Perioperative medicine & outcomes
Development of a machine learning-based risk prediction model for perioperative neurocognitive disorders
Authors' conclusion
This study systematically identified core risk factors for PND in non-cardiac surgical patients using machine learning, and developed both logistic regression-based nomograms and online tools that prioritize interpretability and practicality to support clinical decision-making.
Study context
- Population
- Elderly patients undergoing non-cardiac elective surgery
- Participants
- 2,042 patients
- Comparison
- Ten machine learning prediction models
- Outcome
- Perioperative neurocognitive disorder (PND)
Methodological detailsmachine learning, LASSO regression, SHAP interpretation, decision curve analysis
- Machine learning
- LASSO regression
- SHAP interpretation
- Decision curve analysis
Main finding
As reported in the abstract
Predictive performance for PND
| Group | AUC |
|---|---|
| Neural network | 0.877 (95 % CI: 0.839-0.916) |
Neural network
AUC0.877 (95 % CI: 0.839-0.916)
- AUC range across the ten machine learning models
- 0.615 to 0.877
- Perioperative neurocognitive disorder (PND) incidence
- 12.5 % (255/2042)
Further findings
Key risk factors (SHAP analysis)
- Hyperlipidaemia (highest SHAP value), smoking, ASA classification III and low education level identified as key risk factors
Timing of PND onset (survival analysis)
- Smoking, ASA classification III and hypertension associated with earlier onset of PND
- P < 0.05
- Log-rank test
Original publication
Development of a machine learning-based risk prediction model for perioperative neurocognitive disorders. J Clin Anesth. Published 25 September 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →