EHR-Driven Readmission Prediction: Transparent ML Models For Clinical Decision Support
Main Article Content
Abstract
We obtain de-identified electronic health record data from a prominent hospital with a substantial patient population. Furnish comprehensive clinical data, encompassing age, gender, ethnicity, and socioeconomic status (if accessible). ICD diagnosis encoded. Electronic Health Records and hospital dosages History of hospitalization and medical procedures Laboratory findings including blood tests and imaging results. Data is sanitized and organized post-collection. Machine learning retrieves absent categorical data by mean/median imputation, forward filling, and outlier identification and rectification. Incorporating characteristics enhances model efficacy. Charlson The Comorbidity Index (CCI) values may also indicate patient comorbidity.
We analyze readmission risk prediction by multidimensional interpretable machine learning. Rule-based models employ comprehensible logic, exemplified as designating "high risk" for CHF patients having a pneumonia hospitalization history during the preceding 6 months. Rules-based models analyze complex data but exhibit limited flexibility.
Decision trees classify data utilizing qualities. Clinicians may assess risk using decision trees.
LIME elucidates patient predictions from any black-box model. Modeling the behavior at a data point highlights the advantages of the patient's prediction.
We employ interpretable methodologies to integrate model precision with clinically relevant risk assessment.
Prediction performance and calibration serve as criteria for model assessment. AUROC typically assesses a model's ability to differentiate between readmission and non-readmission. Elevated AUROC levels indicate superior discrimination. Evaluate genuine positives (high-risk readmissions) and true negatives. PPV signifies a high risk of readmission. K-fold cross-validation mitigates overfitting and offers a comprehensive estimation of model performance for thorough evaluation.
Article Details
References
Matheny, M. E., Ricket, I., Goodrich, C. A., et al. (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. https://doi.org/10.1001/jamanetworkopen.2020.35782
Rajkomar, A., Oren, E., Chen, K., et al. (2018). Scalable and accurate deep learning with electronic health records. NPJ Digital Medicine, 1(1), 1-10.
Hao, S., Wang, Y., Jin, B., et al. (2015). Development, validation and deployment of a real time 30 day hospital readmission risk assessment tool in the Maine healthcare information exchange. PloS One, 10(10), e0140271.
Kansagara, D., Englander, H., Salanitro, A., et al. (2011). Risk prediction models for hospital readmission: a systematic review. JAMA, 306(15), 1688-1698.
Vyas, P., Bojja, G. R., Ambati, L. S., Liu, J., & Ofori, M. (2021). Prediction of Patient Willingness to Recommend Hospital: A Machine Learning-Based Exploratory Study. In AMCIS.
Shickel, B., Tighe, P. J., Bihorac, A., & Rashidi, P. (2018). Deep EHR: A survey of recent advances in deep learning techniques for electronic health record (EHR) analysis. IEEE Journal of Biomedical and Health Informatics, 22(5), 1589-1604.
Amarasingham, R., Moore, B. J., Tabak, Y. P., et al. (2010). An automated model to identify heart failure patients at risk for 30-day readmission or death using electronic medical record data. Medical Care, 48(11), 981-988.
Futoma, J., Morris, J., & Lucas, J. (2015). A comparison of models for predicting early hospital readmissions. Journal of Biomedical Informatics, 56, 229-238.
Caruana, R., Lou, Y., Gehrke, J., et al. (2015). Intelligible models for healthcare: Predicting pneumonia risk and hospital 30-day readmission. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1721-1730).
Xiao, C., Choi, E., & Sun, J. (2018). Opportunities and challenges in developing deep learning models using electronic health records data: a systematic review. Journal of the American Medical Informatics Association, 25(10), 1419-1428.
Shameer, K., Johnson, K. W., Yahi, A., et al. (2017). Predictive modeling of hospital readmission rates using electronic medical record-wide machine learning: a case-study using Mount Sinai heart failure cohort. In Pacific Symposium on Biocomputing (pp. 276-287).
Jha, R. K., Bag, S., Koley, D., Bojja, G. R., & Barman, S. (2023). An appropriate and cost-effective hospital recommender system for a patient of rural area using deep reinforcement learning. Intelligent Systems with Applications, 18, 200218.
Goldstein, B. A., Navar, A. M., Pencina, M. J., & Ioannidis, J. P. (2017). Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review. Journal of the American Medical Informatics Association, 24(1), 198-208.
Golas, S. B., Shibahara, T., Agboola, S., et al. (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(1), 1-17.
Rudin, C. (2019). Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nature Machine Intelligence, 1(5), 206-215.
Rehman, Z. U., Zia, M. S., Bojja, G. R., Yaqub, M., Jinchao, F., & Arshid, K. (2020). Texture based localization of a brain tumor from MR-images by using a machine learning approach. Medical hypotheses, 141, 109705.
Lundberg, S. M., Nair, B., Vavilala, M. S., et al. (2018). Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nature Biomedical Engineering, 2(10), 749-760.
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).