Harnessing Predictive Analytics for Nursing: A PyHealth-Driven Approach to Hospital Readmission Prediction Using EHR...

Health care systems face considerable challenges of hospital readmission, which bear adverse implications for patient outcomes and compromise productivity and efficient use of system resources. In this study, PyHealth, an open-source health care machine-learning tool, and the RETAIN model, a reverse-time attention neural network, were used to predict hospital readmissions within 30 days after discharge. The analysis was performed on the publicly available MIMIC-III demo data set that was additionally deidentified for patient privacy. The predictability of hospital readmissions within 30 days after discharge was evaluated using nursing-centric clinical components such as diagnosis, procedures, and prescriptions. The results revealed a strong potential of open-source tools such as PyHealth in enhancing nursing support for workflows such as discharge planning, education, and postdelivery interventions. Accordingly, integrations like PyHealth can assist in improving patient care before discharge so as to minimize the number of hospital readmissions.