During the twentieth century, great strides were made in reducing disease and improving the health of individuals and populations. Public health measures such as sanitation, improved hygiene, and vaccines led to major reductions in mortality and morbidity (Turnock, 2001). Increased attention to the hazards of the workplace resulted in reduced injuries and better health for workers (IOM, 2003a). Advances in biomedical research helped expand knowledge of disease and spurred the development of new clinical and pharmaceutical interventions. More recently, the sequencing of the human genome has provided information that holds the promise for further improving human health.

Over the years a large body of evidence has emerged indicating that social and behavioral factors such as socioeconomic status, smoking, diet, and alcohol use are important determinants of health (Berkman and Kawachi, 2000; IOM, 2000; Marmot and Wilkinson, 2006). Recent studies also suggest that examining interactions among genetic and social-environmental factors could greatly enhance understanding of health and illness. For example, Caspi and colleagues (2003) found “evidence of a gene-by-environment interaction, in which an individual’s response to environmental insults is moderated by his or her genetic makeup.” In a study showing how the social environment can influence biological response, Manuck et al. (2005) found that the socioeconomic status of communities is associated with variations in central nervous system serotonergic responsivity, which may have implications for the prevalence of psychological disorders and behaviors such as depression, impulsive aggression, and suicide.

As part of a strategy to determine how best to integrate research priorities to include an increased focus on the impact on health of interactions among social, behavioral, and genetic factors, the National Institutes of Health (NIH), Office of Behavioral and Social Sciences Research, in conjunction with the National Human Genome Research Institute and the National Institute of General Medical Sciences, requested that the Institute of Medicine undertake a study to examine the state of the science on gene-environment interactions that affect human health, with a focus on the social environment.¹ The goal of the study was to identify approaches and strategies to strengthen the integration of social, behavioral, and genetic research and to consider the relevant training and infrastructure needs. More specifically, NIH requested the following:

1. Review the state of the science on the interactions between the social environment and genetics that affect human health.

2. Develop case studies that will demonstrate how the interactions of the social environment and genetics affect health outcomes; illustrate the methodological issues involved in measuring the interactions; elucidate the research gaps; point to key areas necessary for integrating social, behavioral, and genetic research; and suggest mechanisms for overcoming barriers.

3. Identify gaps in the knowledge and barriers that exist to integrating social, behavioral, and genetic research in this area.

4. Recommend specific short- and long-term priorities for social and behavioral research on gene-social environment interactions; identify mechanisms that can be used to encourage interdisciplinary research in this area.

5. Assess workforce, resource, and infrastructure needs and make actionable recommendations on overcoming barriers and developing mechanisms to accelerate progress.

Chapter 2 of this report explores the impact of the social and cultural environment on health, examining what we know about the influences of these factors on health, and identifying the limitations of current research. Genetic factors and their impact on health are examined in Chapter 3, which focuses on what is known or theorized about the direct link between genes and health and what still must be explored to understand the environmental interactions and relative roles among genes that contribute to health and illness. The impact of behavioral factors on health is explored in Chapter 4.

While research on the impact of interactions has the potential to further the understanding of disease risk and aid in the development of effective interventions to improve the health of individuals and populations, there is a dearth of research that encompasses all three domains. Much remains to be learned about how these factors interact to impact health, including the most basic concept of defining interaction and how it can be characterized. Because greater etiological understanding is needed to identify future clinical research and develop effective interventions aimed at improving health outcomes, the committee focused its efforts on etiological research.

The recommendations discussed below are designed to explicate and facilitate research on the impact on health of interactions among social, behavioral, and genetic factors. Each recommendation is followed by a chapter number in which additional discussion related to the recommendation can be found.

Contributions from research conducted over the past few decades, including the sequencing of the human genome, are pushing scientists to move beyond examining single agents of health and disease to a broader systems view, which is based on the understanding that health outcomes are the result of multiple determinants and their interactions (Lalonde, 1974; Evans and Stoddard, 1990; Kaplan et al., 2000; IOM, 2003a; IOM, 2003b). Understanding the associations between health and interactions among social, behavioral, and genetic factors requires research that embraces the systems view and includes an examination of the interactive pathways through which these factors operate to affect health.² Such research requires the participation of scientific investigators from a variety of different fields and a shift in focus from efforts that are dominated by single disciplines to research that involves collaborative participation of scientists with various expertise at all stages of the research process. While interdisciplinary research focuses on answering questions of mutual concern to those from various disciplines and multidisciplinary research involves research questions of both mutual and separate interest to participating investiga-

tors, transdisciplinary research “implies the conception of research questions that transcend the individual departments or specialized knowledge bases because they are intended to solve research questions that are, by definition, beyond the purview of the individual disciplines” (IOM, 2003b). Therefore, the committee makes the following recommendation:

Key social variables which have been linked consistently and robustly to health outcomes include educational attainment, income and wealth, occupational status, social networks/social support, and work conditions. Well-established behavioral and psychological variables that affect health outcomes include tobacco/alcohol/drug use, eating behavior, physical activity, temperament, perceived stress and coping, perceived social support, emotional state, and motivation. Essential genetic factors affecting health include the DNA sequence variation, structural chromosomal changes, gene expression, epigenetic modifications, and downstream targets of gene expression.

In the search for a better understanding of genetic and environmental interactions as determinants of health, certain fundamental aspects of human identity (i.e., sex/gender and race/ethnicity) pose both a challenge and an opportunity for clarification. However because sex/gender and race/ ethnicity are more complicated than they appear, they need to be considered and analyzed from a variety of perspectives, including social, cultural, psychological, historical, political, genetic, and geographic/ancestral.

Relevant physiological measurements and pathways should also be considered. Understanding the pathways through which interactions operate will aid in identifying links between major levels of organization of living systems: social groups of individuals, individuals composed of physiological systems, physiological systems composed of cells, and cells composed of molecules, especially DNA. Ultimately the results of such research may help to identify where to intervene along the causal chains and pathways between the social world and genes that cause disease in order to improve health outcomes.

Many determinants of health are not static, that is, they influence health in a variety of ways throughout the life course. For example, poverty may differentially and independently affect the health of an individual at different stages of life (e.g., in utero, during infancy and childhood, during

pregnancy, or during old age). Personality traits and psychological status are also known to change over the lifespan and have potential to affect health. In addition to these well-established factors, a growing body of research has documented associations between cultural factors and health (Berkman and Kawachi, 2000). The influence of social, behavioral, and genetic factors on health involves dimensions of both time (critical stages in the life course and the effects of cumulative exposure) and the context or culture within which variables operate to influence health outcomes. Therefore, the committee makes the following recommendation:

For the most part research has taken a linear approach when examining the link between a particular set of variables (e.g., social-environmental or genetic variables) and health. Yet, there remains the need to connect and integrate knowledge across multiple determinants of health in order to understand the mechanisms of integration (for example, how social factors are translated into physiological effects on cellular responses, including changes in gene expression). Future studies should recognize that a linear approach does not reflect the integrated nature of how health outcomes are generated. Therefore, the committee recommends the following:

With approximately 30,000 genes in the human genome, most genes are likely to serve different functions at different times in different environments (McClintock et al., 2005). The ability to measure and evaluate differential gene expression has the potential to provide important insights into the study of health and disease. Alterations in DNA sequence and gene

expression can be modified at different points throughout the life course, dictating variation in protein levels and functionality, as well as subsequent levels of metabolic products that are associated with those proteins. These factors can be measured through the use of genomic, transcriptomic, proteomic, and metabonomic technologies. However, further development of these technologies is needed to allow researchers to accurately study the molecular systems that interact with social and behavioral variables to influence health outcomes (see Chapter 6 for a more detailed discussion of these technologies). Thus, the committee recommends the following:

The context or culture within which individuals exist also is known to exert influence on health outcomes. Relevant social and cultural environments include not only an individual’s immediate personal environment (e.g., his/her family), but also the broader social contexts such as the community in which a person resides. Health psychologists are increasingly calling attention to the critical role of sociocultural context, a necessary factor to consider if efforts to modify risk behaviors are to be effective. Different subgroups may have different genetic backgrounds, as well as varying cultural or socioeconomic characteristics that influence patterns of behavior, thereby creating a correlation between genotype and environmental exposure. Furthermore, it may be found that polymorphisms occurring in genotypes that act as destructive or protective factors for disease and health may be created, modified, or triggered by cultural and contextual factors. It is important to determine if research findings are applicable beyond a small population, and to capitalize on unique gene-environment interactions that could contribute to a broader understanding of factors, mechanisms, and processes. Therefore, the committee recommends the following:

Animal research studies are an important complement to clinical and community-based research because they can serve as models for gene-environment interactions and pathways of human disease. Animal models can be used to conduct studies in which different aspects of social, behavioral, and genetic variables can be controlled, standardized, or manipulated to a significantly larger extent than can be accomplished in human studies. These models also allow for the invasive examination of organ-, tissue-, and region-specific mechanisms at the physiological, cellular, and molecular levels. Animals with short reproductive cycles and life spans provide an invaluable tool for conducting developmental and life-span studies, as well as breeding experiments and genetic manipulation that facilitate the elucidation of inherited traits and genetic effects. In some cases, animal models provide opportunities to establish causality through studies examining the temporal sequence of events, or studies involving removal followed by the add-back of hypothesized mediators at the genetic, protein, physiological, behavioral, or social environment level. Therefore, the committee makes the following recommendation:

A clear formulation of the concept of “interaction” and an understanding of research designs that can be used to test for interactions are central to making progress in assessing the impact on health of interactions. Statistical tests for interaction are entirely dependent on the measurement scale (e.g., additive or multiplicative) used to evaluate the effects of different factors on health. Use of different measurement scales can lead to substantively different conclusions about whether or not interaction is present, and therefore, to different recommendations for intervention. Thus, determining the measurement scale is critical to the design of future studies and to the interpretation of their results. The choice of measurement scale should not be based on statistical convenience. Instead, it should be based on a theoretical model for disease causation that is more closely tied to biology.

Epidemiologists have built a conceptual framework for interaction based on the counterfactual model and the sufficient-component cause model (see Chapter 8 for discussion of these models). Beginning with this conceptual framework, defined at the level of an individual, it is possible to predict patterns of risk in the population when interaction is or is not present. Such an analysis leads to the conclusion that an additive scale for testing interaction more closely reflects the underlying biology than a multiplicative scale. That is, when two factors participate in the same sufficient cause (interaction defined conceptually), disease risks in individuals with both risk factors will be greater than expected from the additive effects of each risk factor alone.

Currently, most of the statistical software commonly used for epidemiologic analysis includes tests for interaction on a multiplicative scale but not on an additive scale. Thus testing for interactions on an additive scale requires the development of new, accessible statistical software. Also, tests for interaction require extremely large sample sizes; hence multisite collaborations may be required in order to assemble databases of sufficient size that are needed to assure adequate statistical power. Additionally, given the complexity in defining interaction and testing for it, new, efficient study designs should be developed for testing interaction. Therefore, the committee recommends the following:

Research conducted to elaborate the impact of interactions among social, behavioral, and genetic factors on human health places several demands on the research infrastructure. This infrastructure includes the human infrastructure (e.g., education and training), data, and incentives and rewards.

The foundation of the research enterprise is the education of its researchers. Given that advances in genomics have been recent—and the challenge of incorporating genetic research with behavior and social factors

is even more recent—it is likely that there are many current researchers who have gaps in their scientific training. Furthermore, training is needed for pre- and postdoctoral students. While universities (and high schools), NIH, and other funders of research training share responsibility for educating researchers, NIH, as the major funder of biomedical and behavioral research, is poised to make major contributions to training a cadre of researchers to conduct transdisciplinary research. Several existing mechanisms could be used as is or modified to facilitate the education of investigators in transdisciplinary research. Therefore, the committee recommends the following:

The study of interactions presents a significant need for datasets that provide information across multiple disciplines, thus allowing the evaluation of gene-environment interactions. Datasets to study such interactions are typically large, difficult to collect, and costly. Therefore, it is important to support the development and use of datasets that can be shared among a wide audience of researchers.

Datasets that already include biological and genetic measures could be augmented to include social and behavioral variables. However, these additions must not only be feasible, but more importantly, they must be scientifically compelling. Alternatively, new datasets with the necessary variables could be developed. For example, health conditions or diseases could be identified for which there is a suspected or known genetic contribution, behavioral factors are likely to be involved, and hypotheses have been formed regarding the role of social factors.

Because there is a significant need for datasets that provide information for the three domains discussed (social, behavioral, and genetic factors), the committee recommends the following:

The report Facilitating Interdisciplinary Research (NAS/NAE/IOM, 2004) outlined several key conditions for effective interdisciplinary research, including “sustained and intense communication, talented leadership, appropriate reward and incentive mechanisms (including career and financial rewards), adequate time, seed funding for initial exploration, and willingness to support risky research.” The committee believes that these same conditions apply to transdisciplinary research. Although aspects of university functioning, such as rewards and incentives, are not within the purview of NIH, they may ultimately affect the ability of NIH to find researchers who can conduct the kind of transdisciplinary research that is envisioned here.

One major challenge is acknowledging multiple investigators on team projects. The recent NIH announcement of plans to recognize multiple Principal Investigators represents a significant advancement in providing external recognition for members of research teams. As NIH explores such new approaches, the next step would be for universities to use that information in ways that would ensure that the impact of the incentives and rewards are felt at the campus level, such as the credit toward promotion and tenure that accrues to those who participate in such projects.

Scientific peer review of research applications also is a key step involved in the support of any area of research. It is not uncommon to hear investigators lament that transdisciplinary projects have difficulty in undergoing the peer review process. An important goal, therefore, is to ensure that transdisciplinary work is fairly reviewed and truly valued throughout the review process. It is not enough to simply place people from different disciplines on a review group. Specific steps need to be taken to ensure that reviewers will be able to appreciate the transdisciplinary nature or goals of a proposal. These steps include selecting reviewers who have engaged in transdisciplinary work and training reviewers about its importance and the differences between transdisciplinary research and other types of research.

Approaches to advance the field of transdisciplinary research need to be systematically applied toward the goal of fostering a type of research that has inherent scientific challenges and that faces specific institutional hurdles. Therefore, the committe makes the following recommendation:

Several important ethical and legal issues need to be addressed when considering information produced by research assessing the impact on health of interactions among social, behavioral, and genetic factors. Although these issues apply to all types of research, they are especially sensitive when considering the transdisciplinary research discussed in this report. First is the issue of conveying complex scientific findings accurately to the public, policymakers, and other researchers. Claims about scientific findings are at times simplified and even exaggerated, sometimes because of the complexity of the concepts or because of economic and social pressures to emphasize the significance of findings in easily understandable terms. These difficulties are compounded by the fact that the media, understandably, prefers straightforward, easy-to-deliver messages. However, failure to convey the limitations and complexity of scientific findings has a significant impact, because beliefs about causation of health and disease affect the allocation of responsibility and resources, which has ethical and social implications.

Another issue of concern is the development of policy based on scientific findings. The array of factors that must be considered in deciding how to use the knowledge gained from research on gene-environment interactions in developing social policy is very broad and extends far beyond the science itself and into a variety of social and ethical considerations. To address difficulties in how individuals and groups understand complex scientific findings, as well as the potential impact such findings could have on policy development, the committee makes the following recommendations:

In addition to research assessing the impact on health of interactions among social, behavioral, and genetic factors and their interactive pathways (i.e., physiological), improving health also requires individuals to act upon research findings. Therefore, the committee recommends the following:

According to the Criteria for IRB Approval of Research (45 CFR § 46.111(a)(7) (2006)), Institutional Review Boards (IRBs) are responsible for ensuring, where appropriate, the protection of research participants’ privacy and the data regarding them. Studying the impact of interactions among social, behavioral, and genetic factors on health requires the collection of information about relevant DNA variants as well as clinical or other phenotypic information. This often includes sensitive personal behavior information and social factors. The risk to research participants could be substantial if such information is accessed by people and institutions outside the study. Given the sensitivity of such research and its implications, it is of primary importance to address the issues of data sharing and informed consent. Therefore, the committee makes the following recommendations:

This report is intended to encourage and facilitate the growth of transdisciplinary research on the impact on health of interactions among social, behavioral, and genetic factors. Such research could further understanding of disease risk and aid in the development of effective interventions to improve the health of individuals and populations. Yet, achieving such understanding is not a short-term effort. Immediate priorities for action include training investigators in transdisciplinary research, expanding and developing datasets that include social, behavioral, and genetic variables (measured over the life course), developing new research strategies, and attending to the important ethical, legal, and social implications of such research. Such steps will facilitate the conduct of hypothesis-generating research to identify high-priority areas for study, which will then lead to targeted studies of interactions focused on specific health outcomes.

Health outcomes are multidetermined and result from complex interactions of many factors over time. Yet, the study of health outcomes has been driven primarily by disciplines that focus upon their own unique areas of expertise. If the study of health outcomes is to advance, investigators must break out of these disciplinary “silos” and attack the determinants of health in concert.

Berkman L, Kawachi I, editors. 2000. Social Epidemiology. New York: Oxford University Press.

Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, McClay J, Mill J, Martin J, Braithwaite A, Poulton R. 2003. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301(5631):386-389.

Evans R, Stoddard G. 1990. Producing health, consuming health care. Social Science Medicine 31(12):1347-1363.

IOM (Institute of Medicine). 2000. Promoting Health: Intervention Strategies from Social and Behavioral Research. Washington, DC: National Academy Press.

IOM. 2003a. The Future of the Public’s Health in the 21st Century. Washington, DC: The National Academies Press.

IOM. 2003b. Who Will Keep the Public Healthy? Educating Health Professionals for the 21st Century. Washington, DC: The National Academies Press.

Kaplan G, Everson S, Lynch J. 2000. The contribution of social and behavioral research to an understanding of the distribution of disease: A multilevel approach. In: Smedley B, Syme S, editors. Promoting Health: Intervention Strategies from Social and Behavioral Research. Washington, DC: National Academy Press. Pp. 37-80.

Lalonde M. 1974. Perspective on the Health of Canadians. Ottawa, ON: Ministry of Supply and Services.

Manuck SB, Bleil ME, Petersen KL, Flory JD, Mann JJ, Ferrell RE, Muldoon MF. 2005. The socio-economic status of communities predicts variation in brain serotonergic responsivity. Psychological Medicine 35(4):519-528.

Marmot MG, Wilkinson RD, editors. 2006. Social Determinants of Health. 2nd edition. Oxford, England: Oxford University Press.

McClintock MK, Conzen SD, Gehlert S, Masi C, Olopade F. 2005. Mammary cancer and social interactions: Identifying multiple environments that regulate gene expression throughout the life span. Journals of Gerontology: Series B 60B(Special Issue 1):32-41.

NAS/NAE/IOM (National Academy of Sciences/National Academy of Engineering/Institute of Medicine). 2004. Facilitating Interdisciplinary Research. Washington, DC: The National Academies Press.

Turnock BJ. 2001. Public Health: What It Is and How It Works. Gaithersburg, MD: Aspen Publishers, Inc.