Ranking sociodemographic, health behavior, prevention, and environmental factors in predicting neighborhood cardiovascular health: A Bayesian machine learning approach

Cardiovascular disease is the leading cause of death in the United States. While abundant research has been conducted to identify risk factors for cardiovascular disease at the individual level, less is known about factors that may influence population cardiovascular health outcomes at the neighborhood level. The purpose of this study is to use Bayesian Additive Regression Trees, a state-of-the-art machine learning approach, to rank sociodemographic, health behavior, prevention, and environmental factors in predicting neighborhood cardiovascular health. We created a new neighborhood health dataset by combining three datasets at the census tract level, including the 500 Cities Data from the Centers for Disease Control and Prevention, the 2011-2015 American Community Survey 5-Year Estimates from the Census Bureau, and the 2015-2016 Environmental Justice Screening database from the Environmental Protection Agency in the United States. Results showed that neighborhood behavioral factors such as the proportions of people who are obese, do not have leisure-time physical activity, and have binge drinking emerged as top five predictors for most of the neighborhood cardiovascular health outcomes. Findings from this study would allow public health researchers and policymakers to prioritize community-based interventions and efficiently use limited resources to improve neighborhood cardiovascular health.