Diabetic patient readmission predictive analysis: a comparative study of machine learning models of hospital readmissions

Hospital readmissions are a major concern in healthcare due to their impact on patient morbidity, hospital workload, and increased healthcare costs. This study aims to predict the readmission of diabetic patients within 30 days postdischarge, addressing the critical need for efficient resource allocation and improved patient care. The motivation stems from the high prevalence of diabetes and the associated risk of complications leading to frequent readmissions. Using a comprehensive dataset from 1999 to 2008, encompassing 130 US healthcare facilities, the research employs advanced machine learning techniques to develop a predictive model. Data preprocessing and feature engineering are meticulously applied to enhance model accuracy. Various classifiers, including Decision Tree, K-Nearest Neighbors, AdaBoost, Naive Bayes, Random Forest, and Logistic Regression, are evaluated against standard performance metrics. Results indicate that the Naive Bayes model achieves the highest F1 score of 0.81, outperforming other models. The study concludes that predictive modeling can significantly enhance clinical decision making and optimize healthcare resources, underlining its importance in managing diabetic patient care.