This study presents a machine learning–based predictive framework for identifying patients at high risk of 30-day hospital readmission among individuals with diabetes. Using the UCI Diabetes dataset comprising over 101,000 hospital encounters, we developed and evaluated Random Forest and XGBoost models, achieving a recall of 76% at an optimized threshold. Feature importance and SHAP analysis identified key predictors including prior inpatient visits, medication burden, and length of stay. Building on these findings, we propose a CDC HI-5 aligned intervention incorporating risk stratification, care coordination, medication management, and a real-time clinical dashboard. The results demonstrate the feasibility of integrating predictive analytics into population health strategies to reduce preventable readmissions.
Mohammed Mustafa Khan (Sat,) studied this question.