Statistical Prediction of Hospital Readmission from Structured Clinical Data
Keywords:
Combining PC-1; PC-2 with ML Ridge regression; SVM Linear; lasso; AUC-ROC curve; area under the precision-recall; maximal accuracy Ya.Yu. et al.; the goodness of fit was evaluated and an internal; external validation of the risk score Risk factors; sensitive; specific; especially; AUC-ROC 3829 clinical/biobanked; 110706 clinically well-characterized; 7400 clinical; 165237 individuals.Abstract
Increasing emphasis on reducing hospital readmission rates has led health systems to develop predictive models that allow for targeting of high-risk patients prior to discharge. Structure data from the electronic health record are often used to create these predictive models; however involving only a subset of clinical features where prediction is likely to be accurate may yield superior model performance at the expense of generalizability to other patient populations or care settings. Although logistic regression is seen as a natural choice for this modeling task due its probabilistic structure, modern statistical learning has proposed a variety of methods such as support vector machines and gradient-boosted trees that provide comparable or improved performance on standard testing criteria; with large amounts of patient data being collected, these approaches can be applied simply. Nevertheless, the rationale for categorical or non-linear methods over properly regularized logistic regression remains insufficiently understood.
A comprehensive predictive modeling framework for identifying patients at high risk of hospital readmission based entirely on structured clinical data. Five methods—logistic regression with and without lasso and ridge regularization, support vector machines with a linear or radial basis function kernel, and gradient-boosted trees—were used and their predictions compared. Models were trained on admissions from a 3-year period and external validation performed using the 4th year. A stepwise variable selection approach was also investigated, aiming to identify a restricted subset of clinical data for which prediction would be accurate. Model performance was evaluated using area under the receiver-operator characteristic curve, calibration plots, and positive predictive value. Results indicated that predictive models for hospital readmission could be developed on the entire patient population in a general hospital, while maintaining sufficient performance characteristics for use in practice.
References
1. Matheny, M. E., Ricket, I., Goodrich, C. A., Shah, R. U., Stabler, M. E., Perkins, A. M., Dorn, C., Denton, J., Bray, B. E., Gouripeddi, R., Higgins, J., Chapman, W. W., MacKenzie, T. A., & Brown, J. R. (2021). Development of electronic health record–based prediction models for 30-day readmission risk among patients hospitalized for acute myocardial infarction. JAMA Network Open, 4(1), e2035782.
2. Huang, Y., Talwar, A., Chatterjee, S., & Aparasu, R. R. (2021). Application of machine learning in predicting hospital readmissions: A scoping review of the literature. BMC Medical Research Methodology, 21, 96.
3. Mattaparthi, R. (2023). Deep Learning-Driven Combustion Anomaly Detection in Diesel Powertrains: A Multi-Sensor Fusion Approach for Real-Time ECM Adaptation. International Journal of Intelligent Systems and Applications in Engineering, 11, 1084.
4. Zhao, P., Yoo, I., & Naqvi, S. H. (2021). Early prediction of unplanned 30-day hospital readmission: Model development and retrospective data analysis. JMIR Medical Informatics, 9(3), e16306.
5. Saha, P., Sircar, R., Bose, A., & others. (2021). Using hospital Admission, Discharge & Transfer (ADT) data for predicting readmissions. Machine Learning with Applications, 5, 100055.
6. Ryu, B., Yoo, S., Kim, S., & Choi, J. (2021). Thirty-day hospital readmission prediction model based on common data model with weather and air quality data. Scientific Reports, 11, 23313.
7. Yandamuri, U. S. (2023). An Intelligent Analytics Framework Combining Big Data and Machine Learning for Business Forecasting. International Journal Of Finance, 36(6), 682-706.
8. Lo, Y.-T., Liao, J. C., Chen, M.-H., Chang, C.-M., & Li, C.-T. (2021). Predictive modeling for 14-day unplanned hospital readmission risk by using machine learning algorithms. BMC Medical Informatics and Decision Making, 21, 288.
9. Sutter, T., Roth, J. A., Chin-Cheong, K., Hug, B. L., & Vogt, J. E. (2021). A comparison of general and disease-specific machine learning models for the prediction of unplanned hospital readmissions. Journal of the American Medical Informatics Association, 28(4), 868–873.
10. Nguyen, O. K., Washington, C., Clark, C. R., Makam, A. N., et al. (2021). Man vs. machine: Comparing physician vs. electronic health record–based model predictions for 30-day hospital readmissions. Journal of General Internal Medicine, 36(9), 2555–2562.
11. Reddy, V. A. R. (2022). Designing Fault-Tolerant Data Ingestion Pipelines for High-Volume Healthcare Transactions. Frontiers in Health Informatics, 11, 861-889.
12. Van Grootven, B., Jepma, P., Rijpkema, C., Verweij, L., Leeflang, M., Daams, J., Deschodt, M., Milisen, K., Flamaing, J., & Buurman, B. (2021). Prediction models for hospital readmissions in patients with heart disease: A systematic review and meta-analysis. BMJ Open, 11(8), e047576.
13. Johnson, A. E., Zhu, J., Garrard, W., Thoma, F. W., Mulukutla, S., Kershaw, K. N., & Magnani, J. W. (2021). Area deprivation index and cardiac readmissions: Evaluating risk-prediction in an electronic health record. Journal of the American Heart Association, 10(13), e020466.
14. Dreyer, R. P., Raparelli, V., Tsang, S. W., D'Onofrio, G., Lorenze, N., Xie, C. F., Geda, M., Pilote, L., & Murphy, T. E. (2021). Development and validation of a risk prediction model for 1-year readmission among young adults hospitalized for acute myocardial infarction. Journal of the American Heart Association, 10(18), e021047.
15. Mangalampalli, B. M. (2023). AI-Driven Anomaly Detection in Healthcare Claims Data: A Business Intelligence Perspective. Journal of Rare Cardiovascular Diseases.
16. Grek, A. A., Rogers, E. R., Peacock, S. H., Hartjes, T. M., White, L. J., Li, Z., Naessens, J. M., & Franco, P. M. (2022). REadmission PREvention in SepSis: Development and validation of a prediction model. Journal for Healthcare Quality, 44(3), 161–168.
17. Omary, C., Wright, P., Kumarasamy, M. A., Franks, N., Esper, G., Mouzon, H. B., Barrolle, S., Horne, K., & Cranmer, J. (2022). Using routinely collected electronic health record data to predict readmission and target care coordination. Journal for Healthcare Quality, 44(1), 11–22.
18. Wang, S., & Zhu, X. (2022). Predictive modeling of hospital readmission: Challenges and solutions. IEEE/ACM Transactions on Computational Biology and Bioinformatics, 19(5), 2975–2995.
19. Wang, S., & Zhu, X. (2022). Nationwide hospital admission data statistics and disease-specific 30-day readmission prediction. Health Information Science and Systems, 10, 25.
20. Davis, S., Zhang, J., Lee, I., Rezaei, M., Greiner, R., McAlister, F. A., & Padwal, R. (2022). Effective hospital readmission prediction models using machine-learned features. BMC Health Services Research, 22, 1415.
21. Kolla, S. H., & Loganathan, R. (2023). Cloud-Native Deep Learning Architectures For Secure Generative AI Deployment In Enterprise Workflow Platforms. Journal of International Crisis and Risk Communication Research, 603-618.
22. Michailidis, P., Dimitriadou, A., Papadimitriou, T., & Gogas, P. (2022). Forecasting hospital readmissions with machine learning. Healthcare, 10(6), 981.
23. Le Lay, J., Alfonso-Lizarazo, E., Augusto, V., Bongue, B., Masmoudi, M., Xie, X., Gramont, B., & Célarier, T. (2022). Prediction of hospital readmission of multimorbid patients using machine learning models. PLOS ONE, 17(12), e0279433.
24. Morrison, J. M., Casey, B., Sochet, A. A., Dudas, R. A., Rehman, M., Goldenberg, N. A., Ahumada, L., & Dees, P. (2022). Performance characteristics of a machine-learning tool to predict 7-day hospital readmissions. Hospital Pediatrics, 12(9), 824–832.
25. Verma, V. K., & Lin, W.-Y. (2022). Machine learning-based 30-day hospital readmission predictions for COPD patients using physical activity data of daily living with accelerometer-based device. Biosensors, 12(8), 605.
26. Zubillaga, A., Laccourreye, P., Kerexeta, J., Larburu, N., Alonso, E., Gómez, D. J., Martínez, F., & Alonso-Arce, M. (2022). Hospital readmission prediction via keyword extraction and sentiment analysis on clinical notes. Studies in Health Technology and Informatics, 295, 339–342.
27. Syed, S. (2023). Shaping The Future Of Large-Scale Vehicle Manufacturing: Planet 2050 Initiatives And The Role Of Predictive Analytics. Nanotechnology Perceptions, 19(3), 103-116.
28. Soh, J. G. S., Mukhopadhyay, A., Mohankumar, B., et al. (2022). Predicting and validating 30-day hospital readmission in adults with diabetes whose index admission is diabetes-related. The Journal of Clinical Endocrinology & Metabolism, 107(10), 2865–2873.
29. Goodman, D. M., Casale, M. T., Rychlik, K., Carroll, M. S., Auger, K. A., Smith, T. L., Cartland, J., & Davis, M. M. (2022). Development and validation of an integrated suite of prediction models for all-cause 30-day readmissions of children and adolescents aged 0 to 18 years. JAMA Network Open, 5(11), e2241513.
30. Gao, X., Alam, S., Shi, P., Dexter, F., & Kong, N. (2023). Interpretable machine learning models for hospital readmission prediction: A two-step extracted regression tree approach. BMC Medical Informatics and Decision Making, 23, 104.
31. Li, M., Cheng, K., Ku, K., Li, J., Hu, H., & Ung, C. O. L. (2023). Modelling 30-day hospital readmission after discharge for COPD patients based on electronic health records. npj Primary Care Respiratory Medicine, 33, 16.
Additional Files
Published
Data Availability Statement
None
