Religion, a social determinant of mortality? A 10-year follow-up of the Health and Retirement Study.

The social determinants of health framework has brought a recognition of the primary importance of social forces in determining population health. Research using this framework to understand the health and mortality impact of social, economic, and political conditions, however, has rarely included religious institutions and ties. We investigate a well-measured set of social and economic determinants along with several measures of religious participation as predictors of adult mortality. Respondents (N = 18,370) aged 50 and older to the Health and Retirement Study were interviewed in 2004 and followed for all-cause mortality to 2014. Exposure variables were religious attendance, importance, and affiliation. Other social determinants of health included gender, race/ethnicity, education, household income, and net worth measured at baseline. Confounders included physical and mental health. Health behaviors and social ties were included as potential explanatory variables. Cox proportional hazards regressions were adjusted for complex sample design. After adjustment for confounders, attendance at religious services had a dose-response relationship with mortality, such that respondents who attended frequently had a 40% lower hazard of mortality (HR = 0.60, 95% CI 0.53-0.68) compared with those who never attended. Those for whom religion was "very important" had a 4% higher hazard (HR = 1.04, 95% CI 1.01-1.07); religious affiliation was not associated with risk of mortality. Higher income and net worth were associated with a reduced hazard of mortality as were female gender, Latino ethnicity, and native birth. Religious participation is multi-faceted and shows both lower and higher hazards of mortality in an adult US sample in the context of a comprehensive set of other social and economic determinants of health.