Urban form and cardiovascular health: Decoupling hierarchical heterogeneity in built environment impacts

Longitudinal impacts of the built environment and transit accessibility on neighborhood-level cardiovascular disease (CVD) prevalence are not well explored. Further, little is known about the magnitude of heterogeneity in the longitudinal links between social and built environment features and CVD prevalence. This study utilized a longitudinal ecological study design covering 26,631 neighborhoods (census tracts) in the 500 largest American cities. A novel data infrastructure is harnessed by integrating time-varying neighborhood-level data on transportation, the built and social environments, and chronic disease prevalence at two time points. In a full Bayesian framework, Markov Chain Monte Carlo-based grouped correlated random parameter models are formulated to simultaneously account for unobserved and correlated heterogeneity impacts. More walkable neighborhoods, characterized by greater density, street connectivity, and land-use mix, had lower CVD and hypertension prevalence over time. Greater transit accessibility was also correlated with lower disease prevalence. Conversely, neighborhoods with higher social vulnerability had greater CVD and hypertension rates over time. A detailed post hoc neighborhood-level assessment revealed significant unobserved and correlated heterogeneity in the impacts of social and built environment features across both neighborhoods and cities. Insights into this heterogeneity, as well as the determinants of neighborhood-level CVD and hypertension prevalence, can help public health officials, engineers, and policymakers implement localized community-based behavioral interventions for smarter and healthier cities. Implications for national disease surveillance systems are discussed.