic. Rather, they are representatives of complex adaptive systems that are innately unpredictable. To understand the unpredictable nature and other features of such systems, it will be necessary to develop a disease theory similar to what is found in treating nonlinear phenomena.89

What is needed to supplement genetic theory is a theory of biological complex adaptive systems.90 If living systems were seen as complex adaptive systems, then we would not be surprised at the previously mentioned failure of prediction. In fact, they would be expected since such systems are unpredictable and actually seek out alternative pathways when perturbed by new information, either from the outside world or from within. The results of alternative pathway selection would include epigenetic change in the genome and resulting change in pattern of gene expression. In the case of mutation, however, we need to remember that it is taking place within an epigenetic framework. As explained by James Shapiro,91 part of the adaptive response of bacterial cells to stress is

the activation of a whole family of enzymes whose job is to remodel the genome, a job that will predictably increase the frequency of mutation. But in this case the mutation would not necessarily be the result of random process but part of an epigenetic response process that has become maladaptive through chronic stress. One might begin the merger of genetic reductionism and epigenetic complexity with those areas where multigenic systems are known to be coordinated by higherorder cellular responses to environmental conditions. Nobel laureate Barbara McClintock, who described mobile genetic elements long before they were discovered by molecular biology, had always been preoccupied with mechanisms that rapidly reorganize the genome. In one of her last reviews, she wrote of the significance of responses of the genome to challenge. She ended that article as follows: “We know about the components of genomes. … We know nothing, however, about how the cell senses danger and initiates responses to it that often are truly remarkable.” 92

Statements concerning cells “sensing danger” and “initiating responses” strike biologists today more as poetic meanderings than as statements with scientific value. But clearly cells and multicellular organisms display these holistic behaviors. Natural selection operates not merely at a genetic level but at all levels of biological organization,93,94 including whole-cell and organismal behavior. Much evidence exists for the idea that cells do sense danger and respond in a manner that is not explored by reductionistic thinking. For example, when tissues are exposed to X rays, there is a tissuewide, or “field,” response in the cellular population as a whole, a response that is quite separate from gene mutation but that actually induces persistent hypersensitivity to future mutation.95 Cells exposed to the stress of removal from normal tissue constraints in vivo when explanted to cell culture adopt many morphological changes that are heritable and that, when continued over a period of time, give rise to transformation and to mutation.96 The flip side here is the demonstrated ability of liver architecture to constrain tumor growth in an age-dependent fashion. Tumor cells explanted into young rat livers generate tumors at a much lower frequency compared to that seen when these cells are transplanted into older livers.97 Findings like this are difficult to reconcile with singe gene mutation causality but are predicted by an epigenetic theory of cancer that locates control of single cell growth in higher (tissue and organ) levels of biological organization.

Perhaps the most important new insight into cancer is the one having

to do with the source of the many (hundreds and thousands of) mutations that may be found in human tumor cells.98 Cancer cells in culture show a hypermutation phenotype but, surprisingly, not under conditions of rapid growth. Only under conditions of nutritional stress or contact inhibition do these cells recruit an epigenetic response capable of “reorganizing the genome” through increased mutation rates. The result of this reorganization is the production of useful mutations (useful from the cancer cell point of view) allowing cancer cells to escape the various constraints that normally contain tumor expansion and metastasis. Single-gene mutation would be amply buffered by epigenetic mechanisms, including redundancy, but with stress producing 1,000- to 5,000-fold increases in mutation rates, the “cause” has to be located in the epigenetic mechanisms responsible for this extraordinarily high rate.

One may conclude that (1) higher-level epigenetic management and constraint on tumor growth delays clinical cancer and may even reverse it and that (2) epigenetic mechanisms at the intracellular level are responsible for generating increased mutation rates necessary for escape of these cells from the higher-level constraints. In other words, cancer is a cellular, epigenetic disease and not the result of single-gene mutation.

An approach to complex analysis of heart disease with multigenetic causality linked to interactive environments is the work of Sing and his group as mentioned previously (see the section “Atherosclerosis”). At these levels above the cell, for complex physiological systems, chaos theory builds on epigenetic thinking and already is providing new ways to think about complex systems.8–10 This is particularly true for cardiac function, where sinus arrhythmia, long thought to be low-level noise or random fluctuation in heart rate, is now seen as high-order chaos.99 Coupling of heart rate to brain function and thus to experience has long been appreciated as an observable patterned occurrence but was mostly inexplicable through standard physiological experiment.100 Chaos theory is an old story in physiology in general 101 and is able to provide a method of revealing generic patterns in what was thought to be random variation. Recognition of these patterns allows new insights into brainheart physiology and may even allow prediction of sudden cardiac death among patients at risk.99


This chapter draws much from previously published materials by the author. See references section for the exact citations.



Different forms of the same gene.
A protein recognized by the immune system.
Programmed cell death.
A technology used to scan whole bodies for diagnostic purposes.
All genomes, including the human genome, tend to remain constant over long periods of time because of the presence in cells of formidable DNA repair systems. Mutations, for example, occur, but most are repaired.
General term for that part of the cell surrounding the nucleus.
Generally refers to the inheritance of factors and processes that are in addition to genes. Also refers to changes in the genome that do not involve sequence changes in DNA.
Interaction between genes.
When the two alleles are different.
all genes come in allelic pairs; “homozygous” refers to cases in which both alleles are the same.
An international effort to identify every gene in the total of 70,000 to 100,000 genes thought to be present in all humans.
A linear or direct representation of one thing by another.
A technical term used to indicate linkage of a gene to a phenotype.
Defined by inheritance pattern when studied in families.
An epigenetic change involving addition of a chemical group (methyl group) to DNA, thus changing gene expression without changing DNA sequences within the gene.
A phenotype (disease) said to be caused by a single gene mutation.
Polymerase chain reaction; a technique that measures extremely small samples of DNA.
What the organism looks and behaves like; its morphology.
A gene or a protein having many effects.
A phenotype shaped by many genes acting in concert.
Characteristic of old age.
A disease (cancer) of the retina.
An experimental procedure in which genes are randomly made mutant.


1. Strohman, R. C.1994. “Epigenesis: The missing beat in biotechnology?” Bio/Technology12, 156–164.


2. Polanyi, M.1968. “Life's irreducible structure.” Science160, 1308–1312.

3. Nurse, P.1997. “The ends of understanding.” Nature387, 657.

4. Williams, N.1997. “Biologists cut reductionist approaches down to size.” Science277, 476–477.

5. Miklos, G. L. G.1993. “Emergence of organizational complexities during metazoanevolution: Perspectives from molecular biology, palaeontology and neo-Darwinism.” Memoirs of the Australasian Association of Palaeontology15, 7–41.

6. Miklos, G. L. G., and Rubin, G.1996. “The role of the Genome Project in determining gene function: Insights from model organisms.” Cell86, 521–529.

7. Goodwin, B. C.1994. How the Leopard Changed Its Spots: The Evolution of Complexity.New York: Scribner's Sons.

8. Webster, G., and Goodwin, B.1997. Form and Transformation.Cambridge: Cambridge University Press.

9. Golderberger, A.1996. “Non-linear dynamics for clinicians: Chaos theory, fractals, and complexity at the bedside.” The Lancet347, 1312–1314.

10. Schipper, H., Turley, E. A., and Baum, M.1996. “A new biological framework for cancer research.” The Lancet348, 1149–1151.

11. Golub, E. S.1997. The Limits of Medicine.Chicago: University of Chicago Press.

12. Weiss, K. M.1995. Genetic Variation and Human Disease.Cambridge: Cambridge University Press.

13. Sing, C. F., Haviland, M. B., and Reilly, S. L.1996. “Genetic architecture of common multifactorial diseases.” Ciba Foundation Symposium197, 211–232.

14. Weatherall, D. J.1982. The New Genetics and Clinical Practice.London: Nuffield Provincial Hospitals Trust.

15. Strohman, R. C.1993. “Ancient genomes, wise bodies, unhealthy people: Limits of genetic thinking in biology and medicine.” Perspectives in Biology and Medicine37 (1), 112–145.

16. Rose, M. R.1991. Evolutionary Biology of Aging.Oxford: Oxford University Press.

17. Finch, C. E.1990. Longevity, Senescence, and the Genome.Chicago: University of Chicago Press.

18. Fries, J. F., and Crapo, L. M.1981. Vitality and Aging: Implications of the Rectangular Curve.New York: W. H. Freeman.

19. Tsai, S. P., Lee, E. S., and Hardy, R. J.1978. “The effect of a reduction in leading causes of death: Potential gains in life expectancy.” American Journal of Public Health68, 966–971.

20. Olshanky, S. J., Carnes, B. A., and Cassel, C.1990. “In search of Methuselah: Estimating the upper limits to human longevity. Science” 250, 634–640.

21. Lander, E. S., and Schork, N. J.1994. “Genetic dissection of complex traits.” Science265, 2037–2048.

22. Plomin, R., Defries, J. C., and McClearn, G. E.1990. Behavioral Genetics. 2d ed. New York: W. H. Freeman.

23. Goodwin, B. C.1985. “What are the causes of morphogenesis?” Bioessays3, 32–36.


24. Nijhout, H. F.1990. “Metaphors and the role of genes in development.” Bioessays12, 441–446.

25. Hood, L.1992. In The Code of Codes. Edited by D. J. Kevles and L. Hood. Cambridge, Mass.: Harvard University Press.

26. Casky, T.1992. In The Code of Codes. Edited by D. J. Kevles and L. Hood. Cambridge, Mass.: Harvard University Press.

27. Wilkins, A. S.1993. Genetic Analysis Of Animal Development. 2nd ed. New York: Wiley-Liss.

28. Brenner, S., Dove, W., Herskowitz, I., and Thomas, R.1990. “Genes and development: Molecular and logical themes.” Genetics126, 479–486.

29. Tautz, D.1992. “Redundancies, development and the flow of information.” Bioessays14, 263–266.

30. Jablonka, E., and Lamb, M. J.1995. Epigenetic Inheritance and Evolution.New York: Oxford University Press.

31. Holliday, R.1990. Philosophical Transactions of the Royal Society of LondonB326, 329–338.

32. Provine, W. B.1971. The Origins of Theoretical Population Genetics.Chicago: University of Chicago Press.

33. McKusick, V. A.1964. Human Genetics.Englewood Cliffs, N. J.: Prentice Hall.

34. Yoxen, E. J.1984. “Constructing genetic diseases.” In Cultural Perspectives on Biological Knowledge, edited by T. Duster and K. Garrett. Norwood, N. J.: Ablex.

35. Wahlsten, D.1990. “Insensitivity of the analysis of variance to heredityenvironment interaction.” Behavior and Brain Science13, 109–161.

36. McKeown, T.1979. The Role of Medicine: Dream, Mirage or Nemesis?Princeton, N. J.: Princeton University Press.

37. McKeown, T.1988. The Origins of Human Disease.New York: Basil Blackwell.

38. Lander, E. S., and Botstein, D.1989. “Mapping Mendelian factors underlying quantitative traits using Rflp maps.” Genetics121, 185–199.

39. Jacob, H. J., Linkpaintner, K., Lincoln, S. E., et al. 1991. “Genetic mapping of a gene causing hypertension in the stroke-prone spontaneously hypertensive rat.” Cell67, 213–224.

40. Jeunemaitre, X., Lifton, R. P., Hunt, S. C., et al. 1992. “Absence of linkage between the angiotensin converting enzyme locus and human essential hypertension.” Nature Genetics1, 72–75.

41. Cambien, F., Poirier, O., Lecref, L., et al. 1992. “Deletion polymorphism in the gene for angiotensin-converting enzyme is a potent risk factor for myocardial infarction.” Nature359, 641–644.

42. Myers, M. M., Brunelli, S. A., Squire, J. M., et al. 1989. “Maternal behavior of shr rats and its relationship to offspring blood pressures.” Developmental Psychobiology22 (1), 29–53.

43. Ayala, F. J., and Kiger, J. A., Jr.1984. Modern Genetics. 2nd ed. Menlo Park, Calif.: Benjamin/Cummings.

44. Urata, H., Healy, B., Stewart, R. W., et al. 1990. “Angiotensin: Ii-forming

pathways in normal and failing human hearts.” Circulation Research66, 883–890.

45. Peterson, L. H.1972. In Neural and Psychological Mechanisms in Cardio vascular Disease.Milan: Casa Editrice.

46. Kurtz, T. W.1992. “The ace of hearts.” Nature359, 588–589.

47. Kurtz, T. W.1992. “Genetic models of hypertension.” The Lancet344, 167–168.

48. Higgins, M., and Higgins, R. V.1989. “Trends and determinants of coronary heart disease mortality: International comparisons.” International Journal of Epidemiology18, 3–12.

49. Ross, R.1993. “The pathogenesis of atherosclerosis.” Nature, 362, 801–809.

50. Speir, E., Modali, R., Huang, E.-S., et al. 1994. “Potential role of human cytomegalovirus and p53 interaction in coronary restenosis.” Science265, 391–394.

51. Marx, J.1994. “Cmv-p53 interaction may help explain clogged arteries.” Science265, 320.

52. Davignon, J., and Roy, M.1993. “Familial hypercholesterolemia in French Canadians: Taking advantage of the presence of a founder effect.” American Journal of Cardiology72, 6d–10d.

53. Davignon, J., Dufour, R., and Cantin, M.1983. “Atherosclerosis and hypertension.” In Hypertension, edited by J. Genest, O. Kuchel, P. Hamet, and M. Cantin. New York: McGraw-Hill.

54. Sing, C. F., Haviland, M. B., Templeton, A. R., and Reilly, S. L.1994. “Alternative genetic strategies for predicting risk of atherosclerosis.” In Proceedings of the Xth Atherosclerosis Symposium. “Excerpta Medica International Congress Series.” Amsterdam: Elsevier Science Publishers.

55. Sing, C. F., Haviland, M. B., Templeton, A. R., et al. 1992. “Biological complexity and strategies for finding DNA variations responsible for interindividual variation in risk of a common chronic disease, coronary artery disease.” Annals of Medicine24, 539–547.

56. Prehn, R. T.1994. “Cancers beget mutations versus mutations beget cancers.” Cancer Research54, 5296–5300.

57. Rubin, H.1990. “On the nature of enduring modifications induced in cells and organisms.” Am. Physio. Soc.L19–L24.

58. Farber, E., and Rubin, H.1991. “Cellular adaptation in the development of cancer.” Cancer Research51, 2751–2761.

59. Lijinsky, W.1989. “A view of the relation between carcinogenesis and mutagenesis.” Environmental and Molecular Mutagenesis14 (16), 78–84.

60. Clark, W. H.1994 “What is inherited in neoplastic systems? Animal models of cutaneous malignant melanoma.” Laboratory Investigation71, 1–4.

61. Levine, A. J., Momand, J., and Finlay, C. A.1991. “The p53 tumour suppressor gene.” Nature351, 453–456.

62. Jacks, T., Faneli, A., Schmitt, E. M., et al. 1992. “Effects of an rb mutation in the mouse.” Nature359, 295–300.

63. Donehower, L. A., Harvey, M., Slagle, B. L., et al. 1992. “Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumors.” Nature356, 215–221.


64. Baker, S. J., Markowitz, S., Fearon, E. R., et al. 1990. “Suppression of human colorectal carcinoma cell growth by wild-type p53.” Science249, 912–915.

65. Pietenpol, J. A., and Vogelstein, B.1993. No room at the p53 inn.Nature356, 17–18.

66. Dutta, A., Ruppert, J. M., Aster, J. C., and Winchester, E.1993. “Inhibition of DNA replication factor rpa by p53.” Nature365, 79–82.

67. Dowdy, S. F., Hinds, P. W., Louie, K., et al. 1993. “Physical interaction of the retinoblastoma protein with human D cyclins.” Cell73, 499–511.

68. Knudson, A.1985. “Hereditary cancer, oncogenes, and antioncogenes.” Cancer Research45, 1437–1443.

69. Benedict, W. F., Banetjee, A., Mark, C., and Murphee, A.1983. “Nonrandom chromosomal changes in untreated retinoblastoma.” Cancer Genetics and Cytogenetics10, 311–333.

70. Gardner, H. A., Gallie, B. L., Knight, L. A., and Phillips, R. A.1982. “Multiple karyotypic changes in retinoblastoma tumor cells: Presence of normal chromosome no. 13 in most tumors.” Cancer Genetics and Cyotgenetics6, 201–211.

71. Bookstein, R., Shew, J. Y., Chen, P. L., et al. 1990. “Suppression of tumorigenicity of human prostate carcinoma cells by replacing a mutated rb gene.” Science247, 712–715.

72. Xu, H. J., Sumegi, J., Hu, S. X., et al. 1991. “Intraocular tumor function of rb reconstituted retinoblastoma cells.” Cancer Research51, 4481–4485.

73. Pritchard, J., and Hickman, J. A.1994. “Why does stage 4s neuroblastoma regress spontaneously?” The Lancet344, 869–870.

74. Farber, E.1994. “Programmed cell death: Necrosis versus apoptosis.” Modern Pathology7, 605–609.

75. Rubin, H.1994. “Epigenetic nature of neoplastic transformation.” In Devel opmental Biology and Cancer, edited by G. M. Hodges and C. Rowlatt. Boca Raton, Fla.: CRC Press.

76. Colditz, G. A., Willett, W. C., Hunter, D. J., et al. 1993. “Family history, age, and risk of breast cancer.” Journal of the American Medical Association270 (3), 338–343.

77. Rubin, H.1985. “Cancer as a dynamic developmental disease.” Cancer Research45, 2935–2942.

78. Hall, J. M., Lee, M. K., Newman, B., et al. 1990. “Linkage of early-onset familial breast cancer to chromosome 17q21.” Science250, 1684–1689.

79. Miki, Y., Swenson, J., Shattuck-Eidens, D., et al. 1994. “A strong candidate for the breast cancer and ovarian cancer susceptibility gene Brca1.” Science266, 66–71.

80. Futreal, P. A., Liu, Q., Shattuck-Eidens, D., et al. 1994. “Brca1 mutations in primary breast and ovarian carcinomas.” Science266, 120–121.

81. Nowak, R.1994. “Breast cancer gene offers surprises.” Science265, 1796–1799.

82. Kamb, A., Nelleke, A. G., Weaver-Feldhaus, J., et al. 1994. “A cell cycle regulator potentially involved in genesis of many tumor types.” Science264, 436–439.

83. Xu, L., Sgroi, D., Sterner, C. J., et al. 1944. “Mutational analysis of Cdkn2

(Mts1/P16ink4) in human breast carcinomas.” Cancer Research54, 5262–5264.

84. Black, W. C., and Welch, H. G.1993. “Advances in diagnostic imaging and overestimation of disease prevalence and the benefits of therapy.” New England Journal of Medicine328 (17), 1237–1243.

85. Thomas, C., Mullis, K. B., Ellison, B. J., and Johnson, P.1993. “Why there is still an HIV controversy.” Unpublished manuscript.

86. Duesberg, P. H.1992. “Aids acquired by drug consumption and other noncontagious risk factors.” Pharmacology and Therapeutics55, 201–277.

87. Kuhn, T.1996. The Structure of Scientific Revolutions. 3rd ed. Chicago: University of Chicago Press.

88. Strohman, R. C.1997. “The coming Kuhnian revolution in biology.” Nature Biotechnology15, 194–200.

89. Waldrop, M. M.1994. Complexity.New York: Simon & Schuster.

90. Gell-Mann, M.1994. The Quark and the Jaguar.New York: W. H. Freeman.

91. Shapiro, J.1992. “Natural genetic engineering in evolution.” Genetica86, 99–111.

92. McClintock, B.1984. “The significance of responses of the genome to challenge.” Science226, 792–801.

93. Gould, S. J.1993. In The Logic of Life, edited by C. A. R. Boyd and D. Noble. Oxford: Oxford University Press.

94. Kaufmann, S.1993. The Origins of Order.New York: Oxford University Press.

95. Frank, J. P., and Williams, J. R.1982. “X-ray induction of persistent hypersensitivity to mutation.” Science216, 307–308.

96. Rubin, H.1993. “Cellular epigenetics: Effects of passage history on competence of cells for “spontaneous” transformation.” Proceedings of the National Academy of Sciences USA90, 10715–10719.

97. McCullough, K. D., Coleman, W. B., Smith, G. J., and Grisham, J. W.1997. “Age dependent induction of hepatic tumor regression by the tissue microenvironment after transplantation of neoplastically transformed rat epithelial cells into the liver.” Cancer Research.

98. Richards, B., Zhang, H., Phear, G., and Meuth, M.1997. “Conditional mutator phenotypes in hMSH2–deficient tumor cell lines.” Science277, 1523–1525.

99. Skinner, J. E., Molnar, M., Vybiral, T., and Mitra, M.1992. “Application of chaos theory to biology and medicine.” Integrative Physiological and Behavioral Science27, 39–53.

100. Bond, W. C., Bohs, C., Ebey, J., and Wolf, S.1973. “Rhythmic heart rate variability (sinus arrhythmia) related to stages of sleep.” Conditional Reflex8 (2), 98–107.

101. Rossler, O. E., and Rossler, R.1994. Integrative Physiological and Behavioral Science29 (3), 328–333.




Part 2 presents an array of novel approaches to human wellness promotion research and focuses attention on specific topic areas selected to exemplify how innovative methods may illuminate new solutions. Stokols (chapter 6) focuses on environmentally oriented approaches to disease prevention and health promotion. Using many examples to illustrate the potential for enhancing health by modifying the physical environment, he argues that this approach should be more consistently incorporated into wellness program planning and implementation. Sanders-Phillips (chapter 11) directs her attention to another aspect of the environment, namely, the effect of living in a community characterized by frequent violence. She points out that individuals, frequently minorities, who reside in neighborhoods where violence in both the home and the street is common face particular barriers to wellness promotion. On a more micro level, East (chapter 8) presents data that suggest that younger sisters of childbearing teens may be at high risk for becoming teenage mothers themselves. She suggests that the home environment may both directly and indirectly predispose these younger sisters to become pregnant and calls for interventions that target this at-risk population. In an interesting examination of pregnancy outcomes among immigrants, Guendelman (chapter 9) exposes the finding that despite their typically “disadvantaged” profile, certain immigrant groups actually have better pregnancy outcomes that native-born Americans. She explains this finding largely in terms of the cultural environment that these

immigrants bring with them and shows how their relatively positive pregnancy outcomes tend to erode with acculturation to the typical lifestyle of native-born Americans.

In the last three chapters of part 2, the authors argue that innovative strategies of data collection and data analysis can point to important revelations in evaluating wellness promotion strategies. Roach (chapter 10) turns attention to the differences in cancer epidemiology between Whites and Blacks. He points out that preexisting societal expectations influence programs, policies, and research. These expectations frame the approach to data collection, analysis, and interpretation in such a way as to make certain conclusions more likely and may perpetuate the status quo with respect to health promotion. Similarly, Ganiats and Sieber (chapter 12) show how attachment to conventional economic analysis leads to a systematic underestimation of the value of health promotion programs. The authors provide examples of how the practice of discounting future outcomes may influence both health policy and individual behavior and call for wider incorporation of time preference as a factor in models of decision making. Finally, Birckmayer and Weiss (chapter 7) describe the use of theory-based evaluation (TBE) as a tool in wellness promotion. Compared to standard program evaluation, in which the focus is on program outcomes, TBE emphasizes the evaluation of intermediate outcomes (e.g., indicators of program implementation and mediators of behavior change) and encourages evaluators to address not only whether an intervention is effective but also why or why not. As a group, the chapters in part 2 make a case for challenging the status quo and looking beneath the surface to discover the underlying dynamics of the social and health problems that wellness professionals typically work to alleviate.



Implications for Theory and Research

Daniel Stokols

Chapter 1 by Stokols and chapter 2 by Breslow suggest that social ecological perspectives are becoming increasingly influential as a basis for wellness promotion research, practice, and policy. One indication of this trend is the growing emphasis in research and intervention programs on linking individually focused, small-group–organizational, and community environmental approaches to wellness promotion (cf. Breslow, 1996; McLeroy et al., 1988; Stokols et al., 1996; Weiss, 1991; Winett et al., 1989). At the same time, however, the delineation of specific environmental leverage points for wellness promotion at each level of analysis remains an important task. This chapter gives particular attention to those sociocultural and physical-environmental qualities of organizations, institutions, and community settings that are especially health promotive. To address these issues, a new unit of analysis for wellness promotion research, practice, and policy is proposed: the healthpromotive (or wellness-promotive) environment. This unit of analysis highlights the interdependencies that exist among sociocultural, political, economic, spatial, and technological features of environmental settings, ranging from homes, neighborhoods, workplaces, and schools to regional and global environments that influence personal and collective well-being.



For the most part, health promotion research has focused on identifying and modifying personal behaviors that enhance physical health and reduce the risk of illness (e.g., Belloc and Breslow, 1972; Cataldo and Coates, 1986; Green, 1984; O'Donnell and Ainsworth, 1984). Examples of health-promotive behaviors are maintaining high-fiber/low-fat diets, engaging in regular aerobic exercise, using vehicle safety belts, refraining from smoking, and avoiding excessive alcohol consumption. From an ecological perspective, however, health promotion is viewed not only in terms of the specific health behaviors enacted by individuals but more broadly as a dynamic transaction between individuals, groups, and their sociophysical milieu. The social ecological approach to health promotion requires explicit analysis of the interplay between environmental resources available in an area and the particular health habits and lifestyles of the people who occupy the area (Lindheim and Syme, 1983).

As a starting point for analyzing transactions between environmental qualities, behavioral patterns, and health outcomes, it is first necessary to specify features of the environment that promote personal and collective well-being. Some suggested dimensions and criteria of health promotive environments are listed in Table 6.1, which offers a preliminary portrait of health-promotive environments and reflects certain of the previously mentioned assumptions associated with the ecological perspective on health promotion.

A basic assumption underlying the ecological perspective is that healthfulness is a multifaceted phenomenon, encompassing physical health, emotional well-being, and social cohesion. Accordingly, these different facets of healthfulness are presented in the three rows of Table 6.1, ranging from individually oriented assessments of physiological health to organizational and community-level analyses of social cohesion and health status. Explicit recognition of the multiple facets of healthfulness has important implications for ecologically oriented analyses of health promotion. For example, because environments can influence personal and collective well-being along several different “paths,” the health-promotive capacity of an environment must be defined in terms of the multiple health outcomes resulting from peopleenvironment transactions over a specified time interval. Thus, for any environmental context of behavior, it becomes important to specify key environmental resources or constraints likely to influence personal and collective well-being among members of the setting.

Resources and
Resources and
Psychological, and
Physiological Outcomes
Facets of Health fulness
Injury-resistant design,
economically sound
design, physically comfortable,
nontoxic and
Physiologic health, absence
of illness symptoms
and injury, perceived
comfort, genetic
and reproductive health
Mental and
Environmental controllability
and predictability,
novelty and challenge,
nondistracting, aesthetic
qualities, symbolic
/spiritual elements
Sense of personal competence,
challenge, and
fulfillment; developmental
growth; minimal
experience of emotional
distress; strong
sense of personal identity
and creativity; feelings
of attachment to
one's physical and social
Dimensions of Well-being
Social Cohesion
at Organizational
and Community
Availability of social
support networks, participatory
design and
management processes,
organizational flexibility
and responsiveness,
economic stability,
low potential for inter-group
conflict, health-promotive
media and
High levels of social contact
and cooperation,
high levels of commitment
to and satisfaction
with organization and/
or community, productivity
and innovation
at organizational and
community levels, high
levels of perceived quality
of life, prevalence
of health-promotive,
injury-preventive, and
environmentally protective

The second column in Table 6.1 lists various environmental resources that can exert a positive influence on individual and group well-being, from microlevel features of the physical environment (e.g., ergonomically sound and injury-resistant design, absence of toxic substances) to more molar or composite aspects of the sociophysical milieu (e.g., presence of pro-social environmental symbols, positive social climate, organizational programs and media to encourage health-promotive behaviors). The third column in Table 6.1 includes several behavioral,

psychological, and physiological indices that can be used to assess health outcomes of people-environment transactions at different levels of analysis (e.g., absence of physiological disorders and illness symptoms; personal feelings of competence, creativity, and commitment; high levels of job satisfaction and perceived quality of work life within organizational settings).

By firmly linking the analysis of health promotion to multiple dimensions of the environment and correspondingly diverse indices of health, some important issues for future research and community intervention are raised. First, whereas scientific research on behavior change strategies and environmental protection programs generally have remained separate, the proposed ecological view of health promotion suggests the efficacy of combining these perspectives in the design and management of environmental settings (see also Geller et al., 1982, for a behavioral approach to the design of environmental protection programs). For example, automobile manufacturers can enable individuals to reduce their risk of serious injury from car crashes by installing air bags and safety belts in their vehicles. Similarly, environmental designers, facility managers, and urban planners can incorporate a variety of physical features within new or renovated settings to promote healthfulness, including the installation of physical fitness facilities on-site or adjacent to the setting to encourage health-promotive exercise regimens among occupants of the area, the specification of ergonomically sound and injury-resistant materials in the design and construction of the setting to reduce occupants' risk of injury, and the avoidance of toxic materials and potential sources of psychosocial stress (e.g., poor lighting and air conditioning systems in buildings, insufficient shielding from noise and other distractions) to minimize environmentally induced illness and discomfort.

Given the diversity of environmental conditions present in most settings, it is likely that the relationships between those conditions and multiple health indices will be quite varied and sometimes discordant. For example, the potential health benefits of a well-designed physical environment may go unrealized if the interpersonal or intergroup relationships within a setting are chronically conflicted and stressful. On the other hand, a socially supportive family or organization may enable setting members to cope more effectively with physical constraints (e.g., high spatial density, aesthetically drab surroundings, resource shortages), thereby avoiding the negative behavioral and health outcomes sometimes associated with those conditions. These examples highlight the importance of examining both physical and social dimensions of

health-promotive or health-impairing environments and their joint influence on personal and collective well-being.

Similarly, several studies suggest that when environments are personally controllable and predictable, individuals' physical and emotional well-being are enhanced (e.g., Cohen et al., 1986; Gardell and Johansson, 1981; Glass and Singer, 1972; Karasek and Theorell, 1990; Sauter et al., 1989). However, to the extent that environments are too predictable and controllable, they can become so boring and unchallenging that they constrain opportunities for coping creatively with novel situations, thereby impeding developmental growth (Aldwin and Stokols, 1988; Kaplan, 1983). Thus, the same qualitative dimensions of an environment (e.g., its controllability and predictability) can be associated with contradictory health effects, depending on their magnitude (e.g., moderate vs. excessive levels of predictability) and duration (e.g., chronic vs. short-term exposure to unpredictable or excessively predictable situations).

Just as environmental conditions can vary in their magnitude and duration, health outcomes differ on these dimensions as well. For example, carcinogenic substances present in an environment may remain invisible and undetected, yet their cumulative impact on physical health can be disastrous. On the other hand, more salient short-term encounters with environmental stressors, such as uncontrollable noise or periodic crowding, may be associated with acute but nonpersisting episodes of emotional stress. Therefore, to gauge adequately the health promotive capacity of an environment, it is necessary not only to specify relevant environmental dimensions and health outcomes but also to differentiate health outcomes in terms of their severity, duration, and overall importance to members of the setting. Since many environments produce a mixture of positive and negative health outcomes, the health-promotive quality of a setting ultimately depends on its capacity to support those health outcomes most desirable and important to its members while eliminating or ameliorating those most clearly negative and detrimental to individual and social well-being.

Determining which health outcomes are of greatest importance to the occupants of a setting is not always a simple matter. Whether an individual or group places greater value on the comforts of a predictable environment or the challenges of coping with a novel one may vary in relation to their age, economic resources, and exploratory tendencies (Stokols et al., 1982). Also, residents of historically significant areas often give greater priority to the symbolic and psychological benefits

of environmental preservation than to the tangible economic gains that would result from neighborhood redevelopment projects (e.g., Firey, 1945; Stokols and Jacobi, 1984). In this case, the symbolic and material benefits associated with the same environmental resources are divergent rather than compatible. Another example of voluntary trade-offs between alternative environmental arrangements and health benefits is the frequent choice of urban residents to live in a highly desirable neighborhood despite the inconveniences and strains of a long-distance commute between home and work rather than reside closer to work in a less desirable area (e.g., Campbell, 1983; Stokols and Novaco, 1981).

The environmental resources and health outcomes shown in Table 6.1 are all highly positive. This emphasis on the positive is consistent with the goals of applied environmental and health research, namely, to optimize or enhance environmental quality and human well-being (Stokols, 1978). Yet, the preceding examples of trade-offs among environmental amenities and costs serve to remind us that most situations are characterized by a mixture of positive and negative environmental circumstances and health outcomes. Thus, an important challenge for future research is to assess the overall health-promotive capacity of environments on the basis of a cumulative analysis and weighting of their positive and negative features as they affect occupants' well-being.


The ecological perspective emphasizes not only the many intrasetting factors that can influence occupants' health but also the ways in which multiple situations and settings (e.g., homes, workplaces, schools, institutional environments) jointly affect the well-being of community members. The scale of environmental units relevant to individual and collective well-being ranges from specific stimuli and situations occurring within a given setting to the more complex life domains that are, themselves, clusters of multiple situations and settings. Situations are sequences of individual or group activities occurring at a particular time and place (Forgas, 1979; Pervin, 1978). Settings are geographic locations in which various personal or interpersonal situations occur on a regular basis (Barker, 1968; Stokols and Shumaker, 1981). Life domains are different spheres of a person's life, such as family, education,

spiritual activities, recreation, employment, and commuting (Campbell, 1981; Stokols and Novaco, 1981). An even broader unit of contextual analysis is the individual's overall life situation (Magnusson, 1981), consisting of the major life domains in which a person is involved during a particular period of his or her life. The environmental dimensions most relevant to individual and collective well-being may vary considerably across these different levels of analysis.

The potential influence of multiple environmental settings on health outcomes raises an important question regarding the appropriate contextual scope of health promotion research. Just as environmental units can be arrayed along a continuum of scale or complexity, contextual analyses can be compared in terms of their relative scope. The contextual scope of research refers to the scale of the contextual units included in the analysis (Stokols, 1987). For example, the spatial scope of an analysis increases to the extent that it represents places, processes, and events occurring within a broad rather than a narrow region of the individual's (or group's) geographic environment. Similarly, the temporal scope of an analysis increases to the extent that it represents places, processes, and events experienced by the individual or group within an extended rather than narrow time frame. Finally, the sociocultural scope of an analysis increases to the extent that it describes behaviorally relevant dimensions of an individual's or group's sociocultural environment. It is important for health promotion researchers to be explicit about the range of settings and time periods encompassed by their analyses and the possible ways in which environmental conditions within multiple settings jointly influence individual and collective health outcomes.

Consider, for example, the challenge of preventing alcohol-related injuries. One strategy for reducing such injuries is to provide employee assistance programs at the workplace that facilitate workers' efforts to decrease their consumption of alcoholic beverages. Alternatively, a multisetting approach to this public health problem would combine employee assistance and treatment programs at the workplace with responsible beverage service programs for restaurant personnel, community-wide media campaigns to increase public awareness of alcohol-related injuries and prevention opportunities, legislative initiatives to raise the minimum age of purchase, and enforcement programs to reduce the illegal sale of alcohol to minors (Geller, 1990; Russ and Geller, 1987). The latter approach is based on a spatially broader analysis of injury prevention than the one focusing exclusively on the workplace, as it incorporates several different intervention programs

implemented in multiple community settings within an extended geographic area.

Wellness promotion analyses and interventions also can be characterized in terms of their temporal scope. For example, work-site health strategies typically emphasize the provision of employee assistance and lifestyle modification programs oriented toward individual workers. They often ignore, however, the design and equipment purchasing decisions made during the construction of the work site. Yet, the physical design and furnishings of the workplace can have long-term and substantial impacts on employees' health once the facility is occupied. For example, the use of formal-de-hyde-laden construction materials, the installation of ineffective air conditioning and ventilation systems, poorly designed stairwells, nonadjustable seating and work surfaces, and space plans that expose workers to excessive crowding and noise all can have deleterious effects on employees' physical and mental well-being (e.g., Archea, 1985; Greenberg, 1986; Hedge, 1989; Makower, 1981; Mendell and Smith, 1990; Stellman and Henifin, 1983). Clearly, the environmental foundations for wellness promotion (or health impairment) begin to take shape far in advance of employees' direct involvement with the workplace and continue to influence their well-being once they have occupied that environment. By explicitly considering the design and construction phase prior to occupancy as well as the postoccupancy phase, the temporal scope of work-site wellness promotion is expanded to include a broader and more robust array of intervention strategies than those focusing exclusively on employee assistance and health education.

An emphasis on the temporal dimensions of people-environment transactions suggests the importance of defining the health-promotive capacity of a setting not only in terms of its immediate impact on occupants' well-being but also in terms of the potential existing within the setting for promoting and maintaining improved levels of health over extended time intervals. Just as assessments of individual health status must take into account current states of well-being as well as the prognosis for future illness or health (Kaplan, 1990; see also chapter 3), environmentally based health promotion programs must distinguish between the immediate and potential capacity of a particular setting, organization, or community to promote health among its members.

Finally, the dimension of sociocultural scope is directly relevant to research on environment-health relationships and the design of wellness

promotion programs. The sociocultural scope of wellness promotion research and interventions is broadened to the extent that they encompass social and cultural factors within community settings that influence personal and collective well-being (e.g., socioeconomic status, gender, ethnicity, cultural norms about health and illness, supportive social relationships and organizational climate). Contextually oriented health research would involve comparative studies of organizational and community settings that vary across these important social and cultural dimensions. For example, Guendelman (chapter 9), Margen and Lashof (afterword), Roach (chapter 10), Sanders-Phillips (chapter 11), Syme (chapter 4), and Wallack (chapter 19) all suggest the importance of broadening the sociocultural scope of research and policy initiatives in the field of wellness promotion.

An important direction for future research is to identify the ways in which social-structural qualities of organizations and communities exert positive or negative effects on members' well-being. Several earlier studies indicate that supportive interpersonal relationships can enhance individuals' emotional and physical well-being and reduce the stressful consequences of negative life events (Berkman and Syme, 1979; Cohen and Syme, 1985; Sarason and Sarason, 1985). Social-structural qualities of settings also may play a key etiologic role in promoting social cohesion and physical and emotional well-being among setting members. For example, extensive efforts have been made to conceptualize and measure the social climate of organizations (Moos, 1976, 1987), and a number of studies have suggested a positive relationship between dimensions of social climate and the mental and physical health of setting members (e.g., Holahan and Moos, 1990; Moos, 1979). Moreover, certain organizations may be structured in ways that permit the smooth resolution of interpersonal conflicts, whereas others lack the capacity to resolve such tensions when they arise. In the former settings, shared goals among members provide a structural basis for cooperation, even when occasional conflicts develop. Also, such settings are likely to incorporate both informal and formal mechanisms of dispute resolution. In “conflict-prone” organizations, however, the positive interdependencies among members are weaker, and effective mechanisms of dispute resolution are unavailable (Stokols, 1992). To the extent that organizations promote chronic conflict among setting members or provide few resources to resolve such conflicts when they arise, they are more likely to impair the health of their members.



In view of the dominant focus of earlier wellness promotion programs on modifying personal health habits and lifestyles, several theorists have called for a redirection of the field based on ecological models of research and community intervention (e.g., Geller, 1987; McLeroy et al., 1988; Winett et al., 1989). Green (1984), for example, noted a “psychological bias” in the health promotion field, in that illness-preventive interventions typically are directed at individuals in a counseling or small-group mode of delivery, with little or no theoretical input from the fields of sociology, anthropology, economics, and political science. Similarly, Syme (1990) emphasized the cost-ineffectiveness of individually oriented wellness promotion programs (e.g., the MRFIT intervention to reduce cardiovascular disease among high-risk individuals) and advocated a stronger community and environmental focus in public health research.

The conceptualization of health promotive environments offers a valuable adjunct to the individual-behavioral focus of earlier health promotion research. Yet, the social ecological approach to wellness promotion encompasses more than just the analysis of environmental factors in health and illness. The social ecological perspective requires a broader analysis of the transactions between individual and collective behavior and the various constraints and resources for health that exist within specific sociophysical environments. Thus, it is important at this point in the chapter to extend our analysis of healthy environments toward a more interactive analysis of the relationships among behavioral and environmental factors in wellness and wellness promotion.


The term “salutogenesis” has been used by Antonovsky (1979) to refer to etiologic processes that enhance emotional and physical well-being. The salutogenic orientation is distinctive in its focus on the etiology

of health, as compared to more traditional pathogenic models that emphasize the development of illness. Antonovsky's research has focused primarily on psychogenic factors in health, especially individuals' “sense of coherence,” which enables them to resist the potentially negative health consequences of stressful life events. Construed more broadly, however, the salutogenic perspective encompasses not only psychological resistance resources but also a wide array of biological, behavioral, and environmental processes that reduce vulnerability to illness and promote enhanced levels of well-being.

Several categories of personal and environmental factors that play either an etiologic or a moderating role in human health are shown in Table 6.2. The personal factors include a variety of biogenetic, psychological, and behavioral processes that promote or undermine well-being. The environmental factors include several facets of the sociophysical environment, such as geographic, architectural/technological, and sociocultural processes that influence health. Thus, both natural and humanmade features of the physical environment are included, as are multiple dimensions of the sociocultural milieu (e.g., social-structural, cultural, economic, legal, and political processes).

Much research in the field of health psychology has focused on the direct links between specific dispositional factors and personal health. For example, several studies indicate the close relationship between personal orientations, such as hostility, optimism, sense of coherence, personal hardiness, and coping efficacy, and individual well-being (e.g., Antonovsky, 1979; Barefoot et al., 1983; Friedman, 1990; Kobasa et al., 1982; Scheier and Carver, 1985; Taylor and Brown, 1988; Watson and Pennebaker, 1989). Other researchers, working from a “biopsychosocial” model of health (e.g., Engel, 1976; Schwartz, 1982), have examined the interplay between psychological dispositions, interpersonal behavior, and physiological processes underlying health and illness. Examples of this research include recent studies of the psychophysiological underpinnings of the coronary-prone and cancer-prone behavior patterns (e.g., Krantz et al., 1987; Temoshok, 1985) and the links between personal dispositions, social behavior, and susceptibility to infectious disease (e.g., Cohen and Williamson, 1991).

What have been omitted from much earlier research on psychological and behavioral factors in health are structural features of the sociophysical environment that affect individual and collective well-being, either directly or interactively in conjunction with biopsychobehavioral

Biopsychobehavioral Factors
Biogenetic Psychological Behavioral
Genetic constitution
and biological resources
or challenges:
family history of illness,
exposure to infectious
(e.g., viruses, bacteria),
competence, inoculation
and medication
history, congenital
disability, disabling
injuries, cardiovascular
reactivity, chronological
age, development
stage, gender
Personal dispositions:
sense of coherence,
psychological hardiness,
creativity, optimism,
pessimistic explanatory
style, health
locus of control, interpersonal
skills, extroversion,
(Type A)
orientation, cancer-prone
(Type C) orientation,
anxiety, hostility/
Dietary regimens, alcohol
smoking, exercise
patterns, sleep patterns,
safety practices
(e.g., use of vehicular
safety belts, bicycle
helmets, safe sexual
and prenatal behaviors),
in health promotion
programs, compliance
with prescribed
medical regimens, use
of community health
resources, health-relevant
and actions made
on behalf of others
factors. These envirogenic processes in health and illness subsume geographic, architectural, and technological features of the physical environment as well as sociogenic qualities of the social and cultural environment that influence the etiology of health and illness.

An important direction for future wellness promotion research is to identify the specific mechanisms by which geographic, architectural/technological, and sociocultural factors influence health and illness. For example, five health-related functions of the sociophysical environment are outlined in Table 6.3. First, both the physical and the social environment can function as media for disease transmission, as exemplified by the occurrence of waterborne and airborne diseases, illnesses resulting from food contamination, and the spread of contagious disease through interpersonal contact. Second, the environment can operate as a stressor, evidenced by the emotional stress and physical debilitation resulting from chronic exposure to uncontrollable environmental demands, such as noise, abrupt economic change, or interpersonal conflict (e.g., Cohen et al., 1986; Dooley and Catalano, 1984; Evans, 1982;

Sociophysical Environmental Factors
Geographic Architectural
and Technological
Climatic and geologic
risks (e.g., earthquakes,
floods, hurricanes,
drought, temperature
extremes), ground-water
radon contamination
of soil, environmental
sources of radioactivity,
radiation, atmospheric
ozone depletion,
global warming,
health consequences
of reduced biodiversity,
restorative potential
of wilderness
and other natural
Injury-resistant architecture,
construction materials
in buildings, ergonomic
design of work
areas and other environmental
environmental aesthetics,
indoor and
outdoor air pollution
(e.g., “sick building
syndrome”), effective
design of health care
facilities, vehicular
and passenger safety,
noise pollution, electromagnetic
water quality
and treatment systems,
solid waste
treatment and sanitation
Socioeconomic status
of individuals and
groups; social support
versus isolation
or social conflict,
bereavement; social
climate in families,
organizations, and institutions;
and conformity processes;
cultural and
religious beliefs and
practices; organizational
or political
stability; economic
changes (job loss and
related stressful life
events); health communications
and media;
health promotion
programs in organizations
and communities
(e.g., health
education); health-promotive
and building codes;
environmentally protective
availability of health
insurance and community
health services
Rook, 1984). On the other hand, exposure to certain environmental conditions, such as natural, aesthetic, and symbolic amenities, can alleviate stress and promote physical and emotional well-being (e.g., Hartig et al., 1991; Kaplan and Kaplan, 1989; Stokols, 1990). Third, the environment functions as a source of safety or danger, as reflected in the health consequences of natural and technological disasters, air and water pollution, occupational hazards, interpersonal violence, and crime (e.g., Baum et al., 1983; Edelstein, 1988; Fielding and Phenow, 1988; Greenberg, 1987; Makower, 1981; Mendell and Smith, 1990). Fourth,
the environment can be viewed as an enabler of health behavior, exemplified by the installation of safety devices in buildings and vehicles, geographic proximity to health care facilities, and exposure to interpersonal modeling or cultural practices that foster health-promotive behavior. Fifth, the environment serves as a provider of health resources, such as high-quality community sanitation systems, organizational and community health services, and legislation protecting the quality of physical environments and ensuring citizens' access to health insurance and community-based health care. These health-relevant functions of the environment are closely intertwined and can operate concurrently in specific environmental contexts (e.g., high rates of crime may generate increased perceptions of physical danger, physiological symptoms of chronic stress, reduced use of community health services among neighborhood residents; cf. Taylor, 1987).

Another important challenge for future research is to develop integrative models that address the joint influence of personal and environmental factors in wellness promotion and disease etiology. Some specific issues for future study suggested by the categories of variables shown in Table 6.2 are the following: (1) the prevalence of negative health effects among low socioeconomic-status groups resulting from their disproportionate exposure to geographic, architectural, and technological hazards (e.g., Bullard, 1990; Lindheim and Syme, 1983; Syme and Berkman, 1976; U.S. Public Health Service, 1991; Vaughan, 1993); (2) the relationship between individuals' age, gender, developmental stage, and their increased vulnerability to certain categories of environmental health threats, such as lead poisoning (Florini et al., 1990; Needleman et al., 1990), fatalities resulting from exposure to community violence, injuries from motor vehicle crashes and alcohol abuse among adolescents and young adults, and fatalities from the complications of falls among older adults (Sanders-Phillips, chapter 11; U.S. Public Health Service, 1991; Wallack, chapter 19); (3) the psychosocial underpinnings of high-risk behaviors (e.g., smoking, unsafe sexual practices, overexposure to ultraviolet radiation, failure to use vehicular safety belts) that predispose certain groups in the population to higher rates of illness, injury, or unwanted pregnancy (e.g., Christopherson, 1989; East, chapter 8; Hofmann, chapter 20; Jeffery, 1989; Keesling and Friedman, 1987; Robertson, 1987; Weinstein, 1987); (4) the ways in which environmental factors (e.g., geographic, architectural, and sociocultural conditions) contribute to the development, modification, and mainte

Dimensions of the Environment
  Physical Environment Social Environment
as Medium
of Disease
Waterborne and airborne
disease, microbial contamination
of food
Spread of contagious
disease through interpersonal
as Stressor
Negative affective states
resulting from exposure
to physical stressors
such as uncontrollable
noise and technological
risks, negative health
consequences of residential
Vulnerability to health
problems resulting from
chronic social conflict,
isolation, organizational
instability, and/
or abrupt economic
as Source
of Safety or
Exposure to climatic and
geologic risks, injury-resistant
design, mutagenic effects
of toxic environments,
Risk of personal injury resulting
from intergroup
conflict, violence, and
as Enabler of
Health Behavior
Geographic accessibility
of health care facilities in
the community, installation
of health behavioral
supports in buildings
and vehicles (e.g., smoke
detectors, seat belts)
modeling of health-promotive behavior
and safety practices,
cultural and religious
Health-Related Functions of the Sociophysical Enviroment
as Provider
of Health
Healthful lighting and
air quality in buildings,
community sanitation
Legislation pertaining to
public health and safety,
availability of organizational
and community
health service
nance of health-promotive behavior (e.g., Centers for Disease Control and Prevention, 1998; Sallis and Hovell, 1990; Winett et al., 1989); and (5) the processes by which psychological dispositions and sociophysical stressors jointly influence emotional and physical well-being (cf. Cottington and House, 1987).

The social ecological view of health promotion has important implications

not only for theory development and basic research but also for public policy, community intervention, and program evaluation. We turn now to a consideration of these policy-related concerns.


The environmental and personal factors in health and illness, summarized previously, offer several leverage points for health-promotive policies and community interventions at municipal, regional, national, and international levels. Examples of these environmental design and public policy options for health promotion are summarized in Table 6.4 in relation to various categories of etiologic factors (i.e., biopsychobehavioral factors and sociophysical features of the environment). The social ecological approach to wellness promotion emphasizes the integration of person-focused and environment-focused strategies to enhance well-being.

Policies and interventions to promote wellness can be arrayed along a continuum ranging from microenvironmental settings (e.g., corporate or institutional facilities) to more molar environmental contexts (e.g., metropolitan and international regions). Each level of analysis poses opportunities for integrating person-focused and environment-focused interventions for health enhancement. For example, the advantages of combining health-promotive environmental design and management policies at the work site with behaviorally oriented programs to modify employees' health practices were noted earlier. At the community level, health promotive urban design and planning strategies (e.g., ensuring geographic accessibility of health care settings and appropriate siting of buildings away from toxic or seismic hazards) can be implemented in conjunction with effective sanitation systems and other health services (e.g., public education and risk-screening programs) to enhance the healthfulness of urban environments.

Because local and more distant environments are linked (both spatially and organizationally) within nested hierarchical systems (e.g., specific behavior settings exist as components of broader institutional, urban, and regional contexts) and are becoming increasingly interdependent because of global technological and social changes, opportunities for designing health-promotive environments at local levels will be more and more influenced by the regulatory and economic policies implemented within municipal, regional, and international contexts. Thus, an architect or facility planner working on the design of a corporate

Focus of Health-Promotive
Examples of Health-Promotive Policies and Programs
Person Focused
Preventive public health programs for risk screening,
genetic counseling, inoculation; medical treatment
regimens (e.g., medication, surgery)
Individual counseling and psychotherapeutic interventions
Health behavior modification (lifestyle appraisal and
modification pertaining to diet, exercise, smoking,
safety practices)
Environment Focused
Health and safety-oriented urban planning (e.g., site
planning to reduce toxic or seismic hazards); land use
policy and environmental law at municipal, regional,
and international levels (e.g., NEPA, CEQA); strategic
siting of health care facilities in the community
Ergonomic and safety-oriented environmental design
and facilities management, design of safe and health-promotive
products (e.g., passenger constraints in
automobiles), community sanitation systems (water
treatment, air filtration)
Organizational development and conflict resolution,
corporate health promotion programs, community
health education and media programming, health-promotive
legislation (e.g., regulation of health-damaging
industries, health insurance and delivery
of health services) and building code
facility, neighborhood playground, apartment complex, hospital, or residential facility for the elderly will need to have knowledge of several disciplines, including environmental law (e.g., the regulations intended to mitigate negative impacts of proposed environmental developments), life-span human development (e.g., the specialized health and safety needs of different age-groups), and ergonomics and public health (e.g., the potential health consequences of poorly designed, toxic, or injury-prone environments). In response to the complex health challenges of the 21st century, there will be a growing need to develop broad-based, interdisciplinary graduate training programs for aspiring environmental designers, facility managers, urban planners, and public health professionals.


Among the topics likely to become more prominent in training programs for environmental planners and public health researchers are the legislative and economic strategies that have been initiated in recent years to protect environmental quality and public health. Commenting on the powerful impact of legislative interventions to enhance public health, McKinlay (1975) noted,

One stroke of effective health legislation is equal to many separate health intervention endeavors and the cumulative efforts of innumerable health workers over longperiods of time. … Greater changes will result from the continued politicization of illness than from the modification of specific individual behaviors. There are many opportunities for a reduction of at-riskness, and we ought to seize them. (p. 13)

The following sections of the chapter examine legislative initiatives and other community interventions that either have been implemented or could be adopted at local, state, national, and international levels to enhance environmental quality and public health.

Municipal, state, and national contexts of health promotion. In an effort to reduce the devastating personal and public health consequences of smoking (e.g., Eriksen et al., 1988; Fielding and Phenow, 1988; U.S. Public Health Service, 1979), several local governments have enacted legislation to ban smoking in public places. In California alone, 172 municipalities and counties had passed ordinances restricting smoking in workplaces and commercial settings, and nearly 400 such ordinances had been enacted nationwide by September 1989 (Pertschuk and Shopland, 1989; see also Bureau of National Affairs, 1986). In addition to protecting nonsmokers from passive smoke exposure, these actions have made smoking less socially acceptable, thereby prompting more smokers to attempt quitting. Other nonlegislated interventions to promote wellness in local communities include media campaigns to encourage heart-healthy behaviors (e.g., Farquhar et al., 1985; Maccoby and Alexander, 1980), elementary school education programs to promote bicycle helmet use among children (DiGuiseppi et al., 1989), and corporatebased programs to increase vehicle safety belt use (Geller, 1984).

At state levels, several legislative actions have been found to reduce injury and fatality rates associated with automobile crashes. These include laws mandating the use of child safety seats in automobiles (e.g., Fawcettet al., 1987; Insurance Institute for Highway Safety, 1987), laws requiring servers of alcoholic beverages to have intervention training to reduce customers' risk of alcohol-impaired driving (Geller, 1990; Russ and Geller, 1987), and laws raising the legal minimum drinking age

(Williams et al., 1983b) or the drivers' licensing age (Williams et al., 1983a). At the national level, lowering the maximum speed limit from 70 to 55 miles per hour in 1973 was associated with a substantial decrease in automobile accident injuries and fatalities throughout the United States (National Safety Council, 1987).

A notable strength of several of the local, state, and national interventions cited here is that their actual influence on public health and safety has been documented through carefully designed quasi-experi-mental studies. Rigorous evaluations of health promotive legislation and community interventions are essential for estimating the scientific validity and practical utility of both existing and proposed programs (e.g., Campbell, 1969; Evans, 1988; Geller, 1990; Syme, chapter 4; Wallack, chapter 19; Birckmayer and Weiss, chapter 7).

Another health-promotive strategy that has been widely used at national, state, and local levels is the enactment of legislation designed to protect natural resources and the quality of public environments. Examples of environmentally protective legislation undertaken at national levels include the 1969 National Environmental Policy Act (NEPA), the 1970 Clean Air Act, the 1972 Clean Water Act, and the 1976 Toxic Substances Control Act in the United States and the 1971 Town and Country Planning Act in Great Britain. NEPA, instituted by the U.S. Congress, requires all federal agencies to prepare detailed written statements about the potentially negative impacts that could result from any of their actions relating to the environment and proposed strategies for avoiding or mitigating those outcomes. The California Environmental Quality Act (CEQA) is one of several state analogues of NEPA that has been implemented in the United States over the past 20 years. CEQA requires that municipal and state agencies not approve a proposed environmental project unless the potentially adverse effects of the development are identified in an environmental impact report (EIR) and all feasible alternatives or mitigation measures to reduce those impacts have been incorporated into the project plans. Today, about half the states in the United States have emulated the NEPA process, and environmental impact assessment is now an established legal process in several nations (e.g., Australia, Canada, the European Community, and Great Britain; CEQA, 1986; Robinson, 1990).

International efforts to protect environmental quality and to promote public health. International efforts directed toward environmental protection and health promotion also have increased substantially in recent years. Growing public concern over global environmental problems

has stimulated greater international collaboration in economic and legal matters (Silver and DeFries, 1990; Wilson and Peter, 1988; World Commission on Environment and Development, 1987). A recent example of intercity and cross-national cooperation in health promotion is the World Health Organization's Healthy Cities Project (e.g., Ashton et al., 1986; Hancock and Duhl, 1985; World Health Organization, 1984). As part of the Healthy Cities Project, public health professionals from several different countries have worked together in developing and implementing intersectoral city health plans. In support of these collaborative efforts, WHO staff provide technical assistance and resource materials to the participating cities. One product of this collaboration is a European television series on the healthy city.

An important defining attribute of healthy cities is that they continually create and improve physical and social environments conducive to the health of their residents (Duhl, 1996; Hancock and Duhl, 1985). At least 14 criteria for assessing the healthfulness of a city have been proposed, including epidemiologic indices of illness and mortality, levels of public safety, quality of the physical and social environment, quality of public health services, the degree of intersectoral collaboration in developing health policies, and the state of the local economy including unemployment levels. These criteria provide a broad framework for establishing coordinated public health plans and objectives among the participating cities.

A central concept that will guide future environmental and health promotive legislation is the notion of sustainable development. According to Robinson (1990), sustainable development is “the emerging cluster of policies by which we manage the use of the Earth's environment and natural resources to ensure the optimal level of sustainable benefits for present and succeeding generations” (p. 16). Growing concern about the sustainability of global resources highlights the crucial importance of public health forecasting, environmental simulation strategies, and the temporal dimensions of health promotion (see Table 6.2). Now more than ever, individually focused and environmentally focused efforts to enhance human health must anticipate the cumulative consequences of seemingly remote processes and distant events, for example, (1) the potential exacerbation of health problems among the elderly by elevated temperatures associated with global warming, (2) increased prevalence of cutaneous melanoma and other diseases resulting from global ozone depletion and heightened exposure to ultraviolet radiation, (3) the biogenetic consequences of exposure to toxic by-products of modern technologies,

(4) the implications of reduced ecosystem biodiversity for human health and medical treatment and research programs, and (5) the ever-present threat of global nuclear war and the health consequences of nuclear weapons testing.

Amidst these somber projections of public health problems and challenges for the 21st century, the previously cited examples of municipal and international cooperation toward health promotion and environmental protection are impressive in their scope and offer a basis for optimism about the willingness of governments to work collaboratively to promote world health. Collaborative international efforts to protect the global environment and promote the well-being of the world's population give new meaning to the concept of “health behavior.” Future health promotion programs must influence not only the behaviors of individuals that enhance or undermine their own well-being but also the decisions they make and the actions they take on behalf of others—ranging from small groups to urban populations—in their roles as voters, environmental planners, corporate executives, and community leaders. This distinction between personal and other-directed health behavior has important implications for the design of effective wellness promotion programs (cf. Stokols, 1996).


The challenge of creating and maintaining healthy environments raises several complex theoretical, methodological, and public policy questions. For example, how shall we conceptualize healthy environments, and by what observable criteria can we determine the extent to which an environment is health promotive? Is the healthfulness of an environment defined primarily by its physical quality, or is it defined in terms of the joint influence of its material and symbolic features on the emotional and physical well-being of its occupants? Does the concept of environmental health refer to the present condition of the environment and its occupants, or does it refer to the potential that exists within a setting for promoting and maintaining improved levels of well-being over an extended period?

To address these issues, a social ecological model of health promotive environments was proposed, emphasizing the interactions among physical-material and social-symbolic features of environments as they affect the emotional, physical, and social well-being of individuals and groups. Health status was analyzed along a continuum ranging from individuals

to larger aggregates and populations and in relation to microlevel, local settings (e.g., homes, offices, neighborhoods) as well as larger-scale and more distant environments (e.g., geographically and politically bounded regions). The temporal dimensions of environmental health were examined with particular emphasis on the stability or instability of healthful conditions within a setting or region and the factors that influence the healthfulness of an environment over extended periods. Finally, several directions for both basic research and the evaluation of policy initiatives to protect environmental quality and promote public health were examined at organizational, municipal, and international levels.


Portions of this chapter were presented as part of the University of California Wellness Lectures Program in 1991 and are adapted from a subsequent article:


Aldwin, C., and Stokols, D. (1988). “The effects of environmental change on individuals and groups: Some neglected issues in stress research.” Journal of Environmental Psychology, 8, 57–75.

Antonovsky, A. (1979). Health, stress, and coping.San Francisco: Jossey-Bass.

Archea, J. C. (1985). “Environmental factors associated with stair accidents by the elderly.” Clinics in Geriatric Medicine, 1, 555–569.

Ashton, J., Grey, P., and Barnard, K. (1986). “Healthy cities: WHO's New Public Health Initiative.” Health Promotion, 1, 319–324.

Barefoot, J. C., Dahlstrom, W. G., and Williams, R. B. (1983). “Hostility, CHD incidence, and total mortality: A 25-year follow-up study 255 physicians.” Psychosomatic Medicine, 45, 59–63.

Barker, R. G. (1968). Ecological psychology: Concepts and methods for studying the environment of human behavior.Stanford, Calif.: Stanford University Press.

Baum, A., Fleming, R., and Davidson, L. M. (1983). “Natural disaster and technological catastrophe.” Environment and Behavior, 15, 333–354.

Belloc, N., and Breslow, L. (1972). “Relationship of physical health status and health practices.” Preventive Medicine, 1, 409–421.

Berkman, L. F., and Syme, S. L. (1979). “Social networks, host resistance, and mortality: A nine-year follow-up study of Alameda County residents.” American Journal of Epidemiology, 109, 186–204.


Breslow, L. (1996). “Social ecological strategies for promoting healthy lifestyles.” American Journal of Health Promotion, 10, 253–257.

Bullard, R. D. (1990). Dumping in Dixie: Race, class, and environmental quality.Boulder, Colo.: Westview Press.

Bureau of National Affairs. (1986). Where there's smoke: Problems and policies concerning smoking in the workplace.Washington, DC: Author.

Campbell, A. (1981). The sense of well-being in America.New York: McGraw-Hill.

Campbell, D. T. (1969). “Reforms as experiments.” American Psychologist, 24, 209–219.

Campbell, J. M. (1983). “Ambient stressors.” Environment and Behavior, 15, 355–380.

Cataldo, M. F., and Coates, T. J., eds. (1986). Health and industry: A behavioral medicine perspective.New York: John Wiley and Sons.

“Centers for Disease Control and Prevention.” (1998). “Panel discussion on policy and environmental actions to promote physical activity.” Atlanta: Physical Activity and Health Branch, Centers for Disease Control and Prevention.

“CEQA—The California Environmental Quality Act” (1986). Sacramento: Governor's Office of Planning and Research, State of California.

Christophersen, E. R. (1989). “Injury control.” American Psychologist, 44, 237–241.

Cohen, S., Evans, G. W., Stokols, D., and Krantz, D. S. (1986). Behavior, health and environmental stress.New York: Plenum.

Cohen, S., and Syme, S. L., eds. (1985). Social support and health.Orlando: Academic Press.

Cohen, S., and Williamson, G. M. (1991). “Stress and infectious disease in humans.” Psychological Bulletin, 109, 5–24.

Cottington, E., and House, J. S. (1987). “Occupational stress and health: A multivariate relationship.” In A. Baum and J. E. Singer, eds., Handbook of psychology and health:Volume 5. Pp. 41–62. Hillsdale, N. J.: Lawrence Erlbaum Associates.

DiGuiseppi, C. G., Rivara, F. P., Koepsell, T. D., and Polissar, L. (1989). “Bicycle helmet use by children: Evaluation of a community-wide helmet campaign.” Journal of the American Medical Association, 262, 2256–2261.

Dooley, D., and Catalano, R. (1984). “The epidemiology of economic stress.” American Journal of Community Psychology, 12, 387–409.

Duhl, L. (1996). “An ecohistory of health: The role of “healthy cities.”” American Journal of Health Promotion, 10, 258–261.

Edelstein, M. R. (1988). Contaminated communities: The social and psychological impacts of residential toxic exposure.Boulder, Colo.: Westview Press.

Engel, G. L. (1976). “The need for a new medical model.” Science, 196, 129–136.

Eriksen, M. P., LeMaistre, C. A., and Newell, G. R. (1988). “Health hazards of passive smoking.” In L. Breslow, J. E. Fielding, and L. B. Lave, eds., Annual Review of Public Health, 9, 47–70. Palo Alto, Calif.: Annual Reviews, Inc.

Evans, G. W., ed. (1982). Environmental stress.New York: Cambridge University Press.


Evans, R. I. (1988). “Health promotion—Science or ideology?” Health Psychology, 7, 203–219.

Farquhar, J. W., Fortman, S. P., Maccoby, N., et al. (1985). The Stanford Five City Project: Design and methods. American Journal of Epidemiology, 63, 171–182.

Fawcett, S. B., Seekins, T., and Jason, L. A. (1987). “Policy research and child passenger safety legislation: A case study and experimental evaluation.” Journal of Social Issues, 43, 133–148.

Fielding, J. E., and Phenow, K. J. (1988). “Health effects of involuntary smoking.” New England Journal of Medicine, 319, 1452–1460.

Firey, W. (1945). “Sentiment and symbolism as ecological variables.” American Sociological Review, 10, 410–418.

Florini, K. L., Krumbhaar, G. D., and Silbergeld, E. K. (1990, March). Legacy of lead: America's continuing epidemic of childhood lead poisoning: A report and proposal for legislative action.Washington, D. C.: Environmental Defense Fund.

Forgas, J. P. (1979). Social episodes: The study of interaction routines.New York: Academic.

Friedman, H. S., ed. (1990). Personality and disease.New York: John Wiley and Sons.

Gardell, B., and Johansson, G., eds. (1981). Working life: A social science contribution to work reform.New York: John Wiley and Sons.

Geller, E. S. (1984). “Motivating safety belt use with incentives: A critical review of the past and a look to the future. SAE Technical Paper Series 840326.” Warrendale, Pa.: Society of Automotive Engineers.

Geller, E. S. (1987). “Applied behavior analysis and environmental psychology: From strange bedfellows to a productive marriage.” In D. Stokols and I. Altman, eds., Handbook of environmental psychology:Volume 1. Pp. 361–388. New York: John Wiley and Sons.

Geller, E. S. (1990). “Preventing injuries and deaths from vehicle crashes: Encouraging belts and discouraging booze.” In J. Edwards, R. S. Tindale, L. Heath, and E. J. Posavac eds., Social influence processes and prevention.Pp. 249–277. New York: Plenum.

Geller, E. S., Winett, R. A., and Everett, P. B. (1982). Preserving the environment: New strategies for behavior change.New York: Pergamon.

Glass, D. C., and Singer, J. E. (1972). Urban stress.New York: Academic.

Green, L. W. (1984). “Modifying and developing health behavior.” Annual Review of Public Health, 5, 215–236.

Greenberg, M. R. (1986). “Indoor air quality: Protecting public health through design, planning, and research.” Journal of Architectural and Planning Research, 3, 253–261.

Greenberg, M. R. (1987). Public health and the environment: The United States experience.New York: Guilford.

Hancock, T., and Duhl, L. (1985). “Healthy cities: Promoting health in the urban context.” A background working paper for the Healthy Cities symposium. Lisbon, 1986. Copenhagen: WHO.


Hartig, T., Mang, M., and Evans, G. W. (1991). “Restorative effects of natural environment experiences.” Environment and Behavior, 23, 3–26.

Hedge, A. (1989). “Environmental conditions and health in offices.” International Review of Ergonomics, 2, 87–110.

Holahan, C. J., and Moos, R. H. (1990). “Life stressors, resistance factors, and improved psychological functioning. An extension of the stress-resistance paradigm.” Journal of Personality and Social Psychology, 58, 909–917.

“Insurance Institute for Highway Safety (1987). Status Report, December” , 1987. Washington, D. C.: Author.

Jeffery, R. W. (1989). “Risk behaviors and health: Contrasting individual and population perspectives. American Psychologist” , 44, 1194–1202.

Kaplan, R. M. (1990). “Behavior as the central outcome in health care.” American Psychologist, 45, 1211–1220.

Kaplan, S. (1983). “A model of person-environment compatibility.” Environment and Behavior, 15, 331–332.

Kaplan, R., and Kaplan, S. (1989). The experience of nature: A psychological perspective.New York: Cambridge University Press.

Karasek, R., and Theorell, T., eds. (1990). Healthy work: Stress, productivity, and the reconstruction of working life.New York: Basic Books.

Keesling, B., and Friedman, H. S. (1987). “Psychosocial factors in sunbathing and sunscreen use.” Health Psychology, 6, 477–493.

Kobasa, S. C., Maddi, S. R., and Kahn, S. (1982). “Hardiness and health: A prospective study.” Journal of Personality and Social Psychology, 42, 168–177.

Krantz, D. S., Lundberg, U., and Frankenhaeuser, M. (1987). “Stress and Type-A behavior: Environmental and biological factors.” In A. Baum and J. E. Singer, eds., Handbook of psychology and health:Volume V. Pp. 203–228. Hillsdale, N. J.: Lawrence Erlbaum Associates.

Lindheim, R., and Syme, S. L. (1983). “Environments, people, and health. Annual Review of Public Health” , 4, 335–354.

Maccoby, N., and Alexander, J. (1980). “Use of media in lifestyle programs.” In P. O. Davidson and S. M. Davidson, eds., Behavioral medicine: Changing health lifestyles.Pp. 351–370. New York: Brunner/Mazel.

Magnusson, D. (1981). “A psychology of situations.” In D. Magnusson, ed., Toward a psychology of situations: An interactional perspective.Pp. 9–32. Hillsdale, N. J.: Lawrence Erlbaum Associates.

Makower, J. (1981). Office hazards: How your job can make you sick.Washington, D. C.: Tilden Press.

McKinlay, J. B. (1975). “A case for refocusing upstream: The political economy of illness.” In A. J. Enelow and J. B. Henderson, eds., Applying behavioral science to cardiovascular risk.Washington, D. C.: American Heart Association.

McLeroy, K. R., Bibeau, D., Steckler, A., and Glanz, K. (1988). “An ecological perspective on health promotion programs.” Health Education Quarterly, 15, 351–378.

Mendell, M. J., and Smith, A. H. (1990). “Consistent pattern of elevated symptoms in air-conditioned office buildings: A reanalysis of epidemiologic studies.” American Journal of Public Health, 80, 1193–1199.


Moos, R. H. (1976). The human context.New York: John Wiley and Sons.

Moos, R. H. (1979). “Social ecological perspectives on health.” In G. C. Stone, F. Cohen, and N. E. Adler, eds., Health psychology: A handbook.Pp. 523–547. San Francisco: Jossey-Bass.

Moos, R. H. (1987). The social climate scales: A user's guide.Palo Alto, Calif.: Consulting Psychologists Press.

“National Safety Council.” (1987). “Accident facts.” Chicago: Author.

Needleman, H. L., Schell, A., Bellinger, D., Leviton, A., and Allred, E. N. (1990). “The long-term effects of exposure to low doses of lead in childhood.” New England Journal of Medicine, 322, 83–88.

O'Donnell, M. P., and Ainsworth, T., eds. (1984). Health promotion in the workplace.New York: John Wiley and Sons.

Pertschuk, M., and Shopland, D. (1989, September). Major local smoking ordinances in the United States. National Institutes of Health Publication 90479. Washington, D. C.: U. S. Government Printing Office.

Pervin, L. A. (1978). “Definitions, measurements, and classifications of stimuli, situations, and environments.” Human Ecology, 6, 71–105.

Robertson, L. S. (1987). “Injury prevention: Limits to self-protective behavior.” In N. D. Weinstein, ed., Taking care: Understanding and encouraging selfprotective behavior.Pp. 280–297. New York: Cambridge University Press.

Robinson, N. (1990). “Sustainable development: An introduction to the concept.” In J. O. Saunders, ed., The legal challenge of sustainable development.Pp. 15–34. Calgary: Canadian Institute of Resources Law.

Rook, K. S. (1984). “The negative side of social interaction: Impact on psychological well-being.” Journal of Personality and Social Psychology, 46, 1097–1108.

Russ, N. W., and Geller, E. S. (1987). “Training bar personnel to prevent drunken driving: A field evaluation.” American Journal of Public Health, 77, 952–954.

Sallis, J. F., and Hovell, M. F. (1990). “Determinants of exercise behavior.” In J. O. Holloszy and K. B. Pandolf, eds., Exercise and sport sciences reviews, 18. Pp. 307–330. Baltimore: Williams and Wilkins.

Sarason, I. G., and Sarason, B. R., eds. (1985). Social support: Theory, research, and applications.Dordrecht: Martinus Nijhoff.

Sauter, S. L., Hurrell, J. J., and Cooper, C. L., eds. (1989). Job control and worker health.Chichester: John Wiley and Sons.

Scheier, M. F., and Carver, C. S. (1985). “Optimism, coping, and health: Assessment and implications of generalized outcome expectancies.” Health Psychology, 4, 219–247.

Schwartz, G. E. (1982). “Testing the biopsychosocial model: The ultimate challenge facing behavioral medicine.” Journal of Consulting and Clinical Psychology, 50, 1041–1053.

Silver, C., and DeFries, R. (1990). One earth, one future: Our changing global environment.Washington, D. C.: National Academy of Sciences.

Stellman, J., and Henifin, M. S. (1983). Office work can be dangerous to your health: A handbook of office health and safety hazards and what you can do about them.New York: Fawcett Crest.

Stokols, D. (1978). “Environmental psychology.” In M. R. Rosenzweig and L. W.

Porter, eds., Annual Review of Psychology, 29, 253–295. Palo Alto, Calif.: Annual Reviews, Inc.

Stokols, D. (1987). “Conceptual strategies of environmental psychology.” In D. Stokols and I. Altman, eds., Handbook of environmental psychology, Volume 1. Pp. 41–70. New York: John Wiley and Sons.

Stokols, D. (1990). “Instrumental and spiritual views of people-environment relations.” American Psychologist, 45, 641–646.

Stokols, D. (1992). “Conflict-prone and conflict-resistant organizations.” In H. Friedman, ed., Hostility, coping, and health.Pp. 65–76. Washington, D. C.: American Psychological Association.

Stokols, D. (1996). “Translating social ecological theory into guidelines for community health promotion.” American Journal of Health Promotion, 10, 282–298.

Stokols, D., Allen, J., and Bellingham, R. L. (1996). “The social ecology of health promotion: Implications for research and practice.” American Journal of Health Promotion, 10, 247–251.

Stokols, D., and Jacobi, M. (1984). “Traditional, present-oriented, and futuristic modes of group-environment relations.” In K. Gergen and M. Gergen, eds., Historical social psychology.Pp. 303–324. Hillsdale, N. J.: Lawrence Erlbaum Associates.

Stokols, D., and Novaco, R. W. (1981). “Transportation and well-being: An ecological perspective.” In J. Wohlwill, P. Everett, and I. Altman, eds., Human behavior and environment: Advances in theory and research, Volume 5: Transportation environments.Pp. 85–130. New York: Plenum.

Stokols, D., and Shumaker, S. (1981). “People in places: A transactional view of settings.” In J. Harvey, ed., Cognition, social behavior and the environment.Pp. 441–488. Hillsdale, N. J.: Lawrence Erlbaum Associates.

Stokols, D., Shumaker, S., and Martinez, J. (1983). “Residential mobility and personal well-being.” Journal of Environmental Psychology, 3, 5–19.

Syme, S. L. (1990). “Health promotion: Old approaches, new choices, future imperatives.” Presented at Conference on “The New Public Health: 1990,” Los Angeles, April.

Syme, S. L., and Berkman, L. F. (1976). “Social class, susceptibility and sickness.” American Journal of Epidemiology, 104, 1–8.

Taylor, R. B. (1987). “Toward an environmental psychology of disorder: Delinquency, crime, and fear of crime.” In D. Stokols and I. Altman, eds., Hand book of environmental psychology.Pp. 951–986. New York: John Wiley and Sons.

Taylor, S. E., and Brown, J. D. (1988). “Illusion and well-being: A social psychological perspective on mental health.” Psychological Bulletin, 103, 193–210.

Temoshok, L. (1985). “Biopsychosocial studies on cutaneous malignant melanoma: Psychosocial factors associated with prognostic indicators, progression, psychophysiology, and tumor-host response.” Social Science and Medicine, 20, 833–840.

“U. S. Public Health Service” (1979). Healthy people: The Surgeon General's report on health promotion and disease prevention. DHEW Publication (PHS) 79-55071. Washington, D. C.: U. S. Government Printing Office.


“U. S. Public Health Service.” (1991). Healthy People 2000: National health promotion and disease prevention objectives. PHHS Publication (PHS) 91-50212. Washington, D. C.: U. S. Government Printing Office.

Vaughan, E. (1993). “Individual and cultural differences in adaptation to environmental risks.” American Psychologist, 48, 673–680.

Watson, D., and Pennebaker, J. W. (1989). “Health complaints, stress, and distress: Exploring the central role of negative affectivity.” Psychological Review, 96, 234–254.

Weinstein, N. D., ed. (1987). Taking care: Understanding and encouraging selfprotective behavior.New York: Cambridge University Press.

Weiss, S. M. (1991). “Health at work.” In S. M. Weiss, J. E. Fielding, and A. Baum, eds., Perspectives in behavioral medicine: Health at work.Pp. 1–10. Hillsdale, N. J.: Lawrence Erlbaum Associates.

Williams, A. F., Karpf, R. S., and Zador, P. F. (1983a). “Variations in minimum licensing age and fatal motor vehicle crashes.” American Journal of Public Health, 73, 1401–1403.

Williams, A. F., Zador, P. L., Harris, S. S., and Karpf, R. S. (1983b). “The effect of raising the legal minimum drinking age on fatal crash involvement.” Journal of Legal Studies, 12, 169–179.

Wilson, E. O., and Peter, F. M., eds. (1988). Biodiversity.Washington, D. C.: National Academy Press.

Winett, R. A., King, A. C., and Altman, D. G. (1989). Health psychology and public health: An integrative approach.New York: Pergamon.

“World Commission on Environment and Development.” (1987). Our common future.New York: Oxford University Press.

“World Health Organization.” (1984). Health promotion: A discussion document on the concept and principles. Health Promotion, 1, 73–76.



Investigating the How and
Why of Wellness Promotion Programs

Johanna Birckmayer and Carol Hirschon Weiss

Agencies and their sponsors undertake evaluation to find out how effectively a program achieves the goals set by its various publics. The emphasis has traditionally been on outcomes: How well does the program accomplish desired ends? Attention is sometimes paid to unanticipated consequences, too: What unintended outcomes appear that either negate or amplify desired outcomes?

Evaluation studies use a variety of research designs to address these questions. When some potential participants are randomly assigned to the program and others to a control group who do not receive the program, program recipients can be compared with the control group after the program is over. Since the randomization process results in equivalent groups at the start, differences between the two groups at the end can be confidently attributed to the program. The evaluator can draw conclusions with considerable confidence about how much change the program “caused.” Other designs, which do not lead to such firm causal attributions, are nevertheless appropriate under certain circumstances, for example, when the evaluator has little control over assignment or when questions of causality are not salient.

In recent years, evaluators have paid increasing attention to examining program process as well, that is, how activities are implemented. The reasons for studying program process are to find out such things as the extent to which implementation follows the prescribed course or is modified to suit local circumstances, the “quality” of the implementation (atleast

in terms of frequency and “dosage” of service provided), characteristics of staff who deliver the program and clients who receive it, and obstacles that arise during the course of implementation that interfere with planned activities. Quantitative or qualitative data on questions of this sort provide important understanding about how the program is actually carried out.

A combination of outcome evaluation and process evaluation can produce even greater learning. Evaluations that include both kinds of data show how programs are conducted and what kinds of outcomes they generate. When outcomes are analyzed in terms of the activities and experiences offered by the program, the evaluation indicates which features of the program (and its participants) are associated with better outcomes. Thus, it may emerge that weight-loss activities offered in nonformal settings are associated with better outcomes than those offered in regular school classes.

Such evaluations have made vital contributions to the understanding of program interventions. But for all their contributions, they have not often been able to say how and why the program produced the outcomes observed. They can say such things as, “activities with more experienced staff tend to have better outcomes” or “participants who receive service on a face-to-face basis do better than those who receive service through telephone contact.” But the reader has to speculate about how staff experience translates into more effective service or why face-to-face service works better than other forms of communication.

In many cases, the information gained is sufficient for practical purposes. A program faced with the findings in the previous paragraph can seek to employ staff with more years of experience and design activities that work in in-person modes.

But a later study may find that experienced staff do not have an edge in terms of outcomes or that telephone service that is responsive to the immediate needs of the client is actually more effective the next time around. It was not necessarily the years of experience that made staff more effective but something they did that novices did not do—something that not all staff with long experience do as a matter of course. And face-to-face service may be advantageous under some circumstances, for example, when the client has to be convinced of the service provider's expertise or good faith, but not necessarily under different circumstances.

What to do? One approach is to take a “theory-based” approach to evaluation. This chapter discusses theory-based evaluation (TBE)—what it is, what it does, and what it is assumed to contribute to more traditional

evaluation fare. It is particularly fitting that we discuss TBE in wellness promotion programs because it is the field that has seen the most extensive use of TBE. Next we review six evaluation studies that have taken a theory-based approach. We look briefly at the programs, how the evaluators conducted the evaluations, and the findings that emerged. Finally, we seek to derive some lessons for the next round of TBE in wellness promotion.


Theory-based evaluation (TBE) is a mode of evaluation that is built around the explicit or implicit assumptions on which the program is based (Suchman, 1967; Weiss, 1972, 1995, 1997, 1998; Chen and Rossi, 1987, 1992; Costner, 1989; Finney and Moos, 1989; Lipsey and Pollard, 1989; Bickman, 1990; Chen, 1990; Lipsey, 1993). The evaluator focuses on the mechanisms by which the program expects to achieve its effects. For example, a smoking prevention program offers activities that stress the perils of smoking (e.g., the known elevated risks of cancer and emphysema). The mechanism by which the program expects to keep young people from smoking is not the activities per se; it is how the participants react to the activities. Thus, one assumption is that participants gain knowledge about possible negative health effects and that the knowledge engenders fear or distaste. If this is the case, the mechanism that intervenes between program activities and refraining from smoking is assumed to be participants' increased knowledge and increased fear.

The evaluation then investigates the extent to which participants do in fact gain knowledge about the negative effects of smoking and the extent to which those with more knowledge become worried or concerned. (For further examples of TBE, see Feindler et al., 1984; Pentz et al., 1989; Donaldson et al., 1994; Goodman and Wandersman, 1994).

In sum, process evaluation collects quantitative or qualitative data on the way in which activities were carried out. Outcome evaluation collects data on the extent to which desired outcomes, in this case abstaining from smoking, were reached. The theory-based part of the evaluation focuses on the mechanisms that link process to outcome.


The benefits that advocates of TBE claim for this approach are of three kinds: advantages to program planning and improvement, advantages

for the growth of knowledge about human behavior and behavior change, and advantages for the planning and conduct of the evaluation of the specific program.

Advantages for Program
Planning and Program Improvement

Theory-based evaluation provides information about the mechanisms that intervene between program activities and the achievement (or nonachievement) of expected results. When the theory on which the evaluation is based is fine grained, the evaluation can track each link in the chains of assumptions. The results of such an evaluation will show which chains of assumptions are well supported by the data collected, which chains of assumptions break down, and where in the chain they break down. For example, a program theory may be that the program increases participants' knowledge, that knowledge leads them to change their attitudes toward the risk behavior, and that attitude change leads to behavior change. Evaluation data may show, as it often does, that the program does increase knowledge and that knowledge often is associated with a change in attitude but that attitude change is not associated with change in the risk behavior. Program developers and redevelopers will then have to understand the obstacles that interfere with behavior change, that is, why people who now agree that they should not engage in a risk behavior nevertheless do so. If they find, for example, that the new attitudes tend to lapse over a short period of time, they may have to build into the program a set of ancillary activities to sustain the changed attitudes.

Advantages for Knowledge Development

Advocates of TBE hope that better knowledge about the mechanisms of change will benefit not only the specific program (or type of program) studied. The hope is that the knowledge will generalize to a wider array of change efforts. For example, if change in knowledge is not sufficient for behavior change in smoking prevention or dietary interventions, expectations for knowledge-based change may not make sense in other kinds of programming as well. Even though each evaluation study is prisoner of the unique characteristics of its setting (e.g., time, place, staff, and participants), repeated evaluations over time may be able to build a corpus of knowledge about which mechanisms work well and which work poorly. When the same kinds of theories are explored for different kinds of programs in different kinds of contexts, the findings may expand

our knowledge about effective means for promoting wellness in different populations.

Such hopes are no doubt optimistic. As social science research has demonstrated over the decades, findings do not always generalize well from one setting or population to another. Furthermore, findings tend to become outmoded as times change.

But human behavior, while complex, is not random. Themes, relationships in the data, and “stories” can often be identified and, over time, may be codified into social science theories. When evaluation can test the relationships between program processes and program outcomes and the mechanisms that link the two, it can yield findings with greater relevance and staying power.

Advantages for Planning the Evaluation Study

Most immediately, a theory-based approach highlights the elements of program activity that deserve attention in the evaluation. The evaluator uses the program's assumptions as the scaffolding for the study. The evaluator can choose to collect data on the linkage mechanisms assumed to be operative in one theory or in several theories, or she can select one set of particularly central (or problematic) assumptions and direct the evaluation toward investigating that specific link in the theory chain. For example, one of the program planners' key assumptions may be that a program must be offered on a community-wide basis rather than on an individual basis, so that it changes what the relevant community regards as acceptable behavior. The theory is that individuals will not accept or sustain new behaviors unless the social groups with which they interact adopt these behaviors as their norm. Program sponsors, program planners and staff, and evaluators may choose to examine this one theoretical assumption—that the program can change community norms and that the widespread acceptance of “wellness” norms will lead to changes in individuals' behavior.


To advance the discussions of TBE, we briefly describe six evaluations of health promotion programs that use a theory-based approach. A summary of each program's design, theory, measures, and findings is presented in Table 7.1. These studies represent fairly large, well-funded

Design Theory Measures Findings
Murray et al.
(1994), antismoking
9th-grade students in Minnesota
compared to students
in Wisconsin; random
survey of 3,600
students in both states
conducted each year
from 1986 to 1990.
Increase in antismoking
activities in school and
media → increase in exposure
to antismoking messages
→ change in beliefs
about health risks → decrease
smoking rates.
Self-reports of exposure to
pro- and antitobacco messages,
antitobacco beliefs
(health consequences to
others, passive smoking
hazard, personalized
health risk), and tobacco
use; expired air test for
carbon monoxide to
confirm smoking self-reports.
90% of schools implemented
activities, but implementation
was short
term; 95% of youth saw
or heard antitobacco ad in
1989-90, on average saw
50 ads per year; Minnesota
increased exposure
to antitobacco messages
over Wisconsin (TV programming,
net change
+6.9%, p = .007; radio,
net change +18.8%, p <
.001; newspaper and
magazine ads, net change
+4.6%, p = .006; billboards,
net change 3.8%,
p = .09); no difference in
antitobacco beliefs; 2.4%
decrease in tobacco use in
Minnesota, not significant
compared to Wisconsin
(P = 32).
Flay et al.
47 schools randomly assigned
to one of five conditions:
social resistance
curriculum (SR), TV campaign,
SR and TV combined
(SR/TV), information
only curriculum
IOC → increase in knowledge
of effects of smoking
→ decrease in smoking;
SR, TV, and SR/
TV → increase in knowledge
of effects of smoking,
awareness of influences to
Self-reports of smoking
knowledge, awareness of
social influences to smoke,
knowledge of resistance
skills, refusal self-efficacy,
efforts to resist smoking,
smoking prevalence estimates
Higher smoking knowledge
in IOC group at T2-T4
(p < .001); positive effect
on social influences awareness
and resistance skill
knowledge in SR (p <
.001) and TV (p < .03)


(IOC), and no-treatment
(NOT); 340 7th-grade
classrooms in study; students
surveyed before intervention
(T1), after (T2),
1 year after (T3), and
2 years after (T4).
smoke, skills to resist, efforts
/confidence to refuse
to smoke, and decrease
perception of smoking acceptability
(norm) → decrease
in behavioral intentions
to smoke → decrease
in smoking.
(as indicator of
smoking norms), approval
of parental smoking, intentions
to smoke-, and
past and current smoking.
groups and SR/TV at T2-T4
compared to NOT, SR,
and TV; greater effect on
social influences and resistance
knowledge at T2 than
SR/TV, difference faded
by T4; no differences
across groups on refusal
self-efficacy; TV/SR positive
effect on efforts at T2
(p < .009) and T (p <
.02), effect faded by T4;
smoking prevalence estimates
lowest for SR (p <
.001) and TV (p < .006) at
T2, effect continued for SR
at T4 (p < .007); no difference
across groups in intentions
to smoke; no difference
in smoking rates.
Eisen et al.
(1992), sex
Random assignment to either
Health Belief Model
(HBM) curriculum or
agencies' usual sex ed
program; 1,444 13- to
19-year olds attending six
family planning services
and one school surveyed
before (T1), after (T2),
and 12 months after (T3)
Increase awareness of probability
and negative consequences
of pregnancy,
benefits of delayed sexual
activity or contraceptive
use, and decrease perceptions
of barriers to
abstinence or contraceptive
use → decrease sexual
activity and increase contraceptive
use → decrease
in pregnancy.
Self-reports of sexual and
contraceptive knowledge,
attitudes, and behavior.
Knowledge in both HBM
and comparison class increased;
HBM class greater
increase (p < .05); health
perceptions in both HBM
and comparison group increased
(p < .01); no difference
between groups;
no difference in continued
abstinence between
groups. T1 female virgins
in comparison group more
likely to use effective contraceptives


      at T2-T3; T1
male virgins, no difference
between programs on outcomes
at T2-T3; both
programs increased contraceptive
efficiency, T1-T3
for nonvirgins; males
in HBM group greater increase
in contraceptive efficiency
than comparison
program (p < .05); no difference
between programs
for females.
Brug et al.
Random assignment to program
or general nutrition
information; 347 employees
in the Netherlands
Royal Shell laboratory
surveyed before and three
weeks after intervention.
Change in attitudes toward
desired diet, perceived social
influences, self-efficacy
beliefs, and intentions to
change diet → reduce
fat intake and increase
vegetable and fruit
Self-reports of attitudes,
social influences, and self-efficacy
expectations; dietary
fat, fruit, and vegetable
Program group more positive
attitudes to vegetable
and fruit consumption
(p < .01); no effect on
self-efficacy or social influences;
program group
strongest intention to
change consumption in
recommended direction
(p < .01); program group
lower fat scores (26.9 vs.
27.2 grams of fat, p < .01);
greatest effect among pre-intervention
high fat eaters
(9% vs. 3% fat reduction
(p < .01)); both groups
increased vegetable but
not fruit consumption.


Puska et al.
heart disease
Program community
compared to contiguous
comparison community
in Finland; independent
random samples of adults
before (T1), 5 years (T2),
and 10 years (T3) after intervention.
Change in individuals'
knowledge and attitudes
and providing social and
environmental support for
behavior change → reductions
in individuals' risk
factors (smoking, serum
cholesterol, and blood
pressure) → reduced mortality
at population level.
Self-reports of knowledge,
attitudes, and behavior;
physical exam measuring
height, weight, blood pressure,
and cholesterol.
Knowledge of risk factors
increased, but change was
only slightly higher in program
community; no major
changes in health attitudes;
net reduction in
smoking for program community
(27%, p < .001);
net reduction in serum
cholesterol for program
men (p < .01); net reduction
for blood pressure
(p < .01); from 1974 to
1979, coronary heart disease
mortality decreased
(22% in program vs. 12%
in control, p. < .05).
Holder et al.
(1997), prevention
of alcohol
injuries and
Three experimental communities
compared to
matched comparison communities
in United States
over five-year period from
1992 to 1997.
Five main program components
→ produce short-term
effects (e.g., alcohol
server training results in
fewer drinks served to intoxicated
patrons, alcohol
law changes results
in fewer sales to youth) →
combined effect of activities
reduce alcohol availability
and consumption,
youthful drinking and
drunk driving, etc. →
reduced alcohol-related
morbidity and mortality.
Random-digit telephone
surveys of adults, youth
telephone and school
surveys, roadside surveys,
and media coding analysis
measured mediating variables;
emergency room
surveys, records of hospital
admissions, community
surveys, and death
certificate data measured
alcohol-related trauma
and death.
Results not yet available.

efforts by social scientists to evaluate programs in which the program theory is developed during the planning process and tested during the evaluation. At the end of this section, we discuss what was learned from the TBE approach. We then discuss the application of TBE in smaller, less formal programs for which program theory is not clearly defined before the program is implemented.

Murray et al. (1994)

Murray et al. (1994) used a relatively simple theory to evaluate the effects of a statewide antismoking campaign. A tax initiative passed in 1985 by the Minnesota State Legislature designated the use of tax money from tobacco products for antitobacco programs. The initiative was expected to spur local communities and schools to develop a variety of tobacco control programs targeting school-age children. Concurrently, the Department of Health was funded to implement a statewide antismoking mass media campaign. While schools and groups receiving money from the initiative were asked to develop activities using a social influences model of smoking, programs varied across the state.

The evaluators tracked whether the initiative resulted in increased antismoking activities in schools. Exposure to these activities coupled with the media campaign was expected to change beliefs about the health risks of smoking. Changes in beliefs were hypothesized to result in lower smoking rates among youth.

A process evaluation found the Department of Health implemented a mass-media campaign from 1986 to 1990. However, schools made only short-term and erratic increases in antismoking activities.

The evaluators compared Minnesota ninth-graders to Wisconsin ninth-graders in terms of exposure to school-based and mass-media programming, beliefs, and tobacco use each year from 1986 to 1990. Minnesota students reported greater increases in exposure to antismoking messages than Wisconsin students. However, there was no difference between the two states in the pattern of change in antitobacco beliefs. Ninth-graders in both states expressed strong antitobacco beliefs at the start that remained stable over time. Reports of tobacco use in Minnesota declined, but the decrease was not significantly different from that in Wisconsin.

The tax initiative increased students' exposure to antismoking messages in mass media. However, the increase in exposure did not lead to changes in beliefs or behavior. Because of the lack of full implementation

of the school-based program, the authors cannot fully determine whether or how a combined school/mass-media campaign of this type can work. They did learn that a mass-media campaign, while clearly reaching children on its own, does not produce changes in attitudes associated with smoking.

Flay et al. (1995)

In a second antismoking program, Flay et al. (1995) assessed the combined effects of a school and media smoking prevention and cessation program for seventh-graders. Schools were randomly assigned to one of five conditions: a social resistance curriculum only (SR), a television campaign only (TV), a combination of the social resistance curriculum and television (SR/TV), an information-only curriculum (IOC), and a no-treatment control group.

The program theory was defined at the outset. The IOC, SR, TV, and combined SR/TV programs were expected to increase knowledge about the effects of smoking. In the IOC program, this knowledge alone was expected to lead to changes in smoking behaviors. The SR, TV, and combined SR/TV programs included, in addition, activities designed to increase youths' awareness of factors that influence smoking and to increase youths' skills to resist smoking. In turn, students' efforts and confidence in their ability to refuse to smoke were expected to increase. Further, the program sought to change smoking norms. The combined effect of changes in these factors was expected to increase behavioral intentions to reduce, quit, or not start smoking, which in turn were hypothesized to lead to reduced smoking.

Questionnaires were used to measure all mediating and outcome measures before the intervention, immediately after, and one and two years after. For the SR, TV, and combined SR/TV programs, the evaluators found significant positive effects on health and resistance skills knowledge, estimates of the prevalence of smoking (which was assumed to reflect community norms), and efforts to quit. The knowledge and prevalence estimates effects decayed partially but remained significant at the two-year follow-up. The effect of the programs on efforts to quit did not persist at the one- and two-year follow-ups. The programs had no effect on confidence to quit or smoking intentions. None of the programs was related to actual smoking at any posttest. The only predictors of smoking at posttest were smoking at pretest and strong intentions to smoke in the future.


The authors have some difficulty interpreting the results of this study. First, they note that the group targeted, seventh-graders, have very low rates of smoking and strong antitobacco beliefs, effectively reducing the chances of finding a program effect on attitudes or behavior. They suggest that programs for this group either are not necessary or need to continue into later grades. Second, a process evaluation found variability in the curriculum delivery and poor execution of the television programming. However, because of the strong effects of the program on certain mediators, the authors are reluctant to attribute the shortcomings to implementation problems alone. They point out that the program affected knowledge about resistance skills and prevalence estimates. While these effects decayed slightly, they persisted over time. Effects on skills were smaller and decayed more quickly. On the basis of these findings, they suggest that an effective preventive program will need to commit to long-term reinforcement of both knowledge and skills. They conclude, however, that given the lack of program impact on smoking rates, the field has more to learn about how programs work.

Eisen et al. (1992)

A well-known individual-level theory of health behavior, the Health Belief Model (HBM), posits that a person's behavior is influenced by perceptions of his/her own susceptibility to the effects of an action, the potential seriousness of these effects, and perceived benefits and barriers to action. Applying the HBM to the issue of adolescent pregnancy, Eisen et al. (1992) developed a school curriculum to target the mediating factors identified in the theory. The curriculum developed was intended to increase students' awareness of the probability of becoming pregnant or causing a pregnancy, the serious negative consequences of pregnancy, and the benefits of delayed sexual activity or contraceptive use. It was also designed to decrease participants' perceptions of barriers to abstinence or contraceptive use. Changes in these mediating factors were expected to lead to changes in contraceptive and sexual behavior. Past research led the authors to believe that males and females and virgins and nonvirgins would respond differently to the program.

The evaluation was designed to assess the impact of the program on sexual behavior and the importance of the mediating factors in the HBM. Six family planning services and one school were selected to participate in the project. Youth aged 13 to 19 participating in these agencies'

programs were randomly assigned to receive the HBM curriculum or the agencies' usual sex education program.

The evaluation measured sexual and contraceptive knowledge, the sexuality-related beliefs identified by the HBM model, and contraceptive and sexual behavior (the targeted outcomes) before, immediately after, and 12 months after program completion. Knowledge in both the experimental and the control group increased at posttest, with the experimental group showing a greater increase. Participants' health perceptions pre- to postintervention improved in both the control and the experimental group; the experimental group showed no greater increase than the control group.

All behavioral results were presented by gender and previous sexual experience. The HBM program produced significantly greater positive changes among males who were sexually active before the program in terms of their self-reported “contraceptive efficiency” (a measure of the consistency with which teens used effective birth control methods). Females in the comparison group who became sexually active after the initial baseline survey reported significantly more effective contraceptive use than those in the program group.

Similar to the Flay et al. study, the authors cite the low rate of sexual activity among the age-group included in the study as one possible explanation for the limited postintervention differences found between the two programs (more than half the teens were still virgins at follow-up). In addition, the authors note that the number and length of sessions of the comparison and intervention program were similar and acknowledge that the two programs may have had the same potential for impact. In part because of these design problems, the authors do not reject the potential importance of the HBM beliefs. While the evaluators did not find a difference in postintervention HBM beliefs between the HBM curriculum and comparison groups, they found that those who perceived fewer barriers and greater benefits of birth control use before the intervention were more likely to be abstinent after the intervention (if they were virgins) and more effective users of contraception (if they were sexually active). This finding suggests that at least some of the HBM beliefs are important in determining sexual behavior. The authors suggest that program theory may operate differently for the specific groups in the study. They hypothesize that the HBM program may have been more effective than the comparison program for sexually active males because it focused on increasing males' awareness of the risk of pregnancy. They

suggest that in contrast, females may be saturated with messages about the threat of pregnancy and so learned less from the HBM program. The authors conclude that differences in program impact by previous sexual experience and gender indicate that programs must be tailored to specific groups.

Brug et al. (1996)

A study by Brug et al. (1996) presents a theory-based approach to evaluating a nutrition education program in the Netherlands. The program aimed to reduce fat intake and increase fruit and vegetable consumption among study participants by programming computers to ask questions about participants' dietary intake and determinants of and barriers to changing their behaviors. The computer then generated messages specifically tailored to individuals' needs. The program theory posited that the tailored messages would change individuals' attitudes toward the desired diet, perceived social influences, beliefs about their ability to perform the behavior (self-efficacy), and intentions to make dietary changes. In turn, these changes would lead participants to reduce their fat intake and increase their vegetable and fruit consumption.

In the evaluation, employees in a Royal Shell laboratory were randomly assigned to receive the program or to receive general nutrition information. Questionnaires were used pre- and three weeks postintervention to assess changes in dietary fat, fruit and vegetable intake, and changes in attitudes, social influences, and self-efficacy expectations.

At posttest, the program group that had received tailored messages had more positive attitudes toward increasing consumption of vegetable, fruits, and fat than the comparison group. No effects were found for self-efficacy or social influences. Program participants expressed stronger intentions to change their consumption of fat and fruit in the recommended direction than the comparison group. At posttest, the program group had significantly lower fat scores than the comparison group. Both groups increased their reported vegetable but not their fruit consumption with no difference found between the changes in the two groups.

The authors conclude that interventions targeting psychosocial beliefs may be more effective than programs providing dietary feedback alone. They suggest that general information regarding recommended consumption of fruit and vegetables may be adequate to induce change,

whereas messages tailored to the individual may be more effective for influencing the more complex issue of fat intake.

Puska et al. (1985)

The North Karelia project in Finland is a much cited study of a community-based approach to the reduction of heart disease. The project team implemented mass-media campaigns and worked with health and community groups to provide health education information to individuals and to initiate changes in the social and physical environment to motivate and maintain behavior change. The basic theory of the program was that targeting individuals' knowledge and attitudes and providing social and environmental support for behavior change would lead individuals to reduce their risk factors for cardiovascular disease. In turn, reductions in individual risk factors, specifically smoking, serum cholesterol, and blood pressure, would lead to reduced morbidity and mortality at the population level. The authors state that the theory of how the program would work was only partially developed at the outset of the program because of the paucity of theories related to community interventions. Thus, the evaluation tracked individuals' knowledge, attitudes, and behavior, the most developed part of the program theory, while documenting program-related activities in the community through a process evaluation.

Independent random samples of adults were surveyed before the intervention in 1972 and again five and 10 years later. Results for North Karelia were compared to those of a comparison community. The surveys asked questions about health knowledge, attitudes, and behavior and included an exam to measure height, weight, blood pressure, and cholesterol. Finally, mortality rates for the two communities were compared.

The evaluation showed that knowledge of risk factors increased somewhat but that the change was only slightly higher in North Karelia. The evaluators reported no major changes in health attitude measures with little difference between the two communities. A net reduction in smoking and blood pressure was found in North Karelia over the 10-year period. A net reduction in serum cholesterol was found for men in North Karelia. Coronary heart disease mortality decreased by 22% compared to 12% in the comparison community.

The lack of change in individual's knowledge and attitudes led the

authors to attribute the success of the project more to its community organization aspects than to the project's efforts to target individuals. They believe that health education messages disseminated through the media and opinion leaders gradually created an environment promoting healthier lifestyles and that these environmental changes gradually influenced individual behavior.

Holder et al. (1997)

The Community Prevention Trial provides a final example of a theorybased approach to evaluation. The study has only recently concluded, and, as study results were not yet available, the authors' paper described here reviews only the evaluation plan. The project used a combination of community awareness and policy-related activities to reduce alcoholinvolved injuries and deaths. The program consisted of five main components that were expected to reinforce one another: working with community groups to plan and implement prevention activities and develop public awareness of alcohol-related trauma; working with alcohol beverage servers and retailers to design and implement safer beverage service policies; developing community programs to reduce underage drinking; working with law enforcement, retail establishments, and others to reduce drinking and driving; and using municipal control policies to reduce alcohol availability.

The hypothesized program model is presented in Table 7.2. Each component of the program is hypothesized to lead to certain short-term effects. The combined effects of the components are expected to lead to changes in intervening variables, such as alcohol availability and consumption, youthful drinking, and drinking and driving. These changes in turn are expected to lead to reductions in alcohol-related morbidity and mortality.

A process evaluation monitored the extent of implementation of the different components within each community in order to assess the timing and quality of the intervention activities. Surveys and media content analysis were used to measure changes in the mediating variables in Table 7.2. Alcohol-related trauma and death were measured using hospital and community records and surveys. Results from the three experimental communities will be compared to matched comparison communities for a five-year period from 1992 to 1997.

Results of the trial are not yet available. Therefore, we cannot yet tell how well the program theory represents the actual processes that were

Component Program
values, and
Media advocacy,
Media coverage of
beverage service
Establishment of
RBS standards,
server training
Changes in server
behavior, reduction
in DWI
Underage drinking In-school programs,
Restricted youth
access, decreased
youth drinking
Risk of drinking
and driving
Increased DWI
Lowered BACs
among drivers,
reductions in
Access to alcohol Density restrictions,
Decreased density,
decreased adult
implemented or how well the program theory produces the expected results. The authors make several statements relevant to the use of TBE. First, they state that community-based approaches are complex efforts and that little is known about how they work. They argue that a theorybased approach is essential to build a better understanding of these processes. Second, they argue that when the processes by which an intervention work are not well understood, it is preferable to do an in-depth study of a few communities rather than a larger, multicommunity study.


While each of the evaluations has some modicum of theory involved, the authors are not always explicit about what they learned from TBE over and above what they would have learned without it. Here we draw conclusions about findings from the theory-based approach to evaluation

based on our own interpretation of the published evaluation reports. One of the incidental learnings we derive from this review is that authors do not always make explicit the relation of their data to the theory of the program. They report in traditional ways without necessarily emphasizing insights about mechanisms of change that the theory-based approach provides. At times our interpretation ranges beyond the authors' reported conclusions (with all the perils of such extrapolation) because of this omission. We offer the following interpretations in the spirit of examining the types of learning TBE can provide. We highlight particular lessons from each study (with the understanding that more than one study may illustrate the point).

The Difficulties of Going to Scale

The antismoking program evaluated by Murray et al. was an effort to take positive evaluation findings from small-scale antismoking programs and “scale up” to the state level. Minnesota appropriated tobacco tax money for the purpose. The evaluation found that a key component of the intervention was not well implemented, and thus the learnings from the theory-based approach were limited.

The study revealed that schools did not conduct strong or consistent antitobacco interventions. Although the media component reached the ninth-grade audience, this intervention alone did not change beliefs or behavior. Because schools' enthusiasm for antismoking curricula was short-lived, the full theory could not receive an adequate test at the state level.

Finding Intermediary Changes

The study by Flay et al. confronted the endemic question of the appropriate age to begin prevention activities, an issue common to smoking, sex, and drug and alcohol programs. An outcome evaluation of the smoking program evaluated by Flay would have discovered no impact on the smoking rates of seventh-graders. This finding, coupled with low rates of smoking in the study age-group, led the authors to question the utility of targeting such a young group. However, as the authors note, results from the TBE indicating a change in attitudes among the youth suggest that programs targeting this age-group may have future effects on smoking. The additional finding that skills and knowledge have differential attrition rates suggests the need to reinforce attitude

changes over time. While the two findings do not conclusively answer the question of whether seventh grade is the appropriate age to begin smoking prevention efforts, they suggest that some effects do appear in seventh-graders and that perhaps program efforts need to be reinforced over time.

Raising Questions

The theory-based evaluation of the sex education course evaluated by Eisen raises many questions about how both the HBM curriculum and the comparison program work. An outcomes evaluation that broke down study results by gender and previous sexual experience would have discovered the differential impact of the programs. Given the explicitly theory-based design of the program, it would have been reasonable to assume that participants' health beliefs were differentially influenced by the program. However, the assessment of beliefs indicates that this was not the case and leads us to question the underlying mechanisms by which both programs worked. Was the HBM curriculum simply inadequate to change important HBM beliefs? If indeed the preintervention HBM beliefs found to be associated with positive outcomes in both groups were able to be changed by the program, would the program outcomes have been different? What accounts for the differences by gender and past sexual experience when little difference was found in health beliefs? What are possible alternative intervening factors that led to the changes in behaviors found?

Identifying Possibly
Unnecessary Program Components

The Brug et al. evaluation of the computer-generated nutrition program found that certain mechanisms of change identified by the theory were affected among program participants but that others were not. Nonetheless, participants made significant improvements in their consumption of fat. An outcome evaluation would have indicated only that the program was successful. The additional findings contributed by TBE suggest the need to explore whether those program elements that were not associated with the dietary improvement can be eliminated in future programs. The authors also suggest the need to study longer-term effects of tailored nutrition programs.


Contributing to a Paradigm Shift

The North Karelia study was one of the first large-scale health promotion studies to attempt to influence the health of individuals through community change efforts. The authors state that the program theory was only partially developed at the start of the project because of an absence of theories about community-level change. The evaluators systematically tracked one of the proposed mechanisms of change at the individual level—and discovering little change. Although the hypothesized individual-level mechanisms were not affected, they nevertheless discovered an overall positive impact of the program on health outcomes. The findings supported a shift from changing only individual knowledge and attitudes to focusing on changing the social and cultural environments in which individuals live.

Providing Clarity and Focus for Evaluation

Twenty years later, in an evaluation of a community-level health promotion effort, Holder et al. explicitly discuss the importance of laying out a program theory for the evaluation of large-scale community-based programs. With the results not yet in, the benefits of the TBE approach for exploring outcomes in this study cannot be assessed. However, the delineation of the program theory of such a complex change effort provides the evaluator with clear guidelines for data collection and analysis and will undoubtedly assist in the clarity of interpreting results and drawing conclusions.


The evaluations described here represent well-funded, primarily re-search-oriented projects. The programs themselves were developed based on theory, although rather rudimentary theory in some cases, and the evaluations test the theory embedded in the program. However, most programs are not explicitly based on theories of change, and many are not well-defined at the planning stage. They are often small programs, based on practitioner experience, and funded at relatively low levels. Is TBE relevant to them? Can it be applied in

these settings, and what can be gained? It is our belief that theory-based evaluation can be applied in these settings and that it has the potential to provide benefits that match those provided for programs with well-articulated theory. Even if the evaluators do not adopt the language of TBE, they can incorporate elements of it into their studies.

Here are some kernels of advice gleaned from this review that we believe apply to even small, atheoretical, marginally funded programs.

Consider Theory Development
as a Stage in the Evaluation

An emphasis on theory development at the start of the evaluation may be, in and of itself, the most beneficial aspect of the theory-based approach.

In many fields, programs are planned on the basis of experience, professional savvy, intuition, and beliefs in fashion in the field. There is little in this melange that is easily characterized by the name of theory. But all programs have a theoretical basis no matter how weakly the assumptions are articulated. Program people make some assumptions about why the set of activities they plan will lead to desirable outcomes. When the assumptions are tacit rather than consciously expressed, the evaluator has the task of eliciting or constructing the theoretical assumptions underlying the program. The evaluator usually undertakes this theorysurfacing exercise in conjunction with program planners and program staff.

When program staff describe their assumptions about the mechanisms by which the program will bring about change, the evaluator has to see whether the theories offered are operative. She has to learn enough about program activities—real activities in action and not just espoused ideas of activities—to figure out whether practitioners' theories are being operationalized in the program. If the program claims that its prevention goals will be realized by increasing participants' sense of selfefficacy but the evaluator finds no activities dedicated to increasing selfefficacy, she will doubt that the theory is “real” for this program. She will have to seek an alternative theory that explains why the real activities are expected to lead to the intended effects. Or she may call the absence of efficacy-building activities to the attention of program staff and perhaps encourage them to alter the program to fit their theory (or alter their theory to fit the program).

Possible sources of program theories are social science theories and research, prior evaluations, planner and practitioner expectations, the

evaluator's knowledge and experience with programs of similar type, and her own logical thinking. Often the evaluator will cycle through several of these sources—asking program planners and program managers, reviewing existing theories in the field, reviewing previous research and evaluations, hypothesizing a theory on the basis of this information, and then negotiating her formulation with program managers and staff to come to an agreement that accords with their thinking.

When people do not agree on their assumptions, the evaluator may incorporate several different chains of reasoning into the study. For example, staff in a pregnancy prevention program may expect the program to work because it provides information about contraception, or because it teaches young women to be more assertive in demands on their partners, or because it makes chastity more socially acceptable within the program group, or for a number of other reasons. Rather than demanding that program staff agree on a single theory, the evaluator can design the study to collect data on several different assumptions. The study can then show which of the theories, if any, is supported by the data.

Identifying poorly defined, implausible, or hotly debated links in how a program is expected to work can force practitioners and managers to define and agree on what they believe they are doing. With greater clarity about what they are trying to accomplish, and how, program people should do a better job of developing and improving programs. And with a greater understanding of how the program is expected to work, the evaluator can better structure the evaluation to answer relevant questions and interpret study results. Even if the evaluator does not have adequate resources to collect data on all the mediating variables identified in a program theory, developing the program theory as the starting point in evaluation can be of value for program development, improvement, and evaluation.

Do Not Expect Theories to Be Completely Right

Remember when conducting a theory-based evaluation that at this stage the theory is primarily a guide to the evaluation. Managers, practitioners, and evaluators of small programs who attempt retrospectively to define program theory should not expect the theories they develop of the program to be exactly right. In fact, in none of the five theory-based evaluations described here was the original theory borne out. (The Holder et al. study of alcohol-related traffic accidents is still in progress.)

Some of the programs activated the mechanisms that were assumed to trigger desired outcomes, but the desired outcomes did not appear. Some of the programs produced the desired outcomes but not through the mechanisms that were assumed to be causal; there were no significant changes in the mechanisms. Obviously, one lesson from these theory-based evaluations is that program planners have limited understanding of how and why programs work or fail to work. Much more effort is needed to understand what it takes to get people to reduce risk behaviors and adopt positive lifestyles. The processes of change are complex, dependent on many variables and on higher-order interactions among variables. Concerted research is required to unravel the many contingencies in behavioral change.

In the studies reviewed, we have come across cases where program theories were not well supported by the data but where the evaluators were reluctant to abandon the theories. No sin inheres in such caution. One single null finding should not lead to the rejection of a sensiblesounding theory; any one study may provide misleading evidence. However, if repeated evaluations show the same pattern, if a theory is disconfirmed again and again, there seems little reason to cling to it. Evaluators should join with program planners to seek more effective theories, develop programs that embody the theories, and conduct evaluations to test them.

One reason for staying with a well-known and well-respected program theory, even when the evidence is slim, is that it often seems so intuitively right. Another reason why people sometimes cling to a familiar, if ill-supported, theory is that few alternatives seem to be waiting in the wings. If programs do not lead to desired outcomes through the mechanisms of knowledge, attitude change, social support, revisions in community norms, or whatever else the theory posits, it is usually hard to figure out what mechanisms would be effective. This represents a challenge to the whole field of wellness promotion.

Include a Process Evaluation

A TBE study should include a description of program implementation, generally called a process evaluation. Some programs do not work because planned activities are not carried out regularly or not carried out well. If an evaluation fails to collect data on the processes of the program, it will be unable to distinguish between “program failure” (the program was not carried out well and therefore did not lead to the desired

effects) and “theory failure” (the idea underlying the program was wrong and therefore expected results did not materialize) (Suchman, 1967). If the study of implementation is forgone, the evaluator loses the ability to tell whether this particular realization of the program or the basic theory is at fault.

One possible reason for the observed outcomes of the TBE studies we have reviewed is that programs work differently for different subgroups. Or outcomes may differ according to the type and frequency of activities to which participants are exposed. Again, process evaluation is essential. The evaluator is well advised to collect data on those characteristics of participants, staff, activities, settings, and time that are likely to be salient to success. In the analysis, the evaluator can examine outcomes in light of the features of people and program activities and thus come to estimates of which activities work best for which groups under which conditions.

When a program is poorly implemented, there may not be a great need to delve deeply into all the hypothesized causal links in the theory chain. Sometimes even with a poorly conducted program, some later expectations are realized. It is difficult to explain positive outcome data by assuming that they were brought about by a weak or incompetent program. Other information is needed. On the other hand, if the outcomes of a poorly conducted program are disappointing, as one might expect, the immediate lesson is to do a better job of conducting the program—and evaluate the program in its superior version.

Use the Information TBE Can Provide

While many funders of small-scale programs ask for outcome evaluations, the results of these evaluations are rarely used to make go/no-go decisions about the program's future. Rather than using evaluative information for continuing or discontinuing the program, funders and program managers usually want information about how to improve the program. The results of theory-based evaluation may prove to be more useful than evaluations of outcomes only. Knowledge of how or why a program is failing provides managers with leads to how to improve the existing program. They can use the information to discuss why programs should work and where they are breaking down. This type of information communicates easily to policy makers and may be more convincing than the results of outcome evaluations only.

Theory-based evaluation has a longer tradition among academic researchers

developing and testing program approaches than among program evaluators called in after a program is in process. We noted that traditional methods of reporting results sometimes got in the way of clearly associating the evaluation results to the theory in the evaluations reviewed. As TBE becomes more widely used in practice, evaluators will need to develop effective ways to report on the findings.


How well the evaluator is able to test a program theory is likely to be a function of three factors. The first is how well the theory is defined. The evaluator may have to consult many stakeholder groups to identify and secure agreement on the definition of the theory underlying the program. Second is how well program activities reflect the assumptions embedded in the theory. Third is the grubby matter of money and time. If TBE is carried out in full detail, it is apt to be an expensive and timeconsuming enterprise.

As several of the examples show, even if an evaluation is able to track a specific theory, the interpretation of results may not be straightforward. Findings that indicate that the program altered some of the hypothesized intervening factors but not others do not lead to easy interpretation. Assessing exactly where the hypothesized theory breaks down, and why, calls for more finely grained study than most evaluations have yet included. A basic limitation is the state of knowledge in the field about which factors are effectual in bringing about change and how factors interact to change human beings and human communities. Advocates of TBE hope that over time results from TBE will lead to cumulative knowledge of change processes and consequently the development of more effective wellness programs. As some of the examples show, even within its limitations TBE is contributing to a growth in knowledge in the health promotion field.


Bickman, Leonard, ed. (1990). Advances in Program Theory. New Directions for Program Evaluation 47. San Francisco: Jossey-Bass.

Brug, J., Steenhuis, I., Van Assema, P., and H. De Vries. (1996). “The Impact of a Computer-Tailored Nutrition Intervention.” Preventive Medicine, 25, 236–242.


Chen, Huey-tsyh. (1990). Theory-Driven Evaluation.Newbury Park, Calif.: Sage.

Chen, Huey-tsyh, and Rossi, P. H. (1987). “The Theory-Driven Approach to Validity.” Evaluation and Program Planning, 10, 95–103.

Chen, Huey-tsyh, and Rossi, P. H. eds. (1992). Using Theory to Improve Program and Policy Evaluations.New York: Greenwood Press.

Costner, H. L. (1989). “The Validity of Conclusions in Evaluation Research.” Evaluation and Program Planning, 12, 345–353.

Donaldson, S. I., Graham, J. W., and Hansen, W. B. (1994). “Testing the Generalizability of Intervening Mechanism Theories: Understanding the Effects of Adolescent Drug Use Prevention.” Journal of Behavioral Medicine, 17, 195–216.

Eisen, M., Zellman, G., and McAlister, A. (1992). “A Health Belief Model–Social Learning Theory Approach to Adolescents' Fertility Control: Findings from a Controlled Field Trial.” Health Education Quarterly, 19 (2), 249–262.

Feindler, E. L., Marriott, S. A., and Iwata, M. (1984). “Group Anger Control Training for Junior High School Delinquents.” Cognitive Therapy and Research, 8, 299–311.

Finney, J. W., and Moos, R. H. (1989). “Theory and Method in Treatment Evaluation.” Evaluation and Program Planning, 12, 307–316.

Flay, B. R., Miller, T. Q., Hedeker, D., et al. (1995). “The Television, School and Family Smoking Prevention and Cessation Project, VIII: Student Outcomes and Mediating Variables.” Preventive Medicine, 24, 29–40.

Goodman, R. M., and Wandersman, A. (1994). “FORECAST: A Formative Approach to Evaluating Community Coalitions and Community-Based Initiatives.” Journal of Community Psychology, special issue, 6–25.

Holder, H. D., Saltz, R. F., Treno, A. J., et al. (1997). “Evaluation Design for a Community Prevention Trial.” Evaluation Review, 21 (2), 140–166.

Lipsey, M. W. (1993). “Theory as Method: Small Theories of Treatments.” New Directions for Program Evaluation, no. 57, 5–38.

Lipsey, M. W., and Pollard, J. A. (1989). “Driving toward Theory in Program Evaluation: More Models to Choose From.” Evaluation and Program Planning, 12, 317–328.

Murray, D. M., Prokhorov, A. V., and Harty, K. C. (1994). “Effects of a Statewide Antismoking Campaign on Mass Media Messages and Smoking Beliefs.” Preventive Medicine, 3, 54–60.

Pentz, M. A., Dwyer, J. H., MacKinnon D. P., et al. (1989). “Primary Prevention of Chronic Diseases in Adolescence: Effects of the Midwestern Prevention Project on Tobacco Use.” American Journal of Epidemiology, 130, 713–724.

Puska, P., Nissinen, A., and Tuomilehto, J. (1985). “The Community-Based Strategy to Prevent Coronary Heart Disease: Conclusions for the Ten Years of the North Karelia Projects.” Annual Review of Public Health, 6, 147–193.

Suchman, E. A. (1967). Evaluative Research.New York: Russell Sage Foundation.


Weiss, C. H. (1972). Evaluation Research: Methods for Assessing Program Effectiveness.Englewood Cliffs, N.J.: Prentice Hall.

Weiss, C. H. (1995). “Nothing as Practical as a Good Theory: Exploring Theory-Based Evaluation for Comprehensive Community Initiatives for Children and Families.” In J. P. Connell, A. C. Kubish, L. Schorr, and C. H. Weiss, eds., New Approaches to Evaluating Community Initiatives: Concepts, Methods and Contexts.Washington, D.C.: Aspen Institute.

Weiss, C. H. (1997). “How Can Theory-Based Evaluations Make Greater Headway?” Evaluation Review, 21, 501–524.

Weiss, C. H. (1998). Evaluation: Methods for Studying Programs and Policies.Upper Saddle River, N.J.: Prentice Hall.



Patricia L. East

This chapter describes a previously unrecognized target population for adolescent pregnancy prevention: the younger sisters of childbearing teens. Data from several studies show that the younger sisters of adolescent mothers have teenage childbearing rates two to six times higher than women in the general population (Cox et al., 1993; Friede et al., 1986; Goldfarb et al., 1977). These studies involved large samples—and, in two cases, statewide samples—of Black and non-Black adolescents in both urban and rural settings, so that such findings appear to be robust (East and Felice, 1992). The younger sisters of adolescent mothers have also been shown to have higher rates of adolescent sexual activity (East et al., 1993; Hogan and Kitagawa, 1985), to be younger at first sexual intercourse and first pregnancy (Hogan and Kitagawa, 1985), and less likely to use contraception when sexually active (Hogan et al., 1985) than other girls of their race and social class.

Although these studies are useful for identifying a population at increased risk of early sexual activity and early pregnancy, they fail to address how or why the sisters of childbearing teens become vulnerable to such outcomes. This chapter describes two groups of causal factors that predispose the younger sisters of pregnant and parenting teens to early pregnancy. The first group of causal factors concerns sisters' shared background and consists of factors equally present for all siblings within a family. These shared preexisting predispositions include the parenting the sisters received, their ethnic identity and socioeconomic status,


Figure 8.1. Sisters' shared vulnerability to early child bearing.

their exposure to shared community-neighborhood norms, a shared biological predisposition, and a common peer-friendship network (see Figure 8.1). The second group of causal factors that would increase a younger sister's risk for early pregnancy are factors that result from the older sister's pregnancy and childbearing and involve interactions between the childbearing teen and her younger sister and family during the postpartum period. These processes include social modeling of the older sister, socialization to early motherhood through child care provided to the sister's child, diminished quality of parenting by the sisters' parents, and increased family stress and family economic hardship resulting from the older sister's pregnancy.

It is further argued in this chapter that these unique effects resulting from the pregnancy of the older sister contribute to younger sisters' preexisting predispositions toward early childbearing to render younger sisters even more vulnerable to early pregnancy and early childbearing. Data are then reviewed that highlight the high level of problem and delinquentlike behavior and the relatively high rate of sexual activity among the younger sisters of childbearing teens. These behavioral proclivities are

significant because they point to several transition behaviors that likely lead to teenage pregnancy and childbearing. Finally, specific considerations for intervening with this unique target population are discussed.


Shared Parenting Influences

One line of reasoning that can be used to explain sisters' similarity in adolescent sexual initiation and permissiveness—and, therefore, risk for adolescent pregnancy and parenthood—is that because all children within a family experience similar parenting styles and disciplinary techniques, they are equally likely to engage or not to engage in sexual activity during adolescence. It has consistently been shown that lack of parental control and discipline are related to adolescents' permissive sexual attitudes and sexual behavior (Baumrind, 1983; Ensminger, 1990; Miller, 1998; Miller et al., 1986). For example, it has been found that sexually active adolescent females have more permissive family rules about the time they are expected home on school nights (Ensminger, 1990) and have lower overall parental supervision (Miller et al., 1986) than nonsexually active girls. Most research on parents' discipline styles characterizes a parent as having a general parenting style (i.e., authoritarian, permissive, or authoritative) that is applied similarly to all children in the family (Baumrind, 1983). To the extent that parenting style is consistent for all children in the family—that is, that all children in the family are controlled and disciplined similarly—then all children within the family would be equally at risk for early sexual behavior and, consequently, early pregnancy and early childbearing.

Related to parents' discipline style, parent-adolescent communication has also been shown to be important for delaying adolescent sexual activity and for pregnancy prevention (Fox, 1980; Fox and Inazu, 1980; Newcomer and Udry, 1985). For example, girls who are the most effective contraceptive users prior to pregnancy are more likely than nonusers to have mothers who communicate information about how and where to obtain contraception (Moore et al., 1986a). Another study reported that daughters who talked with their mothers about sexual and contraceptive matters were more likely than others to postpone sexual activity or to employ effective contraceptive methods if they did become sexually active (Fox and Inazu, 1980). If it is the case that mothers share the

same communication openness with all their daughters, then such effects would be consistent across all girls within the family. That is, all daughters of noncommunicative mothers would be equally at risk for sexual permissiveness and pregnancy, and all daughters of communicative mothers would be equally likely to postpone sexual initiation or to use effective contraception.

Another aspect of parental behavior that affects adolescent sexual activity is parents' marital status. Much research has shown that adolescents from single-parent families (i.e., in which the parent is divorced or was never married) are more likely to engage in nonmarital sexual intercourse than adolescents from two-biological-parent families (Hogan and Kitagawa, 1985; Jemmott and Jemmott, 1992; Kinnaird and Gerrard, 1986; Miller, 1998; Miller and Bingham, 1989; Newcomer and Udry, 1987). For females, the amount of time spent in a single-family household was also found to be associated with a young onset of sexual intercourse (Miller et al., 1997). Most studies examining parents' marital status as a correlate of adolescent sexual behavior conceptualize marital status not just as a family structure variable but as a proxy of parental control, family functioning, and family disruption, with adolescents in single-mother households experiencing less parental control and supervision than teens in two-parent households (Newcomer and Udry, 1987). Given that single parenthood is consistent for all children in the family, this may be another circumstance that encourages a resemblance in sibling sexual behavior during adolescence. That is, because sisters are equally exposed to the loss of parental control and the diminished parental monitoring associated with single-parent households, they would be equally at risk for permissive sexual behavior and, thus, teenage pregnancy.

Shared Societal Risks

A second conceptualization for understanding sisters' shared vulnerability to adolescent pregnancy and childbearing is that sisters within a family share a socioeconomic status, social class, and an ethnic identity that are highly correlated with early sexual initiation and early pregnancy. Key socioeconomic indicators related to adolescent pregnancy and childbearing are parental education and income, with higher maternal education correlated with later sexual initiation (Hayward et al., 1992), a greater likelihood of contraceptive use at first intercourse (Kahn et al., 1990), and a higher probability of abortion given pregnancy (Plotnick,

1992). Regarding race and ethnicity, Black and Hispanic women have teenage birth rates 2.7 times higher than White teenage women: the rates for 1994 were 108 births per 1,000 females aged 15 to 19 for both Hispanics and non-Hispanic Blacks and 40 births per 1,000 women aged 15 to 19 for non-Hispanic Whites (Moore et al., 1997). It has also been shown that teenagers from high-risk environments—characterized as low socioeconomic status, inner-city residence, and a nonintact family—have pregnancy rates as much as 8.3 times higher than girls from low-risk environments—or being of upper to middle class, residing in a suburban neighborhood, and being from an intact family (Crane, 1991; Hogan and Kitagawa, 1985). Hogan and colleagues (1985) discuss this difference as likely reflecting the lack of knowledge about and access to effective birth control in high-risk neighborhoods. Other characteristics of the neighborhood itself—such as employment opportunities, neighborhood quality, and social disorganization—also have been shown to relate to adolescent sexual activity (Brewster et al., 1993).

Several social scientists have discussed the conditions of economic uncertainty and the poor prospects for marriage and for a stable job that are endemic to poor minority youth as contributing to many teenagers' choice of early parenthood as a pathway to adulthood and adult status (Dash, 1989; Hamburg, 1986; Hayes, 1987; Luker, 1991, 1996; Moore et al., 1986b; Williams, 1991). Poverty, uncertain economic conditions, and low prospects for advancement discourage many young women from making the transition to adulthood through educational or career achievements or through parenthood after marriage. Instead, the lack of opportunity leads to frustration and the search for adult status via more accessible routes, such as through sexual activity, pregnancy, and parenting.

Sociologist Kristin Luker states that

teen pregnancy is less about young women and their sex lives than it is about restricted horizons and the boundaries of hope. It is about race and class and how those realities limit opportunities for young people. Most centrally, however, it is typically about being young, female, poor, and non-white and about how having a child seems to be one of the few avenues of satisfaction, fulfillment, and self-esteem. (Luker, 1991, p. 83)

The disproportionate number of pregnant and parenting teenagers observed among poor, minority, inner-city residents may well reflect the absence of employment, education, or marriage options that typically mark the path from adolescence to adulthood. Such life-course experiences

would be similar for all sisters in the family and thus may act to create similar early childbearing behavior among them.

Shared Community-Neighborhood Norms

Many scholars of teenage pregnancy have discussed the strong socialization pressures and multiple role models of young unmarried mothers as powerful contributors to both the acceptability of teenage pregnancy and its prevalence (Crane, 1991; East and Felice, 1992; Hogan and Kitagawa, 1985; Moore et al., 1986b). The social norm framework, derived from the family life-course perspective, states that the scheduling of life-course events associated with the transition to adulthood (e.g., moving out of one's parents' house, getting married, having a child) is relatively well defined within communities, with clear and socially recognized norms for the timing and sequencing of transitional events (Elder, 1975; Haraven, 1978; Marini, 1984; Neugarten et al., 1965). In their ethnographic studies of Black urban and rural neighborhoods, Stack and Burton (1993) discuss the community norms for early childbearing as powerful socialization pressures, with “early childbearing often considered a necessary activity” (p. 159). These expected norms, as well as community sanctions against abortion, undoubtedly reflect the high value that family and kin place on having a child of one's own. Moreover, teenage mothers within both Black and Hispanic communities are often cared for and supported by extended family to a greater extent than are White teenage mothers, with “kin-keeping” the primary means of sustenance among poor urban families (Brindis, 1992; Burton, 1990; Duany and Pittman, 1990; Ladner, 1988; Melville, 1980; Stack, 1974). Such strong family, extended-family, and community support systems work to reduce the stigma associated with nonmarital teenage childbearing and make it a more acceptable lifestyle for Black and Hispanic women.

Shared Biological Predispositions

For heuristic purposes, it is also important to consider that a shared biological predisposition may be contributing to sisters' similarity in sexual and birth timing. Numerous studies show that the timing of puberty (often measured as the age at menarche for girls) is related to the timing of first sexual intercourse, first marriage, and first birth (Presser, 1978; Udry, 1979; Udry and Cliquet, 1982). Furthermore, the timing of puberty is known to be correlated across generations, with age at menarche

highly correlated for mothers and daughters and between sisters (Garn, 1980). Thus, it follows that sisters would have a similar proclivity toward early pregnancy and early sexual behavior because they share the same maturational predisposition for early puberty, a predisposition at least partly inherited from their mother. Some have argued that the intergenerational pattern of teenage childbearing may simply reflect this biological predisposition for early maturation (Kahn and Anderson, 1992), and Newcomer and Udry (1984) were able to confirm links between mothers' pubertal development, daughters' pubertal development, and daughters' sexual onset. Thus, there is reason to suspect that the increased vulnerability for early childbearing among the younger sisters of childbearing teens may have biological origins and that such influences should be carefully considered in future research.

Shared Peer-Friendship Network

Finally, it is possible that sisters have a common peer group or friendship network that is inducing both sisters to engage in early sex and early parenting. At least three different processes of friendship influence may be at work in creating behavioral similarity between sisters who share a common peer network. First, it is possible that the friendship network, common to both sisters, might be exerting strong and pervasive pressures for both sisters to engage in early sexual behavior (East and Shi, 1997). It is well known that friends can be powerful reference groups and socialization agents for adolescent sexual activity (East et al., 1993; Mirande, 1968). Thus, sister pairs might be similar behaviorally because both are influenced (i.e., pressured) by the shared friendship network to have sexual relationships.

Second, it is possible that the older sister actively initiates or facilitates her younger sister to join in with the older sister's friends, friends who are likely to be older and sexually active. Thus, for example, the older sister might take her younger sister to older adolescent events or parties where sexual activity will occur or where potential partners will be available. In this case, older sisters are actively accelerating the sexual experience of their younger sisters by exposing them to an older peer group. This type of facilitative process was tested specifically for White and Black older sister–younger sister pairs by Rodgers et al. (1992), who found, however, little empirical support for this type of sibling influence.

Finally, it is possible that sister pairs who share common friends have similar behavioral attributes with similar inclinations for deviance (East

and Shi, 1997). In this case, the younger sister might act in ways similar to the older pregnant or childbearing sister (i.e., in problem behavior and sexual behavior) because both share similar behavioral inclinations, with the shared friendship network only a reflection of their shared behavioral dispositions. Certainly, these and other possible shared peerfriendship group influences might be at work in enhancing the behavioral similarity between sister pairs, or a combination of these processes might be operating.

The preceding discussion suggests that because sisters receive similar parenting, share societal risks, are socialized by the same community norms, have a common biological predisposition, and may share a common peer-friendship network, they are equally likely to become or not to become pregnant as teens. For the younger sister, however, there is an additional factor to consider: the effect of her older sister's pregnancy and childbearing on her and her family of origin. That is, the predispositions toward early pregnancy described previously may interact with or contribute further to the effects of an older sister's childbearing, thereby rendering younger sisters even more vulnerable to early sexual activity and early pregnancy (East, 1998b). These intrafamilial dynamics that occur after the older sister has had her child are described in the following.


Based on theoretical considerations and drawing from various bodies of literature, Figure 8.2 presents a descriptive model of how an older sister's early childbearing affects her younger sister. Principal assumptions of the model are as follows:

  1. A younger sister is prone to modeling her older sister's behavior.
  2. A younger sister's child care involvement with her sister's baby socializes her for early parenting.
  3. The older sister's childbearing diminishes the quality of her parents' parenting of their children and increases parental acceptance of teenage childbearing.
  4. A younger sister is affected by the increased family financial hardship and increased family stress that her sister's childbearing may provoke.


Figure 8.2. How an older sister's early childbearing contributes to her younger sister's risk of early pregnancy.

The next section also takes up another assumption of the model: that a younger sister's attitudes, future expectations, and problem behaviors serve as important mediators to her sexual and childbearing behavior. Each of these assumptions is discussed in turn.

Younger Sisters Model Their Older Sisters' Behavior

Social learning theory (Bandura, 1977) and modeling explanations of behavior would suggest that because older siblings operate as important role models and socialization agents for younger siblings (Bryant, 1982; Cicirelli, 1982) older siblings have the capacity to shape the attitudes, values, and norms—and, ultimately, the behavior—of younger children in the family. Moreover, because large numbers of pregnant teenagers choose to keep their babies and remain at home within their family of origin (Furstenberg, 1980; Trent and Harlan, 1994), many younger siblings witness firsthand their older sister in the role of mother and, most often, of unwed mother. Furthermore, because motherhood establishes a role for the older adolescent—a role associated with status and femininity—a younger sister may become envious or jealous of her and, consequently, attempt to identify with her (East, 1996a; East and Felice, 1992). Hogan and Kitagawa (1985) discuss this socialization process whereby girls who see their sisters become teenage mothers are more likely to (1) accept adolescent parenthood as normative and (2) accept single parenthood as a

way to achieve adult status. These attitudinal changes are thought to precede an increased willingness to engage in early sexual activity (Plotnick and Butler, 1991). Such beliefs may be especially likely to develop among early adolescent younger sisters who are forming attitudes about schooling, employment, marriage, and family formation (Hamburg, 1974). Thus, an older sister's early childbearing can have a profound impact on the development of her younger sisters by exposing them to a role model that appears attractive because of its associated adult status. Research on sibling influence shows that modeling is enhanced under two conditions: when siblings are particularly close (Rowe and Gulley, 1992; Rowe et al., 1989) and when siblings are particularly rivalrous or competitive with one another (East, 1996c). For example, East and Shi (1997) found that the extent of rivalry and competition between a pregnant or parenting teen and her younger sister was the best predictor of whether the younger sister had engaged in delinquent-like behavior and sexual activity. Rivalry was actually a better predictor of girls' behavior than the sisters' closeness or the conflict between them. One must also remember that sibling competition among pregnant and parenting teens and their sisters is typically occurring on a playing field of few resources and limited opportunities. Thus, where one of the few avenues to status and adulthood is through childbearing, it would be very understandable to see a younger sister also become pregnant in attempts to—consciously or subconsciously—rival the older sister.

A Younger Sister's Child Care
Involvement Socializes Her for Early Parenting

When a teenager has a baby, typically all available family resources and personnel are pooled to help parent her child (Burton, 1995, 1996a, 1996b; Burton and Bengtson, 1985; Stack, 1974, 1975). Younger siblings often participate in the care of their teenage sister's child out of social and economic necessity (Ladner, 1988). Child care assistance provided by early adolescent younger sisters may be favored over other strategies because it best utilizes available family personnel and because it fosters child care skills and parent training for girls who themselves will someday assume parenting responsibilities for their own children.

In the author's own study on the siblings of parenting teens, it appears that many younger siblings, particularly younger sisters, help in the care of their older sister's child (East, 1996c). It also appears that younger sibling child care is a cooperative arrangement between the older and

younger sister, with the younger sister co-participating in her niece's or nephew's care by performing instrumental child care tasks but clearly taking direction from the older sister. Such secondary instrumental care typically includes temporarily occupying the baby, getting a toy, getting diapers, or helping to prepare food for the baby. This pattern is similar to one observed by Weisner (1982), who noted that when mothers give older children caretaking responsibilities for their younger siblings, the mothers still retain primary control over child care.

How does helping to care for an older sister's child impact younger sisters? Many anthropologists who study caregiving by older siblings to younger ones in non-Western societies have noted that sibling caretakers gain a great deal of felt competence in their caregiving role and experience a positive sense of autonomy by carrying out their co-caregiving responsibilities (Draper, 1976; Weisner, 1982, 1986; Whiting, 1983; Whiting and Whiting, 1975). Additionally, Whiting and Whiting (1975) observed that sibling care systems involve intricate authority hierarchies and subtle patterns of rank and deference. Thus, the younger sisters, in addition to learning compliance and prosocial behavior, may learn keen positional awareness and sensitivity.

Sibling caretaking as it occurs within teenage childbearing families has been a completely neglected area of study. (Notable exceptions include Burton, 1995, and Burton, 1996b. In these works, however, Burton describes the intergenerational caregiving patterns in teenage childbearing families wherein the teen's sister provides primary care to her nieces and nephews, with the teen mother completely unable to cope with parenting.) In the author's preliminary work, girls who provided extensive child care assistance to their teenage sisters had more pessimistic expectations of graduating high school, had more positive intentions of an early childbearing, and exhibited more permissive sexual behavior than girls who provided little or no child care help to their teenage sisters (East and Jacobson, 2000). Thus, taking part in the care of the older sister's child may have negative ramifications for girls' development. But several questions remain. For example, what is the nature and extent of child care provided by the younger sisters of teen mothers? To what extent does such caretaking socialize younger sisters for early parenthood? Does the perceived child care competence gained by the younger sister lead her to minimize the hardships associated with early parenting or to become less diligent in pregnancy prevention? What are the effects of caretaking the older sister's baby for the younger sister's role aspirations and her attitudes about early parenting? These questions represent potentially fruitful lines for

future research on how an adolescent sister's childbearing may socialize her younger sister for early pregnancy and early parenting.

Adolescent Childbearing Diminishes
the Quality of Parents' Parenting and
Increases Parents' Acceptance of Teenage Parenting

In discussing the burdens and benefits of early childbearing for the ado-lescent's family of origin, Furstenberg (1980) observed that the teenager's parents' own parenting likely suffers because of the increased stress and time constraints of helping to parent their daughter's child. Typically, the teenager's mother provides much hands-on child care for her grandchild (Brooks-Gunn and Chase-Lansdale, 1991; Burton, 1996a, 1996b; Stack, 1975). Because these grandparenting duties can be extensive, they can interfere with the new grandmother's ability to monitor or supervise her own children. With a new baby in the home, the grandmother may find it particularly difficult to monitor her children's out-of-home, peer-related activities. This problem is compounded by the fact that these women are usually parenting and grandparenting alone, without a co-resident adult (Burton, 1996a, 1996b; Stack, 1975). This inattentiveness or unavailability to her own children creates an opportunity for younger siblings to engage in problem or delinquent behaviors, behaviors that are likely to lead to teenage sexual activity and, consequently, teenage pregnancy (Elliot and Morse, 1989; Ensminger, 1987).

There are also several reasons to expect that the older daughter's early childbearing increases her parents' acceptance and tolerance of early nonmarital childbearing in general and of a younger sister's pregnancy and childbearing in particular. Having already experienced the older daughter's early childbearing, parents—as well as other family members—would likely view a teenage pregnancy by another daughter in the family with less stigma and less disgrace (East and Felice, 1994). Moreover, by the sheer fact that the older sister has had a child as a teen, she has in effect “opened the floodgates” or “broken the barrier” of what constitutes allowable and tolerated behavior for other siblings within the family.

Second, the older daughter's early childbearing may signify to her parents the lack of alternate life options available to their children, particularly their daughters. Consequently, parents may rationalize their daughter's early childbearing as a reasonable and acceptable response to the disadvantaged socioeconomic circumstances in which they live. Such tacit parental acceptance of early nonmarital parenthood has been shown

to precede adolescent sexual activity and pregnancy-risk behaviors (Moore et al., 1986a; Thornton and Camburn, 1987).

To understand the consequences of adolescent pregnancy and childbearing for the family, East studied mothers in families in which only one teenager was currently pregnant and this was the first teenage pregnancy to occur within the family (East, 1999). Mothers were assessed twice, 13 months apart. Results indicated that mothers monitored and communicated less with their other children and were more accepting of teenage sex after the older daughter gave birth. These results suggest that both mothers' parenting and their attitudes may be affected by an adolescent's pregnancy and birth.

Adolescent Childbearing Increases
Family Economic Hardship and Stress

Although virtually no research has examined the economic hardship resulting from teenage childbearing on the adolescent's family of origin, there has been much discussion of the negative socioeconomic consequences experienced by the teenage mother herself. These studies, which take into account the socioeconomic disadvantage that preceded the teen's childbearing, show substantial and qualitatively important socioeconomic costs associated with teenage childbearing (Geronimus and Korenman, 1992, 1993; Hoffman et al., 1993a, 1993b). For example, Hoffman and colleagues found that, after controlling for factors associated with teenage childbearing such as race and economic disadvantage, teenage mothers were more likely to be poor and to receive welfare than were their nonteenage childbearing sisters. These socioeconomic costs were above and beyond what the teen would have experienced had she not become a mother. Given that most teens continue to live with their parents for some time after the birth (Trent and Harlan, 1994), it is likely, then, that the adolescent's family of origin also experiences additional and unique economic costs and financial strains that result directly from the teenager's childbearing.

How would such economic strain on the adolescent's family of origin affect younger sisters? There is an extensive literature that suggests that economic hardship is strongly linked with stressed intrafamilial relations (Conger et al., 1994; Elder et al., 1995), punitive and neglectful parenting (McLoyd, 1990; McLoyd et al., 1994), and adolescents' pessimistic future expectations (Galambos and Silbereisen, 1987; Lempers et al., 1989). For example, under conditions of acute economic decline, mothers

tend to become physically punitive toward their adolescent children and highly critical and pessimistic about their children's future job success (McLoyd et al., 1994), and as the economic decline intensifies, the coercive, hostile parent-child exchanges increase (Conger et al., 1994). When their family experiences a significant loss of income, adolescents also reduce their own expectations about future educational and career attainments (Conger et al., 1994), and they often experience a sudden onset of delinquent-like behaviors (Galambos and Silbereisen, 1987; Lempers et al., 1989; Sampson and Laub, 1994).

The model shown in Figure 8.2 incorporates these findings about the consequences of increased family economic hardship experienced by teenage childbearing families for adverse younger sister outcomes, such as reduced future expectations and increased problem behavior. In addition, the model maps an additional pathway by which economic adversity impinges on parents' socialization of their children (e.g., through punitive parenting and stressed parent-child relations). Thus, the financial strain and hardship within the family provoked by the older sister's early parenting is believed to affect directly the expectations and behavior of younger sisters, as well as indirectly, through parents' parenting practices and expectations for their children.

Other family-related effects that might result from the older sister's childbearing include accommodating the father of baby, who may move in with the teen's family after the birth of the baby (Furstenberg, 1980). This may be more likely to occur if the father of the baby is older and able to help financially support the teen, her baby, and her family (East and Felice, 1996). The father moving in with the teen's family, however, may cause moving other family members (e.g., younger siblings) to live with kin so that there is enough room to house the father. All such household changes—as well as having a young infant in the household—would likely serve to stress all family members and strain family relations.


Several studies have pointed to the high risk for early pregnancy among the younger sisters of pregnant and parenting teens by documenting younger sisters' relatively permissive sexual and childbearing attitudes, higher levels of problem and delinquent-like behavior (e.g., fighting, stealing, destroying property, and being picked up by the police), and higher rates of sexual activity (East, 1996a, 1996b; East et al., 1993).

These characteristics are revealing because they are considered precursors to adolescent pregnancy and childbearing (Elliot and Morse, 1989; Elster et al., 1990; Ensminger, 1987, 1990; Miller and Sneesby, 1988; Moore et al., 1995; Mott and Haurin, 1988; Plotnick and Butler, 1991; Zabin et al., 1984, 1993). Thus, once on a particular pathway, the younger sisters would become prone to engaging in other high-risk behaviors leading to teenage pregnancy, such as frequent and unprotected sexual intercourse (East, 1995). The studies on younger sisters' risk behavior are reviewed briefly in the following.

When compared to early adolescent girls with only nonchildbearing adolescent sisters, East et al. (1993) found that girls of the same age and socioeconomic status who had at least one childbearing adolescent sister had more permissive attitudes about premarital teenage sex and (among virgins only) had more positive intentions to have sexual intercourse in the near future. Results are shown in Figure 8.3. Additionally, girls with a childbearing teenage sister were almost four times more likely to have already had sex (26%) than were girls with only nonchildbearing teenage sisters (7%). This is perhaps more significant considering that girls in the study were, on average, only 13 years old.

Further study of the same sample showed that girls with a childbearing adolescent sister had significantly more permissive attitudes toward childbearing (i.e., they were more accepting of nonmarital teenage childbearing), perceived younger ages for typical life-course transitions (best age to first have sex, get married, have first child), and were more pessimistic about achieving school and career goals than were girls with only nonchildbearing adolescent sisters (see Figure 8.4) (East, 1996c). Moreover, the girls who had a childbearing teenage sister were more likely to have engaged in problem behaviors (with the exception of drug use), such as school truancy and smoking, and had more total problem behaviors (see Figure 8.5).

It should be noted that the differences that emerged between the two younger sister groups cannot be attributed to differences in younger sisters' age, family size, mothers' educational level, family income, or the family's current welfare status. The two sister groups were comparable with regard to all these factors, and subjects' race (which covaried with older sister's childbearing status) was statistically controlled in all analyses.

Additional findings that highlight potential prepregnancy behaviors of the younger sisters of teen mothers derive from a longitudinal study


Figure 8.3. Younger sisters' sexual attitudes, sexual intentions, and sexual status by older sisters' childbearing status. Source: East et al. (1993).

conducted by East that focused on families in which only one teenager in a family was either currently pregnant or had delivered her first child no more than six months previously and in which no other sibling had ever been pregnant (East, 1996a). Thus, the current pregnancy or childbearing was the first to occur within the family. The younger sisters in such families were compared to the same-age, same-race (67% Hispanic, 33% Black) younger sisters of never-pregnant adolescents from families in which no teenage pregnancy had yet occurred.

As with the previous study, the two sets of families were carefully matched along several socioeconomic and demographic characteristics, for example, family size, family structure (e.g., two-parent or motheronly households), and parental educational levels. Nonetheless, the families


Figure 8.4. Younger sisters' attitudes and expectations by older sisters' childbearing status. Source: East 1996c.

of pregnant and parenting teens had lower family incomes and were more likely to be receiving welfare at the time of the study than were the families of never-pregnant teens. Thus, these factors were statistically controlled in all analyses.

Scores for the younger sisters in each family category on several attitudinal and behavioral characteristics are shown in Table 8.1. When compared to girls with a never-pregnant older sister, girls with a pregnant older sister were significantly less optimistic about their future, were more accepting of teenage childbearing, perceived younger ages as appropriate for life-transition norms (i.e., ages for girls to first have sex, marry, and have children), and engaged in more problem behavior at school (e.g., school suspension and disruptive behavior in class) and more delinquent behavior. However, the increase in highly visible delinquent, acting-out behavior may be a short-lived phenomenon in reaction


Figure 8.5. Younger sisters' problem behavior by older sisters' childbearing status. Source: East (1996c).

to an older sister's pregnancy since the frequency of delinquent behavior by the younger sisters of parenting teenagers was comparable to that for the younger sisters of never-pregnant teenagers.

Nonetheless, girls who had a parenting older sister were almost five times as likely to have already had sex (29%) as girls with a neverpregnant older sister, of whom only 6% were nonvirgins. They were also more likely to have had sexual intercourse more frequently and to have engaged in significantly more intimate sexual behaviors (ranging from kissing, to light petting, to heavy petting, to sexual intercourse).

As suggested previously, the differences between the younger sisters of never-pregnant teens and the younger sisters of pregnant or parenting teens may reflect the societal and intrafamily risk factors that may have precipitated the older sister's pregnancy. Consequently, these results need to be cautiously and judiciously interpreted, with the understanding that the differences found in younger sisters' attitudes and behavior

(n = 83)
(n = 29)
(n = 51)
F(2,161) Possible
NOTE: The F-statistic tests for equality of means. If the F is statistically significant (asterisked), the means are not equal. Means with the same-letter superscript in the same row are significantly different (p < 05).
SOURCE: East (1996a) (adapted with permission from the Alan Guttmacher Institute).

*p<.05. **p<.01. ***p<.001.

4.43a,b 4.10a 4.09b 3.40* 1-5
3.44 3.54 3.35 <1


2.99 2.98 2.94 <1 1-5 High
of teenage
1.60b 1.88c 2.39b,c 12.26*** 1-5 High
Acceptance of
teenage sex
2.07b 2.50 2.70b 3.75* 1-5 High
of teenage
1.84a,b 2.16a 2.34b 3.98* 1-5 High
Acceptance of
2.07b 2.30c 2.72b,c 4.12* 1-5 High
norms (years)—
21.94a,b 20.92a 21.08b 3.07* Later
norms (years)—
22.53b 21.65 21.06b 2.77 Later
Status from
1.96b 2.36 2.50b 5.05** 1-5 High status
of early
4.35 4.10 4.04 1.71 1-5 More
of early
1.42b 1.49c 2.03b,c 6.41** 1-5 More definite
Self-esteem 3.80 3.67 3.92 <1 1-5 High
1.94a 2.49a,c 1.95c 6.65** 0-4 More
Drug use 1.55 1.68c 1.41c 2.81 0-4 More
Partying 1.26 1.36 1.34 1.51 0-4 More


1.27a 1.69a,c 1.39c 4.21* 0-4 More frequent
3.25b 4.03 4.88b 3.31* 1-11 More
sexual acts
0.06b 0.17 0.29b 6.49** 0-1 Percentage
Frequency of sexual
intercourse among
0.18b 0.52 0.78b 4.51* 0-4 More frequent
may reflect family and societal factors instead of, or perhaps in addition to, the consequences of having a pregnant or parenting sister.

The differences between the younger sisters of pregnant teenagers and the younger sisters of parenting teenagers, however, would not be biased by selection factors. Since the intergroup social and demographic characteristics were comparable or were statistically controlled, differences between these two groups are likely to indicate how younger sisters are affected specifically by the birth of an older sister's child. Compared to girls with a pregnant older sister, girls with a parenting older sister perceived significantly greater parental approval of teenage childbearing, were more accepting of nonmarital childbearing, and had more definite intentions to have a child themselves at an early age.

Cumulatively, the results of these studies suggest that the younger sisters of childbearing teens are at greater risk for early parenthood than the younger sisters of nonchildbearing teens. Girls with adolescent childbearing sisters are more accepting of nonmarital teenage parenting, perceive younger ages for typical life-course transitions, hold more pessimistic school and career expectations, and are more likely to engage in deviant behaviors—all of which are key risk factors for early sexual behavior and early parenthood (Moore et al., 1995). However, the current findings—based on cross-sectional data—are evidence only of associations between an older sister's childbearing and her younger sister's attitudes and behaviors, not of causal sequences of events. Further longitudinal examination of the changes in younger sisters' attitudes

and behaviors as a function of an older sister's childbearing would be necessary to show a causal relationship.


Because younger sisters of pregnant and parenting teenagers are a singularly high-risk group for early pregnancy, such adolescents are an important and strategic population to target for pregnancy prevention (East, 1996b, 1998a). The design of a prevention program for them must take into account factors related to the identification, accessibility, and participation of this population that, in turn, will likely impact the effectiveness of the intervention.

Identifying the Population

The first characteristic of younger sisters of pregnant and parenting teenagers that should be considered in designing a pregnancy prevention program is that such girls represent a relatively easily identifiable population. Virtually all protocols of teen obstetric clinics include obtaining a family history of every adolescent presenting with pregnancy. In this history, the number of younger sisters, their ages, and their living situations are often noted. In addition, the younger sisters may already be known to health care staff through the prenatal and pediatric visits of the older sister, as well as through well-child checkups of the younger sister herself. Thus, data identifying the younger sisters may be readily available to program personnel.

Involving the Family

Prevention efforts aimed at younger sisters may be optimized if family involvement is encouraged. As noted previously, adolescent childbearing has complex and multifaceted consequences for the adolescent's family of origin, affecting the family dynamics, the family's economic situation, the parents' parenting, and the younger siblings' attitudes, expectations, and behaviors. By intervening at the family level, many of the family's problems and situations can be impacted.

However, clinicians and educators designing pregnancy prevention programs must be cognizant of both the obstacles as well as the advantages of incorporating family participation in the prevention effort. For example, the active involvement of younger sisters may be facilitated by

their and their mothers' motivation to learn about how to prevent a second pregnancy within the family. In addition, the prevention message may be more lasting and long-lived if the family is included in service delivery, as programs are often short term and relatively temporary. Alternately, program designers must be alert to the often strong attitudinal and behavioral influences that older sisters have on their younger sisters. Younger sisters participating in prevention programs that attempt to change their attitudes and values about early sexual involvement (e.g., abstinence-promotion programs) may feel caught in a cross-current: The messages they receive from the program about teen sex and pregnancy may compete or conflict with the messages they continue to receive at home (East and Felice, 1992). In these cases, the extent and type of parental and family involvement in the prevention effort should be carefully considered before the intervention begins. Finally, in situations of suspected abuse or incest, family or parental involvement may not be in the best interest of the younger sister, and a different approach should be sought.

Offering Disincentives to Early Childbearing

Prevention efforts should take into consideration the social reality of girls who model or mimic their older sister's early childbearing. That is, early childbearing may well seem a logical life choice for young women who perceive few alternatives. Programs that do not attempt to alter this perception are likely to fail, as they will be counteracted by messages that the girls receive at home and from peers. Thus, pregnancy prevention efforts aimed at younger sisters require a broader policy agenda that goes well beyond addressing exclusively fertility-related behaviors (e.g., contraception or abstinence). As Lizbeth Schorr (1988) has appropriately argued, at heart what is needed is a comprehensive economic agenda for providing disincentives to early parenting, such as job skills training, reasonable-wage job opportunities, and employment security. If young women are expected to be motivated not to be teen mothers, they need to have a clear vision of what they can do.

Promoting Alternative Role Models

Many younger sisters lack successful role models to help overcome the overwhelming odds for the inter- and intragenerational perpetuation of teenage pregnancy that has preceded them (Cox et al., 1995). The need for alternative role models is especially crucial for the younger sisters

of parenting teens who remain at home with their families of origin. In these cases, the younger sister witnesses firsthand the way in which motherhood establishes a role for the older adolescent—a role associated with status, femininity, and attention—and the choice to emulate her may be particularly seductive. Thus, policy and program development strategies that incorporate caring adult role models and mentors who will inform girls that becoming a mother is not their only avenue to fulfillment and recognition seem uniquely promising for this target group.

Encouraging Group Discussion

A group discussion approach uniting younger sisters may hold many benefits. For example, participants could be asked to discuss the stresses and disruption caused by premarital parenting in general and by their sisters' parenting in particular. They could be encouraged to share their stories of how they and their families have been affected by their sisters' pregnancies and births and how they and their families are coping. As they tell their life stories in a nonthreatening, supportive environment, the younger sisters may develop a broader perspective on their own problems, and the sharing of experiences can provide a source of strength and a positive basis for pregnancy prevention.

Providing Intensive Individualized Attention

A final consideration in program design is to offer younger sisters intensive individualized attention as soon as the older sister learns that she is pregnant and begins seeking prenatal care. Service delivery could be arranged in such a way that the pregnant teen and her sister both receive care at the same site at the same time but from separate care providers, a model based on Teen-Tot Clinics or Two Generation Services (Dryfoos, 1990; Hardy and Zabin, 1991). Certainly, prevention programs that “pay attention” to younger sisters before, rather than after, they become pregnant seem particularly compelling.

Service programs for pregnant teens offer a wide variety of support services, such as individualized counseling, nutritional services, job preparation and placement, case management, special school guidance or instruction, free bus tokens, free food coupons, and sometimes monetary reimbursement or free transportation to attend prenatal clinic visits (Klerman and Horwitz, 1992). As such, it appears that a number

of policies have been established that inadvertently reward young people for early childbearing. Providing personalized, concentrated care for younger sisters could serve to counterbalance the positive inducements that they witness their older sisters receiving as a result of their pregnancies. It will be interesting to see what impact the Federal Welfare Reform Bill—also known as the Personal Responsibility and Work Opportunity Reconciliation Act—will have on the younger siblings within teenage childbearing families. As planned, this act will impose new work requirements on individuals, cut food stamp benefits, and limit the length of time on aid to a maximum five-year lifetime limit. By removing such essential threads in teen parents' safety net, this bill will undoubtedly make it more difficult to support a child while still in school. It may also diminish the attractiveness and apparent ease of having children while one is young.


What are the policy implications of the research presented in this chapter? First and foremost is the need to target the siblings of pregnant and parenting teens for pregnancy prevention services. This has actually started to occur in the form of several statewide policies. For example, in 1996, the State of California—through the Department of Health Services, Maternal and Child Health Branch—legislated $3 million per year to provide Adolescent Family Life case management services specifically to the siblings of pregnant and parenting teens (California Senate Office of Research, 1997). In addition, the California Department of Education implemented a “Teenage Pregnancy Prevention Grant Program,” with $10 million per year given to school programs that, among other things, “target youth who have a sibling who is a teenage parent.” As a result of these initiatives, numerous programs have sprung up across the state that systematically focus prevention efforts on the siblings of pregnant and parenting teens. Although the results of these programs are not yet known, both programs are being intensely evaluated, and their results will help shed light on whether these special prevention services can reduce the teenage pregnancy rate within this population. It is also likely that these programs will serve as model programs for other states that are launching large, statewide sibling pregnancy prevention efforts.

Another trend in pregnancy prevention programs and policies is

the recent emphasis on comprehensive, community-wide collaboratives aimed at reducing the teenage pregnancy rates in particularly high-risk areas and communities (Brindis, 1993). These comprehensive collaboratives are the result of several reports stating that many single-prong approaches to preventing teen pregnancy have been demonstrably unable to show sufficient scale, effectiveness, or sustainability (e.g., Kirby et al., 1997; Miller et al., 1992). For example, The California Wellness Foundation recently started “The Teenage Pregnancy Prevention Initiative,” which funds broad-based, multisector community coalitions that saturate targeted areas with coordinated comprehensive services, ranging from school mentoring and life skills counseling to sexuality education and access to family planning services. These kinds of comprehensive initiatives that help create and then sustain caring communities that can effectively nurture their youth show great promise for significantly lowering California's teenage pregnancy rate—which is the highest in the nation (Henshaw, 1997)—and should be an integral component of any national pregnancy prevention policy.


The younger sisters of pregnant and parenting teens have disproportionately high rates of early childbearing and engage in teenage sexual activity at earlier ages and at higher rates than other girls of their same race, ethnicity, and social class. This chapter has highlighted several reasons why sister pairs within a family share a vulnerability to teenage childbearing, and a model was presented of how a teenager's childbearing further increases her younger sister's risk of early pregnancy. This model can be used to inform community-, school-, and hospital-based pregnancy prevention efforts in addressing the multiple factors that contribute specifically to younger sisters' vulnerability to early pregnancy. The model can also be used to better understand the mechanisms by which the younger sisters become vulnerable to such outcomes so that program developers can design and tailor more effective interventions. As researchers continue to illuminate the factors that predispose these younger sisters to early childbearing, service providers will be better able to construct effective programs to reduce within-family teenage pregnancies. A “younger sister” approach to breaking the cycle of teenage childbearing offers tremendous—and as yet untapped—potential for adolescent pregnancy prevention.



This research was supported by grant R29-HD29472 from the National Institute of Child Health and Human Development, grant APR-000970 from the Office of Population Affairs, and a 1996 Distinguished Wellness Lecturer Award from The California Wellness Foundation and the University of California Office of Health Affairs. The author is very grateful to Erica Johnson, Lisette Lahana, Marijo Villena, and Daisy Barguiarena for their dedication and determination in recruiting and interviewing the study families whose data are presented in this chapter.


Bandura, A. L. (1977). Social learning theory.Englewood Cliffs, N.J.: Prentice Hall.

Baumrind, D. (1983). “Three commentaries on teenage sexuality.” American Psychologist, 36, 528–531.

Brewster, K., Billy, J. O. G., and Grady, W. R. (1993). “Social context and adolescent behavior: The impact of community on the transition to sexual activity.” Social Forces, 71, 713–740.

Brindis, C. (1992). “Adolescent pregnancy prevention for Hispanic youth: The role of schools, families, and communities.” Journal of School Health, 62, 345–351.

Brindis, C. (1993). “Antecedents and consequences: The need for diverse strategies in adolescent pregnancy prevention.” In A. Lawson and D. L. Rhode, eds., The politics of pregnancy: Adolescent sexuality and public policy.Pp. 257–283. New Haven, Conn.: Yale University Press.

Brooks-Gunn, J., and Chase-Lansdale, L. (1991). “Children having children: Effects on the family system.” Pediatric Annals, 20, 467–481.

Bryant, B. K. (1982). “Sibling relationships in middle childhood.” In M. Lamb and B. Sutton-Smith, eds., Sibling relationships: Their nature and significance.Pp. 87–121. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Burton, L. M. (1990). “Teenage pregnancy as an alternative life-course strategy in multigenerational Black families.” Human Nature, 1, 123–143.

Burton, L. M. (1995). “Intergenerational patterns of providing care in African American families with teenage childbearers: Emergent patterns in an ethnographic study.” In V. L. Bengtson, K. W. Schaie, and L. M. Burton, eds., Adult intergen erational relations: Effects of societal change.Pp. 79–96. New York: Springer.

Burton, L. M. (1996a). “Age norms, the timing of family role transitions, and intergenerational caregiving among aging African American women.” The Gerontologist, 36, 199–208.

Burton, L. M. (1996b). “The timing of childbearing, family structure, and the role responsibilities of aging black women.” In E. M. Hetherington and E. A. Blechman, eds., Stress, coping and resiliency in children and families.Pp. 155–172. Mahwah, N.J.: Lawrence Erlbaum Associates.


Burton, L. M., and Bengtson, V. L. (1985). “Black grandmothers: Issues of timing and continuity of roles.” In V. Bengtson and J. Robertson, eds., Grandparenthood: Research and policy perspectives.Pp. 61–78. Beverly Hills, Calif.: Sage.

California Senate Office of Research. (April, 1997). Issue brief: California strategies to address teenage pregnancy.Sacramento, Calif.: Senate Printing Office.

Cicirelli, V. G. (1982). “Sibling influence throughout the lifespan.” In M. Lamb and B. Sutton-Smith, eds., Sibling relationships: Their nature and significance.Pp. 267–284. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Conger, R. D., Ge, X., Elder, G. H., et al. (1994). “Economic stress, coercive family process, and developmental problems of adolescents.” Child Development, 65, 541–561.

Cox, J., DuRant, R. H., Emans, S. J., and Woods, E. R. (1995). “Early parenthood for the sisters of adolescent mothers: A proposed conceptual model of decision making.” Adolescent and Pediatric Gynecology, 8, 188–194.

Cox, J., Emans, S. J., and Bithoney, W. (1993). “Sisters of teen mothers: Increased risk for adolescent parenthood.” Adolescent and Pediatric Gynecology, 6, 138–142.

Crane, J. (1991). “The epidemic theory of ghettos and neighborhood effects on dropping out and teenage childbearing.” American Journal of Sociology, 96, 1226–1259.

Dash, L. (1989). When children want children: An inside look at the crisis of teenage parenthood.New York: Penguin.

Draper, P. (1976). “Social and economic constraints on child life among the !Kung.” In R. Lee and I. DeVore, eds., Kalahari hunter-gatherers: Studies of the !Kung San and their neighbors.Pp. 200–217. Cambridge, Mass.: Harvard University Press.

Dryfoos, J. G. (1990). Adolescents at risk: Prevalence and prevention.New York: Oxford University Press.

Duany, L., and Pittman, K. (1990). Latino youths at a crossroad: Report of the Adolescent Pregnancy Prevention Clearinghouse.Washington, D.C.: Children's Defense Fund.

East, P. L. (1995). “The social contagion model and adolescent sexual behavior.” Directions in Clinical Psychology, 10, 1–10.

East, P. L. (1996a). “Do adolescent pregnancy and childbearing affect younger siblings?” Family Planning Perspectives, 28, 148–153.

East, P. L. (1996b). “Pregnancy prevention opportunities focusing on younger sisters of childbearing teens.” In 1996 Wellness Lectures.Pp. 101–127. Berkeley: University of California Regents, University of California Printing.

East, P. L. (1996c). “The younger sisters of childbearing adolescents: Their attitudes, expectations, and behaviors.” Child Development, 67, 267–282.

East, P. L. (1998a). “Breaking the cycle of teenage pregnancy: Prevention opportunities focusing on the younger sisters of teen mothers.” Education and Urban Society, 30, 157–171.

East, P. L. (1998b). “Impact of adolescent childbearing on families and younger siblings: Effects that increase younger siblings' risk for early pregnancy.” Applied Developmental Science, 2, 62–74.


East, P. L. (1999). “The first teenage pregnancy in the family: Does it affect mothers' parenting, attitudes, or mother-adolescent communication?” Journal of Marriage and the Family, 61, 306–319.

East, P. L., and Felice, M. E. (1992). “Pregnancy risk among the younger sisters of pregnant and childbearing adolescents.” Journal of Developmental and Behavioral Pediatrics, 13, 128–136.

East, P. L., and Felice, M. E. (1994). “The psychosocial consequences of teenage pregnancy and childbearing.” In I. R. Shenker, ed., Adolescent Medicine.Pp. 73–92. London: Harwood Academic Publishers.

East, P. L., and Felice, M. E. (1996). Adolescent pregnancy and parenting: Findings from a racially diverse sample.Mahwah, N.J.: Lawrence Erlbaum Associates.

East, P. L., Felice, M. E., and Morgan, M. C. (1993). “Sisters' and girlfriends' sexual and childbearing behavior: Effects on early adolescent girls' sexual outcomes.” Journal of Marriage and the Family, 55, 953–963.

East, P. L., and Jacobson, L. J. (2000). “The younger siblings of teenage mothers: A follow-up of their pregnancy risk at middle adolescence.” Manuscript submitted for publication.

East, P. L., and Shi, C. R. (1997). “Pregnant and parenting adolescents and their younger sisters: The influence of relationship qualities for younger sister outcomes.” Journal of Developmental and Behavioral Pediatrics, 18, 19–25.

Elder, G. H. (1975). “Age differentiation and the life course.” In A. Inkeles, J. Coleman, and N. Smelser, eds., Annual Review of Sociology.Pp. 165–190. Palo Alto, Calif.: Annual Review Company.

Elder, G. H., Eccles, J. S., Ardelt, M., and Lord, S. (1995). “Inner-city parents under economic pressure: Perspectives on the strategies of parenting.” Journal of Marriage and the Family, 57, 771–784.

Elliot, D. S., and Morse, B. J. (1989). “Delinquency and drug use as risk factors in teenage sexual activity.” Youth and Society, 21, 32–60.

Elster, A. B., Ketterlinus, R., and Lamb, M. E. (1990). “Association between parenthood and problem behavior in a national sample of adolescents.” Pediatrics, 85, 1044–1050.

Ensminger, M. E. (1987). “Adolescent sexual behavior as it relates to other transition behaviors in youth.” In S. L. Hofferth and C. D. Hayes, eds., Risking the future: Adolescent sexuality, pregnancy, and childbearing.Pp. 36–55. Washington, D.C.: National Academy Press.

Ensminger, M. E. (1990). “Sexual activity and problem behaviors among black, urban adolescents.” Child Development, 61, 2032–2046.

Fox, G. L. (1980). “The mother-adolescent daughter relationship as a sexual socialization structure: A research review.” Family Relations, 29, 21–28.

Fox, G. L., and Inazu, J. K. (1980). “Patterns and outcomes of mother-daughter communication about sexuality.” Journal of Social Issues, 36, 7–29.

Friede, A., Hogue, C., Doyle, L., et al. (1986). “Do the sisters of childbearing teenagers have increased rates of childbearing?” American Journal of Public Health, 76, 1221–1224.

Furstenberg, F. F. (1980). “Burdens and benefits: The impact of early childbearing on the family.” Journal of Social Issues, 36, 64–87.

Galambos, N. L., and Silbereisen, R. K. (1987). “Influences of income change and

parental acceptance of adolescent transgression proneness and peer relations.” European Journal of Psychology of Education, 1, 17–28.

Garn, S. M. (1980). “Continuities and change in maturational timing.” In O. G. Brim and J. Kagan, eds., Constancy and change in human development.Pp. 113–162. Cambridge, Mass.: Harvard University Press.

Geronimus, A. T., and Korenman, S. (1992). “The socioeconomic consequences of teen childbearing reconsidered.” Quarterly Journal of Economics, 107, 1187–1214.

Geronimus, A. T., and Korenman, S. (1993). “The socioeconomic consequences of teenage childbearing: Evidence and interpretation.” Demography, 38, 281–290.

Goldfarb, J., Mumford, D., Schum, D., et al. (1977). “An attempt to detect “pregnancy susceptibility” in indigent adolescent girls.” Journal of Youth and Adolescence, 6, 127–144.

Hamburg, B. (1974). “Early adolescence: A specific and stressful stage of the life cycle.” In G. V. Coelho, D. A. Hamburg, and J. E. Adams, eds., Coping and adaptation.Pp. 101–124. New York: Basic Books.

Hamburg, B. (1986). “Subsets of adolescent mothers: Developmental, biomedical and psychosocial issues.” In J. Lancaster and B. Hamburg, eds., School-age pregnancy and parenthood.Pp. 115–145. New York: Aldine de Gruyter.

Haraven, T. K. (1978). Transitions: The family and the life course in historical perspective.New York: Academic Press.

Hardy, J. B., and Zabin, L. S. (1991). Adolescent pregnancy in an urban environment: Issues, programs, and evaluation.Baltimore: Urban and Schwarzenberg Press.

Hayes, C., ed. 1987. Risking the future: Adolescent sexuality, pregnancy, and childbearing.Washington, D.C.: National Academy Press.

Hayward, M. D., Grady, W. R., and Billy, J. O. G. (1992). “The influence of socioeconomic status on adolescent pregnancy.” Social Science Quarterly, 73, 750–772.

Henshaw, S. K. (1997). “Teenage abortion and pregnancy statistics by state, 1992.” Family Planning Perspectives, 29, 115–122.

Hoffman, S. D., Foster, E. M., and Furstenberg, F. F. (1993a). “Re-evaluating the costs of teenage childbearing.” Demography, 30, 1–13.

Hoffman, S. D., Foster, E. M., and Furstenberg, F. F. (1993b). “Re-evaluating the costs of teenage childbearing: Response to Geronimus and Korenman.” Demography, 30, 291–296.

Hogan, D. P., Astone, N. M., and Kitagawa, E. M. (1985). “Social and environmental factors influencing contraceptive use among black adolescents.” Family Planning Perspectives, 17, 165–169.

Hogan, D. P., and Kitagawa, E. M. (1985). “The impact of social status, family structure, and neighborhood on the fertility of black adolescents.” American Journal of Sociology, 90, 825–855.

Jemmott, L. S., and Jemmott, J. (1992). “Family structure, parental strictness, and sexual behavior among inner-city black male adolescents.” Journal of Adolescent Research, 7, 192–207.

Kahn, J. R., and Anderson, K. A. (1992). “Intergenerational patterns of teenage fertility.” Demography, 29, 39–57.


Kahn, J. R., Rindfuss, K. R., and Guilkey, D. K. (1990). “Adolescent contraceptive choice.” Demography, 27, 323–335.

Kinnaird, K. L., and Gerrard, M. (1986). “Premarital sexual behavior and attitudes toward marriage and divorce among young women as a function of their mothers' marital status.” Journal of Marriage and the Family, 48, 757–765.

Kirby, D., Korpi, M., Barth, R. P., and Cagampang, H. H. (1997). “The impact of the postponing sexual involvement curriculum among youths in California.” Family Planning Perspectives, 29, 100–108.

Klerman, L. V., and Horwitz, S. M. (1992). “Reducing the adverse consequences of adolescent pregnancy and parenting: The role of service programs.” Adolescent Medicine: State of the Art Reviews, 3, 299–316.

Ladner, J. (1988). “The impact of teenage pregnancy on the black family.” In H. McAdoo, ed., Black families.Pp. 296–305. Newbury Park, Calif.: Sage.

Lempers, J., Clark-Lempers, D., and Simons, R. (1989). “Economic hardship, parenting, and distress in adolescence.” Child Development, 60, 25–49.

Luker, K. (1991). “Dubious conceptions: The controversy over teen pregnancy.” The American Prospect, 5, 73–83.

Luker, K. (1996). Dubious conceptions: The politics of teenage pregnancy.Cambridge, Mass.: Harvard University Press.

Marini, M. M. (1984). “Age and sequencing norms in the transition to adulthood.” Social Forces, 63, 229–243.

McLoyd, V. C. (1990). “The impact of economic hardship on black families and children: Psychological distress, parenting, and socioemotional development.” Child Development, 61, 311–346.

McLoyd, V. C., Jayaratne, T. E., Ceballo, R., and Borquez, J. (1994). “Unemployment and work interruption among African-American single mothers: Effects on parenting and adolescent socioemotional functioning.” Child Development, 65, 562–589.

Melville, M., ed. (1980). Twice a minority: Mexican American women.St. Louis, Mo.: Mosby.

Miller, B. C. (1998). Families matter: A research synthesis of family influences on adolescent pregnancy.Washington, D.C.: National Campaign to Prevent Teen Pregnancy.

Miller, B. C., and Bingham, C. R. (1989). “Family configuration in relation to the sexual behavior of female adolescents.” Journal of Marriage and the Family, 51, 499–506.

Miller, B. C., Card, J. J., Paikoff, R. L., and Peterson, J. L. (1992). Preventing adolescent pregnancy: Model programs and evaluations.Newbury Park, Calif.: Sage.

Miller, B. C., McCoy, J. K., Olson, T. D., and Wallace, C. M. (1986). “Parental discipline and control attempts in relation to adolescent sexual attitudes and behavior.” Journal of Marriage and the Family, 48, 503–512.

Miller, B. C., Norton, M. C., Curtis, T., et al. (1997). “The timing of sexual intercourse among adolescents: Family, peer, and other antecedents.” Youth and Society, 29, 54–83.

Miller, B. C., and Sneesby, K. R. (1988). “Educational correlates of adolescents'

sexual attitudes and behavior.” Journal of Youth and Adolescence, 17, 521–530.

Mirande, A. M. (1968). “Reference group theory and adolescent sexual behavior.” Journal of Marriage and the Family, 30, 572–577.

Moore, K. A., Miller, B. C., Glei, D., and Morrison, D. R. (1995). Adolescent sex, contraception, and childbearing: A review of recent research.Washington, D.C.: Child Trends, Inc.

Moore, K. A., Peterson, J. L., and Furstenberg, F. F. (1986a). “Parental attitudes and the occurrence of early sexual activity.” Journal of Marriage and the Family, 48, 777–782.

Moore, K. A., Romano, A., and Oaks, C. (1997). Facts at a glance.Washington, D.C.: Child Trends, Inc.

Moore, K. A., Simms, M. C., and Betsey, C. L. (1986b). Choice and circumstance: Racial differences in adolescent sexuality and fertility.New Brunswick, N.J.: Transaction Books.

Mott, F., and Haurin, J. R. (1988). “Linkages between sexual activity and alcohol and drug use among American adolescents.” Family Planning Perspectives, 20, 128–136.

Neugarten, C. N., Moore, J. W., and Lowe, J. C. (1965). “Age norms, age constraints, and adult socialization.” American Journal of Sociology, 70, 710–717.

Newcomer, S. F., and Udry, J. R. (1984). “Mothers' influence on the sexual behavior of their teenage children.” Journal of Marriage and the Family, 46, 477–485.

Newcomer, S. F., and Udry, J. R. (1985). “Parent-child communication and adolescent sexual behavior.” Family Planning Perspectives, 17, 169–174.

Newcomer, S. F., and Udry, J. R. (1987). “Parental marital status effects on adolescent sexual behavior.” Journal of Marriage and the Family, 49, 235–240.

Plotnick, R. D. (1992). “The effects of attitudes on teenage premarital pregnancy and its resolution.” American Sociological Review, 57, 800–811.

Plotnick, R. D., and Butler, S. S. (1991). “Attitudes and adolescent nonmarital childbearing: Evidence from the National Longitudinal Survey of Youth.” Journal of Adolescent Research, 6, 470–492.

Presser, H. (1978). “Age at menarche, socio-sexual behavior, and fertility.” Social Biology, 25, 94–101.

Rodgers, J. L., Rowe, D. C., and Harris, D. F. (1992). “Sibling differences in adolescent sexual behavior: Inferring process models from family composition patterns.” Journal of Marriage and the Family, 54, 142–152.

Rowe, D. C., and Gulley, B. L. (1992). “Sibling effects on substance use and delinquency.” Criminology, 30, 217–233.

Rowe, D. C., Rodgers, J. L., Meseck-Bushey, S., and St. John, C. (1989). “Sexual behavior and nonsexual deviance: A sibling study of their relationship.” Developmental Psychology, 12, 418–427.

Sampson, R. J., and Laub, J. H. (1994). “Urban poverty and the family context of delinquency: A new look at structure and process in a classic study.” Child Development, 65, 523–540.


Schorr, L. B. (1988). Within our reach: Breaking the cycle of disadvantage.New York: Doubleday.

Stack, C. (1974). All our kin: Strategies for survival in a black community.New York: Harper and Row.

Stack, C. (1975). “Who raises black children? Transactions of child givers and child receivers.” In T. R. Williams, ed., Socialization and communication in primary groups.Pp. 183–205. New York: Hague-Mouton.

Stack, C. B., and Burton, L. M. (1993). “Kinscripts.” Journal of Comparative Family Studies, 24, 157–170.

Thornton, A., and Camburn, D. (1987). “The influence of the family on premarital sexual attitudes and behavior.” Demography, 24, 323–340.

Trent, K., and Harlan, C. (1994). “Teenage mothers in nuclear and extended households.” Journal of Family Issues, 15, 309–337.

Udry, J. R. (1979). “Age at menarche, at first intercourse and at first pregnancy.” Journal of Biosocial Science, 11, 433–441.

Udry, J. R., and Cliquet, R. L. (1982). “A cross-cultural examination of the relationship between ages at menarche, marriage and first birth.” Demography, 19, 53–63.

Weisner, T. S. (1982). “Sibling interdependence and child caretaking: A crosscultural view.” In M. Lamb and B. Sutton-Smith, eds., Sibling relationships: Their nature and significance.Pp. 305–327. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Weisner, T. S. (1986). “Socialization for parenthood in sibling caretaking societies.” In J. B. Lancaster, J. Altman, A. Rossi, and L. Sherrod, eds., Parenting across the life span.Pp. 237–270. New York: Aldine de Gruyter.

Whiting, B. B. (1983). “The genesis of prosocial behavior.” In D. Bridgemen, ed., The nature of prosocial development.Pp. 221–242. New York: Academic Press.

Whiting, B. B., and Whiting, J. W. (1975). Children in six cultures: A psychocultural analysis.Cambridge, Mass.: Harvard University Press.

Williams, C. W. (1991). Black teenage mothers: Pregnancy and child rearing from their perspective.Lexington, Mass.: Lexington Books.

Zabin, L. S., Astone, N. M., and Emerson, M. R. (1993). “Do adolescents want babies? The relationship between attitudes and behavior.” Journal of Research on Adolescence, 3, 67–86.

Zabin, L. S., Hirsch, M., Smith, E., and Hardy, J. B. (1984). “Adolescent sexual attitudes and behavior: Are they consistent?” Family Planning Perspectives, 16, 181–185.



Exploring a Paradox

Sylvia Guendelman

Poverty is a well-known determinant of health. Persons of low socioeconomic status have much higher levels of morbidity and mortality than do those of higher status.1–3 Yet recent research revealing positive pregnancy outcomes within poor immigrant groups raises the question of whether poverty is necessarily linked to adverse pregnancy outcomes. Are there protective factors that can buffer against the noxious effects of poverty during pregnancy? If so, what lessons can we learn from immigrant and refugee women about promoting healthy pregnancy outcomes?

To address these questions, this chapter attempts to identify protective factors that may be associated with favorable pregnancy outcomes among Latina and Southeast Asian women. It reports the work by Guendelman and colleagues on Latina women of reproductive age and extends previous analyses to compare Latinas and Southeast Asians. The geographic focus is predominantly on California since it is the most important immigrant-receiving state, absorbing approximately 40% of Latino and Asian newcomers. Favorable pregnancy outcomes among Southeast Asian women and Mexican immigrants have been reported in other states, suggesting that what we learn about these populations residing in California may be applicable elsewhere.



California leads the nation in rapid diversification, moving away from a White “majority” toward a predominantly Asian and Latino population. Immigration and high fertility have fueled the growth of these populations, which are increasing at a rate ten times faster than that of Whites (i.e., White non-Latinos). Newcomers from Mexico and Southeast Asia constitute more than one-third of California's immigration inflow. Immigrants to California from Vietnam, Cambodia, and Laos numbered close to 30,000 in 1993, with the majority coming from Vietnam.4 Besides constituting the largest refugee admissions, Southeast Asians have exhibited the highest fertility rates among ethnic groups in the United States.5 Mexico provides the largest number of Latino immigrants to California: more than 52,000 legal immigrants 4 and as many as 100,000 undocumented immigrants annually.6 Mothers of Mexican descent—who represent the great majority of California's Latinas—also exhibit fertility rates that far surpass those of White women. Although Mexican Americans and Southeast Asians constitute 22.5% and Whites constitute 57% of the state's total population, in 1993 alone, Mexican-American and Southeast Asian women collectively gave birth to 42% (n 243,833, out of which 228,707 were to Mexican Americans) of California's infants, compared to 37% (215,885) of infants borne by White women.7

Aside from their large immigration numbers and high fertility rates, it would seem that Mexican and Southeast Asian women have little in common. Indeed, these ethnic groups have very different histories in California. Mexicans have a long history in the state, while Southeast Asians are recent immigrants. Hence, whereas a large proportion of Mexican origin women of reproductive age are U.S. born, Southeast Asian women are predominantly foreign born. Whereas Mexican women most often come by choice, seeking economic opportunity for themselves or their families, Southeast Asian women have been forced to flee their war-torn native lands. Yet because Southeast Asians are admitted as refugees, they can receive resettlement funds and other assistance such as language instruction and job training. Mexican immigrants, legal or not, are ineligible for such benefits since they do not qualify for “political refugee” status. Such circumstances, which are often exacerbated by anti-immigrant attitudes, create their own set of resettlement stresses for Mexican newcomers. In addition to the differences across these populations, there are

  Mexican-American Cambodian Hmong Laotian Vietnamese Non-Latina
SOURCE: Census data for California, 1990.
Total California
6,118,996 68,190 46,892 58,058 280,223 17,029,126
Total female
2,897,838 35,203 23,041 28,666 132,467 7,800,106
Less than
fifth grade (%)
17.7 47.7 58.9 44.8 12.2 0.9
High school or
higher (%)
48.9 29.7 18.3 31.8 63.9 91.6
median ($)
27,934 17,400 15,978 16,436 32,199 39,564
Median for female
> 15 years ($)
8,991 7,729 7,531 8,059 9,520 14,221
Families below
poverty level (%)
19.3 46.9 59.9 50.4 25.3 5.3
Labor Force
> 16 years
in labor force
55.8 23.3 12.5 26.8 50.9 58.1

important differences in language, culture, and social and political standing within the Mexican and Southeast Asian populations.

Despite these notable differences, similarities do exist at the population level in the general socioeconomic profiles of Mexican Americans and Southeast Asians in comparison to the majority White population of California. Both populations are characterized by low educational attainment and high incidence of poverty, especially among women. With the exception of the Vietnamese, Southeast Asian women have strikingly few years of formal schooling and high rates of unemployment and welfare dependency. They commonly lack English proficiency and transferable job skills.5 Mexican-American women, while more active in the labor force and less likely than Southeast Asian women to have families living below the poverty level, nevertheless have very low incomes in comparison with White women (Table 9.1).

From a health perspective, Mexican-American and Southeast Asian women also share a high-risk profile. Both groups of women experience delayed entry into prenatal care, have large families with short birthspacing intervals, and—with the exception of the Vietnamese—have high rates of teen pregnancy compared to White women (Table 9.2).


These socioeconomic and health risk factors have traditionally predicted adverse pregnancy outcomes in other populations. For instance, African Americans have risk profiles that are similar to immigrants, yet on average their pregnancy outcomes are much worse.3, 8–10 Studies of White women who share similar risk factors also show increased rates of low birthweight.3,11 Surprisingly, Mexican Americans and Southeast Asians enjoy pregnancy outcomes that are comparable to those of the overall White population despite the dramatic differences in risk profiles. In this chapter, we examine this paradox in Mexican Americans and Southeast Asians, using White women as a reference group. Space constraints preclude a consistent analysis of differences among the various Southeast Asian subgroups, but we highlight the differences between foreign-born and U.S.-born Mexican Americans.

California birth cohort files for 1990–92 indicate that infant mortality (birth to 364 days) and postneonatal mortality (28–364 days) rates among all Mexican Americans regardless of nativity states and Southeast Asians are comparable to those of Whites (Table 9.3). The infant mortality

  Mexican-American All
Cambodian Laotian Vietnamese Non-Latina
NOTE: Estimates not available separately for Hmong in this data set.
SOURCE: Birth Cohort Files for california 1990-92.

*Significance compaired with non-latina Whites: p <.001.

aCensus data for California, 1990.

Late entry into
prenatal care
(third trimester
or not at all) (%)
10.9[*] 6.6[*] 7.6[*] 9.6[*] 5.3[*] 3.7
Children born to
women 35-44
(per 1,000
2.9 N.A. 3.6 3.7 2.5 1.5
< 18 years (%)
6.5[*] 3.6[*] 4.4[*] 8.9[*] 1.7[*] 2.5
rates are comparable even with the inclusion of infants weighing less than 500 grams at birth, who are increasingly being saved with improved technology in neonatal intensive care units.

Mexican Americans do have higher neonatal (0–27 days) mortality rates (4.1 per 1,000 live births) than Whites (3.7) or Southeast Asians (3.5); however, when we restrict the comparison to births of Mexicanborn women and exclude U.S.-born Mexican Americans, the neonatal mortality rates (3.9) are similar to those of Whites (Table 9.4). Since foreign-born immigrants are even poorer, less educated, and face more difficulties in access to care than native-born Mexican Americans, their more favorable pregnancy outcomes are especially puzzling.

Birthweight data provide another strong indicator of perinatal health, as infants who weigh 2,500 grams or less at birth have higher-than-average rates of morbidity, neurological impairments, and mortality during the early years of life. Among Mexican Americans in California, the rate of low birthweight is equal to that of Whites (5.1%) despite differences in socioeconomic status (Table 9.3). However, Mexican-born women have significantly lower rates than Whites (Table 9.4). The low birthweight rates among Southeast Asian women appear to be significantly higher than Whites (Table 9.3), possibly because of genetic or biological differences.10 On average, Asian infants weigh one-half pound

Mexican-Born All
Cambodian Laotian Vietnamese Non-Latina
NOTE: Estimates not available separately for Hmong in this data set.
SOURCE: Birth Cohort Files for California, 1990-92.

*Significance compared with non-Latina Whites: p < .001.

Infant mortality per 1,000 live births
Including <500 g 6.6 6.1 6.2 7.4 7.3 5.4 6.4
Excluding <500 g 6.1 5.8 5.7 7.0 6.5 4.9 5.9
Neonatal mortality
per 1,000 live births,
including <500 g
4.1[*] 3.9 3.5 4.4 3.9 3.1 3.7
mortality per 1,000
live births
2.5 2.2 2.7 3.0 3.3 2.3 2.7
Low birthweight (%) 5.1 4.8[*] 6.4[*] 7.6[*] 7.7[*] 5.5[*] 5.1

Mexican-Born Non-Latina
SOURCE: Birth cohort files for California, 1990-92.

aHealth and Nutrition Examination Surveys, 1976-80, 1982-84, NCHS.

*Significance compared with non-Latina Whites: p < .001.

Mean years
of education[a]
11.3 7.8 12.8
Living in
poverty (%)[a]
28.2 38.3 11.3
Late entry
into PNC (%)
7.3[*] 12.3[*] 3.7
< 18 years (%)
11.6[*] 4.6[*] 2.5
Infant mortality per 1,000 live births
Including <500 g 7.8[*] 6.2 6.4
Excluding <500 g 7.1[*] 5.8 5.9
mortality per
1,000 live
4.5[*] 3.9 3.7
mortality per
1,000 live
3.3[*] 2.2 2.7
Low birth
weight (%)
5.9[*] 4.8[*] 5.1
less than White infants.12 Differences in birthweight distribution among Southeast Asians, however, do not appear to affect rates of infant mortality adversely.


Research indicates that there are no straightforward explanations for the epidemiological paradox of positive pregnancy outcomes in immigrant mothers born in Mexico and Southeast Asia. Several hypotheses have surfaced that point to deficits in these populations, such as an underreporting of infant deaths, ethnic misclassification in birth and/or death certificates,

and the possibility that excess fetal deaths might eliminate weaker fetuses before birth. Other hypotheses focus on the positive or “protective” factors that may contribute to healthy outcomes. For instance, selective migration may favor healthy mothers and healthy babies, and immigrant mothers who relocate in California may bring with them certain attitudes, values, and behaviors that protect them against stresses and other adverse conditions associated with poverty and resettlement in a new society. This chapter will examine each of these issues but will emphasize a search for clues in identifying protective factors for positive pregnancy outcomes.


It has been suggested that infant mortality rates among immigrant groups may be artificially low because of underreporting of infant deaths. However, low rates of out-of-hospital births in California, coupled with the fact that the great majority of neonatal deaths occur before the first hospital discharge, make it appear that underreporting of neonatal deaths is not a significant phenomenon for Mexican Americans or Southeast Asians in this state.13–15 Underreporting, if it does occur, is more likely to occur in the postneonatal period (28–364 days), when the child is living at home.15

Infant death may go unreported when a migrant family or mother returns to Mexico following birth.16–18 Crossing the border to give birth is not an uncommon event; in fact, one study found that 10.4% of the women living in the Mexican border town of Tijuana who had given birth between 1982 and 1987 had done so in California.19 Women gave birth across the border, by their own report, principally in order to receive adequate medical care and/or to secure U.S. citizenship for their children. The study participants then returned with the child to live in Mexico, and any postneonatal mortality that may have occurred presumably went unreported in California. A recent comparison showing Mexican-American infant mortality rates to be lower in border than in nonborder states suggests that proximity to the border might facilitate Mexican parents' return to Mexico before a child dies.15 For obvious geographic reasons, such circumstances do not apply to the Southeast Asian populations.

Another possible source of underreporting is misclassification of ethnicity and race. Whereas inaccurate coding of race at birth and at death is low for Whites, it has not been unusual for Asians and Latinos to be misclassified—usually as Whites.15,20,21 Although Latinos can be shown

to have differing infant mortality rates in relation to Whites depending on the definitions used to code infant ethnicity,22 researchers have, nevertheless, shown that the effect of such discrepancies is minor.14,23 Further, the newer standard of linking birth and death certificates, which minimizes reporting inaccuracies, has raised the Mexican-American infant mortality rate only slightly.24 In general, underreporting and misclassification should be considered in evaluating birth outcomes, but the magnitude of these effects for Latinos does not appear to be substantial. Less is known about these factors for Southeast Asian women. Available data are often aggregated into an Asian or Asian Pacific category, lacking information specific to Southeast Asians.

Underestimation of infant mortality cannot, however, explain more favorable birthweight distributions among infants of Mexico-born mothers than White mothers, unless there are selective pressures to return to Mexico when a pregnancy has complications likely to result in adverse pregnancy outcomes. Although this phenomenon has not been formally studied, evidence suggests that Mexican mothers residing in California at risk for pregnancy complications prefer health services on the U.S. side of the border because they are perceived as highly innovative and technical.19 In addition, U.S. citizenship for the U.S.-born child of lowincome immigrant parents ensures Medi-Cal insurance coverage for care after birth. The extent to which welfare reform may change these behaviors because of severe restrictions in eligibility for immigrants will require monitoring in the future.

Excess Fetal Deaths

Another conceivable deflator of the infant mortality rate might be excess fetal deaths among Latinas and/or Southeast Asians, whereby biologically weaker fetuses are eliminated and only healthy ones survive until birth.

Studies of fetal mortality are few, and they offer poor comparability because of different state reporting laws. Examination of available data in California, where the law mandates reporting of fetal deaths after 20weeks' gestation, has not supported the hypothesis that excess late fetal deaths occur in the Latina (predominantly of Mexican birth or descent) and Southeast Asian populations. Guendelman, Chavez, and Christianson studied a large sample of low-income women enrolled in the California Comprehensive Perinatal Program and found that the fetal death rate after 20 weeks' gestation among Latinas (7.8 per 1,000

live births and fetal deaths) was actually lower than the rate among Whites (8.4).25 These ethnic disparities persisted after controlling for sociobehavioral characteristics, such as maternal age and education, support systems, level of acculturation, tobacco use before and during pregnancy, and prenatal care.

One predictor of fetal death after 20 weeks is a history of fetal loss. By self-report, Latina women indicated having had fewer previous fetal losses than White women. Another indicator that compared favorably for Latinas was early fetal deaths. The rate of fetal death before 20 weeks' gestation for Latinas was 8.9 (per 1,000 live births and fetal deaths), whereas for Whites it was 13.4.25 It is important to note that this study sampled a low-income clinic population motivated to seek care, and this disposition may have lowered their risk of unreported fetal death. Future research should focus on women who do not obtain prenatal care since they are at higher risk for fetal death.

Fetal death rates were also found to be lower among Southeast Asian refugees compared to the rest of the population in San Diego County.14 Data from the 1990–92 California birth cohort files indicate comparable rates of late fetal deaths among Southeast Asian and White women

(5.2 vs. 5.0 per 1,000 live births), although there are marked variations among the Southeast Asian groups (3.7 for Vietnamese, 7.0 for Cambodian, and 7.7 for Laotian women). It should be noted that a reluctance to use Western medical services, particularly among the Cambodian and Hmong, may increase lack of ascertainment.26 In addition, the late onset of prenatal care among Southeast Asians and Mexican Americans may skew the fetal death statistics since spontaneous abortions may be occurring at home, without detection.

Clearly, more studies are needed to compare the actual rates of fetal death among our study populations and Whites. Yet the information to date offers little support for the excess fetal death hypothesis as a likely explanation of the epidemiological paradox. In fact, a large multicenter (multistate) study containing 34,350 births recently examined fetal deaths at 20 weeks' gestation or above. According to the findings, the likelihood of a fetal death for Latinos was similar (0.9 vs. 1.0) to that of Whites.27

Selective Migration

A cursory look at birth data from Mexico and the countries of Southeast Asia gives the impression that those who emigrate are not in the

same health pool as those who stay behind. In Mexico, for example, as many as 15% of babies are born at low birthweight, and 47 infants for every 1,000 live births die in their first year.27 Between 1980 and 1985, at the height of Southeast Asian immigration into the United States, the infant mortality rates (per 1,000 live births) for Cambodia, Laos, and Vietnam were 160, 122, and 63, respectively.28,29

Several studies have shown that both economic and cultural selfselection operate in voluntary migration, as from Mexico.30–32 The unpredictability of the economic environment in the sending communities often motivates people to want to take the risks involved in relocation. Immigration is expensive; it includes monetary costs, opportunity costs, and psychic costs, which are reduced if the migrant has network connections in the host country. It is the somewhat more affluent and skilled persons rather than those in the bottom of the socioeconomic hierarchy who are most likely to go in search of higher-paying jobs and a better life.33

Evidence indicates that labor migration decisions are made jointly by family members within households.35 But selection factors may vary according to gender roles and expectations. While Mexican male migrants are pushed out of their communities by lack of employment and pulled to the United States by labor and higher wages and social network ties that facilitate access to employment,33–35 female participation in migration is more often a means of keeping the family together and providing continuity of care.36 This being the case, health selection factors may perhaps be stronger among Mexican men, who are most often the initiators of migration, than among Mexican women, who are often the implementers of household decisions to migrate.

In the case of Southeast Asians, the rates of adverse pregnancy outcomes in refugee relocation camps, while still elevated, were much lower than in the refugees' countries of origin.26 This indicates that perhaps sturdy women were more able and likely to leave their war-torn countries of origin behind. Moreover, health conditions including health care services were likely to be better in the refugee camps. Selection also took place through the health screenings of refugees in the refugee processing centers. These consisted of a general physical exam, a serologic test, a chest X ray, immunizations, and treatment for tuberculosis and venereal disease. Positive screens for infectious diseases resulted in quarantines in the refugee camps. On relocation, diagnostic surveillance continued, and health problems were followed up.37 The hypothesis that migration is selective of a healthier population is further borne out by the obervation

that rates of chronic disease among Southeast Asians in the United States are far lower than those observed in the countries of origin.26

These clues notwithstanding, the selectivity hypothesis has not been empirically tested. For instance, studies that compare the health of nonmigrants and prospective migrants who plan to emigrate from their communities of origin in Mexico with those immigrants who are residing in California are needed to assess selectivity effects.30 Future household surveys may also want to include information from the sisters of the women interviewed in order to assess the differential effects of migration experiences.

Protective Sociocultural Factors

While the hypotheses discussed to this point offer some hints about the health paradox among immigrant mothers, perhaps most compelling from the “prevention” perspective is the idea that immigrants and refugees might be profiting from sociocultural and behavioral factors whose benefits outweigh the risks stacked against them. It appears that newcomers bring to the host society values, attitudes, and behaviors that protect them against the risks of adverse pregnancy outcomes or directly contribute to healthy outcomes.


Several studies have shown that the consumption of tobacco, alcohol, and illicit drugs during pregnancy are associated with poor pregnancy outcomes. Fetal growth retardation has been associated with smoking 38,39 and with moderate to high levels of alcohol use.40,41 Both smoking and drug use have been associated with fetal death, low birthweight, and preterm births.10, 38–43 Substance use also contributes to general pregnancy complications and congenital malformations.3,38,42,43

Cigarette Smoking

Cigarette smoking during pregnancy causes close to 10% of fetal and infant deaths and one-fifth of all low-birth-weight births in the United States and is the single most important known cause of environmentally induced low birth-weight.10 Women who smoke are almost twice as likely to deliver a low-birth-weight baby as are nonsmokers.44 Studies


Figure 9.1. Tobacco use of Asian, Latina, and White maternity patients. Source: Vega et al. (1993 [45]).

consistently show low prevalence rates of tobacco use among Mexican-American, including Mexico-born women, and Southeast Asian women.

Vega et al.,45 in a study of perinatal substance use among almost 30,000 women attending 202 hospitals in California in 1992, found that Latina women, principally of Mexican birth or descent, were far less likely than White women to have reported that they smoked during pregnancy (3.3% vs. 14.8%; Figure 9.1). Foreign-born Latinas were 3.6 times less likely to smoke than native-born Latinas (1.8% vs. 6.6%). Asian women, in comparison with White and Mexican-born women, were the lowest consumers of tobacco (1.7%; Figure 9.1).

Self-reports of tobacco use in the Health and Nutrition Examination Survey (HANES) studies corroborate the finding of Vega et al. that Mexican-American women smoke less during pregnancy than do Whites. Guendelman and Abrams compared 664 Mexican-American women in the Hispanic HANES and 1,156 White women in the second HANES


Figure 9.2. Prevalence of cigarette smoking among low-income Mexican-American and White women. Source: Camilli et al. (1994 [47]).

across stages of the reproductive cycle and found that whereas 23% of Mexican-American women smoked in the interconceptional period, only 8.1% smoked during pregnancy. In contrast, among White women there were nearly twice as many interconceptional smokers (43%), and this level remained high during pregnancy (37.3%).46

Since Guendelman and Abrams's study used crosssectional data, empirical evidence of quitting could not be ascertained. Yet evidence from a recent study by Camilli, McElroy, and Reed shows that Mexican-American smokers are more likely to quit during pregnancy than White smokers. The authors compared 200 Mexican-American and 131 White low-income women seeking prenatal care in a university hospital in Tucson, Arizona. As shown in Figure 9.2, 24% of Mexican-American women had smoked in the year before pregnancy, compared with 51% of White women. Only 19% of Mexican Americans had smoked during any part of their pregnancy, whereas 48% of their White counterparts had done so. Furthermore, on average, Mexican Americans smoked almost five fewer cigarettes per day than Whites (6.9 vs. 11.8). The odds of quitting during pregnancy, as verified by urinary cotinine values, were 4.71 times higher for Mexican Americans (95% CI 1.66–13.38).47

Since quitting even as late as the seventh or eighth month has a positive influence on birth-weight, the health benefits to Mexican Americans are clear.48

Less is known about smoking before and during pregnancy among Southeast Asians. Data from 1989–91 describing women participating in the San Diego Comprehensive Perinatal Program (a state-funded program for pregnant low-income women) indicate no history of smoking for Southeast Asian women in the program.49 In comparison, 6.7% of Mexico-born immigrants and 29.4% of White women reported having smoked previously in their lives. An earlier study of Southeast Asian women in San Diego County found that less than 2% of Southeast Asian mothers were smokers.5 In another study, conducted in Washington State, cigarette smoking among Southeast Asian immigrants during pregnancy decreased slightly from 3% to 2.4% between 1984 and 1986.50 These rates for Southeast Asian women are consistently lower than for Whites.

Since the bulk of evidence shows a clear and consistent association between low birth-weight and infant mortality and smoking, the low rate of smoking in these immigrant populations is clearly advantageous.

Alcohol Use

Alcohol use during pregnancy has been associated with both shortand long-term negative health effects for infants, including congenital malformations and mental retardation.48 Women who consume large amounts of alcohol during pregnancy have higher rates of low-birth-weight babies than do nondrinkers.41 While the evidence is mixed, alcohol use during pregnancy appears to be low among Mexican-American and Southeast Asian women.

Using food frequency data from two Health and Nutrition Examination Surveys (the Hispanic HANES and the second HANES), Guendelman and Abrams compared mean daily servings of beer, wine, and liquor for 664 Mexican-American women and 1,156 White women across four stages of the reproductive cycle.46 On average, pregnant Mexican-American women consumed .02 daily servings of alcohol compared to .08 servings among pregnant White women. Interconceptional, pregnant, lactating, and postpartum Mexican Americans were far less likely than Whites to consume alcohol (Figure 9.3).

Vega et al., in their study of perinatal substance use among 30,000 women, assessed alcohol exposure at the time of delivery, employing


Figure 9.3. Mean daily servings of alcohol for Mexican-American and White women by reproductive stage. Source: Guendelman and Abrams (1994 [46]).

urine toxicology screens.45 The study found that, in comparison with White women, positive alcohol screens were more likely among Latinas (6.1% vs. 6.9%; Figure 9.4). (A woman was considered positive for alcohol use if she had drunk at least 6 ounces of beer, 2 ounces of wine, or 0.5 ounces of distilled spirits in the period immediately before she was admitted as a maternity patient or had drunk larger quantities of alcohol more than a few hours before admission.) The high prevalence rates of alcohol use for both foreign- (6.7%) and native-born (7.3%) Latinas may suggest cultural prescriptions to use alcohol prior to delivery to better cope with labor. Available evidence based on self-reports supports Guendelman and Abrams's findings that Latina women, especially the Mexican born, are lower consumers of alcohol generally than White women.49, 51–54

A study of Southeast Asian women participating in the San Diego Comprehensive Perinatal Program revealed low prevalence of alcohol consumption.49 Furthermore, urine toxicology screens of Asian women at delivery administered in the study by Vega et al. showed fewer positive screens for alcohol compared to Whites (5.1% vs. 6.1%). Anecdotal information suggests that alcohol intake may be restricted to the time of


Figure 9.4. Alcohol use of Asian, Latina, and White maternity patients. Source: Vega et al. (1993 [45]).

delivery. For instance, cultural prescriptions among Cambodian women appear to favor alcohol use both prior to and following childbirth in order to “strengthen the blood.”55

While important, it does not appear that alcohol has nearly as strong an impact on low birth-weight and infant mortality as cigarette smoking.10 However, the low prevalence rates of alcohol consumption during pregnancy in immigrant groups does suggest a reproductive health advantage.

Illicit Drugs

Prenatal use of controlled substances has been correlated with fetal growth retardation, perinatal death, and pregnancy and delivery complications.56–60

In the study by Vega et al.,45 significantly fewer Latinas tested positive for any drug at the time of delivery than did White women (2.8%


Figure 9.5. Drug use of Asian, Latina, and White maternity patients. Source: Vega et al. (1993 [45]).

vs. 6.8%; Figure 9.5). Asians were also more likely to have lower prevalence rates than Whites for licit and illicit drugs, except for opiates (Figure 9.5). As was the case with Latinas, Asian women who were foreign born (which includes the vast majority of Southeast Asians) had much lower prevalence rates of substance use compared to native-born Asians. The only exception was opiate use, which was slightly higher among the foreign born (1.5% vs. 1.3%) and similar for foreign-born Asians and Whites.

Overall, foreign-born Asian and Latina women in this Vega et al. study were far less likely to consume addictive substances than White women during pregnancy, except for alcohol among Latinas. These findings indicating low consumption have been supported by recent studies conducted by Rumbaut and Weeks 49 and Newman et al.61 in San Diego.

Although few pregnant women engage in drug abuse, it appears that those who do are generally in poorer health and obtain limited prenatal

care.48 The much lower prevalence of illicit drug use among immigrants suggests another health advantage.


A nutritious diet helps to meet the changing needs of the pregnant woman and her fetus. Specific nutrients such as calcium, zinc, protein, iron, and vitamins C, A, and E and folic acid have been related to pregnancy outcomes,38,62 and there is no evidence of substantial differences in nutritional requirements among various ethnic groups. Guendelman and Abrams compared the intake of the previously named eight nutrients between White women and Mexican-American women of reproductive age, using data from two Health and Nutrition Examination Surveys.63 For the purpose of comparison, these analyses are extended here to examine the different nutrient intakes of Mexican-American pregnant women in the Hispanic HANES (n = 79), White women in the second HANES (n = 72), and a sample of pregnant Southeast Asian women who participated in a Prenatal Nutrition Project at the University of California at San Diego between similar reference periods (1978–90). The latter were studied by Newman et al., who reported results based on 91 Cambodian, 37 Laotian, and 59 Vietnamese women.61 For all groups, dietary intake was elicited by participant recall of food and beverage consumption during the preceding 24-hour period.

As indicated in Table 9.5, the five study groups did not differ significantly with respect to age or mean number of live births. However, they differed markedly with respect to height, weight, and body mass index (BMI). White women were the tallest, while Mexican Americans had the highest BMI. All three Southeast Asian groups were shorter and lighter and had a lower BMI than either Mexican-American or White women. Despite these differences, the energy intake among the groups was similar (Table 9.6).

A comparison of the mean daily intake of each nutrient relative to the recommended daily allowance (RDA) standards for pregnant women shows that the mean intake of protein was above the RDA for all ethnic groups (Table 9.6). Yet the protein intake was significantly higher for Southeast Asian women (particularly the Vietnamese) in comparison with Mexican-Americans and White women. Mexican-American and White women did not differ significantly in their intake of any of the eight nutrients, and the mean daily intake of vitamins C and A, iron, and zinc relative to the RDA was similar across all ethnic groups. Aside from

(n = 74)
(n = 91)
(n = 37)
(n = 59)
(n = 75)
NOTE: Groups sharing a same-letter superscript are not significantly different from each other at a .05.
SOURCE: Mexican-American data from NCHS, Hispanic Health and Nutrition Examination Survey, 1982–84; Cambodian, Laotian, and Vietnamese data from Newman, Norcross, and McDonald Prenatal Nutrition Project, University of California, San Diego, 1978–90; non-Latina White data from NCHS, Second National Health and Nutrition Examination Survey, 1976–80.
Age (yr) 25a (6) 27a (6) 25a (5) 26a (5) 26a (5)
Height (cm) 159c (7) 153b (5) 149a (5) 152b (7) 164d (6)
Weight (kg) 64c (11) 50b (8) 48a,b (7) 47a (7) 62c (12)
Body mass
25.5d (4.4) 21.4b (3.1) 21.7b (2.9) 20.3a (2.8) 23.3c (4.2)
Previous live
1.6a (1.6) 2.3a (1.8) 2.3a (2.4) 1.6a (1.6) 1.2a (1.6)
protein, then, the findings do not show a better diet for immigrant compared to nonimmigrant pregnant women. (Because the RDAs are estimated to exceed the nutrient requirements of most individuals, intakes below the RDA for a given group are not necessarily inadequate, but they do suggest an increased likelihood of poor dietary intake.) In fact, compared with non-Southeast Asian women, Cambodian women showed a lower intake of folate, vitamin E, and calcium.

These findings were not adjusted for socioeconomic status. White women in the HANES sample had higher incomes than the immigrant groups in either study, and it is possible that after controlling for income, Southeast Asian women, all of whom were at or under 200% of the poverty level in the Newman et al. study, would have had better nutrient intake than Whites. Clearly, more research, utilizing larger samples and controlling for socioeconomic status, is needed to compare the nutrient intake of Southeast Asian and White women.

Somewhat more information is available for Mexican Americans. As with drug and alcohol use, the nutrient intake of Mexican-born women seems to be far better than the intake of U.S.-born women of Mexican descent. According to Guendelman and Abrams in their study of generational differences in nutrition,63 Mexican-born immigrants had significantly higher absolute intake and higher average intake relative to RDA standards for protein, vitamins (A, C, E, and folic acid), and calcium than did second-generation Mexican Americans and Whites (Table 9.7).

Nutrients Mexican-American
(n = 76)
(n = 24)
(n = 24)
(n = 20)
(n = 72)
Energy (Kcal) 1,901a (796) 1,750a (755) 1,630a (576) 1,932a (753) 1,988a (812)
Percentage of Recommended Dietary Allowance
  (n = 79) (n = 91) (n = 37) (n = 59) (n = 72)
NOTE: Groups sharing a same-letter superscript are not significantly different from each other at a .05.
SOURCE: Mexican-American data from NCHS, Hispanic Health and Nutrition Examination Survey, 1982–84; Cambodian, Laotian, and Vietnamese data from Newman, Norcross, and McDonald Prenatal Nutrition Project, University of California, San Diego, 1978–90; non-Latina White data from NCHS, Second National Health and Nutrition Examination Survey, 1976–80.
Protein 135a (66) 173b (76) 176b (76) 187b (86) 134a (67)
Vitamin C 166a (151) 178a (140) 205a (201) 168a (143) 172a (164)
Folate 72b (62) 46a (29) 54a,b (31) 55a,b (35) 68b (55)
Vitamin A 177a (376) 129a (88) 199a (239) 190a (225) 184a (175)
Vitamin E 83b (85) 44a (36) 59a,b (47) 77b (84) 92b (125)
Iron 41a (23) 45a (23) 46a (21) 46a (18) 44a (26)
Calcium 79c (42) 42a (28) 51a,b (31) 67b,c (43) 97c (53)
Zinc 77a (43) 78a (41) 81a (50) 82a (46) 93a (120)

(n = 475)
(n = 898)
(n = 2,326)
NOTE: Data expressed as mean percentage of intake relative to the Recommended Daily Allowance for that nutrient specific to the woman's reproductive state. The NARs were truncated at 1.0.
NOTE: Groups sharing a same-letter superscript are not significantly different from each other at α = .05.
SOURCE: Data for Mexican- and U.S.-born Mexican Americans from NCHS, Hispanic Health and Nutrition Examination Survey, 1982–84; data for non-Latina Whites from NCHS, National Health and Nutrition Examination Survey, 1976–80.
Mean intake,
74.3c (1.7) 68.3b (1.6) 63.9a (0.84)
NAR (SE) 0.92e (0.009) 0.88d (0.006) 0.89d (0.004)
Vitamin A      
Mean intake,
6,347.4b (432.2) 4,240.8a (228.0) 4,596.5a (179.8)
NAR (SE) 0.71f (0.017) 0.61d (0.015) 0.66e (0.008)
Vitamin C      
Mean intake,
mg (SE)
104.1b (4.7) 84.1a (3.9) 87.9a (2.5)
NAR (SE) 0.78e (0.014) 0.69d (0.012) 0.71d (0.008)
Vitamin E      
Mean intake,
Alpha TE (SE)
7.9 (0.5) 7.3 (0.3) 7.5 (0.2)
NAR (SE) 0.70 (0.020) 0.68 (0.012) 0.70 (0.007)
Folic acid      
Mean intake,
mg (SE)
266.5b (12.2) 205.5a (5.2) 200.2a (3.9)
NAR (SE) 0.82e (0.020) 0.75d (0.010) 0.77d (0.006)
Mean intake,
mg (SE)
778.8b (23.8) 644.5a (32.5) 677.7a (16.3)
NAR (SE) 0.70f (0.024) 0.57d (0.014) 0.61e (0.009)


Although this study did not follow women through their pregnancies, the results suggest that nutrition may help to explain the much lower rate of low birth-weight among first-generation Mexican-American women than among U.S.-born women of Mexican descent.64,65 Large epidemiological studies are needed to examine the association between dietary intake, weight gain during pregnancy, and pregnancy outcomes among newcomer populations to help us further unravel the epidemiological paradox.


The role of social factors in explaining the paradox is even more poorly understood than that of health and nutrition habits. Nevertheless, some social factors related to family and social networks seem to provide clues to better reproductive health, even though we do not understand the mechanisms by which they affect pregnancy outcomes.

Close kin networks may confer protection to the pregnant woman and compensate for income deficits by improving access to informational and psychosocial support.66–69 These resources may translate into more knowledge about healthy pregnancies, the encouragement of positive behaviors, and less stress during pregnancy, all of which more directly affect perinatal morbidity and mortality. They may also alter hormonal and immunological responses associated with pregnancy complications.70

Research on Latinos and Southeast Asians has described the centrality of the family in both cultures. Latinos tend to have close kin networks and emphasize the collective needs of the family over individual needs.71–74 Indeed, the family has been described as the single most important institution for Mexican Americans.75,76 Kinship in this case comprises not only relatives but also the Latino compadre system, which establishes “coparents,” in the Catholic tradition, who share broader, less formalized obligations toward the children.77 Recent evidence further suggests that women of Mexican descent appear to have more social network contacts outside of the family 66 compared with Whites as well as enhanced access to psychosocial and informational social support. As noted, these factors may contribute to favorable pregnancy outcomes by making more resources available to the pregnant woman, thereby compensating for economic deficits.

Family studies conducted in Mexico suggest that children in poor

families have a high economic and moral value.78 They are considered one of the few sources of personal achievement and pride. Children are the main reason for marrying, a compensation for any unsatisfactory or broken marriage and a source of companionship and economic support.78

Family is similarly important for each of the Southeast Asian groups.79 After an extensive review of the literature on Southeast Asians, Frye concluded that kinship solidarity is the “lifeline” in these cultures.80 Much as Mexicans do, Southeast Asians view the individual as “subservient to the kinship-based group.” Yet the specific character of Southeast Asian families and traditions varies. Vietnamese have extended patrilineal family systems, Laotians and Cambodians rely more on nuclear family supports, and Hmong have a clan system.

Despite the heterogeneous family structures of Southeast Asians, pregnant women in these cultures consistently tend to receive positive social support from family and kin, as demonstrated by spousal approval of the pregnancy and familial monitoring of the health, diet, and lifestyle of the pregnant woman.14 These immigrant subgroups also seem to share with Mexican Americans the cultural belief that having a child demonstrates one's femininity and that fertility signifies womanhood and a main source of marital satisfaction.81,82 In addition, community assistance from refugee programs in terms of child care, chores, and financial aid represents significant, though dwindling, support.

Family stability also appears to influence reproductive health. For instance, a study by Ramsey et al.69 showed that women who lived alone were at highest risk of having smaller babies, while living with extended family was correlated with higher birth-weights. Living with a husband further increased the likelihood of having a heavier baby.68 These effects might be mediated by such factors as higher income, better nutrition, and less stress.

Compared with Whites, Mexican Americans have a higher proportion of husband-wife families and lower rates of divorce and separation.68 Family stability among Mexican Americans appears to be highest among ever-married women who have the lowest educational level and highest use of the Spanish language. Similarly, among Southeast Asian women, the prevalence of unwed mothers is extremely low. Hopkins and Clarke found that within each Southeast Asian subgroup, less than 1% of mothers were unmarried, compared with over 17% in the general U.S. population.83

Family stability may play a role in the case of teenage pregnancies as

well. Scientific and popular understanding have linked births to teenage mothers (under 18 years) with poverty, welfare dependency, and a host of other social problems, including alienation from family. But teenage pregnancy among Mexican Americans and Southeast Asians appears to follow a different pattern. In both groups, pregnancy at a young age appears to be more common and more culturally acceptable than among Whites, and teenage mothers are often cared for and supported by extended family. While it has been reported that Southeast Asian cultures consider birth out of wedlock to be a disgrace,84–85 early marriage in these groups is encouraged and creates a context for culturally acceptable teenage births.86 In fact, although it does not appear that childbearing begins at an earlier age among Southeast Asians than in other populations, Southeast Asian teens are more likely than their White counterparts to have short birth intervals and low contraceptive use. Hence, teen births in these populations may pose risks not because they signal lack of family supports but because of low socioeconomic status and biological considerations.

Few studies have directly tested the relationship between family networks, family stability, and pregnancy outcomes. The prevalence of strong and stable family networks in immigrant populations suggests that these factors might help to explain the paradox of favorable pregnancy outcomes among at-risk populations.


A corollary to the protective sociocultural hypothesis is the acculturation hypothesis. According to the latter, as immigrants spend more time in the United States or move to the second generation, their healthy behaviors, norms, and attitudes change, resembling those of the White nonimmigrant population or of high-risk groups with which they come into contact. Shifts in health risks coupled with changes in sociodemographic characteristics that occur with acculturation affect pregnancy outcomes.

As discussed in the previous section, alcohol, illicit drugs, and tobacco use are low and family stability is high among recent immigrants. They also adhere to a traditional diet, which seems healthful. With acculturation, healthy behaviors among Mexican Americans are worse: alcohol, tobacco, and drug use during pregnancy increase, and the quality of the diet decreases (Figures 9.1, 9.4, and 9.5; Table 9.7). These

shifts could explain the increased prevalence of low birth-weight and infant mortality in Mexican Americans who are second generation and beyond (Table 9.4).

Furthermore, teen pregnancies under 18 years increase in the second generation for women of Mexican descent. While 4.6% of Mexicanborn mothers are teens, this rate more than doubles to 11.6% among U.S.-born Mexican-American mothers.65 Reynoso et al. found that acculturated pregnant Mexican-American teenagers engaged in sexual behavior at an earlier age than less acculturated teens and reported that they were more likely to consider single parenthood as an option.87

In addition to a loss of protective behaviors, changes in socioeconomic status that occur with acculturation to U.S. society may further affect pregnancy outcomes. Although Mexican immigrants are considered to be of low socioeconomic status in the United States, it has been amply documented that average family earnings are significantly improved over those in Mexico, and a large proportion of newcomers are able to accumulate small savings and send remittances back home.35 In contrast, U.S.-born women of Mexican descent do not appear to enjoy a similar sense of getting ahead in society. Subjective feelings of poverty and discrimination have been shown to negatively affect birth outcomes,11 and many of these women become stalled in poverty and feel oppressed by social discrimination.88 Another insidious influence in ethnic communities is the “corporate targeting” of these groups in alcohol and tobacco advertising. Evidence shows an association between tobacco and alcohol advertisements and an increase in risky behaviors.89–90 A higher consumption of alcohol and cigarettes among nativeborn Mexican Americans is consistent with this evidence.

Recent findings suggest that it may not take a whole generation for changes in the reproductive risk profile of Mexican Americans to become apparent. Guendelman and English found that within five years of moving to this country, there was notable deterioration in the perinatal health of Mexican-born women living in California. Long-term residents had fewer planned pregnancies and were more likely to smoke than newcomers who had lived in the country for five years or less. After controlling for smoking, planned pregnancy, and maternal age, longterm immigrants living in the United States for more than five years were more likely to have pregnancy complications and to deliver preterm and low-birth-weight infants than newcomers.91

Among Southeast Asians, whose immigration to this country is a much more recent phenomenon, more than 95% of women are foreign

born.92 The health effects of acculturation in the second generation and beyond, therefore, remain to be seen, but we can begin to examine the effects of acculturation in the first generation. Preliminary findings suggest that unlike Mexican immigrants, risky behaviors decrease and birth outcomes improve with increased length of stay among Southeast Asians.

Studies by Rumbaut and Weeks 5 and Li et al.50 indicate that the health status of recent cohorts of Southeast Asian immigrants has not improved, as measured by the prevalence of hepatitis B and tuberculosis. However, Li et al.'s analysis of consecutive births to the same parents from 1984 to 1987 showed that low birth-weight declined more than expected during this period among Southeast Asian immigrants. A decline in low birth-weight over a five- to six-year period was also observed in trend analyses of Washington, Massachusetts, and San Diego County births to Southeast Asians.7,26,50 Yip et al. found a similar improvement in low birth-weight rates among low-income Southeast Asian refugees at a national level between 1980 and 1989.93

Although these researchers suggest that initial observations of low birth-weight may have been skewed by health and nutritional deficiencies in refugees coming out of relocation camps, the reduction in low birth-weight may also be related to an increased number of years of residency in the United States. Increased acculturation in the Li et al. study allowed for a change in paternal occupational status from student to employed that was associated with a 27% reduction in the low-birth-weight prevalence, independent of maternal age, infant sex, and prior gravidity.50 As with Mexican Americans, this move into employment among refugees presumably leads to an improvement in socioeconomic status that may be associated with good pregnancy outcomes. Another possibility is that a relatively stable yet declining smoking prevalence during the 1984–86 study period in the Washington State Southeast Asian population may account for these improved outcomes. (The number of births to smokers during pregnancy decreased by 0.6%.) Whether these effects are sustained in the second generation will require investigation in the future.

Changing dietary habits may also alter perinatal health. However, studies of adolescents indicate that while Southeast Asians adopt some U.S. nutritional habits such as drinking more milk and soft drinks, they abstain from a lot of nutritionally weak foods.94 The adherence to a traditional diet may help to preserve favorable birth outcomes in this population.


These preliminary findings indicating differing effects of acculturation on pregnancy risks and outcomes by race/ethnicity suggest that acculturation is mediated by highly contextual factors. The process of acculturation, in which individuals acquire ways of living, values, attitudes, and behaviors from another culture, is complex and segmented and therefore unlikely to be solely a function of time spent in the United States. Future studies assessing the impact of acculturation on birth outcomes should examine the effects of years of residency as well as the age of entry into the United States, the extent of English language use, changes in socioeconomic status, the presence of extended kin networks, family stability and support, community receptivity to immigrant families, and shifts in traditional roles for women.


The rapidly growing Mexican-American and Southeast Asian populations in California are quite heterogeneous in terms of social and cultural backgrounds. Despite the diversity, both within and across immigrant groups, these populations share a socioeconomic disadvantage compared to White Californians.

Although research has linked low socioeconomic status with a host of health risk factors and adverse outcomes, this relationship does not necessarily hold when examining the pregnancy outcomes of these immigrant women. As this chapter has shown, Mexican-American and Southeast Asian immigrants have favorable pregnancy outcomes despite their socioeconomic disadvantages. This health paradox is more accentuated among foreign-born women, who are even poorer than their U.S.-born counterparts. There is strong evidence to suggest that immigrants bring to the United States values, attitudes, and health behaviors that may protect them from adverse pregnancy outcomes. Among the protective factors, the very low use of addictive substances stands out as an important contributor to healthy outcomes. Other protective factors such as good nutrition, a strong sense of family and social support, and a positive attitude toward childbearing show strong potential for contributing to favorable pregnancy outcomes.

These protective factors may buffer immigrant women from the stresses of poverty or else directly contribute to positive outcomes by bolstering the immune and hormonal systems. Although several studies

have focused on the relationship between these factors and pregnancy outcomes in other populations, remarkably few studies have focused on immigrant Latina and Southeast Asian women. Large epidemiological studies are needed to examine the relationship between pregnancy outcomes and healthy diets, weight gain during pregnancy, healthy habits, family stability, teenage pregnancy within a supportive family system, and networks that provide informational and emotional support and reinforce healthy behaviors among immigrants.

As this chapter demonstrates, the pregnancy outcomes of immigrant women vary according to nativity and increased exposure to American society. Although certain risk factors associated with pregnancy out-comes—such as education, income, and access to prenatal care—improve among U.S.-born, second-generation Mexican Americans, many protective factors become eroded. Compared with first-generation Mexican-American women, the pregnancy outcomes of second-generation women are less favorable.

While it is too early to examine generational changes in birth outcomes among the more recently arrived Southeast Asian population, we can begin to explore the effects of acculturation among foreign-born Southeast Asian women. Research suggests that they may be buffered from many of the negative effects of acculturation, as demonstrated by their improving birth outcomes in recent years. This response contrasts with that of Mexican immigrants who appear to show a marked deterioration in risks and pregnancy outcomes after only five years of residing in the United States. Such differentials may be a result of the different ways in which immigrants adapt to our society. More research is needed to examine the modes of immigrant adaptation and its effect on pregnancy outcomes. We must determine whether the differentials observed between the two immigrant groups are a product of a different community receptivity to these populations or a different sociocultural orientation that immigrants bring to our society.

Recognizing that tremendous gaps in knowledge exist, some preliminary conclusions can be drawn regarding what immigrants can teach us about having healthy babies.

This health paradox demonstrates—contrary to the implications of earlier epidemiological studies—that poverty does not necessarily coincide with unhealthy lifestyles and that a lack of economic resources does not always mean a lack of human and social resources. If we grasp the significance of this paradigm shift, we may be in a better position to design health promotion policies that address immigrants' needs by emphasizing

their sociocultural assets rather than assuming—and often blaming them for—their deficits.

With the advent of California's “majority-minority” population in the 21st century and the increasingly negative stereotypes placed on immigrants, as well as the cutbacks in social programs for the poor, it is incumbent on health care providers, public health planners, and policy makers to recognize the positive health and social aspects of immigrant communities. Such awareness is important in order not only to preserve the health and healthy lifestyles of immigrant women and their children but also to learn ways of transferring this knowledge to promote health in other communities with a high incidence of infant mortality and lowbirth-weight babies. In recognition of these positive and protective factors and the benefits that they provide to all communities, the following steps are recommended:

The reproductive health of California's large immigrant populations is a compelling area for future research and the development of new health promotion strategies. Through increased attention to these groups, we can more fully understand how to optimize maternal and child health for all Americans.


The author gratefully acknowledges Ann Banchoff for assistance in preparing this manuscript, Beate Herrchen for providing valuable birth cohort file data, and Paul English and Christopher Grover for helpful comments on an earlier draft of this chapter. Thank you also to Lora Santiago for clerical support.



1. Haan, H., Kaplan, G., and Syme, L.1989. “Socioeconomic status and health: Old observations and new thoughts. In” Pathways to Health: The Role of Social Factors, edited by J. P. Bunker, D. S. Gumby, and B. H. Kehrer.Pp. 76–133. Palo Alto, Calif.: Henry J. Kaiser Foundation.

2. Syme, L., and Berkman, L.1976. “Social class, susceptibility, and sickness.” American Journal of Epidemiology, 104, 1–8.

3. “Institute of Medicine, Committee to Study the Prevention of Low Birth-Weight.” 1985. Preventing Low Birth-Weight.Washington, D.C.: National Academy Press.

4. “California State Department of Finance Demography Research Unit.” Data from federal fiscal year 1992–93 (personal communication).

5. Rumbaut, R. G., and Weeks, J. R.1989. “Infant health among Indochinese refugees: Patterns of infant mortality, birth-weight and prenatal care in comparative perspective.” Research in the Sociology of Health Care, 8, 137–196.

6. Levy, S.1995. California Population Characteristics.Palo Alto, Calif.: Center for Continuing Study of the California Economy.

7. “California Department of Health Services, Center for Health Statistics.” 1993. California Birth Cohort File.

8. “California Department of Health Services.” 1994, February. Analysis of Health Indicators for California's Minority Populations.Sacramento: Author.

9. “California Department of Health Services, Center for Health Statistics.” 1987. California Birth Cohort File.

10. Shiono, P. H., Behrman, R. E.1995. “Low birth-weight: Analysis and recommendations.” Future of Children, 5 (1), 4–18.

11. Cramer, J. C.1995. “Racial and ethnic differences in birth-weight: The role of income and financial assistance.” Demography, 32 (2), 231–247.

12. Shiono, P. H., Klebanoff, M. A., Graubard, B. U., et al. 1986. “Birth-Weight among women of different ethnic groups.” Journal of the American Medical Association, 255 (1), 48–52.

13. Williams, R. L., Binkin, N. J., and Clingman, E. J.1986. “Pregnancy outcomes among Spanish-surname women in California.” American Journal of Public Health, 76, 387–391.

14. Weeks, J. R., and Rumbaut, J. R.1991. “Infant mortality among ethnic immigrant groups.” Social Science and Medicine, 33 (3), 327–334.

15. “Center for Health Policy Research, George Washington University, and the School of Public Health, University of California, Berkeley.” 1995. Mortality Rates among Infants of Mexican Descent in the United States: Evaluating the Validity of Current Estimates.Washington, D.C.: Author.

16. Selby, M. L.1984. “Validity of the Spanish surname infant mortality rate as a health status indicator of the Mexican American population.” American Journal of Public Health, 74 (9), 998–1002.

17. Teller, C., and Clyburn, S.1974. “Texas population in 1970: Trends in infant mortality.” Texas Business Review, 40, 240–246.

18. Palloni, A.1978. “Application of an indirect technique to study group differentials.

In” Demography and Racial and Ethnic Groups, edited by F. Bean and W. P. Frisbie. New York: Academic Press.

19. Guendelman, S., and Jasis, M.1992. “Giving birth across the border: The San Diego-Tijuana connection.” Social Science and Medicine, 34 (4), 419–425.

20. Yu, E. S. H., and Liu, W. T.1992. “U.S. national health data on Asian Americans and Pacific Islanders: A research agenda for the 1990s.” American Journal of Public Health, 82 (12), 1645–1652.

21. Becerra, J., Hogue, C., Atrash, H., and Perez, N.1991. “Infant mortality among Hispanics: A portrait of heterogeneity.” Journal of the American Medical Association, 265 (2), 217–221.

22. Rogers, R. G.1989. “Ethnic differences in infant mortality: Fact or artifact?” Social Science Quarterly, 70 (3), 642–649.

23. Powell-Griner, E., and Streck, D.1982. “A closer examination of neonatal mortality rates among the Texas Spanish surname population.” American Journal of Public Health, 72 (9), 993–999.

24. Lambert, D. A., and Strauss, L. T.1987. “Analysis of unlinked infant death certificates from the NIMS project.” Public Health Reports, 102 (2), 201–204.

25. Guendelman, S., Chavez, G., and Christianson, R.1994. “Fetal deaths in Mexican American, Black, and White non-Hispanic women seeking government-funded prenatal care.” Journal of Community Health, 19 (5), 319–330.

26. Gann, P., Nghiem, L., and Warner, S.1989. “Pregnancy characteristics and outcomes of Cambodian refugees.” American Journal of Public Health, 79 (9), 1251–1257.

27. Copper, R., Goldenberg, T., DuBard, M., Davis, R., and the collaborative group on preterm birth prevention. 1994. “Risk factors for fetal death in white, black and Hispanic women.” Obstetrics and Gynecology, 84, 490–495.

28. Haub, C., and Yanagishita, M.1992. 1992 World Population Data Sheet.Washington, D.C.: Population Reference Bureau.

29. “Population Reference Bureau.” 1994. UN World Population Prospect 1994 Revision Annex Table A28. Washington, D.C.: Population Reference Bureau.

30. Kasl, S., and Berkman, L.1983. “Health consequences of the experience of migration.” Annual Review of Public Health, 4, 69–90.

31. Dinerman, I.1982. Migrants and Stay-at-Homes: A Comparative Study of Rural Migration from Michoacan, Mexico. Monograph Series no. 5. La Jolla, Calif.: Center for U.S.-Mexican Studies, University of California, San Diego.

32. Hull, D.1979. “Migration, adaptation, and illness: A review.” Social Science and Medicine, 13A, 25–36.

33. Frisbie, W. P., and Bean, F. D.1989. “Mexican immigration to the United States: Trends and implications.” International Review of Comparative Public Policy, 1, 65–95.

34. Portes, A.1983. “International labor migration and national development.” In U.S. Immigration and Refugee Policy: Global and Domestic Issues, edited by M. M. Kritz. Lexington, Mass.: Lexington Books.

35. Massey, D.1990. “Social structure, household strategies, and the cumulative causation of migration.” Population Index, 56,3–26.


36. Guendelman, S.1987. “The incorporation of Mexican women in seasonal migration: A study of gender differences.” Hispanic Journal of Behavioral Sciences, 9 (3), 245–264.

37. Rumbaut, R. G., Chavez, L. R., Moser, R. J., et al. 1988. “The politics of migrant health care: A comparative study of Mexican immigrants and Indochinese refugees.” Research in the Sociology of Health Care, 7, 143–202.

38. “Institute of Medicine.” 1990. Nutrition during Pregnancy.Washington, D.C.: National Academy Press.

39. Abel, E. L.1980. “Smoking during pregnancy: A review of effects on growth and development of offspring.” Human Biology, 50, 593–625.

40. Wright, J. T., Waterson, E. J., Barrison, I. G., et al. 1983. “Alcohol consumption, pregnancy, and low birth-weight.” The Lancet, 1, 663–665.

41. Mills, J. L., Grabaud, B. I., Harley, E. E., et al. 1984. “Maternal alcohol consumption and birth-weight: How much drinking in pregnancy is safe?” Journal of the American Medical Association, 252, 1875–1879.

42. Finnegan, L. P.1988. “Drug addiction and pregnancy: The newborn. In” Drugs, Alcohol, Pregnancy and Parenting, edited by I. J. Chasnoff. Pp.59–71Boston: Kluwer Academic Press.

43. Oro, A. S., and Dixon, S. D.1987. “Perinatal cocaine and methamphetamine exposure: Maternal and neonatal correlates.” Journal of Pediatrics, 111, 571–578.

44. Kramer, M. S.1987. “Determinants of low birth-weight: Methodological assessment and meta-analysis.” Bulletin of the World Health Organization, 65, 663–737.

45. Vega, W. A., Kolody, B., Hwang, J., and Noble, A.1993. “Prevalence and magnitude of perinatal substance exposures in California.” New England Journal of Medicine, 329, 850–854.

46. Guendelman, S., and Abrams, A.1994. “Dietary, alcohol and tobacco intake among Mexican American women of childbearing age: Results from the HHANES data.” American Journal of Health Promotion, 8 (5), 363–372.

47. Camilli, A., McElroy, L., and Reed, K.1994. “Smoking and pregnancy: A comparison of Mexican American and non-Hispanic white women.” Obstetrics and Gynecology, 84, 1033–1037.

48. Chomitz, V. R., Cheung, L. W. Y., and Lieberman, E.1995. “The role of lifestyle in preventing low birth-weight.” The Future of Children, 5 (1), 121–138.

49. Rumbaut, R. G., and Weeks, J. R.1994. “Unraveling a public health enigma: Why do immigrants experience superior perinatal health outcomes?” Paper presented at the 122nd annual meeting of the American Public Health Association, Washington, D.C., November 1.

50. Li, D., Ni, H., Schwartz, S. M., and Daling, J. R.1990. “Secular change in birth-weight among Southeast Asian immigrants to the United States.” American Journal of Public Health, 80 (6), 685–688.

51. Caetano, R., and Medina Mora, M. E.1988. “Acculturation and drinking among people of Mexican descent in Mexico and the United States.” Journal of Studies on Alcohol, 49 (5), 462–471.

52. Markides, K. S., Ray, L. A., Stroup-Benham, C. A., and Trevino, F.1990.

“Acculturation and alcohol consumption in the Mexican American population of the southwestern United States: Findings from the HHANES 1982–84.” American Journal of Public Health, 80 (Suppl.), 42–46.

53. Gilbert, M. J., and Cervantes, R. C.1986. “Patterns and practices of alcohol use among Mexican American women: A comprehensive review.” Hispanic Journal of Behavioral Sciences, 8, 1–60.

54. Holck, S. E., Warren, C. W., Smith, J., and Rochat, R.1984. “Alcohol consumption among Mexican American and Anglo women: Results of a survey along the U.S.-Mexico border.” Journal of Studies on Alcohol, 45, 149–154.

55. D'Avanzo, C. E., and Frye, B.1994. “Culture, stress and substance use in Cambodian refugee women.” Journal of Studies on Alcohol, 55, 420–426.

56. Zelson, C., Rubio, E., and Wasserman, E.1971. “Neonatal narcotic addiction: 10-year observation.” Pediatrics, 48 (2), 178–189.

57. Fricker, H., and Segal, S.1978. “Narcotic addiction, pregnancy, and the newborn.” American Journal of Diseases of Children, 132, 360–366.

58. Lifschitz, M., Wilson, G., Smith, E., et al. 1983. “Fetal and postnatal growth of children born to narcotic-dependent women.” Journal of Pediatrics, 102, 686–691.

59. Robins, L. N., Mills, J. L., Krulewitch, C., and Herman, A. A.1993. “Effects of in utero exposure to street drugs.” American Journal of Public Health, 83,12.

60. Oleske, J.1997. “Experiences with 118 infants born to narcotic-using mothers.” Clinical Pediatrics, 16, 418–423.

61. Newman, V., Norcross, W., and McDonald, R.1991. “Nutrient intake of low-income Southeast Asian pregnant women.” Journal of the American Dietetic Association, 91, 793–799.

62. Abrams, A., and Berman, C.1993. “Women, nutrition and health.” Current Problems in Obstetrics, Gynecology and Fertility, 1, 3–61.

63. Guendelman, S., and Abrams, B.1995. “Dietary intake among Mexican American women: Generational differences and a comparison with white non-Hispanic women.” American Journal of Public Health, 85, 20–25.

64. Guendelman, S., Gould, J., Hudes, M., and Eskenazi, B.1990. “Generational differences in perinatal health among the Mexican American population: Findings from HHANES 1982–84.” American Journal of Public Health, 80 (Suppl.), 61–64.

65. “California Department of Health Services, Center for Health Statistics.” 1987. California Birth Cohort File.

66. Shain, R.1991. “Racial/ethnic differences in adverse pregnancy outcomes.” Paper presented at the NICHD Workshop on Infant Mortality and Low Birth-Weight, Bethesda, Maryland, April25–26.

67. Cramer, J. C., Bell, K., and Vaast, K.1991, March. “Race, ethnicity, and the determinants of low birth-weight in the U.S.” Paper presented at the 1991 annual meeting of the Population Association of America, Washington, D.C.

68. Frisbie, W. P., and Bean, F. D.1995. “The Latino family in comparative perspective: Trends and current conditions.” In Racial and Ethnic Families in the United States, edited by C. Jacobson. Pp. 29–71. New York: Garland.

69. Ramsey, C., Abell, T., and Baker, L.1986. “The relationship between family

functioning, life events, family structure and the outcome of pregnancy.” Journal of Family Practice, 22, 521–526.

70. McClean, D. E., Hatfield-Timajchy, K., Wingo, P. A., and Floyd, R. L.1993. “Psychosocial measurement: Implications for the study of preterm delivery in black women.” American Journal of Preventive Medicine, 9 (Suppl. 6), 39–81.

71. Keefe, S. E., Padilla, A. M., and Carlos, M. L.1979. “The Mexican American extended family as an emotional support system.” Human Organization, 38, 144–152.

72. Swicegood, G., Bean, F. D., Stephen, E. H., and Opitz, W.1988. “Language usage and fertility in the Mexican-origin population of the United States.” Demography, 25 (1), 17–33.

73. Bean, F. D., Russell, L. C., and Marcum, J. P.1977. “Familism and marital satisfaction among Mexican Americans: The effects of family size, wife's labor force participation, and conjugal power.” Journal of Marriage and the Family, 39, 759–776.

74. Triandis, H. C., Kashima, Y., Hui, H., et al. 1982. “Acculturation and biculturalism indices among relatively acculturated Hispanic young adults.” Interamerican Journal of Psychology, 16, 140–149.

75. Alvarez, D., and Bean, F. D.1976. “The Mexican American family.” In Ethnic Families in America, edited by C. H. Mindel and R. N. Haberstein. Pp. 271–291. New York: Elsevier.

76. Murillo, N.1976. “The Mexican American family.” In Chicanos: Social and Psychological Perspectives, edited by C. Hernandez. Pp. 15–25. St. Louis: Mosby.

77. Branch, M. P., and Paxton, P. P.1976. Providing Safe Nursing Care for Ethnic People of Color.New York: Appleton-Century-Crofts.

78. De Oliveira, O.1992. Trabajo, Fecundidad y Condicion Femenina en Mexico. El Colegio de Mexico.

79. Rumbaut, R. G., and Weeks, J. R.1986. “Fertility and adaptation: Indochinese refugees in the United States.” International Migration Review, 20 (2), 428–465.

80. Frye, B. A.1995. “Use of cultural themes in promoting health among Southeast Asian refugees.” American Journal of Health Promotion, 9 (4), 269–280.

81. Manderson, L., and Matthews, M.1981. “Vietnamese behavioral and dietary precautions during pregnancy.” Ecology of Food and Nutrition, 11, 1–8.

82. Kunstadter, P., Kunstadter, S. L., Podhisita, C., and Leepreecha, P.1993. “Demographic variables in fetal and child mortality: Hmong in Thailand.” Social Science and Medicine, 36 (9), 1109–1120.

83. Hopkins, D. D., and Clarke, N. G.1983. “Indochinese refugee fertility rates and pregnancy risk factors, Oregon.” American Journal of Public Health, 73 (11), 1307–1309.

84. D'Avanzo, C. E.1992. “Bridging the cultural gap with Southeast Asians.” Maternal and Child Health Nursing, 17, 204–208.

85. Faller, H. S.1992. “Hmong women: Characteristics and birth outcomes, 1990.” Birth, 19 (3), 144–150.


86. Swenson, I., Erickson, D., Ehlinger, E., et al. 1986. “Birth-Weight, Apgar scores, labor and delivery complications and prenatal characteristics and older mothers.” Adolescence, 21 (83), 711–722.

87. Reynoso, T. C., Felice, M. E., and Shragg, G. P.1993. “Does American acculturation affect outcome of Mexican-American teenage pregnancy?” Journal of Adolescent Health, 14, 257–261.

88. Kreiger, N., Rowley, D., Herman, A., et al. 1993. “Racism, sexism and social class: Implications for studies of health, disease.” American Journal of Preventive Medicine, 9 (Suppl.), 82–122.

89. Maxwell, B., and Jacobsen, M.1989. “Targeting Hispanics. In” Marketing Disease to Hispanics.Pp. 27–46. Washington, D.C.: Center for Science in the Public Interest.

90. Mitchell, O., and Greenberg, M.1991. “Outdoor advertising of addictive products.” New Jersey Medicine, 88, 331–333.

91. Guendelman, S., and English, P.1995. “The effect of United States residence on birth outcomes among Mexican immigrants: An exploratory study.” American Journal of Epidemiology, 142 (Suppl.), S30–S38.

92. “U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census. 1992. 1990 Census of Population.” General Population Characteristics. California. Volume 6. Pp. 1–3.Washington, D.C.: Author.

93. Yip, R., Scanlon, K., and Trowbridge, F.1992. “Improving growth status of Asian refugee children in the United States.” Journal of the American Medical Association, 267 (7), 937–940.

94. Story, M., and Harris, L. J.1988. “Food preferences, beliefs, and practices of Southeast Asian refugee adolescents.” Journal of School Health, 58 (7), 273–276.



Implications for Health Care
Delivery and Wellness Promotion

Mack Roach III


“Race” has been defined as “a subdivision of the human species, characterized by a more or less distinctive combination of physical traits that are transmitted in descent” and “health” as “the general condition of the body or mind with reference to soundness and vigor.” 1 There is no obvious reason for these “concepts” to be linked. However, large differences have been noted in the state of “health” between the various racial groups for many years. Racism and the lack of quality of health care are not unique to African Americans, but because the author is most familiar with their impact on this group, the discussion to follow will emphasize these issues as they relate to them. The author believes that this discussion addresses concerns that are universal and should be of value to all Americans who are forced to live with people with skin of a different color than their own and/or to vote on issues related in any way to race and ethnicity.

African Americans have always had a lower survival rate than Whites in this country. Evidence has also always existed suggesting that “nongenetic” explanations have been the dominant causes. For example, based on a study conducted in 1908, Irish and Italian men living in New York City actually had a higher mortality rate than Black men living at that time.2 Furthermore, for many years it has been known that discrepancies in survival within racial groups varied more by residence (rural vs. urban) than between races.3 It is also noteworthy that recent

studies have demonstrated that White men in the Soviet Union currently have a substantially shorter life expectancy than all American men.4

Despite the kinds of data noted here, a body of literature has been perpetuated for many years in this country arguing that “race” is a biologic phenomenon associated with some less-than-desirable health consequences. James H. Jones summarized the position of prominent 19th-century physicians in his book Bad Blood:5,6

Vociferous advocates of black inferiority such as Dr. Josiah Clark Nott of Mobile and Dr. Samuel A. Cartwright of New Orleans published numerous articles during the 1840s and 1850s on diseases and physical properties thought to be peculiar to blacks. Drs. Nott and Cartwright were merely the best known of a group of southern physicians who helped inflame the controversy over slavery. Among the diseases said to be unique to blacks were Cachexia Africana (dirt-eating) and Struma Africana (“Negro consumption”). Influenced by these physicians, slave holders who wished to treat their bondsmen without benefit of professional help begged southern doctors to write medical manuals on the treatment of blacks. Their requests went unanswered. Instead, physicians simply continued to assert that blacks were medically inferior to whites without offering a plausible medical explanation based on racial differences. Their observations were perfect for polemics but useless for the care of sick blacks.

It was in fact this type of thinking that led to the Tuskegee syphilis experiments. Researchers from the U.S. Public Health Service in the 1930s justified studying untreated syphilis in Blacks because they believed that they knew the natural history of syphilis in Whites (based on an old Scandinavian study) and wanted to prove the hypothesis that syphilis was different in Blacks. Since this atrocity, numerous other papers have been published that imply that being a member of the Black race has a detri-mental/adverse effect on the length of survival. Implicit in many of these studies is the suggestion that race is a “real biologic factor,” meaning that it must be considered as a separate factor from tumor- or treatmentrelated factors.

For more than 20 years there has existed clear documentation of an excess cancer mortality in Black Americans compared to White Americans.5 This excess mortality experienced by Black Americans is associated with a disproportionate financial burden because of the lower incomes, higher unemployment, and the inadequacy of health care resources that are currently available to this community. The impact of the excess mortality rate and the financial burden are magnified by the increased incidence of common cancer sites among Black Americans.

There is no definitive explanation for the discrepancy in survival. Are these differences in health state due to intrinsic genetic tendencies associated with “race,” or does belonging to a racial group impact health by “nongenetic” mechanisms?

The assumption that there are major differences in biologic behavior related to race continues to be a theme of medical doctors even today. For example, in the July 1997 issue of the prestigious peer-reviewed Journal of Urology, Moul et al. published a paper titled “Black Race Is an Adverse Prognostic Factor. …” 7 Later, in an even more widely read journal, Cancer, these investigators published an article describing an equation that could be used to predict the risk of failing a radical prostatectomy that incorporated “Black race” as an unfavorable biologic parameter.8 Both of these papers included relatively small numbers of patients, and neither provided an in-depth discussion of alternative explanations. The successful publication of these papers, despite their failure to discuss other explanations, suggests that the reviewers were in agreement with the authors. Clearly there appears to be a critical mass of researchers who believe that “race” has an intrinsic impact on health. As a result of this established dogma, some researchers have found it difficult to publish papers opposing this notion.

For example, in the 1980s a paper was published in Cancer describing the poor outcome of Blacks (n = 92) treated for laryngeal cancer at Harlem Hospital. In response to this paper, I submitted a paper representing a 20-year experience from a Veterans Administration hospital including more than 300 patients demonstrating that the long-term survival in Blacks and Whites was identical.2 In our paper, I explained that based on the details provided, the care delivered appeared to have been suboptimal in the previously published paper. A major criticism of our paper (resulting in a rejection by Cancer) was that we would need “1200 patients to prove that Blacks did not do worse.” Since there has never been a paper published in the world's literature on laryngeal cancer that included more than approximately 600 patients, we were placed into an obvious “catch-22” situation. The reviewers believed that the burden of proof should lay on our shoulders and that, until proven otherwise, race should be considered a significant independent determinant of outcome. But is this true?

The answer to this question has a number of implications for the delivery of health care. First, the recognition of genetic differences could allow specific populations to be targeted for the delivery of certain types

of health care. After accepting such occurrences as “fact,” interventions could be designed to meet the unique needs of these populations. Costeffective guidelines could be developed for prevention, early detection, and treatment specifically for these populations. Furthermore, social resources might be allocated to support basic research devoted to defining the genetic defects and mechanisms resulting in a worse state of health. Conversely, resources would not be allocated to support basic research devoted to identifying the genetic defects if there was no “genetic defect” to detect. Instead, “nongenetic” causes such as diet, lifestyle, environment, or lack of access to health care might need to be addressed. But do we really want to know?


Before discussing data assessing the merits of genetic and nongenetic causes, there are several questions that should be answered. First, since race has “been around forever,” why do these questions still persist? Is this due to lack of data, or could it be that “we” really do not want to know? Stephen J. Gould recognized this issue and began his book The Mismeasure of Man with the following quote from Charles Darwin's Voyage of the Beagle:9

If the misery of our poor be caused not by the laws of nature, but by our institutions, great is our sin.

If the basis for the excess mortality among certain racial groups is an intrinsic characteristic of the group, some might consider this a sign of “racial inferiority.” Although such a possibility would say nothing about the moral, creative, humanitarian, or other more important features of an individual, members of such a racial group are still likely to be defensive. Conversely, if the excess mortality rate is entirely due to various types of social injustices (such as racism and discrimination, resulting in lack of education, underemployment, and poor access to care, resulting in a fatalistic self-destructive lifestyle), the moral and financial implications would be staggering.

To understand the development of the notion of some sort of inherent tendency for Blacks to be “genetically less healthy,” it is best to assess the sources of this belief. Therefore, it is important to look back at the history of events in the history of this country that might have had

an impact on the health status of and beliefs about African Americans. The financial implications (liabilities) should also be most obvious if viewed from this context.


Just as the issue of “racism in America” is usually perceived as largely a “Black versus White” issue, the issue of “race and health” is often seen in a similar context. For example, annual reports sponsored by the federal government compare outcomes between Blacks and Whites and routinely ignore other groups.5 It is a simple matter to find 20 to 30 publications comparing outcomes between Black women and White women with breast cancer, but similar studies for other groups are lacking. This reality may result from the several facts. First, Hispanics are generally considered as an ethnic group, not as a race. When Hispanics are placed into one of the major three groups (Blacks, Whites, and Asians), they are for the most part considered “White.” Second, for Asians, including Pacific Islanders, and Native American Indians, the details surrounding the impact of racism on their health is both complex and heterogeneous. This is not to suggest that the health issues for these groups are any less important but rather that (1) they are not as well documented, (2) for most health outcome end points (e.g., death due to cancer) the differences are not as large and in some cases favor the Asian populations,

(3) differences in social status have not been as clearly enforced by the laws of the land, and (4) over the last 400 years fewer individuals belonging to this “racial group” have been impacted by racism.

The circumstances surrounding the arrival of African people to this country are likely to explain some of the problems in health status seen today. Between 1501 and 1870, it has been estimated that between 9.5 and 14.6 million African people were brought to America in bondage.10 Furthermore, it is believed that nearly as many African people died, resisting capture, via suicide, or en route, due to hardships. The mortality rates at sea alone have ranged from as high as 33% to as low 12%.11 It is obvious that the first generation of African Americans had a very short life expectancy. Harley has summarized selected historical events reflecting social events of note in the history of African Americans, and some of these are listed in Table 10.1.12 Harsh punishments and intolerance were the rule, and it was more than 250 years before the first Black man (who was a slave at the time) was licensed as a physician. This being the

Years Event Comments
1492 Pedro Alonzo Nino, a navigator
of the Santa Maria, arrives
with Christopher Columbus.
1502 Portugal lands its first cargo of
enslaved Africans in the Western
to 1870
Slavery delivered between 9.5
and 14.6 million African people
to the Americas, and nearly as
many are thought to have died,
resisting capture, via suicide, or
en route due to hardships.
After 369 years of slavery, and
other acts of violence, what
reparations would be due these
people and their offspring?
1692 Virginia enacts law making it
lawful to kill a runaway slave
in the course of apprehension.
1693 Philadelphia: Law permitting
Whites to “take up” any Black
found without a pass.
1762 James Derham becomes the
first Black man licensed to
practice medicine in the United
States and 21 years later purchases
his freedom.
1809 New York law sanctions
marriage within the Black
Married Blacks were not legally
recognized as such before
this law.
1810 19% of the U.S. population is
Black, but only 9% are free.
1863 The Emancipation Proclamation
goes into effect, freeing
slaves held in states in rebellion
against the Union, but not in
portions of Louisiana, Eastern
Virginia, West Virginia, or border
states (3).
Although slaves were freed,
they were not able to vote and
continued to be systematically
1866 The first Civil Rights Act is
passed over President Andrew
Johnson's veto, declaring
Blacks free and nullifying
“Black codes.”
“Black codes” restrict the
rights of freedmen/women.
1868 14th Amendment is passed,
granting Blacks “full citizenship
and equal rights.”
1890 U.S. Supreme Court allows
states to segregate public facilities
and control of elections.
1896 “Separate but equal” facilities
ruled constitutional.

Louisiana, Georgia, North
Carolina, Virginia, Alabama,
and Oklahoma adopt the
“grandfather clause.”
Males could vote only if their
fathers or grandfathers were
eligible to vote.
1932 Tuskegee experiments begun.  
1940 U.S. Congress passes Selective
Training and Service Act.
Includes an antidiscrimination
clause and a 10%
quota system to ensure racial
1957 Civil Rights Act of 1957 passed,
authorizing the federal government
to bring civil suits on the
behalf of citizens.
First Civil Rights Act since
1964 U.S. Congress passes the Civil
Rights Act and establishes the
Equal Opportunity Commission
This law was passed to offset
hundreds of years of systematic
discrimination and racism,
but it failed.
1971 National Cancer Act and SEER
program established.
SEER created to collect, analyze,
and disseminate data.
1994 Black-White Breast Cancer
Study: Race not an independent
prognostic factor.
Having a high poverty index,
lack of insurance, and increased
body mass index, and
being divorced, separated, or
never married associated with
a poor outcome.
1994 RTOG 9202 demonstrates that
Blacks have more advanced
prostate cancer.
More advanced prostate cancer
by virtue of higher PSAs
not in the clinical stage.
1995 Proposition 209 goes into effect
in California, banning Affirmative
Action to compensate
for past discrimination.
How many years of affirmative
action compensate for 369
years of slavery and many
years of systematic discrimination
and oppression?
1996 RTOG 9412 and the A2 demographic
Both demonstrate that Blacks
continue to have lower incomes,
less education, and
more advanced disease.
1997 CALGB 8541, based on a prospective
randomized trial including
1,500 women; race
not an independent factor.
Black women continue to have
an excess mortality rate from
breast cancer.
1997 CA Journal published demonstrating
59% five-year survival
for Whites versus 44% for
This 1.4-times greater risk of
cancer death is the largest recorded
since 1960.
1997 Courts rule against “set-aside
programs” in Philadelphia for
city works programs.
Similar programs struck
down in Columbus, Ohio,
and Miami.

case, it should not be surprising to find that individuals who were denied participation in traditional medicine would have poor health as assessed by this traditional medicine.

Of note, although the Emancipation Proclamation was passed in 1863 freeing some slaves, it did not apply in portions of Louisiana, Eastern Virginia, West Virginia, or border states. Moreover, freed men and women continued to be systematically oppressed. Consequently, three years after the Emancipation Proclamation, “Black codes” (which systematically restricted the rights of freed men and women) were passed over the veto of the U.S. president.

Two years later the 14th Amendment granted African Americans full citizenship and equal rights. However, 30 years later “separate but equal” was ruled constitutional and “grandfather clauses” (allowing males to vote only if one's father or grandfather voted) were upheld in a number of states (1898–1910). Finally, after 369 years of slavery and 87 years of systematic discrimination, the Civil Rights Act of 1957 was passed. Later the U.S. Congress passed the Civil Rights Act of 1964.

In 1971 the National Cancer Act establishing the SEER (Surveillance Epidemiology End Result) program was created to collect, analyze, and disseminate data useful in the diagnosis and treatment of cancer.1 These SEER data are published annually and are considered to be the “gold standard” by physicians throughout this country. From 1973 to 1990 information on approximately 1.6 million cases has been collected. Approximately 9.6 percent of the population of the United States is included in the geographic areas making up the database for the SEER program.1 Since the natural histories of treated and untreated cancers of various types have been well studied, this disease will be considered in some detail to assess the prognostic significance of race.


A close look at the primary cancer sites for which differences between White and Black Americans are most apparent is required to identify the causes for the survival discrepancies. Five-year survivals based on SEER data for all cancer sites among Blacks and Whites is shown in Figure 10.1. Although the five-year relative survival rate is 54.5% for White patients, it is only 39.4% for Blacks.1 These data suggest that if you are Black and diagnosed as having cancer, your risk of dying from cancer within five years is 50% higher than for Whites. Equally alarming is the fact that the percentage change in the mortality from 1973 to 1990

increased 16% for Blacks compared to 6% for Whites. These data underscore the magnitude of the cancer health care crisis for African Americans in this country. A similar comparison for cancer outcome differences for two of the most common cancer sites in men and women is discussed in the following.

Figure 10.2 compares the outcome for Black and White men with carcinoma of the prostate. Only 64.4% of Black men diagnosed with prostate cancer were alive at five years compared to 79.4% of White men. This marked difference in survival appears to be a continuing trend with a greater increase in the mortality in Blacks compared to Whites from 1973 to 1990. Of further interest, a much greater increase in the percentage change in incidence was noted among Whites. This trend probably reflects the more frequent use of the serum marker PSA (prostatespecific antigen) to detect otherwise occult disease in this population. In other words, although the risk of prostate cancer is lower among White men than among Black men, more of the former are systematically being screened.

Figure 10.3 compares the outcome for breast cancer by race. Again a lower survival is noted for Black women compared to White women, with 64.2% and 80.5%, respectively, alive at five years. A 21.4% increase in the percentage change in mortality was noted for Blacks compared to 2% for Whites during this same time period (1973–1990).

The tendency for Black Americans to present with more advanced disease is one of the common explanations offered for these difference in survival.13 However, even after correcting for the stage of disease, many studies still report an excess mortality rate among Blacks.1, 13–15 For selected sites, differences in socioeconomic status (SES) have also been proposed as an explanation for differences in survival.14,16,17 However, some studies failed to demonstrate an effect due to SES when the quality of care was comparable.18,19 Furthermore, a biologic mechanism explaining how SES affects outcome is lacking. The possibility that differences in cancer-related mortality might be due to factors such as the quality of the medical care received has not been adequately evaluated. Several studies document differences in initial treatment, patterns of care, the intensity of services provided, as well as a tendency for racial bias in the inpatient setting.20–24 These last two observations support the notion that lifestyle and nongenetic factors may be the overwhelming determinant of the health status for most people. The fact that recent studies continue to demonstrate changes in mortality in both races as a reflection of lifestyle changes provides additional support for the truism that “you are what


Figure 10.1. SEER race data, all sites.


Figure 10.2. SEER race data, prostate cancer.


Figure 10.3. SEER race data, breast cancer.

you do” (eat, drink, smoke, exercise), while our genetic makeup is trying to help us survive.


It is clear that Blacks have a lower survival than Whites for a number of common cancers. In response to these kinds of data, in late 1993 the National Cancer Institute (NCI) mandated that cooperative groups conducting large prospective randomized trials involving cancer treatment must include in their study design mechanisms to address the issue of race and cancer outcome (if the published literature suggests that race or gender might affect outcome).25 This mandate represents a major health care policy decision that could potentially impact the design and implementation of most randomized trials conducted in the United States because for most of the common cancer sites there is a discrepancy in survival between Blacks and Whites.

In addition to explicitly affecting cancer research protocol design policies, the published literature supporting the intrinsic importance of

Conclusions Results
Cancer survival differences exist for
Blacks and Whites.
Race is assumed to be an independent
prognostic factor. More data
are generated to prove that differences
The burden of proof to the contrary
rests on others attempting to disprove.
Biologic basis is assumed for the
observed differences in survival
by race.
This “biologic” phenomenon should
be studied further.
Cancer researchers are funded to
study the “racial biology.”
There is nothing you can do about a
person's race but …
… perhaps understanding the molecular
basis of cancer wilt be directly
helpful (to researchers).
race has tainted the beliefs of many epidemiologists and health care providers. The current belief construct implied by much of the published literature on race and survival from cancer is shown in Table 10.2. Race is attributed independent prognostic significance for survival from cancer. This belief construct is supported by numerous publications demonstrating differences in cancer survival by race, without data to explain the differences. Otherwise, why have data not been repeatedly published comparing the survival of janitors and physicians or of policemen and preachers? The mind-set of expecting to see racial differences frames the comparisons that are made.

This mind-set is contrary to what should be acknowledged as our social norm. Consistent with our social norms opposing racial stereotypes, race should be implicated only as a diagnoses by exclusion. In other words, only after other plausible explanations have been ruled out should race be implicated. If the survival differences noted were attributed to differences in the stage at presentation or to the quality of cancer care received, race should not be the focus of investigation.


An alternative belief construct to explain the observed differences in survival by race is summarized in Table 10.3. The implications for this alternative

Beliefs Implications
Cancer survival differences between
Blacks and Whites can be explained
by differences in the extent of disease
at diagnosis and quality of care.
Race is not an independent prognostic
The burden of proof to the contrary
rests on others attempting to prove
that race is a prognostic factor.
“Racism” is not perpetuated by
Differences in outcomes should be
studied further to identify causes.
Research should focus on differences
in the knowledge, behavior, and
environment as well as on access
and the quality of care.
There is something that you can do
about lack of jobs, education, environmental
factors, and quality of
health care. Lack of health is a symptom
of social diseases.
Intervention is directly helpful to
the people being studied and experiencing
the excess mortality. Social
changes improve health.
belief construct for future research funding and patient care are also included in this table. This belief construct hinges on the notion that the observed differences in survival can be explained by factors other than race.

An epidemiologic phenomenon that I have chosen to call “extent of disease bias” (EDB) may explain much of the reported survival differences. Staging systems represent somewhat arbitrary ways to separate patients in prognostic groups for the purposes of comparison.26,27 Historically, these systems usually were primarily based on the size of the tumor as determined by palpation and usually referred to as the “T” stage, with categories 1 through 4.27 These systems typically also depend on the extent of lymph node involvement, defined by size, location, and number.27–30 The break points for these staging systems typically reflected whether it was believed that a tumor could be completely resected.

Extent of disease bias results from two major types of shortcomings of the current staging systems. First, the currently used staging systems are overly crude in the degree of absolute separation or “the degree of fineness of separation” of distinct prognostic groups. Second, the current staging systems are designed primarily to answer a question of the relative probability of cure rather than the duration of survival among those patients who are not curable. This shortcoming reflects “the disproportionate predictive priorities” of the current staging systems.



Figures 10.4 to 10.6 compare two hypothetical populations with differences in the distribution of the extent of disease at presentation. Population A is composed of a cohort of relatively well educated individuals, of higher socioeconomic status, who tend to present with earlier-stage disease. Cohort B is composed of individuals of lower socioeconomic status, with less education and a high percentage being uninsured, who tend to present with more advanced disease. In this model it is assumed that for each population of cancer patients there is a “bell-shaped” distribution of disease-specific survivals that correlates with the extent of disease. This disease-specific survival distribution is a direct reflection of, and consequently is proportional to, the extent of disease. The term “extent of disease” as used in this model takes into account the volume of the tumor and the degree to which the tumor has spread. Extent of disease also includes the fact that with time, biologic changes tend to occur in the aggressiveness of tumors, which may not be manifested as differences in the tumor volume or extent of spread.

Figure 10.4 compares the distribution of cancer in populations A and B using a relatively “crude” staging system, with early, intermediate, and advanced disease corresponding to stages 1, 2, and 3, respectively. Of note, a larger percentage of individuals in population B have stage 3 disease and fewer have stage 1, while a similar percentage of members from both populations have stage 2 disease. It would seem appropriate to most observers that a survival comparison of patients with stage 2 disease from populations A and B would be valid. This may not be the case, however.

Figure 10.5 compares the same groups using a “superstaging” system composed of 12 stage categories instead of three. Figure 10.6 is a “magnified view” of the intermediate extent of disease group. This intermediate extent of disease group corresponds to superstages 5 to 8 and the “crude” stage 2 group. Using the superstaging system it is now apparent that more group B members belonging to the intermediate extent of disease group have superstage 8 and that more group A members have superstage 5. Because of these differences, group B members would be expected to do worse than group A members. An analysis using crude staging would suggest that group membership had independent prognostic significance. In contrast, an analysis using the superstaging system is likely to demonstrate that group membership has no prognostic significance independent of the true extent of disease (see Figures 10.5 and 10.6).


Figure 10.4. Extent of Disease Bias Model: Example of distribution differences in two populations of patients with cancer.


Figure 10.5. Extent of Disease Bias Model: Comparing distribution differences in two populations with cancer using superstaging.


Figure 10.6. Extent of Disease Bias Model Example: Superstages 5 to 8 versus crude stage 2.


Examples taken from published literature demonstrating the potential for EDB for Blacks and Whites with prostate and breast cancer are shown in Figures 10.7 and 10.8. Note that in both the examples shown the distribution of disease is similar to that of populations A and B in the EDB crude staging model. The extent of disease in cancer patients is not naturally divided into three or four distinct groups. Rather, there is a continuum. The less precisely populations are defined, the more likely patients with extensive disease will be “lumped” with patients with less extensive disease. Therefore, differences in the distribution characteristics in two populations can confound an analysis of outcome.

Examples of the Potential for EDB from
(Non-SEER-Based) Cancer Treatment Literature

In addition to SEER data, numerous other examples of distribution differences between Blacks and Whites exist in the medical literature.13,14,26 The same pattern is seen in all these studies: The distribution of cancer in Blacks is “right shifted.” The term “right shifted” refers to the relative displacement of the overall shape of the extent of disease distribution toward more advanced stages for Blacks compared to less advanced disease for Whites. The “cruder” the staging system used, the more biased the interpretation is likely to be.


Figure 10.7. SEER prostate cancer: Stage distribution data, 1983 to 1987.


Figure 10.8 SEER breast cancer: Stage distribution data, 1990.


Examples of how the degree of fineness of separation within a given stage can affect prognosis are available throughout the cancer literature. Rosen et al., for example, reported that within the category of patients with T1 NoMobreast carcinoma (stage I), significant differences in prognosis exist based on the size of the primary tumor.28 For example, patients with a primary breast tumor less than 1.0 centimeter had an 83% relapse free survival (RFS) at 10 years compared to a 73% RFS for patients with primaries 1.1 to 2.0 centimeters. This represents an excess relapse risk of 59% at 10 years. Differences of this magnitude are common in the series, which report breast cancer survival differences between Blacks and Whites. If the distribution of the extent of breast cancer was such that a larger percentage of Black women had lesions in the 1.1- to 2.0-centimeter range and a larger percentage of Whites had lesions in the range less than or equal to 1.0 centimeter, it is possible that despite “correcting for stage,” an analysis might suggest that there were survival differences due to race.


The current staging systems are designed primarily to predict surgical curability.27–30 Patients who are not curable tend to be lumped into a category that is composed of a heterogeneous group of individuals with a poor long-term prognosis. Although such patients uniformly have a poor long-term prognosis, there can be significant differences in shortterm survival. These differences can significantly affect the overall survival curve because a large number of patients fall into this category. Examples can be found in the literature where the differences in survival for patients belonging to the same stage may vary from less than 6 weeks to 6 months to a few years.29,30 Thus, the current staging systems have disproportionate predictive priorities in the sense that with an increase in the extent of disease, the staging system becomes progressively less discriminating in terms of identifying various prognostic subgroups.

An example of these phenomena is the ability of the current staging systems used to predict the outcome for the average patients with lung cancer. Although most patients who present with lung cancer do not have curable disease (overall five-year survival is about 10%), the staging system is biased in favor of subcategories of patients who are considered potentially surgically resectable.30 The most widely used staging system is characterized by having four out of six of the major substaging categories dedicated to the small percentage of patients who are considered

to be candidates for surgical resection with curative intent. Since only approximately 10% are alive at five years, it is obvious that this staging system is not designed to be of prognostic value for the subgroups to which the typical patient belongs. For example, Albain et al. demonstrated that among patients with metastatic lung cancer (all the same stage), the survival is noted to have been highly variable.29 In this retrospective analysis a subgroup of patients could be identified with a one-year survival of 27%, while another subgroup could be identified with an expected one-year survival of only 7%. Even larger differences in the range of survivals have been noted among subgroups of patients presenting with certain other types of cancer, such as metastatic prostate and breast cancer. Since these are the two most common primaries occurring in men and women, we will use data for these sites to demonstrate the existence of EDB. Awareness of EDB may allow biases that might result from it to be minimized in future epidemiologic studies.


Although SEER data and numerous studies report a lower survival for Black women with breast cancer, several recent studies strongly suggest that race is not an independent prognostic factor.13,14,16,18,26,31–42 A detailed analysis of this issue was reported by Roach and Alexander.26 These authors reviewed the available published literature comparing the survival of Blacks to Whites with breast cancer treated from 1968 to 1988. The five-year survival for Blacks was less than or equal to that of Whites in all studies (difference range about 0 to 19%). However, in 11 of the 18 studies (61%), the reported differences in five-year survival between Blacks and Whites was less than or equal to 3%. We generated a “reliability scoring scale” based primarily on the level of staging sophistication used (crude staging resulted in lower scores) and the likelihood that the quality of care was comparable. Next the relationship between the reliability score and survival was analyzed. Those studies reporting large differences in five-year survival generally had low reliability scores, while those with high scores tended to have small differences in survival.


Using patients treated on phase III prospective randomized trials avoids potential bias due to differences in the initial staging workup and treatment;

thus, assessments of outcome by race are likely to be more accurate.40–43 Such trials involve recruiting patients with a specific type and stage of cancer to be randomly assigned to receive either the best conventional treatment or a “newer” type of treatment that is believed to possibly be better. Such trials ensure well-defined guidelines for staging, eligibility, and uniformity in treatment and follow-up. These trials are also designed to stratify or balance patients for other factors that could confound the outcome of the study. Three cooperative groups conducting large phase III randomized trials have recently completed analysis of breast cancer treatment outcome by race. The Southwest Oncology Group (SWOG), the Cancer and Leukemia Group B (CALGB), and the National Surgical Adjuvant Breast Program (NSABP) all found that race was not a significant independent prognostic factor for survival from breast cancer.41–43 Thus, based on the data from the major cooperative groups doing prospective randomized trials, race has been shown not to be a significant independent prognostic factor. These data must be viewed as the most accurate data available on which to judge the significance of race and survival from breast cancer.

The findings of these studies are consistent with the EDB model. However, there are some differences in the presentation of breast cancer in Black and White women that remain unexplained. For example, it is well known that breast cancer is less common in Black women, and some studies suggest that there is an earlier age of onset.44–46 These differences may be due to the average age of first pregnancy, diet, intrinsic genetic differences, or other factors yet to be defined. Table 10.4 summarizes the findings reached following a prospective study of Black and White women with breast cancer that demonstrated that poverty, martial status, health insurance status, and body mass but not race correlated with death due to breast cancer.46 Failure to account for differences in these areas may explain at least a portion of the impression reached by some researchers that race was likely to be an independent prognostic factor.


A number of retrospective studies have been published addressing the prognostic significance of race and survival from prostate cancer.12,13,15,17,18,23,47–53 Austin et al. conducted a retrospective study including 914 patients (867 Whites, 47 Blacks) treated with radiation

Factor Relative Risk[a]

aUndajusted hazard ratios (95% confidance intervals), modified from Eley et al. (1994 [46]).

Poverty index > 400 (high income) 0.6 × (0.4-0.7)
Divorced, separated 1.6 × (1.2-2.2)
Never married 1.6 × (1.0-2.5)
No health insurance 2.3 × (1.5-3.5)
High body mass index (“overweight”) 2.2 × (1.5-3.2)
therapy.15 The data were obtained from the Connecticut SEER Tumor Registry and included patients with locally advanced disease (stage T2–T4) who were treated from 1973 to 1987. According to these authors, after multivariate analysis controlling for stage and grade, Blacks had a lower survival. Natarajan et al. also noted a lower overall survival for Blacks with prostate cancer.49 They reported no differences for patients with early stage (T1) prostate cancer and less than a 10% difference in the five-year survival for patients with stages T2 and T4 prostate cancer but larger differences for patients with T3 lesions. Neither the cause specific survival, the comparability of the quality of care, nor an adjustment for the distribution of other known prognostic factors is possible in these two studies. Perez et al. reported similar findings for patients with more advanced disease.50 The small number of Blacks included in two of these studies and the mixed nature of the results in the larger study make it difficult to draw definitive conclusions from these studies.

A number of other investigators have reported race to be of no independent significance after adjustment for other prognostic factors.51–53 In contrast to the series reporting a difference in survival as a function of race, either all these series were likely to have provided a similar quality of care (single institution study or as part of a randomized trial) or an adjustment was made for socioeconomic status (perhaps a predictor for quality care). This is important because of studies showing that Blacks tend to be treated less aggressively.18


The largest prospective database available for assessing the outcome following definitive radiotherapy for clinically localized prostate cancer is possessed by the Radiation Therapy Oncology Group (RTOG). Using

this database Roach et al. evaluated the prognostic significance of race among 1,294 patients who participated in three prospective randomized trials conducted between 1976 and 1985.54 One hundred and twenty (9%) of the patients were coded as Black, while 1,077 (83%) of the patients were coded as White. Two of the three studies (RTOG 7706 and 8307) revealed that race was not of prognostic significance for diseasefree or overall survival. Protocol 7506 revealed that the median survival for Blacks was somewhat shorter (5.4 years vs. 7.1 years, p 0.02). But a higher percentage of Blacks treated on 7506 had an abnormally elevated serum acid phosphatase compared to Whites (p 0.006), and the time to distant failure (primarily spread to bone) tended to be shorter (p 0.07). These findings suggest that Blacks treated on 7506 may have had more extensive disease at presentation. Based on these prospective randomized trials, it is most likely that the lower survival noted for Black Americans with prostate cancer reflects the tendency for Blacks to present with more advanced disease.

Based on the available literature, the preponderance of evidence suggests that after adjustment for the extent of disease, when treatment is comparable, there are no differences in survival from early stage prostate cancer that can be attributed to race.51–54 For more advanced disease, it is likely that the survival differences reflect differences in the distribution of disease at the time of diagnoses that are not accounted for in the staging system used.54,55 Since many of the patients treated for prostate cancer are staged clinically and not pathologically, accurately determining the true extent of disease in prostate cancer patients is more difficult than for breast cancer. These inaccuracies in staging, and the differences in the disease distribution in these two populations, result in EDB. Additionally, there are a multitude of socioeconomic differences that might explain differences in outcome.56 For example, in a large demographic study we noted clear differences in income status and the educational level of Blacks and Whites treated on phase III randomized trials. Clearly, having less support and a poorer understanding of your disease is not likely to be beneficial.


In addition to breast and prostate cancer studies, data from prospective randomized trials and retrospective reviews of a number of other cancer sites fail to support the independent prognostic significance of race.57–61

For example, race was not an independent prognostic factor for survival among patients who participated in lung, brain, or esophageal cancer trials conducted by the RTOG.57–59 The Southwest Oncology Group evaluated the prognostic significance of race and survival from multiple myeloma, a hematologic malignancy that is known to occur more frequently in Blacks (two to one, compared to Whites) and to be associated with a higher mortality rate in Blacks.60 This study included 614 patients who were treated on a prospective randomized trial comparing two different chemotherapy regimens. There was no difference in survival by race, and the authors concluded that “the observed differences in mortality between blacks and whites cannot be attributed to differences in survival … given comparable care.”

A large study of childhood cancers reported by investigators from the St. Jude Children's Research Hospital demonstrated rather conclusively that with “equal access to effective contemporary” care, Black children have the same outcomes as White children. This study included more than 5,000 Black and White children treated for cancer between January 1962 and June 1992. These investigators noted that in the early years, Blacks had a lower survival rate largely due to more advanced disease at the time of diagnosis, but in more recent years there was no difference in outcome by race.61 Thus, as with adults, the preponderance of evidence suggests that race is not an independent prognostic factor for survival from childhood cancer.


Following an extensive review of the available published data, race does not appear to be a major independent prognostic factor for survival from cancer. Instead, an epidemiologic phenomenon, EDB, may explain all or much of the apparent discrepancy in survival. This phenomenon is due to two related factors. First, the available staging systems are crude measures of the true extent or severity of disease. These staging systems are designed primarily to predict whether patients are curable and not specifically to predict the duration of survival. This limitation is most obvious when applied to patients with advanced cancer. Second, differences in the extent of disease distribution in different populations create bias when the groups are compared. This bias persists despite crudely correcting for stage. The EDB tends to create the general impression that race is an independent prognostic factor for survival from cancer. Careful analysis of the available epidemiologic studies

fails to support this assertion when EDB is considered. In other words, race does not appear to be a “real biologic factor” for predicting survival from cancer. There are a number of implications resulting from this conclusion.

The assumption that race is an independent factor (without an adequate scientific basis) technically can be considered as “racist,” just as assuming that Black children score lower on standardized tests because of race is racist. The assumption that race is an independent prognostic factor deprives the “victims” of the opportunity to rectify the real health care problems. If the major problem is actually quality of care, efforts should be directed there. Differences in socioeconomic status and the variations in practice patterns and health outcomes by region of the country create enough doubt about the importance of race to require that any study that proposes to demonstrate differences in outcome by race be reviewed very critically for several reasons.7,20,22–24 First, the bulk of the literature suggests that such a study is likely to be flawed even if the source of the flaw is unclear. Second, thus far, the designation of race as a major health care variable has not resulted in the improvement of care for anyone. Those who have benefited the most from such practices have been health care researchers funded to do research on racial differences. Finally, nothing can be done about an individual's race.

If socioeconomic factors (lack of education, unemployment, and poor-quality health care) are the most important factors, money currently spent studying cancer outcome as a function of race might be better spent addressing these issues.62–64 Studies may also be considered to address the possibility of higher levels of environmental carcinogens in Black communities. Such studies would force us to study the impact of racism on health care delivery as a social disease. The political implications arising from these issues may not be as popular as acknowledging that “Blacks do worse,” as has been done for more than 20 years. Awareness of EDB may allow a more accurate assessment to be made of the impact of lack of education, environmental exposures, socioeconomic status, and other factors, such as diet, that may be associated with differences in survival from cancer.64


It is clear from the previous discussion that the preponderance of evidence suggests that race is not likely to be an independent prognostic

Journal Topic Conclusions
Fang et al.,
1996 (65)
New England
Journal of
Study of the relationship
the place of birth
and the cardiovascular
mortality among
non-Hispanic Black
and White residents
of New York City
Although Blacks
born in the South
had substantially
higher age-adjusted
rates of death from
causes, both
Caribbean- and
Blacks ≥65 years
had lower rates.
and Sidney,
1996 (66)
Journal of
Public Health
Association between
blood pressure and
self-reported experiences
of racial discrimination
and responses
to unfair
Black-White differences
in blood pressure
were reduced
by accounting for
reported experiences
of racial discrimination
and responses
to unfair
Ayanian et al.,
1993 (67)
Journal of
the American
Likelihood of coronary
among Medicare
Part A enrollees
Whites more likely
to receive such
Peterson et al.,
1994 (68)
Journal of
the American
Assessment of cardiac
or revascularization
procedures among
Blacks and Whites
admitted with an
acute myocardial
infarction to Veterans
Blacks 33%, 42%,
and 54% less likely
to undergo cardiac
undergo coronary
angioplasty, and
receive coronary
bypass surgery,
respectively, than
Kahn et al.,
1994 (69)
Journal of
the American
Comparison of hospital
care among elderly
Black or poor
Medicare recipients
using a representative
sample from
9,932 patients
Patients who are
Black or poor have
worse processes of
care and greater
instability at


Geronimus et al.,
1996 (70)
New England
Journal of
Study of the excess
mortality among
Blacks and Whites
under the age of 65.
The probability
of reaching age 65
was only 62% for
Blacks compared
to 77% for Whites,
but there were large
variations nationally
such that
Blacks from the
Queens-Bronx had
a higher likelihood
of reaching age 65
than Whites from
the lower east side,
Detroit, the Appalachians,
and northeastern
factor for survival from cancer and that the differences in the extent of disease at diagnosis explain much or all of the apparent survival differences. What about other major causes of death? As stated earlier, even in the early 1900s it was clear that nongenetic factors dominated as causes of excess mortality in this country. What about more recent observations? Table 10.5 summarizes a selected series of recent articles published in the most prestigious journals in this country.65–70 These articles seem to convey the same message: Race is not a major independent determinant of excess mortality.

What about cardiovascular diseases, the number one cause of death? Fang et al. reported that the risk of death from cardiovascular disease appears to depend more on place of birth than on race.65 But how does one explain the differences seen in the incidence and severity of disease? The study by Krieger et al. may shed light on this matter.66 They reported that experiencing racial discrimination was associated with having a higher blood pressure. Several studies suggest that Whites tend to be treated more aggressively than Blacks with the same severity of disease.67 Recent studies have also shown that race was shown to influence the quality of hospital care received for cardiovascular illness regardless of whether patients were treated covered by Medicare or were treated within the Veterans Administration system.68,69 Still other studies demonstrate

that, when assessed by all-cause mortality, regional variations throughout the United States have been noted to exceed the variations between races.70 These data suggest that it is likely that “nongenetic causes” explain most or all of the observed differences in the health status of the various races and that Blacks and the poor in general receive inferior care in this country.65–72


Herein data have been presented that support the notion that an epidemiologic phenomenon, EDB, is likely to explain the lower survival rates noted for some “races” with cancer of several common sites. Race also does not appear to be an independent prognostic factor for survival from cardiovascular disease or overall mortality when adjustments are made for other confounding variables. It should not be assumed that biologic differences exist between people of different races unless there are very strong data to support this assumption. This conclusion provides a basis for designing strategies for addressing the problem of excess mortality seen in Blacks.

In addition to the fact that the overview of available medical data really does not support the notion that race is an independent prognostic factor, there is another major reason to doubt the notion that race has major independent significance. “Race” is not a true scientific biologic construct but rather a political construct for the purposes of conveniently dividing people. Recent review articles published in mainstream journals leave little to no doubt of this fact.73–77 For example, Paul Hoffman, in an editorial written for Discover magazine, summarized his findings of a series of review articles on race by several authorities: 73–76

On average there is a 0.2 percentage difference in the genetic material between any two randomly chosen people on Earth. Of diversity, 85 percent will be found within any local group of people—say, between you and your neighbor. More than half (9 percent) of the remaining 15 percent will be represented by differences between ethnic and linguistic groups within a given race (for example, between Italians and French). Only 6 percent represents differences between races (for example, between Europeans and Asians). And remember—that's 6 percent of 0.2 percent. In other words, race accounts for a minuscule 0.012 percent difference in our genetic material.

With such a small portion of our genetic makeup, why should anyone expect race to determine one's risk of dying of breast cancer, prostate cancer, hypertension, gunshot wounds, having a lower I.Q., and

being unemployed? Clearly there are genetic differences, such as the incidence of sickle cell anemia in different populations. However, this mutation was acquired to provide resistance to malaria and, as is typical of similar mutations, was selected for a survival advantage early enough in life so as to increase one's chances of procreation. None of the major common illness discussed previously are likely to be affected by natural selection because they occur too late in life to provide any advantages during natural selection. The bottom line appears to have been well put by Sharon Begley in Newsweek magazine when she wrote, referring to the conclusions of the Human Genome Diversity Project, that

genetic variation from one individual to another of the same “race” swamps the average differences between racial groupings. The more we learn about humankind's genetic differences, says geneticist Luca Cavalli-Sforza of Stanford University, who chairs the committee that directs the biodiversity project, the more we see that they have almost nothing to do with what we call race.77

Even today, research directed at continuing to prove the superiority of the “White race” continues to be funded.78 Although in this “free country” it may be reasonable to continue to allow privately funded racist research, it is inappropriate to continue to use taxpayer dollars to do so. Such research only serves to divide us as a nation. The mentality behind this type of race-based research was used to justify slavery some 400 years ago and the Tuskegee experiments more than 50 years ago.


The response of our society to its past moral and ethical failures to address the health needs of some of its people has been to first deny any responsibility for it. Blame the victims of race, and then do not compensate them—not only that, outlaw compensation and accuse them of “reverse discrimination.” Carl T. Rowan brought attention to this issue in his book The Coming Race War in America: A Wake-Up Call, when he wrote,

White male paranoia has become epidemic. This despite the fact that the median net worth of black households in this country is $4,604, or just one tenth the median net worth of white families—$44,408. The comparable figure for Hispanics is $5,345.79

The passage of Proposition 209 in California is but one example of how well-meaning people can be mislead into supporting an unjust cause that

will hinder compensation for past wrongs—this, despite, as Rowan also points out,

On talk shows and elsewhere I am frequently asked why “blacks get all the college scholarships.” The General Accounting Office reports that 96 percent of all the scholarship money in America goes to whites has done little to wipe out white cries of persecution.79

The passage of Proposition 209 in California is likely to do more to continue to widen the gap between “the haves and the have-nots” and consequently to do more to widen the gap between the state of health for Blacks and Whites in California. The matter of Proposition 209 brings to mind the letter written to the clergymen on April 16, 1963, from a Birmingham jail by Martin Luther King Jr.:

There are some instances when a law is just on its face and unjust in its application. For instance, I was arrested Friday on a charge of parading without a permit. Now there is nothing wrong with an ordinance which requires a permit for a parade, but when the ordinance is used to preserve segregation … it becomes unjust.80

Much of what is currently believed about race in this country grows out of the same mentality that supported segregation then and race-based research now. This mentality is largely responsible for the state of health of African Americans today. In what is considered his last, and most radical, Southern Christian Leadership Conference (SCLC) presidential address, Martin Luther King Jr. raised the question in the title of his presentation: “Where Do We Go from Here?” 81


I have argued that “race” is not inexplicably related to health but that racism and poverty are. The history of racial oppression of African Americans in this country is old, deep, and solidly entrenched in our literature, our science, and our culture.82 Racism and poverty have locked the vast majority of this population not in chains but in a complex superstructure whose roots reach back into the 1500s but whose branches still blossom and provide fruit of despair. This fruit of despair brings to mind the saying from the Tao Te Ching of Lao Tzu:

The best lock has no bolt, and no one can open it.
The best knot uses no rope, and no one can untie it.83


What can be more effective than “scientific proof” of racial differences as an explanation for being less healthy, less intelligent, and less hard working, even if this “scientific proof” is not valid?

When attempting to answer the question of where we go from here, Martin Luther King Jr. had this to say:

One night, a juror came to Jesus and he wanted to know what he could do to be saved. Jesus didn't get bogged down in the kind of isolated approach of what he shouldn't do. Jesus didn't say, “Now Nicodemus, you must stop lying.” He didn't say, “Nicodemus, you must stop cheating if you are doing that.” He didn't say, “Nicodemus, you must not commit adultery.” He didn't say, “Nicodemus, now you must stop drinking liquor if you are doing that excessively.” He said something altogether different, because Jesus realized something basic—that if a man will lie, he will steal. And if a man will steal, he will kill. So instead of just getting bogged down in one thing, Jesus looked at him and said, “Nicodemus, you must be born again.”

Here I would contend that similarly America needs to be “born again” with regard to its attitude toward race. We must end racist conjecture. Dr. King went on reflecting on the words of Jesus, saying,

He said, in other words, “Your whole structure must be changed.” A nation that will keep people in slavery for 244 years will “thingify” them—make them things. Therefore they will exploit them, and poor people generally, economically. And a nation that will exploit economically will have to have foreign investments and everything else, and will have to use its military might to protect them. All of these problems are tied together. What I am saying today is that we must go from this convention and say, “America, you must be born again!”

… let us go out with a “divine dissatisfaction.” Let us be dissatisfied until America will no longer have a high blood pressure of creeds and an anemia of city of wealth and comfort. … Let us be dissatisfied until those that live on the outskirts of hope are brought into the metropolis of daily security … and every family is living in a decent sanitary home.81

So, too, let us be dissatisfied as long as the risk of dying of cancer is 50% greater for Blacks compared to Whites. It is not due to race. We must end racism by first acknowledging this fact. Only then can we break the link between “race” and “health.”


The author would like to acknowledge Marion Malack and Pamalar Lewis for editorial support and Dr. Deborah Roach for her patience. A special thanks to

Martin Luther King Jr., whose words of wisdom inspired me to confront the issue of race.


1. The Random House College Dictionary. Revised ed. 1975. New York: Random House.

2. Roach, M., Alexander, M., and Coleman, J.1992. “The prognostic significance of race on survival from laryngeal carcinoma.” Journal of the National Medical Association84, 668–674.

3. Jones, James H.1981. Bad Blood: The Tuskegee Syphilis Experiment.New York: The Free Press, 34.

4. Nortzon, F. C., Komarov, Y. M., Ermakov, S. P., et al. 1998. “Causes of declining life expectancy in Russia.” Journal of the American Medical Association279, 793–800.

5. “National Cancer Institute.” 1993. SEER cancer statistics review: 1973–1990. In NIH Publication 93-2789. Bethesda, Md.: National Cancer Institute.

6. Jones, James H.1981. Bad Blood: The Tuskegee Syphilis Experiment, New York: The Free Press, 38.

7. Moul, J. W., Douglas, T. H., McCarthy, W. F., and McLeod, D. G.1996. “Black race is an adverse prognostic factor for prostate cancer recurrence following radical prostatectomy in an equal access health care setting.” Journal of Urology155, 1667–1673.

8. Bauer, J. J., Connelly, R. R., Sesterhenn, I. A., et al. 1997. “Biostatistical modeling using traditional variables and genetic biomarkers for predicting the risk of prostate carcinoma recurrence after radical prostatectomy.” Cancer79, 952–962.

9. Gould, Stephen J.1996. The Mismeasure of Man. 2nd ed. New York: W. W. Norton.

10. Franklin, J. H., and Moss, A. A.1994. From Slavery to Freedom: A History of African Americans. 7th ed. New York: McGraw-Hill, 41.

11. Curtin, P. D.1969. The Atlantic Slave Trade: A Census.Pp. 275–276. Madison: University of Wisconsin Press.

12. Harley, S.1995. “The timetables of African-American history.” In S. Harley, ed., A Chronology of the Most Important People and Events in African-American History.Pp. 226–275. New York: Simon and Schuster.

13. Satariano, W. A., Belle, S. H., and Swanson, G. M.1986. “The severity of breast cancer at diagnosis: A comparison of age and extent of disease in black and white women.” American Journal of Public Health76, 779–782.

14. Bassett, M. T., and Krieger, N. K.1986. “Social class and black-white differences in breast cancer survival.” American Journal of Public Health76, 1400–1403.

15. Austin, J.-P., Covery, K., and Rotman, M.1990. “Age-race interaction in prostate adenocarcinoma treated with external irradiation.” International Journal of Radiation Oncology Biology Physics19 (Suppl. 1), 200.


16. Dayal, H., Power, R., and Chui, C.1982. “Race and socio-economic status in survival from breast cancer.” Journal of Chronic Diseases27, 675–683.

17. Ernster, V. L., Selvin, S., Sacks, S. T., et al. 1978. “Prostatic cancer: Mortality and incidence rates by race and social class.” American Journal of Epidemiology107, 311–320.

18. Keirn, W., and Meter, G.1985. “Survival of cancer patients by economic status in a free care setting.” Cancer55, 1552–1555.

19. Page, W. F., and Kuntz, A. J.1980. “Racial and socioeconomic factors in cancer survival.” Cancer45, 1029–1040.

20. Ruffer, J. E., Barry, E. E., Terry, P., et al. 1991. “Lower radiation dose may account for decreased survival of blacks with prostate cancer: Results of the 1978 patterns of care study.” International Journal of Radiation Oncology Biology Physics21 (1), 212.

21. Diehr, P., Yergan, J., Chu, J., et al. 1989. “Treatment modality and quality differences for black and white breast cancer patients treated in community hospitals.” Medical Care27, 942–958.

22. Egbert, L. D., and Rothman, I. L.1977. “Relationship between race and economic status of patients and who performs their surgery.” New England Journal of Medicine297, 90–91.

23. Flaherty, J. A., and Meagher, R.1980. “Measuring racial bias in inpatient treatment.” American Journal of Psychiatry137, 679–682.

24. McWhorter, W. P., and Mayer, W. J.1987. “Black-white differences in type of initial breast cancer treatment and implications for survival.” American Journal of Public Health77, 1515–1517.

25. National Institutes of Health.1994. National Institutes of Health Guide for Grants and Contracts23 (11, March 18), 2.

26. Roach, M., and Alexander, M.1995. “The prognostic significance of race and survival from breast cancer.” Journal of the National Medical Association87, 214–219.

27. 1997. AJCC Cancer Staging Manual. 5th ed. New York: Lippincott-Raven.

28. Rosen, P. P., Groshen, S., Saigo, P. E., et al. 1989. “A long term follow-up study of survival in stage I (T1N0M0) and stage II (T1N1M0) breast carcinoma.” International Journal of Radiation Oncology Biology Physics7, 355–366.

29. Albain, K. S., Crowley, J. J., LeBlanc, M., et al. 1991. “Survival determinants in extensive-stage non-small-cell lung cancer: The Southwest Oncology Group experience.” Journal of Clinical Oncology9, 1618–1626.

30. Mountain, C. F.1986. “A new international staging system for lung cancer.” Chest89, 225S-233S.

31. Mittra, N. K., Rush, B. F., and Verner, E.1980. “A comparative study of breast cancer in the black and white populations of two inner-city hospitals.” Journal of Surgical Oncology15, 11–17.

32. Gregorio, D. I., Cummings, K. M., and Michalek, A.1983. “Delay, stage of disease, and survival among white and black women with breast cancer.” American Journal of Public Health73, 590–593.

33. Valanis, B., Wirman, J., and Hertzberg, V. S.1987. “Social and biological factors in relation to survival among black vs.” white women with breast cancer. Breast Cancer Research and Treatment9, 134–144.


34. Polednak, A. P.1988. “A comparison of survival of black and white female breast cancer cases in upstate New York.” Cancer Detection and Prevention11, 245–249.

35. Dansy, R. D., Hessel, P. A., Browde, S., et al. 1988. “Lack of a significant independent effect of race on survival in breast cancer.” Cancer61, 1908–1912.

36. Sutherland, C. M., and Mather, F. J.1988. “Charity hospital experience with long-term survival and prognostic factors in patients with breast cancer with localized or regional disease.” Annals of Surgery207, 569–580.

37. Fields, J. N., Kuske, R. R., Perez, C. A., et al. 1989. “Prognostic factors in inflammatory breast cancer.” Cancer63, 1232–1255.

38. Cella, D. F., Orav, E. J., Kornblith, A. B., et al. 1991. “Socioeconomic status and cancer survival.” Journal of Clinical Oncology9, 1500–1509.

39. Axtell, L. M., and Myers, M. H.1978. “Contrasts in survival of black and white cancer patients, 1960–1973.” Journal of the National Cancer Institute60, 1209–1215.

40. Farrow, D. C., and Hunt, S. J. M.1992. “Geographic variation in the treatment of localized breast cancer.” New England Journal of Medicine326, 1097–1101.

41. Albain, K. S., Green, S., LeBlanc, M., et al. 1992. “Proportional hazards and recursive partitioning and amalgamation analyses of Southwest Oncology node-positive adjuvant CMFVP breast cancer data base: A pilot study.” Breast Cancer Research and Treatment22, 273–284.

42. Roach, M., III, Cirrincione, C., Budman, D., et al. 1997. “Race and Survival from Breast Cancer: Based on Cancer and Leukemia Group B Trial 8541.” The Cancer Journal of Scientific American3, 107–112.

43. Dignam, J. J., Redmond, C., Fisher, B., et al. 1997. “Prognosis among African-American women and white women with lymph node negative breast carcinoma: Findings from two randomized and clinical trials of the National Surgical Adjuvant Breast and Bowel Project (NSABP).” Cancer80 (1), 80–90.

44. Krieger, N.1989. “Exposure, susceptibility, and breast cancer risk.” Breast Cancer and Treatment13, 205–223.

45. Morabia, A., and Wynder, E. L.1990. “Epidemiology and natural history of breast cancer.” Surgical Clinics of North America70, 739–752.

46. Eley, J. W., Hill, H. A., Chen, V. W., et al. 1994. “Racial differences in survival from breast cancer: Results of the National Cancer Institute Black/White Cancer Survival Study.” Journal of the American Medical Association272, 947–954.

47. Nattinger, A. B., Gottlieb, M. S., Verum, B. S., et al. 1992. “Geographic variation in the use of breast-conserving treatment for breast cancer.” New England Journal of Medicine326, 1102–1107.

48. Ragaz, J., Jackson, S. M., Plenderleith, I. H., et al. 1993. “Can adjuvant radiotherapy (XRT) improve the overall survival (OS) of breast cancer (BR CA) patients in the presence of adjuvant chemotherapy (CT)?” 10-year analysis of the British Columbia randomized trial. Proceedings of ASCO12, 70.

49. Natarajan, N., Murphy, G. P., and Mettlin, C.1989. “Prostate cancer in blacks: An update from the American College of Surgeons' patterns of care studies.” Journal of Surgical Oncology40, 232–236.


50. Perez, C. A., Garcia, D., Simpson, J. R., et al. 1989. “Factors influencing outcome of definitive radiotherapy for localized carcinoma of the prostate.” Radiotherapy and Oncology16, 1–21.

51. Roach, M., III, Krall, J., Keller, J. W., et al. 1992. “The prognostic significance of race and survival from prostate cancer based on patients irradiated on Radiation Therapy Oncology Group protocols (1976–1985).” International Journal of Radiation Oncology Biology Physics24, 441–449.

52. Epstein, J. I., Paull, G., Eggleston, J. C., et al. 1986. “Prognosis of untreated stage A1 prostatic carcinoma: A study of 94 cases with extended followup.” Journal of Urology136, 837–839.

53. Levine, R. L., and Wilchinsky, M.1979. “Adenocarcinoma of the prostate: A comparison of the disease in blacks versus whites.” Journal of Urology121, 761–762.

54. Roach, M., III. “Is race an independent prognostic factor for survival from prostate cancer?” Journal of the National Medical Association90 (11, suppl.).

55. Vijayakumar, S., Winter, K., Sause, W., et al. 1998. “Prostate specific antigen levels are higher in African-American patients than whites in a national registration study: Results of RTOG 94–12.” International Journal of Radiation Oncology Biology Physics40, 17–25.

56. Chamberlain, R., Winter, K., Vijayakumar, S., et al. 1998. “Radiation Therapy Oncology Group demographic analysis of subjects in multicenter clinical trials.” International Journal of Radiation Oncology Biology Physics40, 9–15.

57. Graham, M. V., Geitz, L. M., Byhardt, R., et al. 1992. “Comparison of prognostic factors and survival among black and white patients treated with radiation therapy for non-small cell lung cancer.” Journal of the National Cancer Institute84, 731–735.

58. Simpson, J. R., Scott, C. B., Curran, W. J., et al. 1993. “Race, gender and socioeconomic status of brain tumor patients entering multicenter clinical trials—A report from the Radiation Oncology Group.” International Journal of Radiation Oncology Biology Physics26, 239–244.

59. Streeter, O. E., Martz, K. L., Gaspar, L. E., et al. 1999. “Does race influence survival for esophageal cancer?” An analysis of patients treated with chemoradiation on RTOG 85–01. International Journal of Radiation Oncology Biology Physics44 (5), 1047–1052.

60. Modiano, M. R., Villar-Werstler, P., Crowley, J., et al. 1996. “Evaluation of race as a prognostic factor in multiple myeloma: An ancillary of Southwest Oncology Group Study 8229.” Journal of Clinical Oncology14, 974–977.

61. Pui, C.-H., Boyett, J. M., Hancock, M. L., et al. 1995. “Outcome of treatment for childhood cancer in black as compared with white children: The St. Jude Children's Research Hospital experience, 1962 through 1992.” Journal of the American Medical Association273, 633–637.

62. Gabel, L. L., and Weddington, W. H.1993. “Obligation and opportunity: Family practice research regarding race and quality of care.” Family Practice Research Journal13 (2), 101–104.

63. Weddington, W. H., Gabel, L. L., Peet, G. M., et al. 1992. “Quality of care and black American patients.” Journal of the National Medical Association84 (7), 569–575.


64. Gorey, K. M., and Vena, J. E.1994. “Cancer differentials among U.S. blacks and whites: Quantitative estimates of socioeconomic-related risks.” Journal of the National Medical Association86 (3), 209–215.

65. Fang, J., Madhavan, S., and Alderman, M.1996. “The association between birthplace and mortality from cardiovascular causes among black and white residents of New York.” New England Journal of Medicine335 (21), 1545–1551.

66. Krieger, N., and Sidney, S.1996. “Racial discrimination and blood pressure: The CARDIA Study of young black and white adults.” American Journal of Public Health86, 1370–1378.

67. Ayanian, J. Z., Udvarhelyi, S., Gatsonis, C. A., et al. 1993. “Racial differences in the use of revascularization procedures after coronary angiography.” Journal of the American Medical Association269 (20), 2642–2646.

68. Peterson, E., Wright, S. M., Daley, J., and Thibault, G. E.1994. “Racial variation in cardiac procedure use and survival following acute myocardial infarction in the Department of Veterans Affairs.” Journal of the American Medical Association271, 1175–1180.

69. Kahn, K. L., Pearson, M. L., Harrison, E. R., et al. 1994. “Health care for black and poor hospitalized medicare patients.” Journal of the American Medical Association271, 1169–1174.

70. Geronimus, A. T., Bound, J., Waidmann, T. A., et al. 1996. “Excess mortality among blacks and whites in the United States.” New England Journal of Medicine335, 1552–1558.

71. Hafner-Eaton, C.1993. “Physician utilization disparities between the uninsured and insured: Comparisons of the chronically ill, acutely ill and well non-elderly populations.” Journal of the American Medical Association269, 787–792.

72. Worthington, C.1992. “An examination of factors influencing the diagnosis and treatment of black patients in the mental health system.” Archives of Psychiatric Nursing6 (3), 195–204.

73. Hoffman, P.1994. “The science of race.” Discover: The World of Science15 (4, Special Issue), 4.

74. Gould, S. J.1994. “The geometer of race.” Discover: The World of Science15 (4, Special Issue), 64–69.

75. Gutin, J. A.1994. “End of the rainbow.” Discover: The World of Science15 (4, Special Issue), 70–75.

76. Diamond, J.1994. “Race without color.” Discover: The World of Science15 (4, Special Issue), 82–89.

77. Begley, S.1995. “Three is not enough.” Newsweek, February 13, 67–69.

78. Miller, A.1994/1995. “The Pioneer Fund: Bankrolling the professors of hate.” Journal of Blacks in Higher Education, no. 6(winter), 58–61.

79. Rowan, C. T.1996. The Coming Race War in America.Boston: Little, Brown, 17.

80. King, Martin Luther, Jr.1991. “Historic essays: Letter from Birmingham City Jail (written April 16, 1963).” In A Testament of Hope: The Essential Writings and Speeches of Martin Luther King, Jr., edited by James M. Washington. San Francisco: HarperCollins, 294.


81. King, Martin Luther, Jr.1991. “Famous sermons and public addresses: Where Do We Go From Here? (written April 1967).” In A Testament of Hope: The Essential Writings and Speeches of Martin Luther King, Jr., edited by James M. Washington. San Francisco: HarperCollins, 251.

82. Morrison, T.1993. Playing in the Dark: Whiteness and the Literary Imagination.New York: Vintage Books.

83. Brown, B., trans.1995. The Tao Te Ching of Lao Tzu.New York: St. Martin's Press, Saying 27.



The Impact of Exposure to Violence

Kathy Sanders-Phillips


Health, as defined by the World Health Organization, is “physical, mental and social well-being, not merely the absence of disease or infirmity” (Breslow, 1972). This goal calls for the enhancement of both the extent and the quality of life for people in all countries. Unfortunately, the goal is far from being met in many low-income ethnic minority groups in the United States, where poor health outcomes and significantly shorter life spans are associated with ethnic minority status.

There is increasing evidence that the health behaviors of low-income ethnic minorities are significantly influenced by their economic and social environments, which often include a legacy of poverty and disadvantaged social status; daily life experiences in stressful, unpredictable, and unsafe environments; lack of adequate health insurance and access to medical care; and little knowledge of prevention (Syme and Berkman, 1976; McGinnis, 1986; Vega et al., 1988; U.S. Department of Health and Human Services, 1991; Minkler, 1992; Wallerstein, 1992; Franke, 1997). In particular, violence has become a chronic stressor in many ethnic minority communities, where homicide is a leading cause of death (U.S. Department of Health and Human Services, 1986). Exposure to high levels of community violence can result in feelings of alienation and powerlessness, which limit the ability of low-income ethnic minorities to obtain health services and engage in health promotion behaviors (Bullough, 1972).

Health behavior does not develop in a vacuum. It is largely determined

by the social context in which it occurs (Berkanovic, 1976; Ehrhardt et al., 1995). There is growing awareness in the field of public health of the critical interplay between the social environment, health behaviors, and health and of the need for a social ecological approach to health promotion (Freudenberg, 1978; Stokols, 1992; Bloom, 1993; Lillie-Blanton et al., 1993; Williams et al., 1994). This approach acknowledges the individual's role in influencing health outcome while recognizing the important role of the environment in determining healthful behavior and health outcome. Environments may impact health behaviors and outcomes by operating as stressors or as sources of safety or danger. The benefits of a well-designed health promotion program may go unrealized if these environmental stressors are not addressed (Stokols, 1992). Developing health promotion programs for low-income ethnic minorities may require recognition of the underlying role that exposure to violence can play in influencing psychological functioning and health behaviors. This chapter reviews literature on the impact of exposure to violence on psychological functioning, health promotion behaviors, and risk behaviors in urban ethnic minority communities. Psychological and behavioral responses to community violence are reviewed, and the impact of exposure to violence on health behaviors is examined, with particular emphasis on the impact of psychosocial stressors, such as exposure to violence on health decisions and behaviors of women. Recommendations are made for future research on the impact of exposure to violence on health behaviors and for the development of health promotion programs that acknowledge and address the potential role of exposure to violence in precipitating unhealthy behaviors in ethnic minority communities.


The violence that occurs in many ethnic minority communities takes several forms. Family violence, which refers to violence occurring in the home, and community violence, which is defined as violence occurring in the community or neighborhood outside of the home (U.S. Department of Health and Human Services, 1986; Osofsky, 1995) are the most common forms. Both types of violence may involve homicide and sexual assault, but community violence also includes aggravated assault, burglary, and robbery (Resnick et al., 1986, 1993).

Ethnic minority communities are disproportionately impacted by violence,

and African American and Latino males are at particularly high risk for death or injury due to violence. Homicide is a leading cause of death for African American males between the ages of 15 to 44 and for African American females between the ages of 15 to 24 (Centers for Disease Control, 1990; Hammond and Yung, 1993). Until age 70, homicide is second only to heart disease in its contribution to excess deaths among African Americans (U.S. Department of Health and Human Services, 1986). Age-adjusted rates of homicide among Latinos are approximately three to four times higher than those for White males (Smith et al., 1986; Tardiff and Gross, 1986; Centers for Disease Control, 1990), and foreign-born Latinos are at higher risk for homicide than nativeborn Latinos (Sorenson and Shen, 1996). Although less data are available for Native Americans and Asians, statistics indicate that accidents, which may be related to violence, are the third-leading cause of death for Native Americans, followed by homicides as the seventh-leading cause of death (Dinges and Joos, 1988). Homicide is the eighth-leading cause of death for Asian American males and fourteenth for females (Gall and Gall, 1993).

Nonfatal injuries due to violence have been estimated to be at least 100 times more frequent than homicides (O'Carroll, 1988). Statistics from the National Crime Survey (Christofel, 1990) indicate that approximately 1.2 million crimes of violence are not reported to law enforcement. Like homicide, nonfatal violence is higher among low-income ethnic minority groups. African American males have higher rates of injury due to violence than any other male group, and they outnumber African American females by three to one (Guyeret al., 1989). Latinos have the second-highest rates of injury due to violence (Sumner et al., 1986; Guyer et al., 1989; U.S. Department of Justice, 1991).

Family violence, which includes violence against spouses, domestic homicide (i.e., the killing of one spouse by another), and other forms of violence, is also higher in ethnic minority populations (U.S. Department of Health and Human Services, 1986; Bell and Chance-Hill, 1991). African American women are more likely than African American men to be murdered by a family member or an intimate partner (Humphrey and Palmer, 1986; Wilbanks, 1986; Harlow, 1989). Both African American and Latino children may be overrepresented among sexual abuse victims (Kersher and McShane, 1984; Cupoli and Sewell, 1988). School violence, as measured by the number of robberies and assaults in high schools, is also higher in urban ethnic minority communities (U.S. Department of Health and Human Services, 1986; Sheleyet al., 1992). Higher rates of

aggravated assaults and illegal drug use, which are associated with high rates of violence, have also been reported for ethnic minority communities (Dawkins and Dawkins 1983; Lester, 1986; Mercy et al., 1986; Bell, 1987; Weiszet al., 1991; Elliot, 1993).

Members of ethnic minority groups are also more likely to witness violence in their daily lives. Much of the violence is witnessed by children, and, in one study, shootings are cited as the most serious danger in the lives of mothers and children living in a low-income ethnic minority community (Dubrow, 1989). In other studies, more than a quarter of urban schoolchildren had witnessed a person being shot, and almost a third had seen a person stabbed (Garbarino et al., 1991a and 1991b; Freeman et al., 1992; Lorion and Saltzman, 1993). Among lowincome elderly living in urban communities, fear of victimization and personal safety accounts for the largest percentage of variance in quality of life for African Americans versus White Americans (Harel, 1986).


There is increasing evidence that exposure to violence negatively impacts health promotion behaviors and may increase risk behaviors, particularly in low-income African American and Latino groups. In a sample of lowincome African American women and Latinas living in an urban community, having a family member killed or murdered was a significant predictor of health promotion behaviors (Sanders-Phillips, 1996a). That is, women who had a family member killed or murdered were less likely to be engaging in health promotion behaviors, including eating a daily breakfast, sleeping seven to eight hours per night, abstaining from alcohol and tobacco, and exercising at least once per week. Women with a family member beaten by another family member were less willing to change their eating habits to include more healthful foods (Sanders-Phillips, 1994a). Exposure to violence is also related to practicing unsafe dieting and eating foods high in fats among adolescents (Orpinas et al., 1995) and is a predictor of lack of a regular health care provider and delays in seeking medical care among ethnic minority patients (Rask et al., 1994). Exposure to violence is also related to poor health perceptions, functional limitations, chronic disease, and somatic symptoms in ethnic minority groups (Golding, 1994).

Risk behaviors such as the use of illegal drugs and alcohol are also associated

with exposure to violence. Illicit drug use is strongly related to assault by a mate, and tobacco and alcohol use are associated with assault by a member of the family of origin (Berenson et al., 1992). Among Latinos, physical abuse is related to higher levels of tobacco use, while both physical and sexual assault are strongly related to alcohol use among African Americans (Berenson et al., 1992). Exposure to violence is also related to higher levels of current, lifetime, and intended drug and alcohol use among children (Lorion and Saltzman, 1993).

There may be specific relationships between exposure to violence and pregnancy-related risk behaviors in women. Stevens-Simon and Reichert (1994) found that adolescent pregnancy was predicted by a history of childhood sexual abuse. Illegal substance use during pregnancy is also significantly related to exposure to violence (Martin et al., 1996). Exposure to psychosocial stressors, which includes exposure to community violence, is related to behavioral risk factors during pregnancy, including chronic medical disease, poor weight gain, alcohol use, and illegal drug use (Orr et al., 1996). Among African American women, exposure to psychosocial stressors is associated with the birth of low-birth-weight infants. Zapata et al. (1992) have also shown that among pregnant women in Chile, exposure to community violence, in the form of political violence, was also related to lower-birth-weight infants.

Exposure to violence has also been found to impact AIDS prevention and risk behaviors in men and women. Lemp et al. (1994) reported higher levels of AIDS risk behaviors for men exposed to sexual violence. Exposure to community and family violence has been correlated with a greater number of sexual partners among African American males (Durant et al., 1994b). Rotheram-Borus et al. (1996) found a relationship between exposure to sexual violence and increased sexual risk behaviors in adolescents who were also more likely to use alcohol and illegal drugs. Several investigators have reported that exposure to sexual violence is related to HIV and STD sexual risk behaviors and increased alcohol consumption in women (Glaser et al., 1991; Wyatt, 1992; Irwin et al., 1995; Zeiler et al., 1996) and adolescent girls (Orpinas et al., 1995). Kavanaugh et al. (1992), although not specifically examining exposure to violence, found that African American women in an AIDS prevention program experienced high levels of alienation and powerlessness associated with urban life that impacted AIDS-related health behaviors.

These findings strongly suggest that exposure to violence is related to greater involvement in

risk behaviors and decreased involvement in health promotion behaviors. In an effort to identify the mechanisms by which exposure to violence may impact health behaviors, a literature review of studies examining relationships between experiences of violence, psychological functioning, and perceptions of health and well-being was conducted. The databases searched were PsychLIT, Medline, Social Citation Index, Sociofile, and Social Work Abstracts using “health,” “trauma,” “violence,” “alienation,” and “powerlessness” as key words.


Although rates of violence have escalated in all segments of our society (Rosenberg and Fenley, 1991), it is the chronic, pervasive, and random nature of the violence that distinguishes life in urban ethnic minority communities from life in other communities. The violence is life threatening and unpredictable and occurs in public places populated by innocent bystanders (Bell and Jenkins, 1991). It has become a chronic stressor that affects quality of life (Cohen et al., 1982; et al., 1991; Garbarino et al., 1991; Sluzki, 1993). Chronic danger, which is defined as regular and persistent attacks of violence that disrupt day-to-day life, requires significant adjustments, including alterations of personality and major changes in patterns of behavior, that allow for interpretation of the danger and accommodation to the realities of community life (Sonnenberg, 1988; Garbarino et al., 1991; Lorion and Saltzman, 1993; Sluzki, 1993). Exposure to chronic danger affects interpersonal, cognitive, psychological, and behavioral functioning on a day-to-day and long-term basis (Lorion and Saltzman, 1993).

Individuals who have been victims of violence may show symptoms of posttraumatic stress disorder, which is characterized by recurrent memories and dreams of the event, illusions, hallucinations, dissociative flashbacks, and intense psychic pain (Sonnenberg, 1988; Davidson and Smith, 1990; Singer et al., 1996). In the long term, victims of violence may continue to experience fear, terror, and a sense of helplessness (Sonnenberg, 1988; Davidson and Smith, 1990). There is often a diminished interest in significant life activities, a feeling of being detached or estranged from others, a restriction of the ability to feel intensely, and a pessimistic sense regarding the future (Sonnenberg, 1988). Symptoms of depression and anxiety are also common among victims of violence (Freeman et al., 1992; Singer et al., 1996).


Those who witness violence may experience all or some of these symptoms. Children who have witnessed violence show higher levels of anxiety and often manifest a sense of futurelessness characterized by a belief that they will not reach adulthood (Hughes, 1988; Bell and Jenkins, 1991; Martinez and Richter, 1993). Adult witnesses to violence also report feeling that the end of life will come soon (Sonnenberg, 1988) and experience pervasive feelings of fear, vulnerability, and hopelessness (Lorion and Saltzman, 1993). Acting-out behaviors, particularly in adolescents, and self-destructive behaviors are common (Bell and Jenkins, 1991). Women and girls are likely to experience depression, anxiety, and sleep problems in response to community violence, while men and boys experience distress and behavioral problems (Lorion and Saltzman, 1993; Martinez and Richter, 1993).

The cumulative experience of exposure to violence and victimization may have significant effects on psychological functioning. Hinton-Nelson et al. (1996) found that African American adolescents who had witnessed violence but who had not experienced violence had higher levels of hope for the future than those who were victims, although feelings of hopelessness were apparent in both groups. Hope was defined as the ability to move toward a goal in life and develop plans to meet that goal. The experience of victimization was related to lower levels of hope and predictions that one's own death would be violent, while exposure to violence was correlated with predictions that one's death would be nonviolent. Durant et al. (1994a, 1995) also found that adolescents who were exposed to high levels of family and community violence did not expect to live to the age of 25. These findings suggest that individuals in communities of high violence tend to predict a cause of death that is consistent with their environment (Hinton-Nelson et al., 1996). Exposure to chronic violence significantly impacts impulse control and risk-taking behavior (Gardner, 1971; Lorion and Saltzman, 1993). Individuals exposed to chronic violence may show a level of passivity and emotional withdrawal, coupled with difficulties in controlling aggressive impulses, that interfere with learning (Gardner, 1971). Since exposure to chronic violence reinforces the conclusion that achieving lasting or socially approved outcomes is unlikely, socially unacceptable, risky, but rewarding behavior may become highly attractive (Garbarino et al., 1991; Lorion and Saltzman, 1993). In addition, community residents may become desensitized to the threat and consequences of violence and pursue opportunities for risk taking and

confrontation with danger (Garbarino et al., 1991; Lorion and Saltzman, 1993). Without hope for a future, some may attempt to gain control over their lives through repeated encounters with life-threatening situations.


While few studies have specifically examined the relationship between exposure to violence and health promotion behaviors, a number of studies have examined relationships between experiences of violence, subsequent feelings of alienation, anomie, and powerlessness and perceptions of health and well-being. Other investigators have examined relationships between feelings of alienation, anomie, and powerlessness and health promotion behaviors. Collectively, the findings suggest that exposure to violence influences perceptions of control over life and over one's health outcome. Exposure to violence, and the resulting feelings of alienation, anomie, and powerlessness, may specifically impact health promotion and risk behaviors, particularly in low-income urban populations.

Based on the existing literature, a conceptual model of relationships between exposure to violence and health behaviors in low income, urban, ethnic minority communities is presented in Figure 11.1. In this model, exposure to violence serves as an intervening variable that may partially explain relationships between low-income, urban, ethnic minority status; decreased health promotion behaviors; and increased risk behaviors. As illustrated, low-income, urban, ethnic minority status in the United States is associated with exposure to high levels of community and family violence. In turn, exposure to violence may result in feelings of powerlessness, anomie, and alienation that are related to the subsequent development of psychological distress, which may be expressed as hopelessness, decreased self-efficacy, and decreased motivation to seek information. These psychological perspectives may lead to changes in life priorities, including the priority assigned to health, decreased problem-solving skills, and difficulties in social learning that may be necessary for health promotion and disease prevention and decreased utilization of available health services, which may result in decreased health promotion behaviors and increased risk taking regarding health.

Alienation, hopelessness, powerlessness, and anomie are related psychological


Figure 11.1. Conceptual model of the impact of exposure to violence on health promotion behaviors.

constructs that refer to an individual's perceived relationship with the larger society and perceptions of control within that society. Alienation is defined by feelings of powerlessness and social isolation (Morris et al., 1966). It includes feelings that result from rejection of cultural norms for evaluating success and rejection of societal means for achieving success (Wilkerson et al., 1982). Powerlessness is a subjective or perceived expectancy or belief that an individual cannot determine the occurrence of outcomes (Seeman, 1959) and is characterized by feelings that one has no effective control over one's destiny (Morris et al., 1966). Hopelessness is the inability to see or make plans for the future (Hinton-Nelson et al., 1996). In contrast, hope lies in the ability to make linkages between previous desires and to use strategies that meet those
needs (Hinton-Nelson et al., 1996). Anomie is a broader concept that includes feelings of alienation, pessimism, distrust, and hopelessness (Cohen et al., 1982). Alienation, hopelessness, powerlessness, and anomie are related to experiences of violence and have been shown to significantly impact health decisions and behaviors.

Cohen et al. (1982) reported that negative feelings about health among both African Americans and Latinos were related to community measures of stress, including poverty, feelings of fear, social problems, and violent crimes. These community stressors resulted in feelings of anomie. In communities of high violence, feelings of powerlessness and anomie may be coupled with a disinclination to worry about events that cannot be controlled (Dubrow and Garbarino, 1989). These feelings are compounded by poor relationships with police (Dubrow and Garbarino, 1989). Community residents are likely to report that they cannot depend on the police or that they fear retaliation if they talk to the police. According to one woman in a low-income housing project plagued by violence, “If identified … as a caller (to the police), one's entire family would be in jeopardy” (Dubrow and Garbarino, 1989, p. 9).

Feelings of powerlessness, alienation, and anomie are also associated with poor health behaviors and less preventive care. Feelings of powerlessness and alienation have been associated with increased doctor visits and decreased levels of prenatal care among Mexican Americans (Hoppe and Heller, 1975), lower levels of knowledge among patients regarding their disease (Seeman and Evans, 1962), lower levels of selfinitiated preventive care (diet, exercise, alcohol, smoking) and lower levels of belief in the efficacy of early treatment (Seeman and Seeman, 1983), higher rates of pregnancy and lower levels of family planning behavior among low-income women (Groat and Neal, 1967), and low levels of well child care (i.e., immunizations) in low-income mothers, particularly African American mothers (Morris et al., 1966). Alienation, powerlessness, hopelessness, and social isolation significantly decreased the use of preventive services among low-income mothers in Los Angeles, and the effect was most pronounced for African American mothers (Bullough, 1992). Although poverty was a potent predictor of the use of preventive services in this sample, alienation and powerlessness were critical intervening variables in determining health promotion behaviors. More recently, high levels of alienation and powerlessness have been associated with higher rates of AIDS risk behaviors (Kavanaugh et al., 1992).



There are several ways in which exposure to violence, and the resulting feelings of powerlessness, anomie, and alienation, may impact health promotion behaviors. Exposure to violence and perceptions of uncontrollability of aversive and unpredictable events (i.e., powerlessness, helplessness, hopelessness) may result in depression (Green, 1982; Garbarino et al., 1991; Freeman et al., 1992). Depression, in turn, is related to higher levels of smoking and drinking alcohol (Cohen et al., 1991). Other types of psychological distress, including high levels of perceived stress, are related to poor eating habits in the general population and in Mexican Americans (Rakowski, 1988; Vega et al., 1988). These findings suggest that exposure to violence, and subsequent feelings of powerlessness, alienation, and anomie, may result in depression and/or psychological distress that diminish an individual's ability to engage in health promotion behaviors.

Feelings of powerlessness, hopelessness, and alienation may also influence life priorities and affect motivation to seek health promotion and disease prevention information. Motivation to seek information depends on an individual's feelings of powerlessness and the value placed on an outcome (Dodge et al., 1997). High levels of alienation and powerlessness are related to low levels of motivation to seek information, particularly health information, and less value placed on health outcome (Seeman and Evans, 1962; Groat and Neal, 1967). Seeman and Evans and (1962) concluded that “a sense of personal control makes knowledge concerning one's affairs motivationally relevant” and, conversely, that “knowledge acquisition may be irrelevant for those who believe that external forces control the fall of events” (p. 773). “Seeking information is to behave as though it is within one's capacity to control events through knowledge” (Seeman and Evans, 1962, p. 781).

Exposure to violence may also affect social learning and problemsolving skills, resulting in failure to engage in learning that involves planning and control (Seeman and Evans, 1962; Gardner, 1971; Rosenberg, 1987; Ross and Mirowsky, 1989; Spivak and Hausman, 1989; Dodge et al., 1997). Decreased levels of perceived control and problem-solving skills are also related to depression (Ross and Mirowsky, 1989). Thus, the likelihood that a person will seek health information is affected by feelings of control and the value placed on health, and these feelings are significantly impacted by exposure to violence. Perceptions of uncontrollability

and a low value for health may also contribute to increased involvement of minorities in high-risk health behaviors. Feelings of powerlessness and alienation also limit the likelihood that an individual will utilize preventive services that may be available and/or that he/she will persist in engaging in health promotion activities (Seeman and Evans, 1962; Mindlin and Densen, 1971; Bullough, 1972; Dubrow and Garbarino, 1989). For example, low-income African American women who lived in violent neighborhoods and reported feelings of powerlessness were less apt to use the services and programs that were available to them (Dubrow and Garbarino, 1989). Similar groups of low-income ethnic minority women were found to initiate health prevention practices but were less likely to persist in these activities over time (Mindlin and Densen, 1971). Both powerlessness and anomie appear to act as psychological deterrents to making the kind of sustained effort that is necessary to be successful in overcoming obstacles or barriers to a desired end (Bullough, 1967). As a result, powerlessness, hopelessness, and alienation may indirectly affect health promotion behaviors by influencing the extent to which barriers such as lack of money or lack of services may be overcome.

Powerlessness may also directly affect health status. Wallerstein (1992) has shown that powerlessness, or lack of control over one's destiny, is an independent risk factor for disease. Powerlessness has been linked to higher rates of physical, mental, and behavioral health problems (Wallerstein, 1992).

Perhaps one of the primary means by which feelings of alienation, powerlessness, and anomie impact health behaviors, particularly in children and adolescents, is via relationships with hopelessness. Ideally, hope is established during infancy in the context of nurturing caretakerinfant interactions (Snyder, 1994). Hope is then cultivated via a child's interactions with the environment and through reinforcement of the perception that they can act to reach their goals and handle impediments to realizing those goals (Hinton-Nelson et al., 1996). According to Hinton-Nelson et al. (1996), life in chronically violent environments may adversely impact parent-child relationships and destroy children's perceptions of links between goals, actions, and achievements. As a result, many of these children perceive that early death is probable, and both children and parents may feel that life cannot be protected (Rinear, 1988). Under these circumstances, hope is understandably diminished, and there is little rationale to pursue goals of health or life. Hopelessness, coupled with the differential access to coping resources that is a characteristic of

poverty (Durant et al., 1995), may limit perceived life and health options and decrease motivation to engage in life-enhancing or health promotion behaviors.


Health promotion behaviors are voluntary, and the factors influencing voluntary behaviors, particularly those undertaken without recommendation by a physician, differ considerably from those influencing other health behaviors (Berkanovic, 1981–82). Although many theories have been proposed to explain the adoption of health promotion behaviors, there is general consensus that anticipation of a negative outcome and the desire to avoid the outcome, or to reduce its impact, creates motivation for self-protection (Weinstein, 1993). The individual must also weigh the expected benefits in risk reduction against the expected costs of acting (Weinstein, 1993). The value ascribed to health also becomes increasingly important as discretion in the use of a service and/or in engaging in a behavior increases (Berkanovic, 1973, 1981–82). Lastly, an individual's belief in his/her ability or likelihood of changing current patterns of behavior (i.e., self-efficacy/perceived behavioral control) will also impact motivation and the degree to which he or she engages in health promotion behavior (Weinstein, 1973; Berkanovic, 1981–82). It is clear that exposure to violence may impact each of these prerequisites of health promotion behavior. It affects one's anticipation of and ability to control negative outcomes, value ascribed to life and health, motivation to weigh costs and benefits of actions, and perceived efficacy and control in determining one's fate. Particularly for ethnic minorities, a sense of efficacy and control are dependent on the nature of their social experiences and the larger social structure and community in which they function (Groat and Neal, 1967; Hughes and Demo, 1989). The larger community both serves as a channel of communication and information and provides for the maintenance of a sense of personal control (Groat and Neal, 1967). However, powerlessness serves as an intervening variable between an individual's social circumstances and his/her social learning (Groat and Neal, 1967). To the extent that one's social environment reinforces the conclusion that individual destiny cannot be controlled, it becomes doubtful that one will engage in self-protective behavior. According

to Lorion and Saltzman, “It may make little sense to be careful for oneself or others if physical harm or death are deemed inevitable” (Lorion and Saltzman, 1993, p. 57).

Rainwater (1960) has argued that a sense of stability and trust in the future are essential preconditions for health promotion behaviors. At the very least, health promotion behavior requires planning and a sense of control over one's future or a sense that a future exists. It is clear that exposure to chronic violence can destroy these perceptions.


The impact of violence on health behaviors in low-income ethnic minorities may be exacerbated by feelings of powerlessness and anomie that are associated with poverty and ethnic minority status in this country (Berkanovic, 1973; Green, 1982; McLoyd, 1990), regardless of the degree of exposure to violence. A life of poverty is characterized by higher levels of stress, negative life events, and chronic conditions outside of personal control (Liem and Liem, 1978; Kessler, 1979). Chronic poverty severely restricts individual choice in all domains of life, renders a person subject to greater control by others, and weakens an individual's ability to cope with new problems and difficulties (McLoyd, 1990). Thus, those in poverty are more likely to experience feelings of hopelessness, powerlessness, and depression (Green, 1982) that are more severe when catastrophic events are outside of the control of the individual (Liem and Liem, 1978; Kessler and Cleary, 1980).

Even in the absence of violence, the stress of poverty, in conjunction with stressors of ethnic minority status (Ogbu, 1983), may result in feelings of victimization and dissatisfaction (McLoyd, 1990) that influence health behaviors. There is also evidence that poverty is related to decreased access to resources for coping with adversity and increased vulnerability to the long-term effects of violence (Kessler and Magee, 1994). Berkanovic and Reeder (1973) have concluded that both ethnicity and socioeconomic status create different life experiences that are related to value preferences and, subsequently, to variations in health behaviors and use of discretionary health services. The cumulative impact of exposure to violence and poverty may severely limit the motivation, ability, and belief of many low-income ethnic minorities that they can successfully impact their lives and health outcomes.



In the past decade, significant increases have occurred in our knowledge of factors that influence women's health decisions and behaviors as well as barriers to healthy behaviors among women (Brown-Bryant, 1985; Calnan and Johnson, 1985; Makuc et al., 1989; Ruzek et al., 1997b). There is also greater awareness of factors that influence health behaviors in ethnic minority women (Cope and Hall, 1985; Leigh, 1994; Sanders-Phillips, 1994a, 1994b, 1996a, 1996b) and strategies that can be effective in promoting healthier behaviors in ethnic minority women (Eng et al., 1985; Schaefer et al., 1990; Levine et al., 1992; Eng, 1993).

Previous findings indicate that health behaviors are gender specific (Cohen et al., 1982, 1991; Gottlieb and Green, 1984; Rakowski, 1988; Baum et al., 1991; Ruzek et al., 1997a) and impacted by a range of psychosocial factors, including psychological status, exposure to violence, and perceptions of stress. These factors influence health-related self-efficacy and priorities regarding prevention behaviors (Cohen et al., 1982; Broman and Johnson, 1988; Smith, 1992; Sanders-Phillips, 1996a, 1996b) and involvement in health promotion and disease prevention behaviors (Sanders-Phillips, 1994a, 1994b). Although many of the following studies did not examine direct relationships between exposure to violence and health behaviors in women, they demonstrate that women's health behaviors are significantly related to psychological functioning and perceptions of stress. Since exposure to violence is associated with higher levels of stress and depression, it is probable that women's health behaviors are adversely impacted by exposure to violence.

Ferrence (1988) found that women's risk behaviors are significantly related to social status and interactions outside the home, and previous findings indicate that mental health factors may be better predictors of health behaviors for women than men (Lex, 1991). In addition, stressful life events are better predictors of involvement in healthy behaviors for women than for men (Gottlieb and Green, 1984; Rakowski, 1988; Cohen et al., 1991). That is, fewer stressful life events are related to healthier behavior in women, especially in low-income groups. Psychological status, which is related to levels of stressful life events such as exposure to violence, also influences risk behaviors in women (Cohen et al., 1991; Shumaker and Hill, 1991). Poorer health behaviors are more common in women with symptoms of depression (Cohen et al., 1991; Leftwich and Collins, 1994), which is also a correlate of exposure to violence.


Baum and Grunberg (1991) have argued that cigarette smoking, alcohol, and illegal drug use should be viewed as coping behaviors for women, and Gottlieb and Green (1984) concluded that relationships between stress, alcohol, and smoking in women may be strong enough to justify sex-specific norms for smoking and drinking as coping mechanisms for stress. Others have found that both trauma and depression are etiologic factors in women's drug abuse (Fullilove et al., 1992; Singer et al., 1993) and that illegal and prescription drug use among women may serve as forms of self-medication to cope with stresses (Booth et al., 1991). Gender-based differences in the use of substances may be related to differences in perceptions of stress and/or in motivations to regulate mood through substance use. Thus, women may be at greater risk for coping with stress by altering behaviors that impact health (Baum and Grunberg, 1991).

These results underscore the importance of psychological and social variables to women's risk behaviors and support previous findings that psychological and social factors, such as exposure to violence, are significant predictors of risk behaviors among all women, particularly ethnic minority women (Rodin, 1986). Increased awareness of relationships between these stressors and unhealthy behaviors may be critical to our understanding of the development of risk behaviors and to the consequences of high stress for women. These findings also support the conclusion that women's risk behaviors may have common etiologies. Current data also suggest that psychosocial factors may be more important than demographic or biomedical factors in determining health outcomes, especially in low-income ethnic minority groups (Rodin and Ickovics, 1990; Pincus and Callahan, 1995).

In the future, gender-specific models of health behavior are needed that acknowledge the role that factors such as exposure to violence may play in the health and risk behaviors of women. As patterns and profiles of risk factors and illness by gender change, research on psychosocial factors related to women's health will be urgently needed (Rodin and Ickovics, 1990). Increasing rates of mortality and morbidity for women appear to be related to several factors, including the ability of women to adapt and cope with stress and their perceptions of control (Krieger, 1990; Rodin and Ickovics, 1990; Auerbach and Figert, 1995). Understanding relationships between psychosocial factors, health and risk behaviors, and health outcomes in women may greatly improve our ability to develop more effective health promotion programs for women. Greater attention to psychosocial factors, such as experiences of violence

that may specifically influence health behaviors and outcomes for minority women, may also enhance our theoretical understanding of health and risk behaviors in women.


The findings presented in this chapter identify exposure to violence as a potentially critical intervening variable in explaining the relationship between minority status, poverty, and poor health behaviors. Present data suggest that poor health outcomes are but one of the many problems facing low-income ethnic minority families, perhaps not even the most pressing or immediate problem. In communities with high rates of violence, it is unlikely that health will be a priority in the lives of residents, and they may be less likely to seek health information. Poor ethnic minority families must contend with the cumulative effects of violence, poverty, and stressful life events in urban communities. As Berkanovic (1981–82) notes,

We need to recognize that health is only one of many values competing for the time and energy of the individual. It may be perfectly rational for one who believes that a course of action is deleterious to his health to take that course of action if some more important values can be realized. In addition to exploring structural factors that go beyond standard demographic characteristics, therefore, perhaps students of health protective behavior might cast a wider net with respect to the social psychological factors that influence such behavior. (p. 235)

Recent studies have attempted to identify racial, demographic, and socioeconomic correlates of health behavior (Martinez and Lillie-Blanton, 1996). To date, however, we know little of the sociodemographic variables related to health promotion that can be modified (Becker, 1979; Berkanovic, 1981–82). Yet, exposure to violence is a sociodemographic variable that can be an important determinant of health promotion behavior, affects attitudes and beliefs about health, and may be modifiable, if health promotion programs are appropriately developed.

Traditional health promotion programs often rely on the diffusion of ideas: They focus on communicating information to a population (Braithwaite and Lythcott, 1989). Unfortunately, this approach often fails to consider the factors that may prevent people from receiving and/or acting on the information. Health professionals must recognize that

providing education regarding health promotion activities may not be effective in changing attitudes or behaviors of ethnic minorities. “It is easy to assume that people will act in accordance with their attitudes and beliefs provided there are no other factors acting to influence their behavior” (Berkanovic, 1981–82, p. 235). Health interventions for lowincome ethnic minority groups must address social contexts and quality of life and acknowledge the daily experiences that shape individual perceptions and perceived options. Successful intervention with low-income ethnic minority groups may need to offer ways of reducing and/or coping with life stressors, particularly high levels of chronic violence. The stresses of violence and poverty may be so overwhelming that health promotion programs have limited chance of success unless health is presented as a means of coping with stress and regaining control over one's life. Healthy lifestyles, therefore, must provide a means of coping with life rather than be an added burden, and health promotion programs must empower urban communities to address the stressors that impact their lives and health outcomes.

The concept of increasing a sense of personal or community control and power in order to effect health behavior change is an integral component of the community empowerment model, which emphasizes the importance of reclaiming control over one's life and environment (Gottlieb, 1985; Braithwaite and Lythcott, 1989; Minkler, 1992; Wallerstein, 1992). Empowerment is a social-action process that promotes participation of people, organizations, and communities toward the goals of increased individual and community control, political efficacy, improved quality of community life, and social justice (Wallerstein, 1992, p. 198). Community empowerment models have long recognized the relationship between feelings of alienation, powerlessness, and health behaviors. However, future programs may need to assist communities in reducing levels of community violence before issues of health behavior can be effectively addressed. Or, communities must be assisted in addressing both the level of violence and poor health outcomes. Practical solutions to the realities of violence must be provided, and safe and convenient locations for refuge from the violence and/or for conducting health promotion activities are essential. There is evidence that comprehensive approaches to health promotion, which emphasize community empowerment and address the role of violence in the lives of community residents, can be successful in low-income ethnic minority communities (Minkler, 1992).

For example, as we develop health promotion programs for ethnic

minority communities, it will become increasingly important to recognize that violence prevention is indeed a health promotion intervention in ethnic minority communities where there are high levels of interpersonal and community violence. Based on the findings presented in this chapter, high levels of community violence not only impact the wellbeing of victims of violence but also influence the health behaviors of those who are chronically exposed to it. Thus, health promotion programs targeting urban ethnic minority communities must not only acknowledge that the prevalence of violence influences health behaviors but also must strive to reduce the incidence of violent episodes in the community as a means of improving health behaviors and health outcomes. Perhaps the work of Minkler (1992) best exemplifies this approach. In designing health promotion programs for the elderly poor in San Francisco, “safehouses” where residents could seek refuge from community violence were established to increase a sense of safety among community residents before other health promotion programs were initiated. In this urban community, and in many others, a sense of safety is a prerequisite to healthier individual behaviors.

However, as community-wide violence prevention programs are established, it may also be necessary to address the previous impact of exposure to violence on individual health behaviors. In this regard, lay health advisers have been used successfully to promote health behavior change, especially in ethnic minority groups, and address the social and cultural barriers, such as exposure to violence, to healthy behaviors (Hargreaves et al., 1989; Amezuca et al., 1990). The effectiveness of lay health advisers in helping individuals overcome the cultural and social barriers to healthy behaviors, recruiting individuals to health promotion interventions, and increasing health promotion behaviors has been well documented (Warnecke et al., 1975, 1976; Salber, 1979; Brownstein et al., 1992; Levine et al., 1992). The importance of using lay health advisers who are similar to the target population and who conduct interventions in programs where there is a shared sense of identity has also been stressed (Warnecke et al., 1975; Israel, 1982, 1985). The success of health intervention programs that focus on experiences related to ethnicity (DiClemente and Wingood, 1995) also supports the use of indigenous community workers and underscores the importance of addressing issues such as exposure to violence that may be common in ethnic minority groups.

Finally, in order to develop more effective health promotion programs for urban ethnic minority communities, extensive research on the

relationship between exposure to violence and health behavior is also needed. We need to understand how exposure to different types of violence (e.g., murder, domestic violence, physical or sexual abuse) affects health promotion behaviors. Does the time of exposure (i.e., childhood versus adulthood) and/or the level of exposure to violence influence the degree to which health promotion behavior is impacted? Does gender influence the relationship between exposure to violence and health promotion behaviors? And can an individual's response to violence, and the resulting impact on health promotion behaviors, be modified without significant changes in levels of community violence? Answers to these questions are important to our understanding of the relationship between exposure to violence, health promotion behavior, and successful intervention and to the relative need for individual versus community-wide intervention approaches.

Given the growing disparities in health outcomes for many minorities in this country, understanding and examining factors related to ethnic differences in risk factors and health behaviors is crucial. We must move beyond simply identifying ethnic differences in risk and health behaviors to a more comprehensive understanding of what these differences may mean and represent (Martinez and Lillie-Blanton, 1996). Jessor and Donovan (1985), in studying the structure of health and problem behaviors in adolescents, have concluded that risk behaviors are interrelated and can be accounted for by common factors. Their research provides important support for the conclusion that risk behaviors may have common etiologies. It also encourages research on the “risk factors” for the “risk factors” in ethnic minority populations. Orpinas et al. (1995) have also reported evidence of covariation in risk behaviors and suggest that risk behaviors may be linked to psychological and/or environmental factors. These studies provide important support for the conclusion that risk behaviors may have common etiologies. They also encourage research on the “risk factors” for the “risk factors” that impact ethnic minority populations. Link and Phelan (1995) have also emphasized the importance of examining the social conditions that place individuals at risk of risk behaviors.

Martinez and Lillie-Blanton (1996) have also concluded that ethnic differences in health behaviors and outcomes may reflect differences in life experiences and the environments in which these individuals grow and develop. This view is consistent with the findings presented in this chapter on the impact of exposure to violence on health behaviors in ethnic minority populations. Indeed, exposure to community and/or interpersonal

violence may be one of the risk factors for other risk factors, such as smoking, lack of exercise, use of alcohol, and use of illegal drugs, that severely compromise the health outcomes of ethnic minorities in this country. Examination of the mediating influence of exposure to violence on health behaviors in urban ethnic minority groups may also increase our understanding of relationships between socioeconomic status and health in ethnic minority groups and help to explain findings linking socioeconomic status to risk factors for chronic disease (Lowry et al., 1996), prevalence of disease in ethnic minority populations (Breen and Figueroa, 1996), and screening behavior in ethnic minority women (Pearlman et al., 1996). Such studies are needed to understand how race and ethnicity interact with other factors in impacting health behaviors and outcomes.

Research in this area may also have significant implications for the planning and development of health promotion and disease prevention programs for urban ethnic minority communities. A recent survey of health officers in economically stressed urban centers in the United States revealed that the five most important public health prevention goals were reducing the incidence of HIV infection and AIDS, improving maternal and infant health, controlling sexually transmitted diseases, reducing violent and abusive behavior, and immunizing against infectious diseases (Greenberg et al., 1995). Current findings suggest that exposure to violence may contribute to the incidence of the other health problems and that controlling violence in urban centers may be an important vehicle for addressing other health problems. Existing data also suggest that failing to address multiple risk behaviors and their common etiologies in comprehensive programs of health intervention and prevention developed for urban communities may result in marginal effects on health outcomes (Orpinas et al., 1995). Health promotion professionals cannot be defeated in their efforts to impact health behaviors and health outcomes among low-income ethnic minority communities. Programs must be developed that acknowledge and incorporate an understanding of levels of violence and respect for realities of life in these communities. We must refocus on the revised definition of health promotion as “a process of enabling people to increase control over and to improve their health” (World Health Organization, 1986) while working to enhance the capability of individuals and communities to effectively respond to challenges posed by the environment (Minkler, 1992). In sum, considerable effort must be made to reestablish health as a priority in the lives

of low-income ethnic minority groups exposed to violence and to facilitate a sense of empowerment regarding their life and health outcomes.


Amezuca, C., McAlister, A., Ramirez, A., and Espinoza, R.1990. “A su salud: Health promotion in a Mexican-American border community.” In Health Promotion at the Community Level, edited by N. Bracht. Newbury Park, Calif.: Sage Publications.

Auerbach, J. D., and Figert, A. E.1995. “Women's health research: Public policy and sociology.” Journal of Health and Social Behavior36, 115–131.

Baum, A., and Grunberg, N.1991. “Gender, stress, and health.” Health Psychology10, 80–85.

Becker, M. H.1979. “Psychosocial aspects of health-related behavior.” In Handbook of Medical Sociology,3rd. ed., edited by H. Freeman, S. Levine, and L. Reeder. Englewood Cliffs, N.J.: Prentice Hall.

Bell, C.1987. “Preventive strategies for dealing with violence among Blacks.” Community Mental Health Journal23, 217–228.

Bell, C. C., and Chance-Hill, G.1991. “Treatment of violent families.” Journal of the National Medical Association83, 203–208.

Bell, C. C., and Jenkins, E. J.1991. “Traumatic stress and children.” Journal of Health Care for the Poor and Underserved2, 175–188.

Berenson, A. B., San Miguel, V. V., and Wilkinson, G. S.1992. “Violence and its relationship to substance use in adolescent pregnancy.” Journal of Adolescent Health13, 470–474.

Berkanovic, E.1976. “Behavioral science and prevention.” Preventive Medicine5, 92–105.

Berkanovic, E.1981–82. “Who engages in health protective behaviors?” International Quarterly of Community Health Education2, 225–237.

Berkanovic, E., and Reeder, L.1973. “Ethnic, economic, and social psychological factors in the source of medical care.” Social Problems21, 246–259.

Bloom, M.1993. “Toward a code of ethics for primary prevention.” Journal of Primary Prevention13, 173–182.

Booth, M. W., Castro, F. G., and Anglin, M.1991. “What do we know about Hispanic substance abuse?” A review of the literature. In Drugs in Hispanic Communities, edited by R. Click and J. Moore.New Brunswick, N.J.: Rutgers University Press.

Braithwaite, R., and Lythcott, N.1989. “Community empowerment as a strategy for health promotion for Black and other minority populations.” Journal of the American Medical Association261, 282–283.

Breen, N., and Figueroa, J. B.1996. “Stage of breast and cervical cancer diagnosis in disadvantaged neighborhoods: A prevention policy perspective.” American Journal of Preventive Medicine12, 319–326.

Breslow, L.1972. “A quantitative approach to the World Health Organization definition of health: Physical, mental and social well-being.” International Journal of Epidemiology1, 347–355.


Broman, C. L., and Johnson, E. H.1988. “Anger expression and life stress among Blacks: Their role in physical health.” Journal of the National Medical Association80, 1329–1334.

Brown-Bryant, R.1985. “The issue of women's health: A matter of record.” Family and Community Health7, 53–65.

Brownstein, J. N., Cheal, N., Ackermann, S. P., et al. 1992. “Breast and cervical cancer screening in minority populations: A model for using lay health educators.” Journal of Cancer Education7, 321–326.

Bullough, B.1967. “Alienation in the ghetto.” Journal of Sociology72, 469–478.

Bullough, B.1972. “Poverty, ethnic identity, and preventive health care.” Journal of Health and Social Behavior13, 347–359.

Calnan, M., and Johnson, B.1985. “Health, health risks and inequalities: An exploratory study of women's perceptions.” Sociology of Health and Illness7, 55–75.

Centers for Disease Control. 1990. “Homicide among Black males—United States, 1978–1987.” Morbidity and Mortality Weekly Report39, 869–873.

Chatman, L. M., Billups, M. D., Bell, C. C., and Priest, M. L.1991. “Injury: A new perspective on an old problem.” Journal of the National Medical Association83, 43–48.

Christofel, K. K.1990. “Violent death and injury in U.S. children and adolescents.” American Journal of Diseases in Children144, 697–706.

Cohen, P., Struening, E., Muhlin, G., et al. 1982. “Community stressors, mediating conditions and wellbeing in urban neighborhoods.” Journal of Community Psychology10, 377–391.

Cohen, S., Schwartz, J. E., Bromet, E. J., and Parkinson, D. K.1991. “Mental health, stress, and poor health behaviors in two community samples.” Preventive Medicine20, 306–315.

Cope, N., and Hall, H.1985. “The health status of black women in the U.S.: Implications for health psychology and behavioral medicine.” Sage2, 20–24.

Cupoli, J., and Sewell, P.1988. “One thousand fifty-nine children with a complaint of sexual abuse.” Child Abuse and Neglect12, 151–162.

Davidson, J., and Smith, R.1990. “Traumatic experiences in psychiatric outpatients.” Journal of Traumatic Stress3, 459–475.

Dawkins, R., and Dawkins, M.1983. “Alcohol use and delinquency among Black, White and Hispanic offenders.” Adolescence18, 798–809.

DiClemente, R., and Wingood, G.1995. “A randomized controlled trial of an HIV sexual risk-reduction intervention for young African-American women.” Journal of the American Medical Association274, 1271–1276.

Dinges, N. G., and Joos, S. K.1998. “Stress, coping, and health: Models of interaction for Indian and native populations.” In Behavioral Health Issues among American Indians and Alaska Natives: Explorations on the Frontiers of the Biobehavioral Sciences.

Dodge, K. A., Lochman, J. E., Harnish, J. D., et al. 1997. “Reactive and proactive aggression in school children and psychiatrically impaired chronically assaultive youth.” Journal of Abnormal Psychology15, 37–51.


Dubrow, N., and Garbarino, J.1989. “Living in the war zone: Mothers and young children in a public housing development.” Child Welfare68, 3–20.

Durant, R. H., Cadenhead, C., Pendergrast, R. A., et al. 1994a. “Factors associated with the use of violence among urban Black adolescents.” American Journal of Public Health84, 612–617.

Durant, R. H., Getts, A., Cadenhead, C., et al. 1995. “Exposure to violence and victimization and depression, hopelessness, and purpose in life among adolescents living in and around public housing.” Developmental and Behavioral Pediatrics16, 233–237.

Durant, R. H., Pendergrast, R. A., and Cadenhead, C.1994b. “Exposure to violence and victimization and fighting behavior by urban Black adolescents.” Journal of Adolescent Health15, 311–318.

Ehrhardt, A. A., Exner, T. M., and Seal, D. W.1995. “The effectiveness of AIDS prevention efforts.” HIV prevention: State-of-the-science. Commissioned by the Office of Technology Assessment. Compiled and produced by the American Psychological Association Office on AIDS.

Elliot, B. A.1993. “Community responses to violence.” Primary Care20, 495–502.

Eng, E.1993. “The Save Our Sisters Project: A social network strategy for reaching rural Black women.” Cancer72, 1071–1077.

Eng, E., Hatch, J., and Callan, A.1985. “Institutionalizing social support through the church and into the community.” Health Education Quarterly12, 81–92.

Ferrence, R. G.1988. “Sex differences in cigarette smoking in Canada, 1900–1978: A reconstructed cohort study.” Canadian Journal of Public Health79, 160–165.

Franke, N. V.1997. “African American women's health: The effects of disease and chronic life stressors.” In Women's Health: Complexities and Differences, edited by S. B. Ruzek, V. L. Olesen, and A. E. Clarke. Columbus: Ohio State University Press.

Freeman, L. N., Mokros, H., and Poznanski, E.1992. “Violent events reported by normal urban schoolaged children: Characteristics and depression correlates.” Journal of the American Academy of Adolescent Psychiatry32, 419–423.

Freudenberg, N.1978. “Shaping the future of health education: From behavior change to social change.” Health Education Monographs6, 373–377.

Fullilove, M. T., Lown, A., and Fullilove, R. E.1992. “Crack 'hos and skeezers: Traumatic experiences of women crack users.” Journal of Sex Research29, 275–287.

Gall, S. B., and Gall, T. L., eds. 1993. Statistical Record of Asian Americans.Detroit: Gale Research.

Garbarino, J., Kostelny, K., and Dubrow, N.1991a. Children and Youth in Dangerous Environments: Coping with the Consequences of Community Violence.San Francisco: Jossey-Bass.

Garbarino, J., Kostelny, K., and Dubrow, N.1991b. “What children can tell us about living in danger.” American Psychologist46, 376–383.

Gardner, G. E.1971. “Aggression and violence—the enemies of precision learning in children.” American Journal of Psychiatry128, 77–82.

Glaser, J. B., Schachter, J., Benes, S., et al. 1991. “Sexually transmitted diseases in post pubertal female rape victims.” Journal of Infectious Diseases164, 726–730.


Golding, J. M.1994. “Sexual assault history and physical health in randomly selected Los Angeles women.” Health Psychology13, 130–138.

Gottlieb, B.1985. “Social networks and social support: An overview of research, practice, and policy implications.” Health Education Quarterly12, 5–22.

Gottlieb, N., and Green, L.1984. “Life events, social network, life-style, and health: An analysis of the 1979 national survey of personal health practices and consequences.” Health Education Quarterly11, 91–105.

Green, L.1982. “A learned helplessness analysis of problems confronting the Black community.” In Behavior Modification in Black Populations: Psychosocial Issues and Empirical Findings, edited by S. M. Turner and R. T. Jones.New York: Plenum.

Greenberg, M., Schneider, D., and Martell, J.1995. “Health promotion priorities of economically stressed cities.” Journal of Health Care for the Poor and Underserved6, 10–21.

Groat, H. T., and Neal, A. G.1967. “Social psychological correlates of urban fertility.” American Sociological Review32, 945–949.

Guyer, B., Lescohier, I., Gallagher, S. S., et al. 1989. “Intentional injuries among children and adolescents in Massachusetts.” New England Journal of Medicine321, 1564–1589.

Hammond, W. R., and Yung, B.1993. “Psychology's role in the public health response to assaultive violence among young African-American men.” American Psychologist48, 142–154.

Harel, Z.1986. “Older Americans act related homebound aged: What difference does racial background make?” Journal of Gerontological Social Work9, 133–143.

Hargreaves, M. K., Baquet, C., and Gamshadzahi, A.1989. “Diet, nutritional status, and cancer risk in American Blacks.” Nutrition and Cancer12, 1–28.

Harlow, C.1989. “Female victims of violent crime.” Bureau of Justice Statistics Special Report NCJ-126826. Washington, D.C.: U.S. Department of Justice.

Hinton-Nelson, M. D., Roberts, M. C., and Snyder, C. R.1996. “Early adolescents exposed to violence: Hope and vulnerability to victimization.” American Journal of Orthopsychiatry66, 346–353.

Hoppe, S., and Heller, P.1975. “Alienation, familism and the utilization of health services by Mexican-Americans.” Journal of Health and Social Behavior16, 304–314.

Hughes, H.1988. “Psychological and behavioral correlates of family violence in child witnesses and victims.” American Journal of Orthopsychiatry58, 77–90.

Hughes, M., and Demo, D. H.1989. “Self-perceptions of Black Americans: Selfesteem and personal efficacy.” American Journal of Sociology95, 132–159.

Humphrey, J., and Palmer, S.1986. “Race, sex and criminal homicide: Offendervictim relationships.” In Homicide among Black Americans, edited by D. Hawkins.Lanham, Md.: University Press of America.

Irwin, K. L., Edlin, B. R., Wong, L., et al. 1995. “Urban rape survivors: Characteristics and prevalence of human immunodeficiency virus and other sexually transmitted infections.” Obstetrics and Gynecology85, 330–336.


Israel, B. A.1982. “Social networks and health status: Linking theory, roles and practice.” Patient Counseling and Health Education4, 65–77.

Israel, B. A.1985. “Social networks and social support: Implications for natural helper and community level interventions.” Health Education Quarterly12, 65–80.

Jessor, J. E., and Donovan, R.1985. “Structure of problem behavior in adolescence and young adulthood.” Journal of Consulting and Clinical Psychology53, 890–904.

Kavanaugh, K. H., Harris, R. M., Hetherington, S. E., and Scott, D. E.1992. “Collaboration as a strategy for acquired immunodeficiency syndrome prevention.” Archives of Psychiatric Nursing6, 331–339.

Kersher, G., and McShane, M.1984. “The prevalence of child sexual abuse victimization in an adult sample of Texas residents.” Child Abuse and Neglect8, 495–501.

Kessler, R.1979. “Stress, social status, and psychological distress.” Journal of Health and Social Behavior20, 259–272.

Kessler, R., and Cleary, P.1980. “Social class and psychological distress.” American Sociological Review45, 463–478.

Kessler, R. C., and Magee, W. J.1994. “Childhood family violence and adult recurrent depression.” Journal of Health and Social Behavior35, 13–27.

Krieger, N.1990. “Racial and gender discrimination: Risk factors for high blood pressure.” Social Science and Medicine30, 1273–1281.

Leftwich, M. J. T., and Collins, F. L.1994. “Parental smoking, depression, and child development: Persistent and unanswered questions.” Journal of Pediatric Psychology19, 557–570.

Leigh, W.1994. “The health status of women of color.” Joint Center for Political and Economic Studies.

Lemp, G., Hirozawa, A., Givertz, D., et al. 1994. “Seroprevalence of HIV and risk behaviors among young homosexual and bisexual men.” Journal of the American Medical Association272, 449–454.

Lester, D.1986. The Murderer and His Murder: A Review of Research.New York: AMS Press.

Levine, D. M., Becker, D. M., and Bone, L. R.1992. “Narrowing the gap in health status of minority populations: A community-academic medical center partnership.” American Journal of Preventive Medicine8, 319–323.

Lex, B. W.1991. “Some gender differences in alcohol and polysubstance users.” Health Psychology10, 121–132.

Liem, R., and Liem, J.1978. “Social class and mental illness reconsidered: The role of economic stress and social support.” Journal of Health and Social Behavior19, 139–156.

Lillie-Blanton, M., Anthony, J., and Schuster, C.1993. “Probing the meaning of racial/ethnic group comparisons in crack cocaine smoking.” Journal of the American Medical Association269, 993–997.

Link, B. G., and Phelan, J.1995. “Social conditions as fundamental causes of disease.” Journal of Health and Social Behavior36 (Suppl.), 80–94.

Lorion, R. P., and Saltzman, W.1993. “Children's exposure to community violence: Following a path from concern to research to action.” Psychiatry56, 55–65.


Lowry, R., Kann, L., Collins, J. L., and Kolbe, L. J.1996. “The effect of socioeconomic status on chronic disease risk behaviors among US adolescents.” Journal of the American Medical Association276, 792–797.

Makuc, D. M., Fried, V. M., and Kleinman, J.1989. “National trends in the use of preventive health care by women.” American Journal of Public Health79, 21–26.

Manson, S. M., and Dinges, N. G., eds. 1988. “American Indian and Alaska native mental health research.” Journal of the National Center Monograph Series1, 9–64.

Martin, S. L., English, K. T., Clark, K. A., et al. 1996. “Violence and substance abuse among North Carolina pregnant women.” American Journal of Public Health86, 991–998.

Martinez, P., and Richter, J. E.1993. “The NIMH community violence project: II. Children's distress symptoms associated with violence exposure.” Psychiatry56, 22–35.

Martinez, R. M., and Lillie-Blanton, M.1996. “Why race and gender remain important in health services research.” American Journal of Preventive Medicine12, 316–318.

McGinnis, J.1986. “The 1985 Mary E. Switzer Lecture: Reaching the underserved.” Journal of Allied Health15, 293–305.

McLoyd, V. C.1990. “The impact of economic hardship on Black families and children: Psychological distress, parenting, and socioemotional development.” Child Development61, 311–346.

Mercy, J., Goodman, R., Rosenberg, M., et al. 1986. “Patterns of homicide victimization in the City of Los Angeles.” Bulletin of the New York Academy of Medicine62, 427–445.

Mindlin, R., and Densen, P.1971. “Medical care of urban infants: Health supervision.” American Journal of Public Health61, 687–697.

Minkler, M.1992. “Community organizing among the elderly poor in the United States: A case study.” International Journal of Health Services22, 303–316.

Morris, N. M., Hatch, M. H., and Chipman, S. S.1966. “Alienation as a deterrent to well-child supervision.” American Journal of Public Health56, 1874–1882.

O'Carroll, P. W.1988. “Homicides among black males 15–24 years of age, 1970–1984.” Morbidity and Mortality Weekly Report37 (SS-1), 53–60.

Ogbu, J.1983. “Minority status and schooling in plural societies.” Comparative Education Review27, 168–190.

Orpinas, P. K., Basen-Engquist, K., Grunbaum, J. A., and Parcel, G. S.1995. “The comorbidity of violence-related behaviors with health-risk behaviors in a population of high school students.” Journal of Adolescent Health16, 216–225.

Orr, S. T., James, S. A., Miller, C. A., et al. 1996. “Psychosocial stressors and low birth-weight in an urban population.” American Journal of Preventive Medicine12, 459–466.


Osofsky, J.1995. “The effects of exposure to violence on young children.” American Psychologist50, 782–788.

Pearlman, D. N., Rakowski, W., Ehrich, B., and Clark, M. A.1996. “Breast cancer screening practices among Black, Hispanic, and White women: Reassessing differences.” American Journal of Health Promotion12, 327–337.

Pincus, T., and Callahan, L. F.1995. “What explains the association between socioeconomic status and health: Primary access to medical care or mindbody variables?” Advances: The Journal of Mind-Body Health11, 4–36.

Rainwater, L.1960. And the Poor Get Children.Chicago: Quadrangle Books.

Rakowski, W.1988. “Predictors of health practices within age-sex groups: National survey of personal health practices and consequences, 1979.” Public Health Reports103, 376–386.

Rask, K. J., Williams, M. V., Parker, R. M., and McNagny, S. E.1994. “Obstacles predicting lack of a regular provider and delays in seeking care for patients at an urban public hospital.” Journal of the American Medical Association271, 1931–1933.

Resnick, H., Falsetti, S., Kilpatrick, D., and Freedy, J.1986. “Assessment of rape and other civilian trauma-related post-traumatic stress disorder: Emphasis on assessment of potentially traumatic events.” In Stressful Life Events, edited by T. W. Miller.Madison, Conn.: International Universities Press.

Resnick, H., Kilpatrick, D., Dansky, B., et al. 1993. “Prevalence of civilian trauma and posttraumatic stress disorder in a representative national sample of women.” Journal of Consulting and Clinical Psychology61, 984–991.

Rinear, E.1988. “Psychological aspects of parental response patterns to the death of a child by homicide.” Journal of Traumatic Stress1, 305–322.

Rodin, J.1986. “Aging and health: Effects of the sense of control.” Science233, 1271–1276.

Rodin, J., and Ickovics, J.1990. “Women's health: Review and research agenda as we approach the 21st century.” American Psychologist45, 1018–1034.

Rosenberg, M.1987. “Children of battered women: The effects of witnessing violence on their social problem-solving abilities.” The Behavior Therapist10, 85–89.

Rosenberg, M. L., and Fenley, M. A., eds. 1991. Violence in America: A Public Health Approach.New York: Oxford University Press.

Ross, C. E., and Mirowsky, J.1989. “Explaining the patterns of depression: Control and problem solving—or support and talking.” Journal of Health and Social Behavior30, 206–219.

Rotheram-Borus, M. J., Mahler, K. A., Koopman, C., and Langabeer, K.1996. “Sexual abuse history and associated multiple risk behavior in adolescent runaways.” American Journal of Orthopsychiatry66, 390–400.

Ruzek, S. B., Clarke, A. E., and Olesen, V. L.1997a. “What are the dynamics of differences?” In Women's Health: Complexities and Differences, edited by S. B. Ruzek, V. L. Olesen, and A. E. Clarke. Columbus: Ohio State University Press.

Ruzek, S. B., Olesen, V. L., and Clarke, A. E., eds. 1997b. Women's Health: Complexities and Differences.Columbus: Ohio State University Press.


Salber, E. J.1979. “The lay health advisor as a community health resource.” Journal of Health Politics, Policy and Law3, 469–479.

Salber, E. J.1994a. “Correlates of healthy eating habits in low-income Black women and Latinas.” Preventive Medicine23, 781–787.

Salber, E. J.1994b. “Health promotion behavior in low-income Black and Latino women.” Women and Health21, 71–83.

Salber, E. J.1996a. “Correlates of health promotion behaviors in low-income, Black and Latino women.” Journal of Preventive Medicine12, 450–458.

Salber, E. J.1996b. “The ecology of urban violence: Its relationship to health promotion behaviors in Black and Latino communities.” American Journal of Health Promotion10, 88–97.

Schaefer, N., Falciglia, G., and Collins, R.1990. “Adult African-American females learn cooperatively.” Journal of Nutrition Education22, 240D.

Seeman, M.1959. “On the meaning of alienation.” Sociological Review24, 783–791.

Seeman, M., and Evans, J.1962. “Alienation and learning in a hospital setting.” American Sociological Review27, 772–782.

Seeman, M., and Seeman, T.1983. “Health behavior and personal autonomy: A longitudinal study of the sense of control in illness.” Journal of Health and Social Behavior24, 144–160.

Sheley, J., McGee, Z., and Wright, J.1992. “Gun-related violence in and around inner city schools.” American Journal of Diseases of Childhood146, 677–682.

Shumaker, S. A., and Hill, D. R.1991. “Gender differences in social support and physical health.” Health Psychology10, 102–111.

Singer, L., Arendt, R., and Minnes, S.1993. “Neurodevelopmental effects of cocaine.” Clinics in Perinatology20, 245–262.

Singer, M. I., Anglin, T. M., Song, L., and Lunghofer, L.1996. “Adolescents' exposure to violence and associated symptoms of psychological trauma.” Journal of the American Medical Association273, 477–482.

Sluzki, C.1993. “Toward a model of family and political victimization: Implications for treatment and recovery.” Psychiatry56, 178–187.

Smith, J., Mercy, J., and Rosenberg, M.1986. “Suicide and homicide among Hispanics in the Southwest.” Public Health Reports101, 265–270.

Smith, T.1992. “Hostility and health: Current status of a psychosomatic hypothesis.” Health Psychology11, 139–150.

Snyder, C. R.1994. The Psychology of Hope: You Can Get There from Here.New York: The Free Press.

Sonnenberg, S. M.1988. “Victims of violence and post-traumatic stress disorder.” Psychiatric Clinics of North America11, 581–590.

Sorenson, S., and Shen, H.1996. “Homicide risk among immigrants in California, 1970 through 1992.” American Journal of Public Health86, 97–100.

Spivak, H., Hausman, A., and Prothrow-Stith, D.1989. “Practitioner's forum: Public health and the primary prevention of adolescent violence—The violence prevention project.” Violence and Victims4, 203–212.

Stevens-Simon, C., and Reichert, S.1994. “Sexual abuse, adolescent pregnancy, and child abuse: A developmental approach to an intergenerational cycle.” Archives of Pediatrics and Adolescent Medicine148, 23–27.


Stokols, D.1992. “Establishing and maintaining healthy environments: Toward a social ecology of health promotion.” American Psychologist47, 6–22.

Sumner, B., Mintz, E., and Brown, P.1986. “Interviewing persons hospitalized with interpersonal violence-related injuries: A pilot study.” In Report of the Secretary's Task Force on Black and Minority Health, Volume 5.Washington, D.C.: U.S. Department of Health and Human Services.

Syme, S., and Berkman, J.1976. “Social class, susceptibility and sickness.” American Journal of Epidemiology18, 635–643.

Tardiff, K., and Gross, E.1986. “Homicide in New York City.” Bulletin of the New York Academy of Medicine62, 413–426.

“U.S. Department of Health and Human Services.” 1986, January. Homicide, suicide, and unintentional injuries, volume 5. In Report of the Secretary's Task Force on Black and Minority Health.Washington, D.C.: U.S. Department of Health and Human Services.

“U.S. Department of Health and Human Services, Public Health Service, Health Resources and Services Administration.” 1991. Health Status of Minorities in Low Income Groups. DHHS Publication HRS-D-DV 85-1, 3rd ed.

“U.S. Department of Justice.” 1991. Criminal victimization, 1990. Special Report NCJ-122743. Washington, D.C.: Bureau of Justice Statistics.

Vega, W., Sallis, J., Patterson, T. R., et al. 1988. “Predictors of dietary change in Mexican-American families participating in a health behavior change program.” American Journal of Preventive Medicine4, 194–199.

Wallerstein, N.1992. “Powerlessness, empowerment, and health: Implications for health promotion programs.” American Journal of Health Promotion6, 197–205.

Warnecke, R. B., Graham, S., Mosher, W., et al. 1975. “Contact with health guides and use of health services among Blacks in Buffalo.” Public Health Reports90, 213–222.

Warnecke, R. B., Graham, S., Mosher, W., and Montgomery, E.1976. “Health guides as influentials in central Buffalo.” Journal of Health and Social Behavior17, 22–34.

Weinstein, N. D.1993. “Testing four competing theories of health-protective behavior.” Health Psychology12, 324–333.

Weisz, J., Martin, S., Walter, B., and Fernandez, G.1991. “Differential prediction of young adult arrests for property and personal crimes: Findings of a cohort follow-up study of violent boys from North Carolina's Willie M. program.” Journal of Child Psychology and Psychiatry32, 783–792.

Wilbanks, W.1986. “Criminal homicide offenders in the U.S.: Black vs. White.” In Homicide among Black Americans, edited by D. Hawkins. Lanham, Md.: University Press of America.

Wilkerson, J., Protinsky, H. O., Maxwell, J. W., and Lentner, M.1982. “Alienation and ego identity in adolescents.” Adolescence17, 133–139.

Williams, D., Lavizzo-Mourey, R., and Warren, R.1994. “The concept of race and health status.” Public Health Reports109, 26–41.


World Health Organization. 1986. “Report of the working group on concept and principles of health promotion, 1984.” Health Promotion1, 73–76.

Wyatt, G. E.1992. “The sociocultural context of African American and White American women's rape.” Journal of Social Issues48, 77–91.

Zapata, B. C., Rebolledo, A., Atalah, E., et al. 1992. “The influence of social and political violence on the risk of pregnancy complications.” American Journal of Public Health82, 685–690.

Zeiler, S., Wotbeck, B., and Mayer, K.1996. “Sexual violence against women living with or at risk for HIV infection.” American Journal of Preventive Medicine12, 304–310.



Theodore G. Ganiats
and William J. Sieber

As shown throughout this volume, promoting human wellness involves a wide variety of themes. These themes are usually operationalized through either large groups (e.g., health policy) or on a more micro level (e.g., an individual's behavior). Sometimes this involves formal programs (The Great American Smokeout, Medicare approval of screening mammography). At other times wellness is promoted by supporting an individual's decision to change behavior. Those of us interested in promoting human wellness decide where best to place our efforts.

To make sound decisions, whether they be related to health policy or with an individual patient or client, we must consider all relevant factors. Therefore, an examination of both the methodologies used in quantitative health policy analysis and an individual's choices and the behavior resulting from these choices is essential. Quantitative health policy analysis, an explicit method of evaluating benefits and harms, serves as a policy tool for government and health policy agencies. One element of this tool is particularly influential; in fact, analysis can actually present health policy agencies with conflicting conclusions depending on how an analyst values future health outcomes. Similarly, we argue that the valuation of future health outcomes affects an individual's behavior, but the major models of human health behavior do not account for this time preference.

In this chapter, we explore how the valuing of future health affects both policy analysis and individual health behavior. We begin with a review

of the basic principles of quantitative health policy analysis, paying special attention to the concept of discounting (a method of determining the present value of a future outcome). We follow with a review of research on the psychology of discounting. Finally, we provide an overview of two major psychological theories that deal with health behaviors and demonstrate that these theories fail to account for individual differences in the valuation of future health outcomes. In the final part of the chapter, we discuss possible effects of adding time preference to both levels of analysis and propose areas for future research and theory development.


A legislature dictates which health care benefits to offer Medicaid recipients. An insurance company decides which benefits to offer its various health plan subscribers. A hospital develops rules that dictate how and when to use various hospital resources. These are the realities of health care policy, manifestations that touch each of us at one time or another.

Policy formulation reflects a host of decision-making techniques. Sometimes political action groups work to ensure that certain health benefits are legislated. Sometimes experts convene and arrive at a consensus. Another method is to list the goals of a health system and then rank programs, giving the greatest weight to those most likely to meet system goals.

This last method, the quantitative approach to health policy formulation, has several advantages over alternatives. It is less susceptible to special interests, and it is more likely to optimize the health of the target population. In addition, when health care resources are limited, the quantitative approach is most likely to optimize their use.1

Modeling Health and Resource Use

Examining one such quantitative approach demonstrates the principles of quantitative analysis in health policy development. In the General Health Policy Model, a pioneering effort in health policy formulation,1–3 any health program has two outcomes: health status and resource use. Using this model, an analyst assumes that the goal of the health care system is to improve health, measured as both life expectancy and quality of life (alternatively stated as “adding years to life and life to years”).1,4


Figure 12.1. The time-wellness graph in which quality of life and time are combined to produce a new health outcome: the quality-adjusted life-year (QALY).

The system achieves this goal by implementing programs that use health care resources, usually described in dollars. If there are budget constraints, the goal is pursued through the efficient use of limited resources.

An analyst estimating resource use evaluates how much the program costs (e.g., in terms of medications, hospitalization, and complications) and how much the program saves (perhaps measuring decreased future health care, decreased risk of future infection, and increased productivity by returning to work). Similarly, when evaluating health, an analyst estimates how the program improves quality of life (decreased symptoms or increased functioning), decreases quality of life (side effects from medications or complications from surgery), and affects mortality.

The resources consumed and saved are merely opposite sides of the same economic coin. Together, they reflect the net resource utilization. In contrast, the two components of health care outcomes are fundamentally different: Quality of life and life expectancy are not directly additive. These two components of health can, however, be combined to create a single health outcome measure for quantitative analysis.

Figure 12.1 shows the integration of quality of life and life expectancy on a single graph, with “time” on the horizontal axis and “wellness,” or quality of life, on the vertical axis. The healthier the person, the higher the line; the longer the person lives, the longer the line. To see how well a program succeeds in meeting the goal of prolonging life while increasing quality of life, one merely measures the area displayed under this “time-wellness” plot. Larger areas depict longer, more healthy lives. The area is in the units of the “quality-adjusted life year” (QALY).5


Rationing: New Name, Familiar Choices

Each of us hopes for high QALYs. While a quantitative approach to health policy can lead to rationing, such rationing is not a new threat to health care. Our present system relies on rationing to allocate health resources. When a patient needs admission to a full intensive care unit (ICU), a physician decides who will be denied an ICU bed. Because of the short supply of organs for transplantation, organs are routinely rationed among transplant candidates. During the Persian Gulf War, a shortage of gamma globulin required physicians to decide not who needed the gamma globulin but who needed it most. Rationing is a fact of life in health care. A quantitative approach to policy simply illuminates choices related to rationing—choices that are frankly uncomfortable and politically sensitive. This is why we must scrutinize the tenets and perspective of any system used to set health policy.

Perspective: A Societal Vantage Point

Any health policy analysis assumes some perspective. Think of it as varying according to where you sit. Behind a corporate desk, the perspective may be purely financial, aimed at maximizing profit regardless of the health outcome. From a kitchen table, a patient with excellent health insurance may be most concerned with health benefits since his wallet is not involved. As a society, we usually choose a balance between these two perspectives. We want to optimize health for all, but we set limits when resources are strained. The first part of this chapter represents this group perspective; we will assume that the group wishes to use societal resources to improve the average health of the population.


So far, the description of the quantitative approach to health care has been straightforward: We add benefits (positive health outcomes and dollars saved), subtract costs (adverse health outcomes and dollars spent), and evaluate the balance. In practice, however, this is an oversimplification.

Psychologists have long recognized that people have a time preference for outcomes. Simply stated, time preference recognizes that most

Years 5% Discount Rate 10% Discount Rate
1 0.952 0.909
2 0.907 0.826
3 0.864 0.751
4 0.823 0.683
5 0.784 0.621
10 0.614 0.386
15 0.481 0.239
20 0.377 0.149
25 0.295 0.092
30 0.231 0.057
40 0.142 0.022
of us prefer a present reward over a future reward of equal value. Similarly, we usually prefer a future penalty over a present penalty of equal weight. Adjusting for this time preference is a standard element of current quantitative health policy analysis. This adjustment is called “discounting.” 5

A simple example illustrates discounting. As a lottery winner, would you choose a $10,000 cash prize today or in one year? Most of us would choose cash today. Even if the future award were adjusted for inflation (e.g., if inflation were 5%, the $10,000 would grow to $10,500), most of us would still prefer our winnings now. This choice implies that we value a future dollar less than a present one. In other words, we discount the value of future dollars.

The process of calculating the present value of future dollars by discounting is, mathematically, the opposite of calculating the future value of present dollars with interest. For example, $10,000 invested at 5% interest will be worth $10,500 ($10,000 0.05 $10,000) after one year. In contrast, if the discount rate is 5%, then $10,000 a year from now is currently valued at $9,500 ($10,000 [0.05 $10,000]). Just as compound interest offers dramatic returns after a few years, the compounded effects of discounting can have a marked effect on an economic analysis; the further into the future the outcome occurs, the greater the impact of discounting.

Table 12.1 demonstrates this effect. The discount multiplier for 25 years at a 5% discount rate is 0.295. Someone with a 5% discount rate for future dollars would consider $10,000 received in 25 years to have a

current value of 0.295 $10,000, or $2,950. In other words, this person would be indifferent to receiving $2,950 today or $10,000 in 25 years.

Discounting's Economic Rationale

Why do most of us demonstrate this positive time preference for money, preferring $10,000 in our pockets to comparable funds in a year (even after adjusting for inflation)? The answer is multifactorial.5–14 First, we could invest the money and earn more than inflation. Second, we could spend money in hand and enjoy the benefits of the dollars now rather than in the future. Third, many of us do not trust the future. Maybe we are not sure that we will be alive in one year and want to enjoy the money now. Maybe we suspect that the source of the $10,000 may default on the dollars owed.

The second and third factors indicate that our preference for today's dollar is not simply a matter of its investment potential. They demonstrate other human factors at work. Thus, discounting is based on multifactorial preferences for present over future dollars.

Economists, while comfortable with the concept of discounting, do not agree on what this discount rate should be. In cost-effectiveness research, proposed discount rates vary from 0% to 20%.5,11,15,16 Typically, analysts use a discount rate that approximates the prevailing interest earned on investments, and the recent consensus report suggests a discount rate of 3%. Consideration of the previously noted human factors suggests that this is a low estimate.

Discounting Future Health

In a quantitative health policy analysis, one usually assumes that the economic and health discount rates are equal.5,6,15 This assumption causes an apparent prevention program paradox: While most prevention programs are intuitively cost-effective, many do not fare well in a traditional cost-effectiveness analysis. The explanation for the paradox is clear—prevention programs generate expenditures today, but related health outcomes occur in the future. The ratio of full-valued present dollar costs to discounted future health benefits creates the unfavorable costeffectiveness ratios so often calculated for prevention programs.

Table 12.2 shows the effect of discounting on cost-effectiveness in two prevention programs. It demonstrates that discounting has the expected, and at times marked, effect on the calculated cost-effectiveness of prevention

Program Cost-Effectiveness

aFuture years of life were not quality adjusted in this study.

bThe exact assumptions are not important to illustrate the point here. The reader is referred to the original article for the details.

Universal neonatal hepatitis
B immunization
$38,632 per life-year[a] $3,066 per life-year[a]
Universal adult hepatitis
B immunization
$257,418 per life-year[a] $54,524 per life-year[a]
Neonatal circumcision
(Assumption A[b]
$67,402 per QALY $33,156 per QALY
Neonatal circumcision
(Assumption B[b]
N/A (it costs dollars and decreases QALYs) $8,161 per QALY
programs. This calculation can also affect health policy. For example, if we are willing to spend up to $70,000 to extend life by one year (a reasonable assumption given current health policy), the adult hepatitis B immunization program is cost-effective (at $54,000 per life-year) if we do not discount future health. In other words, if we are willing to spend up to $70,000 in order to get one life-year, then we would find a program costing only $55,000 per life-year to be acceptable. On the other hand, the undiscounted program (at $257,000 per life-year) costs more than $70,000 per life-year and should not be considered cost-effective. In the Assumption B circumcision example, the procedure moves from “charlatan” territory (spending money for diminished health) to a highly cost-effective procedure, depending on whether the analyst discounts future health outcomes. These illustrations offer a graphic demonstration that discounting is not an esoteric topic but is central to costeffectiveness analysis.

The most often quoted rationale for keeping the economic and health discount rates the same is the work of Keeler and Cretin.6 They state that if the health discount rate is lower than the monetary discount rate, then no particular program should ever be implemented because the program will become more cost-effective by postponing it for one year. This is the result of the dollar numerator decreasing faster, under the influence of the larger discount rate, than the health outcome denominator: The calculated ratio must be smaller (i.e., more cost-effective) with each successive year. Others note that by keeping the monetary and health discount

rates equal, we ensure that the future relationship between the valuation of health and dollars remains constant and the same as today.5

These arguments may not be valid since there are significant differences between dollars and health. For example, dollars can be invested and grow in value over time. Health not only cannot be invested but in most people remains constant or slowly decreases over time. In addition, wealth can be spent to obtain some other commodity, but you cannot “spend” health. In fact, health and wealth are so fundamentally different that no evidence suggests that people uniformly discount them at a similar rate.

Further investigation reveals other flaws in the arguments for keeping economic and health discount rates identical. In the argument by Keeler and Cretin, health care programs are postponed because next year they will be more efficient. The math may be correct, but the argument is not relevant. Society does not limit its health resources to only the most efficient program. Instead, we put resources into programs that have an acceptable level of cost-effectiveness. Today's cost-effective program may be more cost-effective next year, but it is still acceptable compared with current (as opposed to future) alternatives. We implement sufficiently efficient programs this year, even if they would be more efficient next year.17

The second argument, that the two rates must be equal to preserve a constant relation between dollars and health, is also flawed.18 There is no a priori reason to insist on this constancy. A nation's health and its wealth may be dynamic. A change in either may modify the relationship between the valuation of future health and future dollars. For example, there may be a general perception that the health of the nation is improving but the economy is stable. Here the ratio of the present value of future health to the present value of future dollars may be greater than the ratio of today's valuation of present health to today's valuation of present dollars. If this is the case, the discount rate for future health should be less than the discount rate for future dollars. The converse, of course, holds if the general sense is that the prospects for future health are declining. In addition, there is evidence that as a country's wealth increases, its residents spend more of the gross national product on health care.19 This implies that the monetary valuation of health is not an inherent constant but can vary as a function of gross national product. Without constant health valuation, there is no reason to believe that health and economic discount rates remain equal.

Most research on the psychology of discounting (see the next section), though applicable to individual human behavior, has limited usefulness


Figure 12.2 Example of a paradox: The effect of discounting life expectancy.

for health policy. For example, most studies assess the discount rate of individuals as opposed to that of a population. This approach may have less relevance in health policy, where population-based data may be more important than data that can be applied to only one individual. Another limitation of current research is that most attempts to assess an individual's discount rate are time consuming and not easily applied to large populations. There are elements of the present research that is limited from both the individual health behavior and the health policy perspectives. For example, there is a tendency in current research to dismiss individuals who claim a negative discount rate for health (i.e., those who prefer future health to present health). Finally, this research tends to focus on graduate students and not patients. The implications of this focus are not clear, but it is reasonable to assume that the preferences of a graduate student will differ from those of a patient with a serious acute or chronic condition. For example, it is unlikely that graduate students would display the dynamic inconsistency that Christensen-Szalanski found in pregnant women.30

Finally, Figure 12.2 shows the effect of discounting health outcomes over time and highlights another apparent paradox induced by current models. The current discounting model tells us that the more distant in

the future an outcome occurs, the less it is valued. Applying this premise to current life expectancy in the United States, we can see that the 73-year life expectancy of a newborn has a present value of 20 years, while the 5.3-year life expectancy of an 85-year-old is valued as 4.7 years. In other words, according to the current discounting model, the life expectancy of a newborn is valued only 15 years longer (20 years 4.7 years)—only about four times as long—than the life expectancy of an 85-year-old.


Thus far, this chapter has focused on the impact of discounting on the methods used in quantitative health policy analysis. While discounting can have a major effect on a cost-effectiveness analysis, it may also have a direct effect on individual behavior. Research in economics and psychology has identified several important characteristics of discounting in economics.21

High Rates Subjects often report very high discount rates. While this is certainly possible, it does fly in the face of current economic practice. Perhaps more significantly, applying these high discount rates to the health outcomes of a prevention program all but ensures that the program will not be cost-effective (our earlier discussion demonstrated the accrual of discounted future benefits while consuming resources in the undiscounted present).

Dynamic Inconsistency Our discount rates can change as the time to the reward varies.21–23 For example, one may prefer $200 in eight years over $100 in six years (discount rate 41%). Six years from now, the same person might prefer $100 right away instead of doubling the money in two years (discount rate 41%).21

Magnitude Effect Some investigators have found that discount rates are lower for large magnitude outcomes.22–24 For example, someone may prefer $10 now over $15 in a year but may prefer $15,000 in a year over $10,000 now, even though both involve the same relative change.

Sign Effect The framing of a question may be important since discount rates may be lower for losses than for gains.23–26


Sequence Effect Investigators have found that discount rates are not additive. Subjects confronted with a series of decisions (each with its own discount rate) often do not make the same choice as when the decision is presented as a single problem.25,27,28

In contrast to this body of research in economic discounting, little work has been done to address the psychology of discounting health 14,20,29,30 There is some evidence that dynamic inconsistency exists,20 although this is not a universal finding.29 In one example, Christensen-Szalanski studied labor anesthesia preferences among pregnant women.20 A month before and a month after delivery, the women wanted to avoid anesthesia; during labor, however, many of these same women were much more favorably disposed toward anesthesia. Like economic outcomes, the psychology of discounting health outcomes is likely multifactorial.


A person's time preference affects his choices, and these choices in part determine health behavior. It is therefore surprising to find that dominant psychological models of health behavior fail to consider time preference. In this section we illustrate this failing by highlighting two major psychological models of individual health behavior. We then offer directions for incorporating the concept of discounting into future work in the field.

Over the past two decades, two models have dominated the psychological literature on explaining individual health behavior. The Health Belief Model 31,32 postulates that before an individual engages in a healthpromoting behavior, several factors influence the initiation of action. The model has served as the theoretical basis for hundreds of health education research studies and helped in the formulation of innumerable interventions. The Transtheoretical Model33,34 uses stages of change to help identify where an individual is in the change process and match the processes most effective in each of these stages with the individual to maximize further change. It has been used primarily in designing proactive interventions, particularly assisting in the cessation of behaviors such as cigarette smoking. Better understanding of each of these models and how discounting may affect each model's ability to predict health behavior is needed.

The Health Belief Model has been used across the health continuum,

from prevention to detection to illness behavior. It is appealing to a wide range of professionals in designing and evaluating interventions to alter health behavior. The Health Belief Model was developed in an effort to explain the widespread failure of people to participate in programs to prevent or detect disease 35 but was later extended to address compliance with medical regimens.31 In this model health behavior is seen as resulting from a combination of several necessary factors: an individual's perceived susceptibility to illness, perceived severity of the illness, perceived benefits of action to reduce vulnerability (i.e., health-promoting behavior), and an evaluation of potential barriers to the proposed action. Lastly, a relevant stimuli is required to trigger health action (i.e., perceived body dysfunction, advice from a physician), though this is the least understood component of the model. More recently, Bandura's concept of self-efficacy has been incorporated into the model to help better explain individual differences observed in taking action to improve one's health. However, what is missing is the importance (or value) of the health benefit of taking such action.36

There are components of the Health Belief Model that address this issue indirectly. Namely, perceived illness severity is used as a measure of the aversiveness of the outcome should no change occur in health behavior. This aversiveness may measure motivation to act. However, a more direct measure of the value that an individual places on the health gain at some time in the future may be a more accurate predictor of behavior than rating of disease severity without reference to time. A second component of the model, perceived benefits of action to reduce vulnerability, assesses the perceived effectiveness of the behavior change in reducing the noxious outcome (i.e., disease). Again, the expected efficacy of the behavior on an outcome is not directly related to the value of the outcome. While two individuals may rate the efficacy of a behavior in producing an outcome to be equal, they may differ greatly in how valuable that outcome is. For example, even though a person may see the “severity” of cancer to be significant and the perceived benefit of action (i.e., quit smoking) as being effective in reducing the vulnerability to cancer, the value of life without cancer may be very different for a 70-year-old male with other multiple health problems as compared to the value of a cancer free life as rated by a 30-year-old adult with no other health problems. Also, while a behavior change may be seen as effective in producing some outcome, significant differences in the delay in which benefits are realized may affect the value of that outcome. For example, greater motivation to quit smoking would exist if benefits were derived

within a matter of days or hours as compared to describing only those benefits derived from years of abstinence. Adding the value of the outcome (accounting for discounting) to the Health Belief Model may better assess the motivation of an individual to engage in a health-promoting behavior. Discounting of health may vary between individuals—as well as across time for the same individual—in predictable ways and thus add to our understanding of why some individuals engage in healthpromoting behavior under some circumstances while others do not.

The Transtheoretical Model emerged from a comparative analysis of leading theories of psychotherapy and behavioral change. Most of the early research on this model compared cigarette smokers who successfully stopped cigarette smoking as a result of a structured smoking cessation program to those who were unsuccessful.37 This research, and several studies that followed, revealed that behavioral change unfolds through a series of stages. Innumerable studies have used this model to explain the success (and failure) of specific programs designed to promote healthy behavior. The stages described in this model have been labeled precontemplation, contemplation, preparation, action, and maintenance. Precontemplation is used to describe that period in which a person has no intention to change behavior in the next six months, contemplation involves an intent to change behavior within six months, and preparation involves the intent to change behavior within 30 days and may include some initial change in health behavior. The action stage is defined as having implemented some behavior change within the previous six months, while maintenance is defined as overt behavior change having lasted for more than six months. Prochaska et al. confirmed that the Transtheoretical Model also generalized across 12 problem behaviors, including cessation of negative behavior (e.g., smoking) as well as the acquisition of positive behavior (e.g., exercise, mammography screening).38 This model proposes that people are at different stages in the change process and that matching different processes with the stage in which an individual is at should maximize transition to a more advanced stage of behavior change. Recent studies have demonstrated that this matching does maximize treatment effectiveness (e.g., reference 39).

While matching of the stages to intervention types has shown to be effective, less is known about facilitating transition between stages, particularly from precontemplation to contemplation to preparation.40 While these authors argue that research is needed to determine the consistency of the crossover (one stage to the next) occurring prior to the action stage, gathering evidence suggests that many people transition to

the next stage when the pros of changing the behavior outweigh the cons. In a number of studies, the ratio of pros to cons (or benefits to harms) of performing a certain behavior predicts subsequent behavior. Initial suggestions are that interventions designed to help a person progress from precontemplation to contemplation should target increasing the pros of changing and that only later should the focus be on reducing the negatives of changing. The value of a health outcome may play a significant role in a person's view of the pros of behavior change. That is, while improved health is often assumed to be a valuable benefit of behavior change, differences in the value assigned to health may differ significantly between individuals and over time.

Variables that affect the value of health need to be better understood. For example, most exercise programs are designed for people who have decided to exercise (e.g., preparation or action stage), yet most people do not exercise, and few have little interest in starting.41,42 This may reflect the fact that the cons of beginning an exercise program (e.g., aches, effort) outweigh the pros of exercising (e.g., improved cardiovascular health) because the health outcome is placed in the future where its value is discounted. Namely, if the benefits of exercise focused on are primarily experienced within six months of initiating the exercise program, the value of this benefit is likely to be more motivating than if the benefits focused on would not be realized for five years. One reason we believe that people differ in their engaging in health behavior is that the value of the health benefit differs between individuals and is affected by how far in the future it is to be realized. Transition from one stage in the Transtheoretical Model to the next may be more the result of changing values placed on future health than of reducing barriers to change. Clearly, more must be known about how individuals differ in their perception of benefits (i.e., future health) before we can examine when the balance between the pros and cons favors behavior change.

While these two models have dominated the field of health psychology and have been successfully used to understand the process by which individuals change behavior, both models do not clearly address the motivational component of change. Namely, people change behavior because the pros of change outweigh the cons, and in fact there has been an unchallenged assumption that for individuals to change, they must value health in the future (when change will supposedly result in better health). However, how people value health in some distant future is anything but clear.



It is apparent that the scientific inquiry into time preference has been somewhat neglected. Most health policy research has focused on the limited view of the normative model that future health and future financial outcomes must be discounted at the same rate. Much of the research in the field assumes the normative stance. Further, a large amount of the research focuses on populations that are not representative of the population at large. Major psychological models of human health behavior similarly neglect or minimize the issue of time preference—all of this despite the importance of time preference in individual and group decision making.

Work is needed on the appropriate discount rate for policy analysis, especially of prevention programs. The current recommendations of the U.S. Department of Health and Human Service's Panel of Cost Effectiveness in Health and Medicine 43 are that future health and financial outcomes be discounted at the same rate. While this is the current state of the art, it may not be the final word. Prevention programs are disadvantaged by this approach since the bulk of the expenses (both dollar investments and adverse health effects) occur in the undiscounted present, while benefits (cost savings and improved health) accrue in the discounted future. If in fact people do discount preventive care, this approach is appropriate. However, prevention specialists argue, and the growing number of people living a healthier lifestyle support, that prevention programs may be valued higher than treatment programs.

This point can be illustrated with a discussion of routine Papanicolaou (Pap) smears. The chance that a woman will benefit from any one Pap smear is quite small, as are the risks associated with that Pap smear. Still, women routinely disrupt their daily routines to have this test performed; for these women the chance of future health benefits outweighs the certainty of the current inconvenience. The benefits also include other factors, such as the peace of mind afforded by a negative examination. In many ways these benefits increase with time, so discounting these benefits may be counter to the patient's true preferences. If the analyst does not discount future outcomes, the future benefits of the Pap smear would be magnified. This perspective would vastly improve the apparent cost-effectiveness of most prevention programs. More work is needed in this area before we pursue any single direction.

Several directions may be pursued in incorporating the concept of discounting

into attempts to account for individual behavior change. First, exploration of the best assessment methodology to assess a person's value of future health will need to be developed. Discounting rates can be derived for an individual at present through a rather elaborate and laborious task. Pursuit of establishing discount rates for groups of people may be a worthwhile avenue to pursue. Second, once assessment technologies are developed, the value that an individual places on a health outcome could be assessed in addition to the other components of the Health Belief Model. Given that health benefits may be experienced at various intervals of delay from the change in behavior, assessing the value of the health outcome when it is to be realized would be especially important. Discounting the value of a future heath outcome may be found to be a stable individual trait or more prone to influence by situational variables; this may prove a fertile area for future research. Third, understanding how a health outcome is valued may illuminate why some behavior change strategies are more effective than others. Research has focused on motivation to change behavior but rarely incorporated differences in perceived value of a health outcome as a component of that motivation. Fourth, once valuing (discounting) of health outcomes is better understood, interventions could be tailored for individuals who have high discount rates by focusing more on immediate/short-term benefits, whereas interventions for individuals who discount future health less drastically may make use of motivational strategies that focus on long-term health benefits of behavior change. Identification of these individual differences or the circumstances that influence the phenomenon of discounting may provide valuable information for the designers of health education/intervention programs.


Time preference affects health policy and the study of individual health behavior. In the current economic climate, health care will undergo a radical revision. The values in decision making regarding expenditures for prevention, clinical practice, and what constitutes “basic health care” are being addressed. By better understanding all key elements affecting health policy and behavior, such as time preference, we can best promote human wellness.



Prepared in part for the 1994 University of California/Health Net Wellness Lecture Series under a grant from Health Net. Parts previously published in the American Journal of Preventive Medicine. Used with permission.


1. Chen, M., and Bush, J. W.1976. “Maximizing health system output with political and administrative constraints using mathematical programming.” In quiry13, 215–227.

2. Fanshel, S., and Bush, J. W.1970. “A health status index and its application to health services outcomes.” Operations Research18, 1021–1066.

3. Kaplan, R. M., Anderson, J. P., and Ganiats, T. G.1993. “The Quality of Wellbeing scale: Rationale for a single quality of life index.” In Quality of Life As sessment: Key Issues in the 1990s, 2nd ed., edited by S. R. Walker and R. M. Rosser. Pp. 65–94. London: MTM Press.

4. “World Health Organization.” 1984. Health promotion: A discussion document on the concepts and principles. (!CP/HSR 602(m01) ed.) Copenhagen: WHO Regional Office Europe.

5. Weinstein, M. C., Fineberg, H. V., Elstein, A. S., et al. 1980. Clinical Decision Analysis.Philadelphia: W. B. Saunders.

6. Keeler, E. B., and Cretin, S.1983. “Discounting of life-saving and other nonmonetary effects.” Management Science29, 300–306.

7. Udry, J. R., and Morris, N. M.1971. “A spoonful of sugar helps the medicine go down.” American Journal of Public Health61, 776–785.

8. Sox, H. C., Blatt, M. A., and Higgins, M. C.1988. Medical Decision Making.Boston: Butterworths.

9. Warner, K. E., and Luce, B. R.1982. Cost-Benefit and Cost-Effectiveness Analysis in Health Care: Principles, Practice, and Potential.Ann Arbor, Mich.: Health Administration Press.

10. Sackett, D. L., Haynes, R. B., and Tugwell, P.1985. Clinical Epidemiology—A Basic Science for Clinical Medicine.Boston: Little, Brown.

11. Grabowski, H. G., and Hansen, R. W.1990. “Economic scales and tests.” In Quality of Life Assessments in Clinical Trials, edited by B. Spilker. Pp. 61–70. New York: Raven Press.

12. Wheeler, J. R. C., and Smith, D. G.1988. “The discount rate for capital expenditure analysis in health care.” Health Care Management Review13, 43–51.

13. Messing, S. D.1973. “Discounting health: The issue of subsistence and care in an undeveloped country.” Social Science and Medicine7, 911–916.

14. Lipscomb, J.1989. “Time preference for health in cost-effectiveness analysis.” Medical Care27 (Suppl.), S233–S253.


15. Russell, L. B.1987. Evaluating Preventive Care: Report on a Workshop.Washington, D.C.: The Brookings Institution.

16. Barnum, H.1987. “Evaluating healthy days of life gained from health projects.” Social Science and Medicine24 (10), 833–841.

17. Ganiats, T.1992. “On sale: future health care—The paradox of discounting.” Western Journal of Medicine156 (5), 550–553.

18. Ganiats, T. G.1994. “Discounting in cost-effectiveness research.” Medical Decision Making14 (3), 298–300.

19. Schieber, G. J., and Poullier, J. P.1989. “International health care expenditure trends.” Health Affairs8, 169–177.

20. Chapman, G. B., and Elstein, A. S.1995. “Valuing the future: Temporal discounting of health and money.” Medical Decision Making15 (4), 373–386.

21. Thaler, R.1981. “Some empirical evidence on dynamic inconsistency.” Economic Letters8, 201–207.

22. Benzion, U., Rapoport, A., and Yagil, J.1989. “Discount rates inferred from decisions: An experimental study.” Management Science35 (3), 270–284.

23. Loewenstein, G., and Prelec, D.1992. “Anomalies in intertemporal choice: Evidence and interpretation.” Quarterly Journal of Economics107 (2), 573–597.

24. Loewenstein, G.1987. “Anticipation and the valuation of delayed consumption.” Economic Journal97, 666.

25. Loewenstein, G. F.1988. “Frames of mind in intertemporal choice.” Management Science34 (2), 200–214.

26. Elster, J.1985. “Weakness of the will and the free-rider problem.” Economics and Philosophy1, 231.

27. Loewenstein, G. F., and Sicherman, N.1991. “Do workers prefer increasing wage profiles?” Journal of Labor Economics9, 67–84.

28. Rose, D. N., and Weeks, M. G.1988. “Individual's discounting of future monetary gains and health states.” Medical Decision Making8, 334.

29. Redelmeier, D. A., and Heller, D. N.1993. “Time preference in medical decision making and cost-effectiveness analysis.” Medical Decision Making13 (3), 212–217.

30. Christensen-Szalanski, J. J. J.1984. “Discount functions and the measurement of patients' values: Women's decisions during childbirth.” Medical Decision Making4, 47–58.

31. Becker, M. H., ed. 1974. “The health belief model and personal health behavior.” Health Education Monographs2.

32. Janz, N. K., and Becker, M. H.1984. “The health belief model: A decade later.” Health Education Quarterly11, 1–47.

33. Prochaska, J. O.1979. Systems of Psychotherapy: A Transtheoretical Analysis.Pacific Grove, Calif.: Brooks-Cole.

34. Prochaska, J. O., and DiClemente, C. C.1984. The Transtheoretical Approach: Crossing the Traditional Boundaries of Therapy.Homewood, Ill.: Dow-Jones/Irwin.

35. Rosenstock, I. M.1996. “Why people use health services.” Milbank Memorial Fund Quarterly44 (3), 94–124.


36. Bandura, A.1986. Social Foundations of Thought and Action: A Social Cognitive Theory.Englewood Cliffs, N.J.: Prentice Hall.

37. DiClemente, C. C., and Prochaska, J. O.1982. “Self change and therapy change of smoking behavior: A comparison of processes of change in cessation and maintenance.” Addictive Behavior7, 133–142.

38. Prochaska, J. O., Velicer, W. F., Rossi, J. S., et al. 1994. “Stages of change and decisional balance for 12 problem behaviors.” Health Psychology13, 39–46.

39. Perz, C. A., DiClemente, C. C., and Carbonari, J. P.1996. “Doing the right thing at the right time?” The interaction of stages and processes of change in successful smoking cessation. Health Psychology15, 462–468.

40. Velicer, W. F., Rossi, J. S., and Prochaska, J. O.1996. “A criterion measurement model for health behavior change.” Addictive Behaviors21, 555–584.

41. Marcus, B. H., Banspach, S. W., Lefebvre, R. L., et al. 1992. “Using the stages of change model to increase the adoption of physical activity among community participants.” American Journal of Health Promotion6, 424–429.

42. Marcus, B. H., and Owen, N.1992. “Motivational readiness, self-efficacy and decision-making for exercise.” Journal of Applied Social Psychology22, 3–16.

43. Gold, M. R., Siegel, J. E., Russell, L. B., and Weinstein, M. C., eds. 1996. Cost-Effectiveness in Health and Medicine.New York: Oxford University Press.




Part 3 extends the discussion of innovative wellness promotion strategies into the realm of making recommendations for practice. Minkler (chapter 13) discusses some of the issues that have been inadequately addressed in past wellness programs, including the role of social inequities in determining health status, the empowerment of communities to participate in health planning, and the unanticipated consequences of health promotion efforts. She provides a specific example of a program wherein the health behaviors of elderly residents in a highcrime neighborhood are enhanced by increasing their sense of personal security through community-level changes. The needs of the elderly also are addressed by Duxbury (chapter 15) and Beck (chapter 16), with the former discussing the pros and cons of cancer screening and the latter focusing on the prevention of disability through comprehensive assessment techniques. The arguments put forth by Duxbury against populationwide prostate cancer screening provide an illustration of Minkler's point concerning the unanticipated consequences of health promotion, while Beck's description of the need to include social and environmental assessments in concert with medical assessments to effectively prevent disability in the elderly exemplifies one possible effect of a “person-centered” approach on clinical practice.

Within the context of educational institutions, Slavin and Wilkes (chapter 17) present an approach toward restructuring the medical school curriculum to promote a person-centered approach among physicians,

and Nader (chapter 18) highlights the value of establishing partnerships between universities and communities to promote wellness. The model curriculum described by Slavin and Wilkes centers around problem-based learning and encourages medical students to look beyond vital statistics to familial, social, and environmental influences on health. The partnerships depicted by Nader illustrate how these different levels of influence can be incorporated into wellness interventions targeted toward school-age children. Together, these chapters speak to the pivotal role that educational institutions can play in community wellness promotion.

Finally, Zhu and Anderson (chapter 14) provide an example of an intervention that melds the clinical and public health approaches into an integrated program to promote smoking cessation. The Smokers' Helpline combines the intensity of the clinical approach with the accessibility and low cost associated with the public health approach to achieve meaningful and cost-effective change at the population level. Moreover, the way that the program was conceived and the feature that allows callers to choose from a menu of preferred services reflect Minkler's point concerning the value of involving communities in the planning of programs that are designed to promote their health. Altogether, part 3 serves to illustrate some of the ways that nontraditional perspectives may lead to highly effective health promotion practice.



Challenges and Dilemmas

Meredith Minkler

Recent discussions of the challenges facing health promotion in the United States have focused heavily on the advantages and disadvantages of managed care and where health promotion and disease prevention may fit in our rapidly evolving health care system. Although such deliberations are important, little attention has been devoted in these discussions to how we as a society can make the necessary changes to significantly improve the health of the American people.1 This larger question goes well beyond the nature of our medical care system, and it raises difficult challenges, beginning with how we as a nation should be defining health promotion in the first place.

This chapter begins by briefly examining the dominant vision of health promotion in the United States, its policy level implementation, and its more recent evolution and expansion. I then highlight just a few of the health promotion challenges and paradoxes we face, as health professionals and as a society, at the dawn of a new century. These challenges include, first, the need to continue to broaden the focus of our health promotion efforts in order to better address the profound role of social inequities in influencing health status; second, the importance of giving more than lip service to the rhetoric of empowerment and community participation for health by fully embracing them as a framework for health planning; third, the need to recognize and address some of the unanticipated consequences of our health promotion efforts, for example, in unwittingly reinforcing prejudice against the elderly and disabled and

in inadvertently encouraging the tobacco industry to become even more aggressive in its targeting of youth, people of color, and Third World nations; and fourth, the need to reframe the continuing public health debate between individual autonomy and the common good. I will argue in particular that we need to broaden and considerably deepen our definition of the common good to stress societal interdependence and not merely the collective rights of individual citizens within a society.

Although this chapter can do little more than scratch the surface in these areas, it is put forward in the hope of contributing to a dialogue that will help us think about the challenges we face in health promotion in some new and different ways. Toward this end, the chapter closes with a brief discussion of the recent Canadian framework for health promotion as a means of illuminating potentially useful avenues for further expanding the vision and the reality of health promotion in the United States.


Before looking at where we might head as a society in terms of health promotion, it is important to look at where we are and where we have come from. The dominant view of health promotion in the United States today emerged in the 1970s in response to a growing disillusionment with the limits of medicine, pressures to contain health care costs, and a social and political climate emphasizing self-help and individual control over health.2,3 It is a vision that sees individual behavior as in large part responsible for the health problems we face as a society. In the words of j. K. Iglehart, editor of the journal Health Affairs, this vision suggests that “most illnesses and premature death are caused by human habits of living that people choose for themselves” (emphasis added).4

Ironically, this traditional approach to heath promotion has tended to be disease oriented rather than health oriented. As Wallack and Montgomery 5 have pointed out, it defines health primarily as the absence of disease and sees disease as being associated largely with known and controllable risk factors such as cigarette smoking, poor diet, and heavy drinking. The individual is seen as the appropriate focus for intervention to control risk factors, with those interventions typically consisting of providing knowledge and skills for changing unhealthy behaviors.5 This vision of health promotion was given institutional expression in Canada, with the publication of the Lalonde Report 6 in 1974, and in the United

States, in the Surgeon General's report Healthy People.7 Both of these documents, it should be noted, discussed the role of broader environmental factors in influencing health and did not limit themselves to a discussion of individual lifestyle or personal behavior issues. The Surgeon General's report, for example, argued persuasively that “we are killing ourselves” not only by “our own careless habits” but also by polluting the environment and permitting harmful social conditions to exist.7 Despite their efforts to address some of these broader issues, however, the major contributions of both the Lalonde and the Surgeon General's reports lay in calling attention to the often substantial role individuals can play in modifying their personal behaviors and in other ways improving their health status.8–10

In the United States, the Surgeon General's report was followed by the development of clearly articulated and measurable “Objectives for the Nation.” 11 Developed by the U.S. Public Health Service with the Office of Health Information and Promotion (OHIP) playing a facilitating role, the “Objectives” were designed to help address more specifically the broad goals set forth in Healthy People. They made a major contribution in focusing attention on prevention and health promotion and in providing clear performance indicators and serving as a stimulus to action. The listing of activities for achieving each objective was extremely thorough and included strategies on the levels of institutional change, legislation, and policy and not merely in the realm of personal behavior change. This broad approach further was in keeping with the view of health promotion put forward at this time by OHIP Director Lawrence Green and his colleagues,12 who defined health promotion as “any combination of health education and related organizational, political and economic interventions designed to facilitate behavioral and environmental changes conducive to health.”

In reality, however, implementing this broad vision, particularly in an era of fiscal conservatism, proved difficult indeed. Moreover, as Green 13 has noted, the sharp distinction drawn in U.S. policy between health promotion (focused mainly on behavior and lifestyle issues) and health protection (concerned more with the physical environment) led to a narrower interpretation of health promotion in the United States than in many European nations, which argued that both physical and social environmental factors lay within the purview of health promotion.

In the United States, fully a third of “Objectives” and a special section within Healthy People were devoted to health protection with a focus on environmental concerns within such domains as toxic agent control

and occupational health and safety. As noted earlier, many of the interventions proposed under these categories were far-reaching in scope, yet the very process of creating a dichotomy between health promotion and protection may have had the effect of limiting our vision with respect to a broader view of health promotion. In Green's 13 words, “We Americans allowed our health promotion terrain to be restricted to lifestyle determinants of health, but we also allowed lifestyle to be interpreted too narrowly as pertaining primarily if not exclusively to the behavior of those whose health is in question.”

As a consequence, most of the programs that grew out of the early push for health promotion in the United States tended to focus primarily on the level of personal behavior change. The programmatic emphasis on individual responsibility for health, in short, frequently was not accompanied by attention to individual and community response-ability,14 or the capacity of individuals and communities to build on their strengths and respond to their personal needs and the challenges posed by the environment.

Expanding the Vision

Health promotion in the United States has evolved in important new directions since the early 1980s. Although most work-site health promotions continue to operate primarily on the level of the individual,15 many have significantly broadened their focus. Education on parenting skills, prenatal care, and interpersonal communications is frequently included in U.S. work-site health promotion programs in addition to the more traditional smoking cessation classes and related interventions aimed at modifying objective risk factors. The “Education for Action” training program, through which workers in hospitals and other settings around the nation have been helped to identify and address unsafe working conditions, provides an illustration of the broader approach to work-site health promotion that has gained popularity.16

In the community, innovative health promotion efforts, like the schoolbased Adolescent Social Action Program (ASAP) in Albuquerque, New Mexico, are including empowerment education and community development in both their methodology and their raison d’être.17 A universitycommunity partnership, ASAP involves dozens of public schools in an effort to address alcohol and substance abuse problems on multiple levels while creating the conditions in which youth can become empowered to make healthier choices in their own lives. In the area of HIV prevention,

projects like Stop AIDS in San Francisco, California, similarly are stressing both community development and individual behavior change. Participants in Stop AIDS education and support groups, for example, are asked as part of their involvement to commit themselves both to practicing safer sex and to doing community organizing around the epidemic.18

On another level, statewide legislative initiatives, such as California's Proposition 99, which put a 25-cent tax on cigarettes and allocated 20% of the revenue generated to tobacco education and anticigarette advertising, increasingly are seen as being within the purview of health promotion activities. Finally, as will be noted later, unprecedented new efforts to drastically curtail the power and privileges of the tobacco industry nationwide currently are being attempted and have achieved considerable public support. Such developments are encouraging and have helped widen the scope of health promotion in the United States in important new ways. For the most part, however, even these broader efforts have failed to address the profound influence of broad social inequities on health.

Addressing Social Inequities

The need for developing health promotion programs that address the role of social inequalities in influencing health status is well documented. A voluminous body of evidence, for example, has demonstrated that social class is one of the major risk factors, and perhaps even the major risk factor for disease.19–22 Studies have shown that there is a clear gradient in social class and mortality rates: Not only do people in the highest socioeconomic groups have the lowest mortality rates, but these rates increase at each correspondingly lower rung of the socioeconomic ladder.21 As Syme 22 has noted, the evidence linking social class and illness is indeed so powerful that researchers routinely control for socioeconomic status in their studies of variables influencing health since this single factor would otherwise overshadow most of their other findings.

The need for addressing social inequities in the design and implementation of our health promotion programs is illustrated in a case study from San Francisco's Tenderloin district. Twenty years ago, my students and I began a project in the Tenderloin to help reduce social isolation and powerlessness and thereby improve the physical and mental health of some of the neighborhood's 8,000 low-income elderly residents. The health problems in this area were daunting. Approximately 40% of the elderly residents were malnourished or undernourished, for example,

and although cooking in their rooms was not allowed, they often resorted to illegal hot plates since they could not afford to eat out.23 Many of the Tenderloin residents with whom we worked wanted to improve their diets. But they lived on small, fixed incomes in a neighborhood that afforded little access to fresh fruits and vegetables. Many residents wanted to get more exercise, but they lived in tiny, cramped rooms in an area with the highest crime rate in the city. Taking a brisk walk may be anything but health promoting in such an environment!

For residents of neighborhoods like the Tenderloin, the primary challenge often is not that of having individuals “take more responsibility” for their health” but rather of improving their “response-ability,” by ensuring an adequate income, access to nutritious foods, and, for those with significant functional impairments, access to such coping resources as homemaker services, meals on wheels, and senior escorts. Services like these routinely are made available to the disabled in Canada and western Europe. However, they are effectively rationed in this country by our grossly inadequate funding for such programs and by the fact that they are easy targets for the budget axe in times of fiscal retrenchment.23

Health promotion efforts that fail to address the social context within which people live not only minimize the possibility of success but also risk violating the ethical admonition to “do no harm.” Already alienated individuals may experience an even greater sense of powerlessness when they try to change health-related behaviors and fail, for example, and this in turn may have negative health consequences 24 (see chapter 11). While continuing to acknowledge the critical need for individuals to take more responsibility for their health, a major challenge at the turn of the century, then, is to develop health promotion efforts that include an equally compelling emphasis on changing those broader social and environmental conditions that so often constrain individual choice in matters related to health. Part of incorporating broader contextual considerations into the policies and programs we design involves placing a far greater emphasis on the principles of community participation and empowerment, and it is to these related areas that I now turn.

Community Participation and Empowerment

Empowerment is a much used and abused term, and it has often been coopted by conservative policy makers who have used the rhetoric of empowering communities as a rationale for cutting back on needed health and social services. But if power is “the ability to predict, control

and participate in one's environment,”25 then empowerment is the process by which people and communities are enabled to take such power and act effectively in transforming their lives and their environments.26 Borrowing from earlier feminist conceptualizations of power, it suggests that we reframe old notions of “power over” with newer visions of “power to” or “power with.” 27

The concepts of community empowerment and participation reflect a commitment to such precepts as the common good and shared responsibility for health. Acting on these concepts means enabling communities to participate, in equal partnership with health professionals, in setting the health agenda defining their health problems and helping to develop the solutions to address those problems.

This focus is critical. Sociologist John MacKinlay is reported to have remarked that professionals frequently suffer from an unfortunate malady known as “terminal hardening of the categories.” 28 We get the kinds of answers we are comfortable dealing with because we ask the kinds of questions that will give us those answers. We conduct behavioral risk surveys, for example, that will carefully document heart disease as a major community health problem but will, in all likelihood, miss the fact that very different sorts of issues, like drugs or violence, may be the major health concerns of residents.28

In contrast, an empowering approach to health promotion would “start where the people are” by having them set the health agenda and then work to address the issues they collectively have identified. Such an approach validates the community's ability to assess its needs and strengths and builds on the latter in helping to increase the problem solving ability of both individuals and the larger community. The results can be dramatic. The Tenderloin project referred to earlier is illustrative.

When my students first formed the Tenderloin Senior Organizing Project (TSOP), they organized support groups among the elderly residents of several deteriorating Tenderloin hotels and asked participants what their major health concerns were. In hotel after hotel, the residents responded, “Crime,” and the students politely said, “You misunderstood, we were asking about health problems.” The residents held their ground, pointing out that they could not safely go outdoors without being mugged and therefore could not get to the doctor's office, go for a walk, or get an evening meal. Crime, they argued, was their biggest health problem.23

The students and project staff listened. Then they helped the residents organize a community-wide meeting on the subject of crime and enlist

the support of the mass media. They helped garner resources so that the residents could start an interhotel coalition, the Tenderloin Tenants for Safer Streets. Members of this grassroots coalition subsequently met with the mayor and demanded and got increased beat patrol officers in the neighborhood. The students and staff also helped, but always in the background, as residents began the Safehouse Project, recruiting 40 local merchants and agencies to be places of refuge where residents could go for immediate aid if they were being followed or just needed to sit down because of shortness of breath.

Residents' organizing on crime prevention was given much of the credit for an 18% drop in the crime rate that occurred in this neighborhood in the first 12 months of their mobilization.23 Their organizing efforts also translated into some effective individual-level behavior change. For example, residents' new found feelings of power, self-efficacy, and improved self-esteem led some to successfully quit smoking and cut down on problem drinking.23

Had the students and staff of the Tenderloin Senior Outreach Project failed to pay attention to and support the community's definition of need, they might still be running support groups in hotel lobbies one morning a week—if indeed they were still welcome at all. Instead, by trusting the people to determine and act on their own health agenda, they were able to contribute to something that has had a real and lasting impact on the health of that community.

The Tenderloin project is but one of a number of examples that could be cited of health promotion efforts that have actively involved local communities in identifying and addressing their health problems and in the process building on and reinforcing community strengths. The earlier mentioned ASAP program in New Mexico, through which youth, many of them Native American and Hispanic, are helped to identify and then creatively address substance abuse problems in their community, is an example of another such effort and one that has demonstrated measurable successes in terms of increasing participants' perceptions of high-risk behaviors and their sense of social responsibility.17 The Black Women's Health Project (BWHP), founded in 1981 and based in Oakland, California, is another example of an empowering approach to health promotion. A national network of approximately 100 self-help groups in approximately two dozen states, the BWHP helps African American women gain information, skills, and access to resources while enabling them to work together to identify and analyze health-related concerns and to collectively address these issues.29 On a smaller scale, an innovative

health promotion action research project in an automotive parts plant in Michigan involved workers, in partnership with researchers, in a Stress and Wellness Committee that identified sources of stress in the workplace and then designed and sought implementation of strategies on a variety of levels to address these concerns.30

Projects like these share a commitment to helping communities identify their needs and then working with them in developing health promotion programs and approaches that truly meet those needs. Implementing such a commitment, however, is often far from easy. As Gail Siler-Wells 31 points out, “Behind the euphemisms of empowerment and community participation lay the realities of power, control and ownership.” The very real structural distinctions that exist between professionals and communities, and our very location as professionals in health agencies and bureaucracies, confer a certain power that includes the power to set the health agenda.31

As we enter a new decade and a new century, new and innovative ways must be developed to better enable individuals and communities to take the power they need to bring about health improvements. Professionals committed to facilitating this process may be aided by applying a tool developed by community organizers Herbert and Irene Rubin 32 and called the DARE criteria for empowerment. The DARE criteria would have us answer the following set of questions with respect to any health promotion project in which we are engaged:

Who Determines the goals of the project?
Who Acts to achieve them?
Who Receives the benefits of the actions?
Who Evaluates the actions?

The more we can answer these questions by responding, “The community,” the more likely our health promotion projects are to be contributing to true community empowerment and self-determination.33

Public health departments, hospitals, and other health institutions also can contribute to community participation and empowerment by forming what Ron Labonte 34 describes as “authentic partnerships” with communities and ensuring that our “community based” projects are not merely “community placed.” 34 Such authentic partnerships will take on particular relevance, moreover, as new and complex areas such as violence prevention become an increasing part of public health practice.


A problem such as violence presents a whole new set of challenges, in part because it is widely viewed in this society as a problem to be addressed through law enforcement and criminal justice rather than through efforts directed at economic development, human welfare, and public health. The statistics, of course, tell a different story. As Michael McGinnis and William Foege 35 have demonstrated, firearms rank sixth (after tobacco, diet and exercise, and so on) as an actual cause of death in the United States. And although the overall homicide rate in this country began to decline in the early 1990s,36 we continue to rank first—by a wide margin—among the advanced industrialized nations.37 When we turn our attention to youth violence, an even more disconcerting picture emerges. For despite recent declines, and whether as victims, perpetrators, or witnesses, young people aged 15 to 24 continue to experience disproportionately high rates of violence. A recent comparison of the overall firearms-related deaths among children under 15 in 26 advanced industrialized countries revealed the U.S. rate to be 12 times higher than that of all the other countries combined (1.66 vs. 0.14).38 The firearmsrelated homicide rate in U.S. children was almost 16 times that of the other nations combined (0.94 vs. 0.06) 38 (see Figure 13.1).

Facts and statistics like these provide compelling evidence that the public health system should be a major force in violence prevention efforts.

The suggestion that health professionals and their institutions take on violence as a major public health problem is in no way meant to underplay the critical need for broad societal-level economic and social change if we are truly to make a lasting impact on the violence that plagues our nation. With many of the new jobs created at or below minimum wage and with the gap between rich and poor greater today than at any time since World War II and continuing to expand rapidly,39 it is not hard to understand the increase in violent crime, particularly among unemployed and underemployed youth for whom drug trafficking and related violence may be very lucrative. Similarly, with handguns and semiautomatics within easy reach of our nation's most vulnerable children, and with violence endemic on our movie and television screens, it is unlikely that significant and lasting change can be made at the community level without dramatic changes on our societal landscape.

However, communities can and do have a vital role to play, and by working in partnership with communities, local health departments can lend their resources, skills, and credibility to a frontline attack on violence.


Figure 13.1. Rates of firearm-related homicides among children aged <15 years in 26 industrialized countries. Rates are per 100,000 children aged <15 years and for one year during 1990–1995. In this analysis, Hong Kong, Northern Ireland, and Taiwan are considered countries. Source: Adapted from Morbidity and Mortality Weekly Report, February 7, 1997, Figure 1, p. 104.

A powerful illustration of an empowering and collaborative approach to violence prevention that links health institutions to local communities was developed by Deborah Prothow-Stith and her colleagues 40 at Harvard's School of Medicine and Massachusetts General Hospital. Offered to youth in high-risk neighborhoods through local housing projects, schools, churches, and YMCAs, the project teaches young people new ways of coping with their anger and aggressive feelings. But it also has them explore the root causes of violence in racism, poverty, and a culture where, in Prothow-Stith's words, the most popular heroes have

“Rambo hearts and Terminator heads.” 41 Although this program has been hailed as a success in terms of primary prevention and increasing individual and community responsibility in relation to violence prevention, part of its strength lies in its continued efforts to mobilize other sectors, including government, business, and the mass media, in order to develop broad-based and multilevel attacks on the problem of violence.38,41 These efforts appear to have paid off. Breaking down the traditional turf lines between police, the courts, schools, churches, and nonprofit health and youth agencies, the city of Boston has developed an ambitious combination of after-school activities, job training, police reorientation toward knowing and valuing their neighborhoods, and other violence prevention efforts. The “Boston model” of crime prevention indeed has been credited with the fact that the city recently experienced an unprecedented 29-month period without a single murder among its children and adolescents.42

On a still larger scale, the Violence Prevention Initiative (VPI) of The California Wellness Foundation has been credited with helping to create a new social and political climate for promoting handgun control and other measures that can help reduce violence on a statewide basis.43 Initiated in 1993 with a five-year, $35 million endowment, the VPI was designed to reduce the state's high youth violence rates through a multipronged grant-making program that included community action programs, public education, research, and policy development. A central piece of this ambitious effort—the successful mobilization of a statewide effort to pressure for the banning of handguns—is described in detail by Lawrence Wallack in chapter 19. In his words, the VPI has left “a large footprint … in the legislative landscape” and has paved the way for continued dramatic policy changes in a state where murder remains one of the top two killers of children and youth.44

Empowering community and multisectoral approaches to violence prevention like those described here well illustrate how expanding our definition of health promotion and disease and injury prevention and working in broad partnerships on a wide variety of levels can help address one of the most critical public health problems of our times.

Avoiding Prejudice against the Elderly
and Disabled in Our Health Promotion Efforts

I have discussed so far the need to devote greater attention in our health promotion efforts to the social determinants of health and to do this in

part by taking more seriously the notion of community empowerment for health. While we move in these broader directions, however, we face another challenge as we enter a new century—that is, to reexamine the health promotion efforts already under way in our country to better understand the ways in which these programs and approaches may unwittingly reproduce and transmit such problematic aspects of our dominant culture as gerontophobia and handicappism or stigmatization of and prejudice against the elderly and the disabled.45

With some notable exceptions, the elderly largely have been ignored in most private- and public-sector health promotion efforts. The high costs of such failure are painfully evident. Helen Schauffler and her colleagues at the University of California, Berkeley, School of Public Health,46 for example, conducted the first study ever to demonstrate a relationship between risk factors for heart disease and Medicare payments. They discovered that three triggers for heart disease—high blood pressure, high cholesterol, and cigarette smoking—are costing Medicare at least $16.6 billion per year in extra medical services, yet the program does almost nothing to prevent these conditions (see Figure 13.2).

Although Medicare's failure to cover much health promotion and disease prevention for the elderly reflects in large part a more generalized reluctance in both public- and private-sector health insurance plans to cover such services,47 it may also reflect the widespread belief that such programs have little to offer the elderly, who are deemed “resistant to change” anyway and whose “productive years” are largely behind them.48 Recent changes in the Medicare program take an important step forward in providing coverage for colorectal screening and annual mammograms, increased payments for preventive injections, and coverage of glucose monitoring and other costs associated with the management of diabetes.49 But while our health promotion efforts must include even greater efforts toward the prevention of unnecessary disability and functional impairment, a challenge for the decades ahead is to find ways of doing so that do not in the process stigmatize and devalue those elders who are or may become disabled.

The renewed emphasis on individual responsibility for health in this country has been accompanied by the reemergence of a Victorian-era notion that healthy old age is a just reward for a life of self-control and “right living.” 50 Such a notion opens the door to victim blaming of those elders who dare to become chronically ill or disabled. In David Lewin's words, “Good health has become a new ritual of patriotism, a market place for the public display of secular faith in the power of will.” 51


Figure 13.2. Increased Medicare costs per elderly person per year associated with risk factors for heart disease. Source: Schauffler et al. (1993 [46]).

Within such a vision, where is there a place for the 85-year-old man with a disabling respiratory ailment or for the obese and severely arthritic elderly woman in a wheelchair?

The unwitting tendency of some health promotion efforts to foster this kind of stigmatization can also be seen in many recent injury prevention campaigns. Once again, we are faced with a paradox. On the one hand, it is critical that health promotion include an emphasis on injury prevention, and it is heartening, too, that this has expanded to include such controversial areas as handgun control. On the other hand, we must be concerned when the messages of our prevention campaigns build

on and contribute to fear of disability and in the process further stigmatize those who are already disabled.

As Caroline Wang 52 has pointed out, health promotion approaches to injury prevention that carry the implicit or explicit message, “Don't let this happen to you!” often inadvertently stigmatize people with disabilities, suggesting that they are “inherently flawed” and undesirable. In her words,

If the public health perspective rightly contends that becoming disabled is an unacceptable risk in our society, it paradoxically often fails to acknowledge the stigmatizing notion that being disabled is an unacceptable status in our society.52

Heeding the physicians' admonition to “first, do no harm,” those of us concerned about health promotion and injury and disease prevention must ensure that the campaigns we design and the messages we transmit through them do not further contribute to the handicappism and gerontophobia that already plague our society.

The Continuing Challenge of the Tobacco Industry

In some of the most critical arenas within health promotion, such as alcohol and tobacco, the very success of some of our health promotion efforts is raising new problems and challenges we have only begun to tackle. Recent developments in the area of tobacco marketing and advertising provide an excellent case in point.

Accounting for over 416,000 deaths per year, cigarette smoking has been called by the Centers for Disease Control and Prevention “the most devastating preventable cause of disease and premature death this country has ever experienced.” 53 Yet despite this reality, some impressive strides have been made. Smoking rates overall have dropped from 40% of the population in 1965 54 to about 23% in 1998.55 Public health measures such as increased excise taxes, mandatory warning labels on cigarette packages, restrictions on tobacco advertising, mass-media campaigns, and public and private antismoking rules and ordinances have been given much of the credit for these declines.56

The very success of health promotion efforts in the area of tobacco control, however, has contributed inadvertently to aggressive new efforts by cigarette manufacturers to find new markets at home and abroad and to take steps to increase their ebbing legitimacy. In the United States, increased

targeting of people of color, women, and young people has been one consequence, with direct advertising appeals supplemented by tobacco companies' sponsorship of baseball games, tennis matches, and cultural events like Cinco de Mayo parades.57 Although recently cut back as a result of the tobacco settlement, such efforts have paid off. Among 8th-, 10th-, and 12th-graders, for example, the proportion of youth who smoke daily increased by almost 50% between 1991 and 1996, with 20% of 12th-graders now smoking on a daily basis.58,59

To improve their public image, leading tobacco companies also have become major donors to both large and small nonprofit organizations including, ironically, the Partnership for a Drug Free America.60 Such money often comes with strings attached. Shortly after accepting $150,000 from Phillip Morris, for example, New York City's Coalition for the Homeless was asked to help kill a bill mandating antismoking ads by pressuring the city council to focus on more important issues like homelessness.60 In a time of dwindling public and private funding for a plethora of good organizations and worthy health and social causes, the tobacco companies' substantial role in financing raises troubling ethical dilemmas.33

A combination of developments in the late 1990s seriously damaged the already waning credibility of the tobacco industry and greatly increased political and popular support for stiff new tobacco taxes and other antismoking measures. The release and wide publicizing of credibility harming tobacco industry memos and other secret documents, public revulsion at new evidence of the industry's sophisticated efforts to hook teens and to increase the addictive content of cigarettes, state-level efforts to sue tobacco companies as a means of recovering Medicaid costs for smoking-related illnesses, and one cigarette manufacturer's break with the rest of the industry in acknowledging the addictive nature of nicotine were among the events signaling a potential watershed in antitobacco mobilization in the United States.61 Health promotion efforts such as the well-financed Campaign for Tobacco Free Kids and the earlier mentioned California Tobacco Control Program (which was credited with having helped a million smokers quit the habit in its first three years of operation)62 also began playing an important role in discouraging smoking, especially among youth.

Major legislative events have further helped change the tobacco landscape. After months of heated debate between tobacco industry representatives, state attorneys general, and other parties, a proposed $368.5 billion “tobacco settlement” was announced in June 1997 that

called for sweeping changes in the regulation, sale, and advertising of tobacco.61,63 The settlement was quickly opposed by a plethora of antitobacco forces, including the American Lung Association and the American Public Health Association, which argued that it “let the tobacco industry off the hook” far too easily. The proposed settlement, for example, would have hindered the Food and Drug Administration's (FDA's) ability to regulate nicotine for the next 12 years despite the fact that this regulatory power had already been granted by a federal court.63 In the ensuing months, President Clinton charged Congress with passing tough national tobacco legislation that would include a combination of industry payments and penalties that would increase the cost of a pack of cigarettes by up to $1.50 over the next 10 years. A comprehensive plan to dramatically reduce youth smoking, expanded efforts to restrict access and limit the appeal of cigarettes, and the granting of full authority to the FDA to regulate tobacco products were among the key elements proposed.64 The national tobacco settlement finally reached in 1998 involved payments to 46 states totaling about $206 billion over the next 25 years. Under the settlement, more than 14,000 cigarette billboards were removed, outdoor advertising was banned on public transit and in many public arenas, and such popular cartoon figures as Joe Camel were permanently retired. Of even greater importance, the settlement prohibited the industry from lobbying against tobacco control laws and ordinances.65

The tobacco settlement led to a 45-cent-per-pack increase in the price of cigarettes between 1998 and 1999, with a corresponding decline in total cigarette consumption of 7.5% over the same period.65 As some analysts have noted, however, “the welcome fall in consumption resulting from the price increase may turn out to be temporary unless it is followed up with serious efforts to dissuade kids from smoking.” 65 With just 8% of the tobacco settlement monies to states so far going to antismoking efforts and the great bulk of the funds being used for road construction, tax cuts and the like,66 early hopes for dramatic declines in youth smoking appear unlikely to be realized.

We continue to face other ethical challenges when the activities of American cigarette manufacturers are viewed within a global context. The use of tobacco world-wide has increased dramatically since the mid1960s, with much of this increase occurring in Third World countries. While cigarette smoking now is declining by 1.4% annually in industrialized nations, moreover, the habit continues to grow by 1.7% per year in the developing nations, where tobacco accounted for more than

1.2 million deaths annually by 1995.65 In many Third World nations, smoking has only recently become widespread, and daily consumption is lower than in the developed world largely for economic reasons. Yet as large cohorts of young smokers age and as personal disposable income increases in these nations, substantial increases in daily consumption are anticipated, unless effective tobacco control measures are put into place.66–68

In a provocative and troubling article “Advertising for All by the Year 2000,” Wallack and Montgomery 5 demonstrate how dwindling markets and increasing advertising bans at home have made Third World nations one of the most “promising frontiers” for our nation's tobacco industry. The United States is the world's largest tobacco exporter, with our cigarette exports growing 260% from 1986 to 1996 and two-thirds of cigarettes manufactured in the United States now sold internationally.63,69 American tobacco companies also rank among the leading advertisers in many Third World nations,5 and in Asia and Latin America, researchers have documented an association between increased cigarette advertising and both general increases in smoking and specific increases among women and children.70 Tobacco consumption already accounts for a substantial proportion of overall mortality in many Third World countries, and estimates suggest that by 2025, seven out of every 10 tobacco-related deaths will occur in the Third World.69,71

The policy implications raised by these realities are profound and suggest the need for action on multiple levels. Working through groups like the World Health Organization (WHO), for example, we might promote a massive international antitobacco advertising campaign while at home we continue to press for national tobacco legislation along the lines described previously. We further must require that U.S.-manufactured cigarettes that are marketed in other nations be held to the same standards in terms of warning labels, nicotine content, and the like as those sold domestically. Such efforts are essential if we are to achieve real success in dramatically reducing smoking rates at home while preventing the realization of the WHO's dire projection that tobacco will be the world's leading cause of death by 2020.71

Rethinking “The Public Good”

A final challenge for broadening the focus of health promotion as we begin a new century is perhaps the most difficult, for it involves rethinking conceptualizations of health issues that are deeply embedded in both

Western and uniquely American value systems. Health-related behaviors such as cigarette smoking frequently are discussed in terms of the tension between individual autonomy and the public or community good. This debate has been badly constrained, however, because our dominant notions of justice are impoverished.70,72 In Larry Churchill's 72 words, they are based on “a moral heritage in which answers to the question ‘what is good?’ and ‘what is right?’ are lodged definitively in a powerful image of the individual as the only meaningful level of moral analysis.”

When the ethical notion of public or community good is invoked in arguments for mandatory motorcycle helmet use, for example, “common good” frequently is operationalized in terms of the economic rights of law-abiding citizens. Public or community good, in short, is defined as my right not to pay for your foolish or risky behavior.2 Economic arguments of this sort may have a place. But when we limit our conceptions of the public or common good in this way—when we suggest that “the public good is nothing more … than the protection of every individual's private rights,” 73—we miss the broader meaning of community. In Dan Beauchamp's 74 words,

By ignoring the communitarian language of public health, we risk shrinking its claims. We also risk undermining the sense in which health and safety are a signal commitment of the common life—a central practice by which the body-politic defines itself and reaffirms its values.

Motorcycle helmet laws, for example, are not simply “championing some individuals over others” but rather are upholding “the public or community interest” over the interests of individuals or groups.75 Such laws ideally are saying that when one of us engages in risky behavior, our collective well-being is affected because we are all part of the same community.

Broadening our concept of the public good to embrace a sense of our intimate interdependence, a notion that we are indeed “all in this together,” will not be easy. For in the words of Dan Callahan,76 the dominant culture in America “does not speak easily the language of community.” If we are to move toward broader, more community-oriented policies in the name of health promotion, however, our vision of public good must become considerably more than the rights and obligations of the collection of individuals who happen to occupy the same geographic space.

Contrary to witnessing an enhanced notion of the common good, unfortunately, the late 1990s heralded several new measures that further

restricted America's sense of community. The 1996 welfare reform bill, for example, abolished the nation's 60-year-old commitment to an entitlement of aid for low-income families with dependent children and by some estimates may move an estimated 1.1 million additional children into poverty.77 The Personal Responsibility and Work Opportunity Reconciliation Act, as this measure was named, also called for the elimination of food stamps, SSI (Supplemental Security Income), and a range of other benefits for some 800,000 of the nation's legal immigrants.78 Although many of the proposed cuts were subsequently restored through the Balanced Budget Act, new measures on the state and federal levels continue to capitalize on growing anti-immigrant sentiment and threaten to constrain still further notions of the common good.

The Canadian Framework for Health Promotion

As this chapter has attempted to demonstrate, addressing social inequities, taking seriously the rhetoric of community participation and empowerment, confronting the unanticipated consequences of our health promotion efforts, and rethinking and broadening our notion of the common good constitute major challenges for those concerned with broadening the vision and the reality of health promotion in the United States. Moreover, although health promotion has moved in important new directions since the rebirth of interest in this approach in the 1970s, recent setbacks, including importantly the failure to enact health care reform legislation and the passage of a punitive and potentially healthcompromising welfare reform bill, suggest that much remains to be done.

As we grapple with these and other challenges to health promotion during this unique historical period, a new conceptual framework must be developed that incorporates such underlying principles as a commitment to social justice, empowerment, and a broader notion of the common good. Toward this end, it is useful to review WHO's vision of health promotion and specifically how that vision has been crafted into a conceptual framework for health promotion policy and its implementation in Canada.

WHO radically revised its notion of health promotion in the mid1980s, defining it as “a process of enabling people to increase control over, and to improve their health.” 79 It went on to state that health promotion represents “a mediating strategy between people and their environments, synthesizing personal choice and social responsibility in health.” 79 The principles set forth by WHO as underlying this alternative

vision of health promotion included acting on the determinants or causes of health, eliciting high-level public participation, and using a variety of approaches that go well beyond lifestyle education and that include legislation, organizational change, and community development.79,80

The Canadian approach to health promotion that was developed in the mid-1980s and refined over the subsequent decade provides an illustration of how such a broadened vision may constitute a useful framework for action. After a period of considerable preoccupation with healthy lifestyles and individual responsibility for health, the Canadian government undertook a massive restructuring of its approach to health promotion. Two important points stand out in the Canadian approach. First, the number one challenge set forth for health promotion is reducing inequities between low- and high-income groups, and this is not framed in terms of individual responsibility but of broader societal responsibility.81 Second, three levels of concern are set forth—health challenges, health promotion mechanisms, and implementation strategies—and on each of these levels there is attention to the role of broad institutional or environmental change (see Figure 13.3). Self-care, for example, is advocated within a framework that devotes considerable attention to the creation of healthy environments within which positive personal health behaviors can flourish. Canadian legislation on smoking is among the toughest in the world, with many provinces having developed “healthy public policies” on tobacco that have included changing their policies on marketing, crop substitution, and smoking in the workplace at the same time that they urge individuals to quit the habit.

In several Canadian provinces, premier's councils on health have been established through which government leaders in the different sectors provide advice on health promotion and work together in jointly setting goals for helping to address the social determinants of health.80 In the Northwest Territories, land claims and the development of First Nation's People's rights have been discussed as part of a broadly defined health agenda.82 Finally, across the nation, hundreds of cities have designated themselves “healthy communities,” stressing intersectoral planning, high-level community participation, and reciprocity between the individual and the broader society.83

Hard outcome data that would indicate whether the new Canadian approach to health promotion has resulted in actual declines in morbidity and mortality are not yet available. Further, as Irving Rootman 84 and Lawrence Green 85 predicted, increased government perceptions of a need for cutbacks in social spending have led to some redesign of social


Figure 13.3. A framework for health promotion. Source: Epp (1986 [79]).

programs in ways that are constricting the implementation of health promotion, in much the same way that occurred in the United States under the Reagan and Bush administrations. Although the Canadian Public Health Association's 86 recent “Action Statement for Health Promotion in Canada” reaffirmed the importance of continuing to use the Ottawa Charter as “the framework that defines health promotion in Canada,” it went on to note that the current climate of increasing poverty, environmentally harmful global economic practices, and cuts in the very health and social programs that have defined Canadians “as a caring people” present a stark contrast to “the optimistic days when the Ottawa Charter was first written.” 86 The Action Statement went on to reaffirm those visions and values deemed essential to health promotion, among them an “explicit value base” that includes a commitment to respecting individual liberties but giving priority to the common good, giving priority to people whose living conditions place them at greater risk, and pursuing
social justice “to prevent systemic discrimination and to reduce health inequities.” 86

As noted earlier, even health promotion efforts that are firmly grounded in a values base like this one may stumble as a result of continuing budget cuts, growing power differentials between rich and poor, and the structural distinctions that continue to exist between communities and health and social service professionals. At the same time, however, the Canadian framework for health promotion and the values and principles underlying it stand as an important example of a vision that offers a balanced concern for personal behavior change within the context of broader social change. In its attention to broader environmental issues, moreover, the Canadian approach explicitly addresses the connection between health promotion and the roots of public health in concerns for social justice and improved social and economic conditions as a vital part of public health.

The classic illustration of this connection lies in the well-known story of Rudolph Virchow,87 the founder of cellular pathology, who in 1848 was asked by his superiors to provide recommendations on what to do about a typhus epidemic that was raging in an impoverished area in eastern Prussia. After visiting the area and studying the situation, Virchow prepared a report in which he spoke of the need for land reform, redistribution of wealth and income, and only later of some medical reforms. The report was met with displeasure, and Virchow was accused of proposing a political rather than a medical solution to the epidemic. His often-cited reply was that “medicine is a social science and politics nothing more than medicine on a large scale.” 87

Taking a cue from Virchow, health promotion and the ways in which we define and operationalize it must be seen in inherently political and multidisciplinary terms if we are to be successful in meeting the profound challenges to the public's health that we face in this last decade of the 20th century. As health professionals, we must ask ourselves what our own responsibilities are in relation to health promotion programs and activities. Do we contribute to social equity, or do we reinforce existing inequities? Do we advance the concept of individual and community empowerment, or do we merely use the language of community partnerships and empowerment to blur hierarchical distinctions without changing the status quo? 33 Do our health promotion activities inadvertently contribute to problems such as gerontophobia and handicappism? Do they, in Marshall Becker's 88 words, equate “being ill” with “being guilty?” And finally, is our notion of community good broad enough to reflect an

appreciation of our intimate interdependence? With Larry Churchill,72 are we able to see individual freedoms and social bonds as complementary rather than opposing forms of human well-being?

The story is told of a traveler, hopelessly lost on a dusty dirt road, who stops a local farmer and asks, “Is this the road to Baduka?” The farmer replies, “Mister, if I was going to Baduka, I wouldn't start from here.” The dominant American self-image stressing individualism, self-reliance and progress, and the “ethical individualism” to which it predisposes us 72 makes our starting point on the road to a broader approach to health promotion a difficult one indeed. But as we enter a new century and a new millennium, the time is ripe for considerably broadening the ways in which we think about and approach health promotion.


1. “U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control.” 1993. Public health in the new American health system. Discussion paper, Bethesda, Maryland.

2. Leichter, H. M.1991. Free to Be Foolish: Politics and Health Promotion in The United States and Great Britain.Princeton, N.J.: Princeton University Press.

3. Walker, S. N.1994. “Health promotion and prevention of disease and disability among older adults: Who is responsible?” Generations (spring), 45–50.

4. Iglehart, J. K.1990. “From the editor: Special issue on promoting health.” Health Affairs9 (2), 4–5.

5. Wallack, L., and Montgomery, K.1992. “Advertising for all by the year 2000: Public health implications for less developed countries.” American Journal of Public Health Policy13 (2), 76–100.

6. Lalonde, M.1974. A New Perspective on the Health of Canadians.Ottawa: Government of Canada.

7. “U.S. Surgeon General.” 1979. Healthy People: The Surgeon General's Report on Health Promotion and Disease Prevention.Washington, D.C.: Department of Health and Human Services.

8. Hancock, T.1986. “Lalonde and beyond: Looking back at “A new perspective on the health of Canadians.”” Health Promotion1, 93–100.

9. Terris, M.1984. “Newer perspectives on the health of Canadians: Beyond the Lalonde report.” Journal of Public Health Policy5, 327–337.

10. Newbauer, D., and Pratt, R.1981. “The second public health revolution: A critical appraisal.” Journal of Health Politics, Policy and Law6 (2), 327–337.

11. “U.S. Surgeon General.” 1980. Health promotion Disease prevention: Objectives for the nation.Washington, D.C.: Department of Health and Human Services.

12. Green, L. W., Kreuter, M., Deeds, S. G., and Partridge, K. B.1980. Health

Education Planning: A Diagnostic Approach.Palo Alto, Calif.: Mayfield Publishing.

13. “Personal communication from L. W. Green” , May 17, 1988.

14. “Personal communication from E. Zimmerman” , October 16, 1980.

15. Stokols, D., Pelletier, K., and Fielding, J.1996. “The ecology of work and health: Research and policy directions for the promotion of employee health.” Health Education Quarterly23 (2), 137–158.

16. Weinger, M., and Wallerstein, N.1990. “Education for action: An innovative approach to training hospital employees.” In Essentials of Modern Hospital Safety, edited by W. Charney and J. Schirmer. New York: Lewis Publishers.

17. Wallerstein, N., Sanchez-Merki, V., and Dow, L.1997. “Freirian praxis in health education and community organizing: A case study of an adolescent prevention program.” In Community Organizing and Community Building for Health, edited by M. Minkler. Pp. 195–211. New Brunswick, N.J.: Rutgers University Press.

18. Wohlfeiler, D.1997. “Community organizing and community building around gay and bisexual men: The Stop AIDS project.” In Community Organizing and Community Building, edited by M. Minkler. Pp. 230–244. New Brunswick, N.J.: Rutgers University Press.

19. Berkman, L., and Syme, S. L.1976. “Social class—Susceptibility and sickness.” American Journal of Epidemiology104, 1–8.

20. Kaplan, G. A., Haan, M., Syme, S. L., et al. 1987. “Socioeconomic status and health.” American Journal of Epidemiology125, 989–998.

21. Marmot, M. G., Rose, G., and Hamilton, P. J. S.1978. “Employment grade and coronary heart disease in British civil servants.” Journal of Epidemiology and Community Health32, 244–249.

22. Syme, S. L.1990. “Control and health: An epidemiological perspective.” In Self Directedness: Cause and Effects throughout the Life Course, edited by J. Rodin, C. Schooler, and K. W. Schaie. Pp. 213–229. Hillsdale, N.J.: Lawrence Erlbaum Associates. New York: Wiley (in association with the Commission of European Communities).

23. Minkler, M.1997. “Community organizing among the elderly poor in San Francisco's Tenderloin District: A case study.” In Community Organizing and Community Building for Health, edited by M. Minkler. Pp. 244–258. New Brunswick, N.J.: Rutgers University Press.

24. Minkler, M.1994. “Challenges for health promotion in the 1990s: Social inequities, empowerment, negative consequences, and the common good.” American Journal of Health Promotion8 (6), 403.

25. Kent, J.1970. A descriptive approach to community. Unpublished report, Denver, Colorado.

26. Miller, M.1985. Turning Problems into Actionable Issues.San Francisco: Organize Training Center.

27. French, M.1986. Beyond Power: On Women, Men and Morals.London: Abacus.

28. Labonte, R.1994. “Health promotion and empowerment: Reflections on professional practice.” Health Education Quarterly21 (2, summer), 253–268.

29. Avery, B. Y.1990. “Breathing life into ourselves: The evolution of the Black

Women's Health Project.” In Black Women's Health Book, edited by E. C. White. Pp. 4–10. Seattle: Seal Press.

30. Schurman, S. J., and Israel, B.1995. “Redesigning work systems to reduce stress: A participatory action research approach to creating change.” In Job Stress Intervention: Current Practices and New Directions, edited by G. Keiter, S. Sauter, J. Hurrell, and L. Murphy. Pp. 235–263. Washington, D.C.: American Psychological Association.

31. Siler-Wells, G. L.1989. “Challenges of the Gordian knot: Community health in Canada.” In International Symposium on Community Participation and Empowerment Strategies in Health Promotion, edited by J. Warren Salmon and Eberhard Goepel. Pp. 42–55. Bielefeld, Germany: Center for Interdisciplinary Studies, University of Bielefeld.

32. Rubin, H., and Rubin, I.1992. Community Organizing and Development. 2nd ed. New York: Macmillan.

33. Minkler, M., and Pies, C.1997. “Ethical issues in community organizing and community participation.” In Community Organizing and Community Building for Health, edited by M. Minkler. Pp. 120–136. New Brunswick, N.J.: Rutgers University Press.

34. Labonte, R.1997. “Community, community development and the forming of authentic partnerships: Some critical reflections.” In Community Organizing and Community Building for Health, edited by M. Minkler. Pp. 88–102. New Brunswick, N.J.: Rutgers University Press.

35. McGinnis, J. M., and Goege, W. H.1993. “Actual causes of death in the United States.” Journal of the American Medical Association270 (18), 2207–2212.

36. “Trends in rates of homicide—United States” , 1985–1994. 1996. Morbidity and Mortality Weekly Report45 (22, June 7), 460–464.

37. Rachuba, L., Stanton, B., and Howard, D.1995. “Violent crime in the United States: An epidemiologic profile.” Pediatric and Adolescent Medicine149, 953.

38. “Rates of homicide, suicide and firearm-related death among children—26 industrialized countries.” 1997. Morbidity and Mortality Weekly Report46 (5, February 7), 101–105.

39. Thurow, L.1996. The Future of Capitalism.New York: William Morrow.

40. Prothow-Stith, D.1995. “The epidemic of youth violence in America: Using public health prevention strategies to prevent violence.” Journal of Health Care for the Poor and Underserved6 (2), 95.

41. Prothow-Stith, D.1995. “Violence prevention with youth.” Keynote presentation for the Annual Conference of the American Journal of Health Promotion, Orlando, Florida, March 20.

42. Tucker, C.1997. “Boston shows how to deal with teens.” San Francisco Chronicle, December 20, A20.

43. “RAND and Stanford Center for Research in Disease Prevention.” 1997, June. The California Wellness Foundation Violence Prevention Initiative Mid-Initiative Assessment, Volume 1. Santa Monica, Calif.

44. “California Department of Health Services.” 1997. Injury deaths for homicides by age and cause, California 1995. Sacramento, May 22.


45. Robertson, A., and Minkler, M.1994. “The new health promotion movement: A critical examination.” Health Education Quarterly21 (3), 295–312.

46. Schauffler, H. H., D'Ogostino, R. B., and Kannel, W. B.1993. “Risk for cardiovascular disease in the elderly and associated Medicare costs: The Framingham Study.” American Journal of Preventive Medicine9 (3), 146–154.

47. Davis, K., Bialek, R., Parkinson, M., et al. 1990. “Paying for preventive care: Moving the debate forward.” American Journal of Preventive Medicine6 (4), 7–30.

48. Minkler, M., and Pasick, R.1986. “Health promotion and the elderly: A critical perspective on the past and future.” In Wellness and Health Promotion of the Elderly, edited by K. Dychtwald. Pp. 39–54. Rockville, Md.: Aspen Systems Corporation.

49. Lynch, M., and Minkler, M.1998. “The restructuring of Medicare and Medicaid and its impacts on the elderly: A conceptual framework and analysis.” Critical Gerontology: Perspective from Political and Moral Economy, edited by M. Minkler and C. L. Estes. Pp. 185–201. Amityville, N.Y.: Baywood Publishing.

50. Cole, T.1988. “The specter of old age: History, politics and culture in an aging America.” Tikkun3 (5), 14–18, 93–95.

51. Levin, D.1987. Pathologies of the Modern Self.New York: University Press.

52. Wang, C.1982. “Culture, meaning and disability: Injury prevention campaigns in the production of stigma.” Social Science and Medicine3 (5), 1093–1102.

53. “U.S. Department of Health and Human Services, Office on Smoking and Health.” 1989. Smoking, Tobacco and Health: A Factbook.Washington, D.C.: General Accounting Office.

54. “U.S. Department of Health and Human Services, Office on Smoking and Health.” 1991. Trends in Cigarette Smoking Prevalence in the United States, 1965–1991. Washington, D.C.: U.S. Government Printing Office. (data from the National Health Interview Surveys, 1965–1991, compiled by the Office on Smoking and Health)

55. “State-specific prevalence of current cigarette and cigar smoking—U.S.” 1998. Morbidity and Mortality Weekly Report45 (48, November), 1034–1039.

56. “U.S. Department of Health and Human Services, Office on Smoking and Health.” 1989. Reducing the Health Consequences of Smoking: 25 Years of Progress: A Report of the Surgeon General.Washington, D.C.: U.S. Government Printing Office.

57. Warner, K. E.1986. Selling Smoke: Cigarette Advertising and Public Health.Washington, D.C.: American Public Health Association.

58. “U.S. Department of Health and Human Services.” 1996. National Survey Results on Drug Use from the Monitoring the Future Study, 1975–1995. Volume 1: Secondary School Children, Table 3, 60. Washington, D.C.: U.S. Department of Health and Human Services.

59. Johnston, D. C.1997. “Anti-tobacco groups push for higher cigarette taxes.” New York Times, April 3, A1, A18.


60. Quindlen, A.1992. “Good causes, bad money.” New York Times, November 15, A1.

61. Humphrey, H. H., III. 1997. “Let's take the time to get it right.” Public Health Reports112, 378–385.

62. Skolnick, A.1994. “Anti-tobacco advocates fight “illegal” diversion of tobacco control money.” Journal of the American Medical Association271 (18), 1387–1389.

63. “Tobacco deal: Public health advocates say prlls short.” 1997. The Nation's Health, July, 1, 24.

64. “Clinton responds.” 1997. The Nation's Health, October, 7.

65. “National Association of Attorneys General.” 1999. Tobacco settlement proceeds to be released to states, tobacco sales down during first year since settlement. PRNewsire, November 12.

66. “National Association of Attorneys General.” 1999. Tobacco wars, still. Washington Post, December 29, A26.

67. “World Health Organization, Tobaclth Program.” 1996. The tobacco epidemic: A global public health emergency.Tobacco Alert: Special Issue. World No-Tobacco Day, 1996. Geneva, Switzerland: World Health Organization.

68. Lee, G. A.1995. “Mixed review for environment study: Growing population, cigarette production marked '94.” Washington Post, May 21, A5.

69. Brown, L.1997. State of the World, 1997.New York: World Watch Institute.

70. Chapman, S., and Wong, W.1990. Tobacco Control in the Third World: A Resource Atlas.Penang, Malaysia: International Organization of Consumers Unions.

71. Kadlec, D.1997. “How tobacco firms will manage.” Time, June 30, 29.

72. Churchill, L.1987. Rationing Health Care in America: Perceptions and Principles of Justice.Notre Dame, Ind.: University of Notre Dame Press.

73. Levy, L.1957. The Law of the Commonwealth and Chief Justice Shaw.Cambridge, Mass.: Harvard University Press.

74. Beauchamp, D. E.1985. Community: The Neglected Tradition of Public Health.Hastings-on-the-Hudson, N.Y.: The Hastings Center.

75. Wickler, D.1987. “Who should be blamed for being sick?” Health Education Quarterly14 (1), 11–25.

76. Callahan, D.1987. Setting Limits: Medical Ethics in an Aging Society.New York: Simon and Schuster.

77. Edelman, P.1997. “The worst thing Bill Clinton has done.” Atlantic Monthly, March, 43–58.

78. “Families USA.” 1996. Hurting Real People: The Human Side of Medical Cuts.Washington, D.C.: Families USA.

79. “World Health Organization.” 1984. Report of the Working Group on the Concept and Principles of Health Promotion.Copenhagen: World Health Organization.

80. “World Health Organization.” 1986. Ottawa Charter for Health Promotion.Copenhagen: World Health Organization.

81. Epp, J.1986. Achieving Health for All: A Framework for Health Promotion.Ottawa: National Health and Welfare, Government of Canada.


82. Yazdanmehr, S.1994. “Northwest Territories.” In Health Promotion in Canada, edited by A. Pederson, I. Rootman, and M. O'Neill. Pp. 226–243. Toronto: W. B. Saunders.

83. Pederson, A., Rootman, I., and O'Neill, M., eds. 1994. Health Promotion in Canada.Toronto: W. B. Saunders.

84. “Personal communication from I. Rootman” , January 12, 1994.

85. Green, L. W.1994. “Canadian health promotion: An outsider's view from the inside.” In Health Promotion in Canada, edited by A. Pederson, I. Rootman, and M. O'Neill. Pp. 314–326. Toronto: W. B. Saunders.

86. “Canadian Public Health Association.” 1996, July. Action Statement for Health Promotion in Canada.Ottawa: Canadian Public Health Association.

87. Taylor, R., and Rieger, A.1985. “Medicine as a social science: Rudolf Virchow on the typhus epidemic in Upper Silesia.” International Journal of Health Services15 (4), 547–559.

88. Becker, M.1986. “The tyranny of health promotion.” Public Health Review14, 15–23.



California Smokers' Helpline

Shu-Hong Zhu and Christopher M. Anderson


Smoking is a major risk factor for heart disease, lung cancer, and strokes and is the most important preventable cause of premature death in the United States.1 In California alone, over 42,000 deaths each year are attributable to smoking.2 However, studies have consistently shown that smokers who achieve long-term cessation significantly reduce their risk of disability and early death.3 For that reason, efforts to persuade smokers to quit and to help them with the process have been encouraged on both the federal and state levels.4,5 In recent years, in particular, many states have launched aggressive antismoking campaigns. With revenues from a voter-approved tobacco tax initiative (Proposition 99), California has developed one of the most comprehensive tobacco control programs in the country, with a budget that includes funding for a variety of smoking cessation interventions.6–9

Such interventions have been said to fall into two categories: the clinical approach, with its emphasis on depth of intervention, and the public health approach, which emphasizes broad-based interventions.10,11 The clinical approach is characterized by intensive, multisession interventions; small populations; and high quit rates. The public health approach, on the other hand, is characterized by brief, low-cost interventions; large populations; and relatively low quit rates. There are many studies and service programs that fall within one framework or the other, but programs that successfully combine the intensity of the clinical approach with the breadth of the public health approach are few.


One that does attempt to bridge the two approaches is the California Smokers' Helpline, a telephone-based program that was established as a statewide service in August 1992. Because it is telephone based, the Helpline is accessible to nearly all smokers in the state and is thus in a position to play a strong public health role as both a referral agency and a clearinghouse for cessation materials. In addition, the Helpline also provides intensive, multicomponent, multisession behavioral modification counseling for any who need it. Though intensive, this telephone counseling intervention is focused and brief in its approach, emphasizing costefficiency in its design. In this way, the California Smokers' Helpline strives to put the benefits of a strong clinical program within reach of a much broader population than more traditional programs have been able to do.


One problem with the traditional clinical approach to smoking cessation is that although psychologically intensive programs are effective in helping smokers quit smoking,12 few smokers use them.13,14 One national survey has shown that only 10% of smokers who tried to quit smoking used a program to help them do so.14 The reason for this is unclear; one reasonable guess would be that the remaining 90% do not want any help. But surveys have also shown that most smokers are worried about the prospect of quitting smoking. According to the 1990 California Tobacco Survey, 77.6% of smokers believe they are addicted to cigarettes, and 85.9% consider it important that there be programs to help smokers quit.15 Moreover, there is a striking difference in help-seeking patterns among ethnic groups, with smokers of ethnic minority backgrounds less than half as likely as White smokers to seek help to quit.16

Even smokers who want to quit and who believe they need help may not know where to find that help. For example, a 1990 survey conducted in San Diego County asked 1,049 smokers to name “up to three programs that are helpful for people who want to quit smoking.” Surprisingly, only 39.4% of smokers were able to name any program at all relating to smoking cessation. Furthermore, non-White smokers were even less aware of available programs (31.3%) than White smokers were (41.2%), suggesting one reason why minority smokers are less likely to seek help.17

Even if smokers do know where help can be found, they must still

weigh the costs and benefits of obtaining that help. A decision-making model would suggest that only when the perceived benefits outweigh the perceived costs will smoker PARTs take action to attend a program.18 Attendance fees are not the only cost involved. Smokers may weigh less tangible costs as well, such as time away from home, scheduling difficulties, child care and transportation problems, and loss of anonymity. All of these costs must be weighed against the uncertain benefits of success in quitting. For those smokers who would like help quitting, these costs may lessen the perceived accessibility of that help.


As part of an effort to redress this problem, the Tobacco Control Section of the California Department of Health Services began funding the California Smokers' Helpline in August 1992. Smokers from across the state are referred to the Helpline through a variety of means, including media advertisements and physician referral. When they call, they are offered a range of services according to their preference and their readiness to quit. Smokers who are not yet ready to quit are sent materials designed to spur them along, while those who do feel ready but who prefer to quit on their own receive self-help quit kits. Smokers who would like more intensive help can enroll in the Helpline's free telephone counseling program.

The first way in which the Helpline increases the accessibility of smoking cessation services is by educating smokers on where they can get help to quit. The Helpline is aggressively promoted by the media component of the state's Tobacco Control Section, which includes the Helpline's toll-free numbers in ads urging smokers to quit. These ads, in the six languages in which Helpline services are provided, appear on television, radio, and billboards and in newspapers across the state. The Helpline also provides detailed information about its services to health care providers and volunteer organizations so that they have a dependable referral source. Moreover, the Helpline publicizes other cessation programs by sending each caller a descriptive list of all the available programs in his or her area. In these ways, public awareness of help for smoking cessation is enhanced.

The Helpline also reduces many of the costs associated with getting help to quit smoking. The first of these is the actual financial cost. Supported

by revenues from the state tax on cigarettes and other tobacco products, the Helpline offers help at no charge to the caller.

As mentioned previously, there are other less tangible costs that smokers may incur from traditional programs, including having to wait for cessation classes to form, taking time away from home to attend them, and the effort and expense of arranging for transportation and child care. Even the potential benefit of social support from attending group sessions may be outweighed, for some, by the prospect of facing a roomful of strangers. Some smokers face geographic or language barriers as well. The Helpline reduces all of these costs by enabling smokers to get help without leaving home and by providing services in six of the state's most common languages—English, Spanish, Mandarin, Cantonese, Vietnamese, and Korean. (The Helpline also has a TDD line for the hearing impaired and has recently added a line for smokeless tobacco users.)

The Helpline is also able to increase its accessibility by stretching its resources. The program operates out of a single site, minimizing overhead costs. Also, it employs a stepped-care approach; 19 that is, instead of attempting to provide every caller with its most intensive form of assistance—telephone counseling—it presents a variety of options, including simply receiving materials in the mail or attending one of the cessation groups noted on the referral list. The Helpline lets each caller select the services that he or she feels would be most useful. By allowing callers to “serve themselves” from a menu of different approaches, the Helpline is able to spread its resources among a greater number of people. In just over four years of operation, the Helpline has served more than 41,000 smokers in this way, an average of more than 10,000 people a year. An additional 2,300 nonsmokers have also called the Helpline to get help for their friends and families.

In many cases, it is smokers' own ambivalence that limits their access to programs. A study conducted in southern California when the Helpline was funded to provide counseling only in San Diego County may illustrate this point. In response to calls from more than 700 Los Angeles smokers who said they were planning to quit within a month, the Helpline sent self-help materials and a directory of all the smoking cessation programs in the Los Angeles area. Five weeks later, the Helpline staff called them back and asked if they had attended any of the programs in the directory. Only 6.2% had done so. Given that the first time they called they had appeared motivated to get help to quit, the low rate at which they actually did so suggests that, as a group, they experienced

considerable ambivalence about using the available programs. It seems reasonable to suppose that many in the group found the idea of committing to a program—and thus of being obliged to take action to quit smoking—to be an uncomfortable prospect.17

To counteract this ambivalence and reduce the resulting attrition, the Helpline tested a proactive approach with its counseling clients. As part of a larger study, more than 3,000 smokers who said they were ready to quit within a week and who opted for telephone counseling were told by intake personnel that they would receive a packet of quitting materials in a few days, at which time they should call back to begin the counseling. Subjects were then randomized into two groups. Members of one group were left to call back as instructed; about 34.2% eventually did so and received counseling. In contrast, members of the other group were contacted directly by a counselor. In this group, 74.7% received counseling, demonstrating that a proactive approach to providing service can have a strong counteractive effect on clients' ambivalence.20


With respect to the population that uses the Helpline, two demographic dimensions deserve special attention. One is its ethnic diversity, and the other is the active participation of rural smokers.

Table 14.1 indicates the ethnic diversity of the population using the Helpline. As the table shows, the 1993 California Tobacco Survey (CTS) found that smokers of ethnic minority backgrounds were underrepresented among those who sought help in general to quit smoking. His-panic/Latino smokers, for example, accounted for 18.5% of the smokers in California but only 9.4% of those who sought help to quit. African-American smokers made up 7% of the state's smokers but only 4.3% of those who sought help. Asian-American smokers were likewise underrepresented (5% vs. 2.6%).

On the other hand, ethnic minority smokers were about as well represented among Helpline users as among smokers in general. Hispanic/Latino smokers accounted for 18.2% of Helpline users, a proportion approaching their representation among the state's smokers. African-American smokers were actually overrepresented among Helpline users (11.4% vs. 7%). Asian-American smokers were still underrepresented (2.6% vs. 5%). Taken as a group, however, smokers of ethnic minority

  Smokers in
1992–1993 (%)
Sought Help
≤12 Months
Prior to 1993
CTS (% [±95 CI])
Called the
(% [±95% CI]
NOTE: Because only adult smokers participated in the 1993 CTS, Helpline callers under 18 are excluded from this analysis. SOURCES: 1993 California Tobacco Survey and the California Smokers' Helpline. Adapted from Pierce et al. (1994 [8]) and Zhu et al. (1995 [16]).
N 4,078,306 3,425 39,903
White 67.4 78.9 (4.9) 62.3 (0.5)
Hispanic 18.5 9.4 (3.4) 18.2 (0.4)
Black 7.0 4.3 (3.0) 11.4 (0.3)
Asian 5.0 2.6 (1.9) 2.6 (0.2)
Others 2.0 4.8 (3.5) 5.4 (0.2)
backgrounds were better represented among Helpline users (37.6%) than among smokers seeking help in general (21.1%).

The fact that the Helpline's services are available in several languages, each with its own 1-800 number, contributes to its success in recruiting minority smokers. The high proportion of Hispanic callers, for example, is due largely to the Helpline's Spanish line. In fact, 71% of all Hispanic callers use that line. Conversely, the low percentage of Asian callers may be due in part to the relatively late addition of the Asian lines, 19 months after the English and Spanish lines.

Through targeted advertising an