Heart disease remains the leading cause of death in the United States, and existing predictive models rely almost exclusively on clinical biomarkers such as cholesterol, blood pressure, and family history. This paper presents a novel approach: predicting cardiovascular disease risk using financial stress indicators derived from the CDC Behavioral Risk Factor Surveillance System (BRFSS) 2015 survey (n ≈ 400,000). We introduce the Financial Stress Score (FSS), a composite metric that encodes three dimensions of economic vulnerability — income deprivation, inability to afford healthcare, and unemployment status — into a single continuous 0–1 scale. Three machine learning classifiers are trained and compared — Logistic Regression, Random Forest, and XGBoost. XGBoost achieves the highest AUROC (0.8131). SHAP analysis confirms that the FSS constituent components collectively represent the largest combined predictive signal. These findings demonstrate that socioeconomic data can serve as meaningful proxies for cardiovascular risk.
Hari Priya Vykuntapu (Fri,) studied this question.