Project summary / abstract
Smoking, alcohol consumption, and obesity are the three leading causes of preventable disease and death in
the U.S. Each year tobacco use alone kills nearly 440,000 Americans, who die up to 15 years earlier than
nonsmokers. Risky health behaviors are more prevalent among low socioeconomic status (SES) groups, and
significant sources of the substantial disparities in health between them. Such disparities are formed early in
life and become more pronounced as individuals age.
In this application we seek to test the hypothesis that protective socioeconomic and policy environments
moderate the effects of high-risk genetic variants for smoking, alcohol consumption, and obesity; evaluate how
such gene-by-environment (GxE) interplay evolves over the lifecycle and how it contributes to health
disparities; generate information relevant to decision makers and health professionals; and when possible
investigate the underlying mechanisms. Using results from genome-wide association studies (GWAS), we
focus on specific genetic variants or aggregated genetic scores and specific dimensions of the socioeconomic
and policy environment.
The proposed research is a natural continuation of our project, “From understanding to reducing health
disparities: a model-based evaluation” (R01 AG037398), in which we developed a theory of the formation and
evolution of health disparities between SES groups, and informed by the theory, explored the extent to which
disparities in health and longevity between SES groups are the result of differences in job-conditions, health
behavior, medical care, and labor-force withdrawal. We also investigated the effects of possible policy
interventions on health and health disparities. Two key findings of that research are that: (i) health behaviors
play a very important role in health disparities between SES groups, and (ii) multiple factors (that have a
genetic basis) influence both SES and health, and interact with SES in producing health. This suggests an
important role for interactions between genes and the environment in shaping health behavior. With the very
recent order of magnitude increases in the power of polygenic prediction, and the very recent genetic discovery
results for several new smoking, alcohol consumption, and obesity phenotypes, it is now possible to
incorporate into our analyses the effects of genetic predispositions, in interaction with the socioeconomic and
policy environment, on these three risky health behaviors.
The proposed project will integrate several complementary methods: descriptive analyses to explore
associations for different levels of genetic risk and different measures of the social and policy environment at
different stages of life; estimation of structural lifecycle models to better understand GxE interplay and make
predictions; exploitation of natural experiments to address causality; and counterfactual analyses, using the
structural models, to evaluate intervention alternatives.