OBJECTIVES:

There are research knowledge gaps in economic burden among postmenopausal women with heart failure (HF). This study identified the leading predictors and their associations with economic burden among postmenopausal women with HF using machine learning methods.

METHODS:

This cross-sectional study used data from the 2020 Medical Expenditure Panel Survey (MEPS) and included postmenopausal women (>50 years) with HF (weighted N=600,742). The economic burden was measured with total third-party healthcare expenditures by the payers and out-of-pocket expenditures by the patients. We employed eXtreme Gradient Boosting (XGBoost) regression to determine key predictors. Global and local interpretations of associations were performed using SHapley Additive exPlanations (SHAP). Our predictive model used 21 features such as age, health status including comorbidities, and social determinants of health (SDoH) such as health insurance, education, and poverty. The model building steps included 70% training and 30% testing split of the data, 10-fold cross-validations, and six rounds of optimization using Python 3.9.12. Model performance was evaluated using the test dataset.

RESULTS:

Model performance metrics were: mean absolute errors (0.442,0.310), root mean square errors (0.452,0.342), and coefficients of determination (0.935,0.987) for third-party and out-of-pocket expenditures, respectively. The top 10 leading predictors of third-party expenditures included polypharmacy, age, comorbidities, physical health, and SDoH. The top 10 leading predictors of out-of-pocket expenditures included age, comorbidities, and SDoH. SHAP plots suggested a complex relationship between age and third-party and out-of-pocket expenditures. Polypharmacy and asthma were associated with higher third-party expenditures. Being poor and Latinx identification were associated with lower out-of-pocket expenditures. Finally, comorbidities were associated with higher out-of-pocket expenditures.

CONCLUSIONS:

The leading predictors differed by payer source. SDoH variables were associated with economic burden, suggesting that addressing SDoH may reduce healthcare costs. Cost-containment policies, programs, and interventions at the payer and patient levels need to include effective comorbidity management strategies.