What determines health: a causal analysis using county level data

Abstract

This article revisits the long-standing issue of the determinants of health outcomes. We make two contributions to the literature. First, we use a large and comprehensive US county level health data set that has only recently become available. This data set includes five measures of health outcomes and 24 health risk factors in the categories of health behaviors, clinical care, social and economic factors, and physical environment. Second, to distinguish causality from correlation, we implement an emerging data-driven method to study the causal factors of health outcomes. Among all included potential health risk factors, we identify adult smoking, obesity, motor vehicle crash death rate, the percent of children in poverty, and violent crime rate to be major causal factors of premature mortality. Adult smoking, preventable hospital stays, college or higher education, employment, children in poverty, and adequacy of social support determine health-related quality of life. Finally, the Chlamydia rate, community safety, and liquor store density are three important factors causally related to low birth weight. Policy implications of these findings are discussed.