Perioperative medicine & outcomes
Prospective validation and real-time implementation of an automated machine learning postoperative mortality prediction model
Authors' conclusion
This prospective validation and EHR implementation of a previously published random forest machine learning model predicting postoperative in-hospital mortality demonstrated acceptable real-world performance of the implemented model and feasibility of integrating such a system into clinical practice.
Study context
- Population
- Inpatients undergoing surgery at a single academic medical centre
- Comparison
- Implemented 32-feature model vs 58-feature model vs ASA physical status
- Outcome
- Prediction of postoperative in-hospital mortality (AUROC, AUCPR)
Methodological detailssingle-centre, real-time electronic health record (EHR) implementation, random forest machine learning
- Single-centre
- Real-time EHR implementation
- Random forest machine learning
Performance
As reported in the abstract
Prediction of postoperative in-hospital mortality, prospectively implemented model
AUROC of the prospectively implemented model
0.874
95% CI 0.860–0.887
Area under the curve precision-recall (AUCPR) of the prospectively implemented model
0.111
Secondary outcomes
Feasibility of pilot clinical decision support tool
- Feasibility shown by real-time data updates, automated transfer of model outputs to the EHR, and provider survey entries
Original publication
Prospective validation and real-time implementation of an automated machine learning postoperative mortality prediction model. Br J Anaesth. Published 17 January 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →