Leveraging Machine Learning to Predict Medication Adherence and Cardiovascular Outcomes Under Centralized Pharmaceutical Procurement: A Medicare Part D Analysis

Medication non-adherence remains a persistent challenge in the management of cardiovascular disease (CVD), contributing significantly to preventable hospitalizations, poor health outcomes, and rising healthcare expenditures. Within the framework of Medicare Part D, centralized pharmaceutical procurement has been proposed as a strategic policy instrument to control drug prices and improve access. However, the interaction between centralized procurement mechanisms and individual-level adherence behaviors remains poorly understood. In this study, we leverage machine learning (ML) techniques to model and predict medication adherence and its impact on cardiovascular outcomes among Medicare Part D beneficiaries. Using a nationally representative dataset comprising prescription claims, demographic profiles, clinical risk indicators, and longitudinal cardiovascular event data, we implement supervised learning algorithms—logistic regression, random forests, and gradient boosting machines—to identify high-risk populations with suboptimal adherence. The models are calibrated to reflect socioeconomic, policy, and access-related variables unique to centralized procurement regimes. Key predictors of non-adherence include copayment burden, medication complexity, and prior cardiovascular events. Predictive performance is evaluated using AUC-PR, F1-score, and SHAP-based feature interpretation to ensure transparency and actionable insights. The study highlights the utility of ML as a policy tool to optimize procurement strategies and tailor interventions aimed at improving adherence. It further emphasizes the need for integrating real-time analytics into Medicare infrastructure to preempt adverse outcomes and enhance patient-centered pharmaceutical care under centralized systems.