Title: An AI-Powered Predictive Model for Reducing Hospital Readmissions in Chronic Disease Management Programs
Authors: Bamidele Samuel Adelusi, Damilola Osamika, MariaTheresa Chinyeaka Kelvin-Agwu, Ashiata Yetunde Mustapha, Adelaide Yeboah Forkuo, Nura Ikhalea
Volume: 9
Issue: 4
Pages: 48-68
Publication Date: 2025/04/28
Abstract:
Hospital readmissions pose a significant burden to healthcare systems, particularly in the management of chronic diseases such as diabetes, heart failure, and chronic obstructive pulmonary disease (COPD). This study presents the development and evaluation of an AI-powered predictive model aimed at reducing hospital readmissions within chronic disease management programs. Leveraging machine learning algorithms, including logistic regression, random forest, and gradient boosting, the model analyzes large-scale electronic health records (EHRs), demographic information, clinical indicators, and social determinants of health to identify patients at high risk of readmission. The model was trained and validated using a dataset of over 50,000 patient records collected from multiple healthcare institutions across the United States between 2018 and 2023. Feature engineering and selection techniques were employed to extract relevant variables, including prior hospitalization history, medication adherence, comorbidity indices, and care coordination metrics. Among the evaluated models, the gradient boosting algorithm achieved the highest predictive performance with an AUC-ROC of 0.87, precision of 0.79, and recall of 0.76. The AI model was subsequently integrated into a clinical decision support system (CDSS) to enable healthcare providers to intervene proactively through personalized care plans, follow-up scheduling, and patient education initiatives. Pilot implementation across three hospitals demonstrated a 23% reduction in 30-day readmission rates over six months, with improved care coordination and patient satisfaction scores. The model's explainability was enhanced using SHAP (Shapley Additive Explanations) values, allowing clinicians to interpret individual risk factors and decision-making pathways. This approach fosters trust in AI-driven recommendations and aligns with value-based care objectives by optimizing resource utilization and improving patient outcomes. In conclusion, the proposed AI-powered predictive model demonstrates substantial potential to transform chronic disease management by reducing avoidable hospital readmissions. Future research will focus on expanding model generalizability across diverse populations and incorporating real-time data streams for dynamic risk assessment. The integration of explainable AI in healthcare delivery ensures ethical, efficient, and patient-centered decision-making in chronic care settings.