Geospatial and machine learning analysis of upstream determinants of cardiovascular disease mortality across united states counties
1
Global Health and Health Policy, Harvard University, Cambridge, United States
Popul. Med. 2026;8(Supplement Supplement 1):
ABSTRACT
INTRODUCTION:
Cardiovascular disease (CVD) remains a leading cause of death worldwide, with large geographic disparities that reflect behavioral, environmental, clinical, and structural determinants. We aimed to identify upstream contributors to county-level CVD mortality in the united states using a multi-domain framework that can inform targeted, equity-oriented prevention strategies.
METHODS:
We constructed a dataset for more than 3000 counties from publicly available surveillance sources capturing mortality, chronic disease prevalence, health behaviors, health care access, environmental exposures, and social and economic indicators. We mapped spatial patterns, calculated correlations, and used stepwise multiple regression with cross-validation to estimate joint associations while addressing multicollinearity. A random forest model assessed nonlinear relationships and variable importance. Residual analyses identified counties where observed mortality substantially exceeded model predictions.
RESULTS:
County-level CVD mortality varied more than twofold across the united states, with highest rates concentrated in the southeast and lower mississippi river valley. Chronic obstructive pulmonary disease showed the strongest correlation with mortality (r = 0.70), followed by smoking, short sleep duration, physical inactivity, and fine particulate matter exposure. The multivariable model explained 66.9 percent of mortality variation. The random forest model identified chronic obstructive pulmonary disease, smoking, medication nonadherence, particulate matter exposure, and hypertension as leading contributors.
CONCLUSIONS:
Integrating population indicators with geospatial and machine learning methods clarifies upstream determinants of cardiovascular disease mortality and highlights opportunities for public health action. Findings point to policies and programs focused on tobacco exposure, air quality, chronic disease management, and structural determinants in high-burden communities. This framework offers a scalable approach that can strengthen non-communicable disease surveillance and guide resource allocation.
Cardiovascular disease (CVD) remains a leading cause of death worldwide, with large geographic disparities that reflect behavioral, environmental, clinical, and structural determinants. We aimed to identify upstream contributors to county-level CVD mortality in the united states using a multi-domain framework that can inform targeted, equity-oriented prevention strategies.
METHODS:
We constructed a dataset for more than 3000 counties from publicly available surveillance sources capturing mortality, chronic disease prevalence, health behaviors, health care access, environmental exposures, and social and economic indicators. We mapped spatial patterns, calculated correlations, and used stepwise multiple regression with cross-validation to estimate joint associations while addressing multicollinearity. A random forest model assessed nonlinear relationships and variable importance. Residual analyses identified counties where observed mortality substantially exceeded model predictions.
RESULTS:
County-level CVD mortality varied more than twofold across the united states, with highest rates concentrated in the southeast and lower mississippi river valley. Chronic obstructive pulmonary disease showed the strongest correlation with mortality (r = 0.70), followed by smoking, short sleep duration, physical inactivity, and fine particulate matter exposure. The multivariable model explained 66.9 percent of mortality variation. The random forest model identified chronic obstructive pulmonary disease, smoking, medication nonadherence, particulate matter exposure, and hypertension as leading contributors.
CONCLUSIONS:
Integrating population indicators with geospatial and machine learning methods clarifies upstream determinants of cardiovascular disease mortality and highlights opportunities for public health action. Findings point to policies and programs focused on tobacco exposure, air quality, chronic disease management, and structural determinants in high-burden communities. This framework offers a scalable approach that can strengthen non-communicable disease surveillance and guide resource allocation.
| eISSN: | 2654-1459 |
We process personal data collected when visiting the website. The function of obtaining information about users and their behavior is carried out by voluntarily entered information in forms and saving cookies in end devices. Data, including cookies, are used to provide services, improve the user experience and to analyze the traffic in accordance with the Privacy policy. Data are also collected and processed by Google Analytics tool (more).
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
