In the last several years a spirited debate has developed in the epidemiologic literature about the appropriate focus for epidemiologic research and practice. 1–8 Some have argued for a focus on the distribution and determinants of disease at the population level while others have emphasized the importance of identifying disease risk factors at the individual level. Proponents of the population approach argue that, while risk factors and risky behaviors are observed in individuals, they are generated and reinforced in social contexts. Therefore, interventions should take account of these contexts. Those in support of a risk factor approach agree with this idea in general, but feel that interventions are more realistically directed at changing individual behavior, rather than social structures.
Social epidemiologists have struggled with these same issues. These debates have been strongly influenced by the work of three seminal thinkers: Emile Durkheim, Geoffrey Rose, and John Cassel. Durkheim studied one of the most intimate behaviors imaginable: suicide. 9 He argued that although the causes of suicide must be found in the individual life histories of individuals, there was a patterned regularity in suicide rates among population groups, over time, even as individuals moved in and out of these groups. He suggested that there must be something about the social environment that promoted a characteristic suicide rate for the group. While these population characteristics do not determine which individuals in the group commit suicide, they do help explain the group differences in rates over time.
Geoffrey Rose laid out two fundamental principles not entirely dissimilar from the points made by Durkheim. 10 First, groups of individuals function collectively and are affected by the average functioning of those around them. Second, due to this first principle, we should examine entire population distributions rather than just those at “high risk” because it is the average rate that is of import, not the extremes of the distribution. Rose observed that when a risk factor is normally distributed within a population, lowering everyone’s risk even by a modest amount has a greater overall effect than if only those at high risk change their risk profiles. Change therefore comes about by influencing entire populations and the conditions to which these populations are exposed.
John Cassel 11 suggested that a focus on risk factors was misplaced because many biological risk factors (such as viruses and bacteria) are ubiquitous in the environment: because almost everyone is exposed to them most of the time, exposure to these risk factors cannot help explain why some individuals get sick and some do not. He suggested that we must better understand factors that affect an individual’s susceptibility to disease. He argued that factors affecting susceptibility are to be found not only in an individual’s biologic makeup but also in the social environment in which individuals live and that our research focus should shift to study those environmental forces.
In 1994 and 1995, two pioneering books were published dealing specifically with a population approach to the study of disease:Why are Some People Healthy and Others Not? 12 and Society and Health. 13 Since then, eight additional books have been published. All suggest that a population approach is most appropriate for the study of social relations, social determinants and the prevention of disease. These books are:
1996 -Unhealthy Societies. The Afflictions of Inequality 14
1997 -A Life Course Approach to Chronic Disease Epidemiology 15
1998 -Independent Inquiry into Inequalities in Health 16
1999 -Social Determinants of Health 17
1999 -The Society and Population Health Reader. Volume 1: Income Inequality and Health 18
1999 -The Society and Population Health Reader. Volume II: A State and Community Perspective 19
2000 -Social Epidemiology 20
2001 -Income, Socioeconomic Status and Health: Exploring the Relationships 21
The publication of so many excellent books dealing with the topic of social factors and disease in such a short period of time is a remarkable phenomenon. Each book has its own emphasis but they have in common an effort to understand the importance of social forces in determining the distribution of diseases in the population and the way in which intervention strategies can be developed to change those forces. Nevertheless, in coming to the conclusion that a population approach is the one most appropriate for the study of social determinants of disease, these books bring to the fore several difficult issues that inevitably emerge when such a perspective is embraced: (1) how to properly conceptualize and study populations, (2) how to understand disease at the population level, and (3) how to develop new strategies that focus on populations rather than on individuals. Of course, these issues are of interest not only to social epidemiologists but also to epidemiologists in general.
How to Conceptualize and Study Populations Properly
While epidemiology is concerned with the study of the determinants and distribution of disease in populations, epidemiologic studies often do not actually study populations, but rather individuals living in those populations. To address this issue, all of these books on social epidemiology focus on the study of the relations of people to one another in social institutions such as families, schools, and communities. This focus goes beyond the identification of individual-level risk factors, and it therefore provides some first steps toward a fundamentally different approach to understanding social determinants and possible interventions. The books by Berkman and Kawachi 20 and Amick et al, 13 for example, provide entire chapters on the role of ecological approaches in social epidemiology and on the topic of community and health.
The study of disease in social groups is certainly helpful in shifting the focus of epidemiology away from an exclusively individual risk factor model. There is a caveat, however, for future epidemiologic thinking. Among the increasing numbers of studies examining the distribution of disease in neighborhoods, for example, most treat the neighborhood as the unit of observation and thereby combine data from individuals. For example, health outcomes are regressed on the average income of all individuals within the neighborhood. This tells us something about each individual’s participation in the prevalence of disease in the neighborhood, but it tells us next to nothing about the role of the neighborhood itself in bringing about disease. In contrast, social capital and income inequality are examples of characteristics of communities. These concepts refer to the social features of the community, in contrast to such individual level phenomena as social support and income, and they have been shown to be important determinants of health, independent of these individual characteristics. In order for epidemiology to take the next step away from the individual risk factor paradigm we need to think about ways to measure social relations as more than the sum of aggregated individual attributes.
How to Understand Disease at the Population Level
Everyone knows the difficulty we now have in identifying risk factors for disease. For many diseases, the risk factors we have identified account for only a fraction of the disease that occurs. Clearly more and better research will increase that fraction over time. Another approach to improving our understanding of disease is to rethink the way in which we (1) conceptualize fundamental biologic processes and (2) identify fundamental social determinants of disease. We must therefore reconsider the way we conceptualize both determinants and outcomes.
1) Rethinking Fundamental Biologic Processes
One of the distinctive contributions of the new books on social epidemiology is that they do not organize their material in terms of clinical disease categories. This is helpful because a focus on clinical diseases makes it difficult to understand disease at the population level; a clinical perspective inevitably draws attention to individuals and individual risk factors. As these books point out, social factors are related etiologically not just to one, but to many clinical diseases. It is thus likely that social factors affect fundamental biologic processes (for example, the neuroendocrine, autonomic, and immune systems, as well as central neural circuits) that, in turn, render people susceptible to a wide range of disease outcomes. In this circumstance, it might be useful to study relationships between social factors and fundamental biologic processes that underlie these various expressions of clinical disease.
2) Rethinking Fundamental Social Determinants
The fact that social and behavioral factors are related to many disease outcomes has consequences also for the way we think about social determinants of disease. All of the books reviewed here suggest that there may be elements in common between such variables as social class and race that might compromise population disease resistance; social factors make groups vulnerable to a wide range of diseases (whether it be HIV, cardiovascular disease, or diabetes). It is not a question of adding up the effects of determinants such as class, race, and social support on individuals, but rather a question of understanding what these variables have in common that might affect populations and produce this panoply of outcomes.
How to Develop New Strategies that Focus on Populations Rather Than on Individuals
Most health interventions focus on helping individuals reduce their risk of disease by changing their behavior. We ask people to smoke less, eat differently, exercise more, take prescribed medications, drive more carefully, and so on. There are two difficulties with this approach. One is that it has proven very difficult for people to make changes in their behavior. The second problem is that, even if they did, new people would continue to enter the at-risk population at an unaffected rate because we rarely target those aspects of the environment that cause the problem in the first place. If we could more precisely identify environmental risk factors, we could solve both of these difficulties simultaneously. An example is legislation to improve work conditions rather than to develop stress management programs for individuals at work. Each of these books offers a variety of new perspectives on this issue that help us see practical ways in which policies might be developed to affect the distribution of disease at the population level.
It is a remarkable phenomenon in the history of epidemiology for 10 new books dealing with social factors and disease to have been published within the last 8 years. Fifty years ago, there was almost no research or published works on such issues. The books noted in this review are thoughtful and sophisticated contributions that raise important ideas for all epidemiologists to consider. Obviously, a population perspective on the determinants of disease is important. A major challenge for future work is to develop better trans-disciplinary thinking regarding the study of health and disease, and these books are an important step toward meeting that challenge.
We thank W. Thomas Boyce, Elize Brown and John Frank for their helpful suggestions in the preparation of this book review.
1. Gori GB. Epidemiology and public health: Is a new paradigm needed or a new ethic? J Clin Epidemiol 1998; 51: 637–641.
2. Gori GB. Reply to the preceding dissent. J Clin Epidemiol 1998; 51: 647–649.
3. Pearce N. Epidemiology as a population science. Int J Epidemiol 1999; 28: S1015–S1018.
4. Rothman KJ, Adami H-O, Trichopoulos D. Should the mission of epidemiology include the eradication of poverty? The Lancet 1998; 352: 810–813.
5. Krieger N. Sticky webs, hungry spiders, buzzing flies, and fractal metaphors: On the misleading juxtaposition of “risk factors” versus “social epidemiology.” J Epidemiol Community Health 1999; 53: 678–680.
6. McKinlay JB, Marceau LD. A tale of 3 tails. Am J Public Health 1999; 89: 295–298.
7. Pearce N, McKinlay JB. Back to the future in epidemiology and public health: Response to Dr. Gori. J Clin Epidemiol 1998; 51: 643–646.
8. Susser M. Should the epidemiologist be a social scientist or a molecular biologist? Int J Epidemiol 1999; 28: S1019–S1021.
9. Durkheim E. Suicide. In: Simpson G, ed. A study of Sociology. Glencoe, Il: Free Press, 1951.
10. Rose G. The Strategy of Preventive Medicine. Oxford, England: Oxford University Press, 1992.
11. Cassel J. The contribution of the social environment to host resistance. Am J Epidemiol 1976; 104: 107–123.
12. Evans RG, Barer ML, Marmor TR. Why Are Some People Healthy and Others Not? The Determinants of Health of Populations. New York: Aldine de Gruyter, 1994.
13. Amick BJ III, Levine S, Tarlov AR, Chapman Walsh D. Society and Health. New York: Oxford University Press, 1995.
14. Wilkinson RG. Unhealthy Societies. The Afflictions of Inequality. New York: Routledge, 1996.
15. Kuh D, Ben Shlomo Y. A Life Course Approach to Chronic Disease Epidemiology. Oxford: Oxford University Press, 1997.
16. Acheson Sir D. Independent Inquiry into Health Inequalities Report. London: The Stationary Office, 1998.
17. Marmot M, Wilkinson RG. Social Determinants of Health. Oxford: Oxford University Press, 1999.
18. Kawachi I, Kennedy BP, Wilkinson RG. The Society and Population Health Reader. Volume 1: Income Inequality and Health. New York: The New Press, 1999.
19. Tarlov AR, St. Peter RF. The Society and Population Health Reader. Volume II: A State and Community Perspective. New York: The New Press, 1999.
20. Berkman L, Kawachi I. Social Epidemiology. New York: Oxford University Press, 2000.
21. Auerbach JA, Krimgold BK. Income, Socioeconomic Status and Health: Exploring the Relationships. Washington: National Policy Association, 2001.