Perioperative medicine & outcomes
Postoperative delirium risk estimation and assessment with machine learning: development and validation of the Open-DREAM Model: A peri-operative dataset analysis
Authors' conclusion
An open-access ML-based model using routinely collected preoperative variables can predict POD with high sensitivity and preserved interpretability.
Study context
- Population
- Patients ≥60 years undergoing noncardiac, nonintracranial surgery without preoperative delirium, single secondary care hospital in Switzerland, January 2023 to January 20241,425 screened
- Participants
- 748 patients
- Comparison
- Machine learning (ML) algorithms using preoperative data to predict postoperative delirium (POD)
- Outcome
- POD (Nu-DESC score at least 2 at any peri-operative assessment)
Methodological detailsobservational, prospectively collected routine care data, single-centre
- Observational
- Prospectively collected routine care data
- Single-centre
Performance
As reported in the abstract
Prediction of POD
| Test or model | AUC | Sensitivity | Optimal probability threshold |
|---|---|---|---|
| XGBoost | 0.80 | 0.89 | 0.35 |
XGBoost
AUC0.80
Sensitivity0.89
Optimal probability threshold0.35
Further findings
Feature selection (SHAP-derived rankings)
- Predictor set reduced from 55 to 15 features without loss of AUC
Strongest predictor
- Type of anaesthesia (spinal vs general) was the strongest predictor
Original publication
Postoperative delirium risk estimation and assessment with machine learning: development and validation of the Open-DREAM Model: A peri-operative dataset analysis. Eur J Anaesthesiol. Published 27 April 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →