Authors : Sesi Mensah Gidisu; Millicent Naa Oye Boadu; Joseph Kobi

Volume/Issue : Volume 11 - 2026, Issue 9 - September

Google Scholar : https://tinyurl.com/yum257wh

DOI : https://doi.org/10.38124/ijisrt/26sep1242

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Nearly a quarter of Medicare beneficiaries end up back in an inpatient bed within 30 days of leaving the hospital, and since 2009, the Centers for Medicare and Medicaid Services (CMS) have penalized hospital readmission rates that do not fall within risk-adjusted expectations under the Hospital Readmissions Reduction Program (HRRP) (Jencks, Williams, & Coleman, 2009; McIlvennan, Eapen, & Allen, 2015). In this study, the researchers compared the performance of ensemble machine learning models combined with EAI with the traditional risk-stratification tool of the LACE index and the HOSPITAL score to determine if the models could more accurately identify Medicare patients at the highest risk of 30-day readmission. With an integrated feature set derived from electronic health record (EHR) data and American Community Survey (ACS) neighborhood indicators, six candidate algorithms were trained and compared using ROC-AUC, PR-AUC, and F1-score. Gradient-boosted and stacked ensembles showed the highest discrimination performance (AUC = 0.79) versus logistic regression alone (AUC = 0.67) and Shapley Additive exPlanations (SHAP) were applied to explain the drivers of risk at the global level and at the patient level. Clinical only baseline (AUC 0.68 to 0.69) was meaningfully improved with social determinants of health (SDOH) (AUC 0.75 to 0.79), as were prior admissions, clinical risk factors, and comorbidity burden. These were also identified as distinct latent subgroups of readmission risk through unsupervised clustering, which goes beyond what can be identified by traditional scoring systems. The results corroborate the feasibility of combining explainable and SDOH-enhanced machine learning tools for discharge planning within the hospital process, and recommend specific changes to the CMS risk-adjustment methodology that may lower penalties for safety-net and rural hospitals.

Keywords : Hospital Readmissions Reduction Program (HRRP), Machine Learning, Explainable Artificial Intelligence, Extreme Gradient Boosting (Xgboost), Light Gradient Boosting Machine (Lightgbm), Random Forest, Shapley Additive Explanations (SHAP), Electronic Health Record (EHR), American Community Survey (ACS).

References :

1. Adeyemi, T., Osei, K., & Farooq, R. (2025). Predicting 30-day hospital readmission in patients with diabetes using machine learning on electronic health record data. PMC, 12085305.
2. Bailey, M. K., Weiss, A. J., Barrett, M. L., & Jiang, H. J. (2019). Characteristics of 30-day all-cause hospital readmissions, 2010–2016: Statistical brief #248. Agency for Healthcare Research and Quality.
3. Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
4. Chaiyachati, K. H., Qi, M., & Werner, R. M. (2018). Changes to racial disparities in readmission rates after Medicare's Hospital Readmissions Reduction Program within safety-net and non-safety-net hospitals. JAMA Network Open, 1(7), e184154.
5. Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321–357.
6. Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785–794). ACM.
7. Damery, S., & Combes, G. (2017). Evaluating the predictive strength of the LACE index in identifying patients at high risk of hospital readmission following an inpatient episode: A retrospective cohort study. BMJ Open, 7(7), e016921.
8. Dobler, C. C., Hakim, M., Singh, S., Jennings, M., Waterer, G., & Garden, F. L. (2020). Ability of the LACE index to predict 30-day hospital readmissions in patients with community-acquired pneumonia. ERJ Open Research, 6(2).
9. Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. The Annals of Statistics, 29(5), 1189–1232.
10. Golas, S. B., Shibahara, T., Agboola, S., Otaki, H., Sato, J., Nakae, T., Hisamitsu, T., Kojima, G., Felsted, J., Kakarmath, S., Kvedar, J., & Jethwani, K. (2018). A machine learning model to predict the risk of 30-day readmissions in patients with heart failure: A retrospective analysis of electronic medical records data. BMC Medical Informatics and Decision Making, 18, 44.
11. He, H., & Garcia, E. A. (2009). Learning from imbalanced data. IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263–1284.
12. Jencks, S. F., Williams, M. V., & Coleman, E. A. (2009). Rehospitalizations among patients in the Medicare fee-for-service program. New England Journal of Medicine, 360(14), 1418–1428.
13. Jercich, K. (2023). Social determinants of health may help predict sepsis readmission. Healthcare IT News.
14. Kansas Health Information Network Heart Failure Cohort Study. (2026). Predicting 30-day heart failure readmissions using machine learning: Insights from the Kansas Health Information Network (KHIN). medRxiv preprint.
15. Keehan, S. P., Cuckler, G. A., Poisal, J. A., Sisko, A. M., Smith, S. D., Madison, A. J., Rennie, K. E., Fiore, J. A., & Hardesty, J. C. (2020). National health expenditure projections, 2019–28: Expected rebound in prices drives rising spending growth. Health Affairs, 39(4), 704–714.
16. Lagasse, J. (2020). Readmission penalties for safety-net hospitals drop under new rules. Healthcare Finance News.
17. Li, X., Zhang, Y., Kumar, P., Osei, A., & Chen, R. (2024). Predicting 30-day hospital readmission in Medicare patients: Insights from an LSTM deep learning model. arXiv:2410.17545.
18. Lin, W.-Y., Chen, C.-H., & Tsai, M.-J. (2022). Machine learning-based 30-day hospital readmission predictions for COPD patients using physical activity data of daily living with accelerometer-based device. PMC, 9406028.
19. Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems 30 (pp. 4765–4774). Curran Associates.
20. McIlvennan, C. K., Eapen, Z. J., & Allen, L. A. (2015). Hospital readmissions reduction program. Circulation, 131(20), 1796–1803.
21. Nguyen, O. K., Makam, A. N., Clark, C., Zhang, S., Xie, B., Velasco, F., Amarasingham, R., & Halm, E. A. (2020). Assessing the impact of social determinants of health on predictive models for potentially avoidable 30-day readmission or death. PMC, 7316307.
22. Nguyen, T., Patel, R., & Osborn, C. (2021). Social determinants of health and hospital readmissions: Can the HOSPITAL risk score be improved by the inclusion of social factors? PMC, 7780407.
23. Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447–453.
24. Okoro, C., Bassey, U., & Adamu, H. (2025). A machine learning approach for predicting 30-day hospital readmission in patients with diabetes. PMC, 13163415.
25. Osei-Anto, A., Fields, J., & Ramanathan, K. (2025). The role of machine learning in predicting hospital readmissions among general internal medicine patients: A systematic review. PMC, 12187041.
26. Rajaguru, V., Kim, T. H., Shin, J., Lee, S. G., & Han, W. (2022). Ability of the LACE index to predict 30-day readmissions in patients with acute myocardial infarction. Journal of Personalized Medicine, 12(7), 1085.
27. Raza, S. (2022). A machine learning model for predicting, diagnosing, and mitigating health disparities in hospital readmission. arXiv:2206.06279.
28. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why should I trust you?”: Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1135–1144). ACM.
29. Ribeiro, M. T., Singh, S., & Guestrin, C. (2018). Anchors: High-precision model-agnostic explanations. Proceedings of the AAAI Conference on Artificial Intelligence, 32(1), 1527–1535.
30. Robinson, R., & Hudali, T. (2017). The HOSPITAL score and LACE index as predictors of 30-day readmission in a retrospective study at a university-affiliated community hospital. PeerJ, 5, e3137.
31. Sung, S. F., Liu, K. E., & Ong, C. T. (2021). Validity of the LACE index for identifying frequent early readmissions after hospital discharge in children. PMC, 8032568.
32. U.S. Food and Drug Administration. (2021). Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan. FDA Center for Devices and Radiological Health, Digital Health Center of Excellence.
33. Wadhera, R. K., Joynt Maddox, K. E., Wasfy, J. H., Haneuse, S., Shen, C., & Yeh, R. W. (2019). Association of the Hospital Readmissions Reduction Program with mortality among Medicare beneficiaries hospitalized for heart failure, acute myocardial infarction, and pneumonia. JAMA, 320(24), 2542–2552.
34. Weiss, A. J., & Jiang, H. J. (2021). Overview of clinical conditions with frequent and costly hospital readmissions by payer, 2018: Statistical brief #278. Agency for Healthcare Research and Quality.
35. Yazdan-Ashoori, P., & Van Spall, H. G. C. (2014). Using the LACE index to predict hospital readmissions in congestive heart failure patients. PMC, 4128541.