Religion and Death in the United States: A Meta-Regression Comparative Assessment of Between-County Mortality Heterogeneity in the United States

Abstract

Religion can have a favorable impact on individual-level health. The influence of religion on population health, however, remains less clear. This study investigated the association between religion and mortality at the population-level. Using county data, a meta-regression was performed to examine between-county mortality heterogeneity. The percent heterogeneity associated with religion variables were compared to demographics (i.e., place, race, language, age, and gender) and health factors (i.e., individual behaviors, clinical care, social and economic, and physical environment) as predictors of mortality. Religion was measured in terms of adherence (i.e., prevalence attending/belonging to a congregation), congregation density, and the diversity of adherents and congregation by denominations. Results showed counties with lower mortality were associated with higher proportions of religion adherents and a greater diversity of adherents and congregations. Counties with higher mortality were associated with higher religion congregation density. Religion, as a parsimonious multivariate model with all demographic and health factor predictors, had less added value when controlled for individual variables or constructs. The direction of association between religion and mortality was consistent, even when controlling for demographics and health factors, and thus merits further consideration as a population health determinant, as it may play a critical role in understanding other population health outcomes.