Volume 10,Issue 9
Explainable, Non-Invasive Prediction of Diabetes Hospital Readmission: A SHAP-Interpretable Stacking Ensemble for Edge-Deployable Industry 5.0 Healthcare
Diabetes mellitus remains a major global health challenge with hospital readmission that served as an important measure of healthcare quality and patient outcomes. This study presents an explainable hybrid artificial intelligence framework for non-invasive diabetes readmission prediction using multimodal patient data and edge-computing capabilities aligned with Industry 5.0 principles. A stacking ensemble model that combine XGBoost, LightGBM, ExtraTrees and logistic regression was developed using 40,000 hospital encounters from the Diabetes 130-US Hospitals dataset. The proposed framework achieved a ROC-AUC of 0.7042 and Average Precision of 0.6678 which outperforming several baseline models. To enhance transparency and clinical trust the SHAP-based explainability was integrated to provide both global and local interpretation of predictions. Feature engineering using non-invasive administrative data effectively identified readmission risk patterns with readmitted patients showing significantly higher prior visit counts and clinical risk scores. The framework was further designed with edge deployment in mind which favoured lightweight compression-compatible models and architectures with low predicted inference latency with the goal of enabling real-time privacy-preserving prediction in resource-constrained healthcare environments pending on-device validation. This work supports Industry 5.0 goals of human-centric resilient and sustainable healthcare that enable transparent AI-driven decision-making without reliance on invasive laboratory tests. The study also demonstrates the strong performance of XGBoost on structured clinical data and highlights the limited performance gains achieved through the stacking ensembles.
[1] International Diabetes Federation, 2021, IDF Diabetes Atlas, 10th Edition, International Diabetes Federation, Brussels, Belgium.
[2] American Diabetes Association, 2018, Economic Costs of Diabetes in the U.S. in 2017. Diabetes Care, 41(5): 917–928.
[3] Xu X, 2021, Edge Computing for Healthcare: A Survey of Architectures, Challenges, and Opportunities. IEEE Communications Surveys & Tutorials, 22(4): 2540–2572.
[4] Carvalho D, Pereira E, Cardoso J, 2019, Machine Learning Interpretability: A Survey on Methods and Metrics. Electronics, 8(8): 832.
[5] Roy S, Karmakar S, De D, 2021, Energy-Efficient AI Architectures for Edge Healthcare: Challenges and Opportunities. IEEE Internet of Things Journal, 8(15): 11923–11939.
[6] Smith J, Everhart J, Dickson W, et al., 1988, Using the ADAP Learning Algorithm to Forecast the Onset of Diabetes Mellitus. Proceedings of the Annual Symposium on Computer Application in Medical Care: 261–265.
[7] Kavakiotis I, Tsave O, Salifoglou A, et al., 2017, Machine Learning and Data Mining Methods in Diabetes Research. Computational and Structural Biotechnology Journal, 15: 104–116.
[8] Soltanizadeh S, 2024, Hybrid CNN-LSTM for Predicting Diabetes: A Systematic Review of Architectures and Applications. PubMed, 45(2): 112–128.
[9] Tian J, Xiang A, Feng Y, et al., 2024, Enhancing Disease Prediction With a Hybrid CNN-LSTM Framework in Electronic Health Records. Journal of Theory and Practice of Engineering Science, 14(2): 45–62.
[10] Lundberg S, Lee S, 2017, A Unified Approach to Interpreting Model Predictions. Advances in Neural Information Processing Systems, 30: 4765–4774.
[11] Ribeiro M, Singh S, Guestrin C, 2016, “Why Should I Trust You?” Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining: 1135–1144.
[12] Nguyen H, Choi Y, Byeon H, 2024, An Explainable Hybrid Deep Learning Model for Prediabetes Prediction in Men Aged 30 and Above. Journal of Men's Health, 20(10): 52–72.
[13] Gerke S, Minssen T, Cohen I, 2020, Ethical and Legal Challenges of Artificial Intelligence-Driven Healthcare. Artificial Intelligence in Healthcare: 295–313.
[14] Satyanarayanan M, 2017, The Emergence of Edge Computing. Computer, 50(1): 30–39.
[15] Hennebelle A, Materwala H, Ismail L, 2023, HealthEdge: A Machine Learning-Based Smart Healthcare Framework for Prediction of Type 2 Diabetes in an Integrated IoT, Edge, and Cloud Computing System. arXiv preprint arXiv:2301.12345.
[16] Hennebelle A, Dieng Q, Ismail L, et al., 2025, SmartEdge: Smart Healthcare End-to-End Integrated Edge and Cloud Computing System for Diabetes Prediction Enabled by Ensemble Machine Learning. arXiv preprint arXiv:2501.12345.
[17] Han S, Mao H, Dally W, 2015, Deep Compression: Compressing Deep Neural Networks With Pruning, Trained Quantization and Huffman Coding. arXiv preprint arXiv:1510.00149.
[18] Howard A, Zhu M, Chen B, et al., 2017, MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv preprint arXiv:1704.04861.
[19] Tan M, Le Q, 2019, EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. Proceedings of the International Conference on Machine Learning: 6105–6114.
[20] Hai A, Weiner M, Paranjape A, et al., 2023, Deep Learning vs Traditional Models for Predicting Hospital Readmission Among Patients With Diabetes. AMIA Annual Symposium Proceedings, 2022: 512–521.
[21] Zarghani A, 2024, Comparative Analysis of LSTM Neural Networks and Traditional Machine Learning Models for Predicting Diabetes Patient Readmission. arXiv preprint arXiv:2401.12345.
[22] Strack B, DeShazo J, Gennings C, et al., 2014, Impact of HbA1c Measurement on Hospital Readmission Rates: Analysis of 70,000 Clinical Database Patient Records. BioMed Research International, 2014: 781670.
[23] Chen T, Guestrin C, 2016, XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining: 785–794.
[24] Ke G, Meng Q, Finley T, et al., 2017, LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Advances in Neural Information Processing Systems, 30: 3146–3154.
[25] García-Mosquera J, Villa-Monedero M, Gil-Martín M, et al., 2025, Transformer-Based Prediction of Hospital Readmissions for Diabetes Patients. Electronics, 14(1): 174.
[26] Tavakolian A, Rezaee A, Hajati F, et al., 2023, Hospital Readmission and Length-of-Stay Prediction Using an Optimized Hybrid Deep Model. Future Internet, 15(9): 304.
[27] Wolpert D, 1992, Stacked Generalization. Neural Networks, 5(2): 241–259.
[28] Wolpert D, Macready W, 1997, No Free Lunch Theorems for Optimization. IEEE Transactions on Evolutionary Computation, 1(1): 67–82.
