Social, Developmental and Genetic Epidemiology of Alcohol Use Disorders

This project seeks to clarify how environmental adversity at the individual, family and community level and genetic risk jointly contribute to the development of alcohol use disorders. To accomplish this goal, we will apply advanced statistical methods to a unique set of epidemiological resources available in the country of Sweden. The findings will be relevant to the US and other developed countries, with potential impacts on prevention, treatment and policy.

Despite the importance of both genetic and environmental contributions to alcohol use disorders (AUD), we still have limited knowledge about the dynamic and causal relationships between family, peer and neighborhood contexts and genetic risks in the development of AUD over the life course. To further our understanding of specific social and genetic effects on AUD, we propose to take a life course perspective to improve knowledge of causal mechanisms.

Our specific aims are:

1. to assess effects of the neighborhood environment, peer context and family system over the life course during sensitive developmental periods and examine the accumulated impact of neighborhood environments (using objective measures that eliminate same-source bias); to disentangle the stress (adverse environments cause AUD) vs. drift (AUD cause downward social mobility) hypotheses;
2. to examine mediators (chains of risk) and effect modifiers in population subgroups; to distinguish family-level environment and genetic effects on AUD; and to determine the degree to which genetic risk factors moderate sensitivity to the pathogenic effects of environmental adversity.

We propose to use comprehensive data from multiple nationwide data sources in Sweden. This will allow us to assess cumulative neighborhood exposures beginning in 1970 for the entire Swedish population and to conduct follow-up analyses of AUD until 2010. Our database will contain nationwide data on 11.8 million men and women whose neighborhoods of residence are geocoded and defined based on social (e.g. deprivation, crime) and physical (alcohol availability) factors. With careful ethical safeguards, national registries permit us to construct database by linking census data, family relationship data, neighborhood-level social and physical environmental records, crime data, military conscript data, cause of death records, inpatient and outpatient hospital records, and all prescription medication records. AUD diagnoses are available beginning in 1973 and individual- and neighborhood-level factors beginning in 1970. We will account for individual mobility and neighborhood change over time by using latent class growth modeling and marginal structural models. We will use propensity score matching and co- relative control designs to control for selective migration and thereby improve the ability to determine causality. Furthermore, we will produce refined assessments of neighborhood exposures from advanced GIS analytic techniques and study gene-environment interactions, which will provide a more robust basis for policy interventions and health promotion via an integrated genetics and environmental cross-disciplinary approach.

Applying our expertise in human development and social and genetic epidemiology to a uniquely powerful sample, we expect this study to have important implications for AUD research, prevention and policy.