Liangyuan Na, Kimberly Villalobos Carballo, Jean Pauphilet, Ali Haddad-Sisakht, Daniel Kombert, Melissa Boisjoli-Langlois, Andrew Castiglione, Maram Khalifa, Pooja Hebbal, Barry Stein, Dimitris Bertsimas
2023.5.25INFORMS Journal on Applied Analytics
tlooto Summary
A significant reduction in the average length of stay is observed following the adoption of a machine learning tool that predicts the probabilities of next 24-hr/48-hr discharge and intensive care unit transfers, end-of-stay mortality and discharge dispositions.
Abstract
We build and deploy machine learning models that accurately predict short- and medium-term inpatient outcomes for Hartford HealthCare, including 24–48 hour discharge, ICU transfer, mortality, and discharge disposition (AUC 76%–93%). More than 200 clinicians currently use these predictions in daily rounds, leading to earlier discharge planning, shorter length of stay (0.63 days per patient), and substantial financial benefits (between $52 and $67 million annually) for the healthcare system.
Citation format
NA, Liangyuan, et al. Patient outcome predictions improve operations at a large hospital network [preprint]. arXiv, 2023. arXiv:2305.15629.