Uncovering county-level drivers of cardiovascular mortality through integrated geospatial and machine learning analysis

Background Cardiovascular disease (CVD) mortality displays large regional variation, but the combined influence of clinical, behavioral, environmental, and socioeconomic determinants remains insufficiently characterized at the population level. Comprehensive, fine-scale assessments are needed to support targeted prevention strategies. Methods A nationwide dataset integrating demographic, clinical, behavioral, environmental, and socioeconomic indicators was assembled at the population-unit level. Associations with CVD mortality were examined using correlation analysis, spatial patterning, stepwise multivariable regression, and machine learning models. Predictors included chronic disease burden, medication adherence, lifestyle behaviors, air-pollutant exposure, and indicators of social vulnerability. Results Marked disparities in CVD mortality were observed across demographic subgroups, with increased burden in older individuals, men, and historically underserved populations. Higher prevalence of chronic pulmonary disease, hypertension, tobacco use, short sleep duration, medication non-adherence, fine particulate matter exposure, poverty, and limited educational access showed strong associations with elevated mortality. In contrast, higher household income, greater physical activity, and participation in cardiac rehabilitation were associated with reduced mortality. Multivariable regression explained 66.9% of mortality variation and identified environmental and behavioral factors among the leading contributors. Random Forest models confirmed the dominant influence of chronic pulmonary disease, medication non-adherence, tobacco use, and sleep duration. Conclusions Population-level CVD mortality is shaped by intersecting clinical, behavioral, environmental, and socioeconomic determinants. Integrated analytical approaches that combine spatial assessment, multivariable modeling, and machine learning can help identify high-risk areas and inform targeted prevention strategies addressing air quality, tobacco control, chronic disease management, and structural inequities.Spatial patterns of CVD mortalityFor image description, please refer to the figure legend and surrounding text. Dominance of risk factorsFor image description, please refer to the figure legend and surrounding text.