Related Experiment Videos
Development and Validation of a Prediction Model for Postoperative Delirium
Laixi Li1, Ningning Qiao2, Xiaojuan Yang3
1Department of Anesthesiology, 541 General Hospital, Yuncheng, Shanxi Province, People's Republic of China.
Neuropsychiatric Disease and Treatment
|June 15, 2026
Summary
A machine learning model accurately predicts postoperative delirium (POD) risk. Key predictors include cognitive status, nutrition, and demographics, enabling early intervention for surgical patients.
Area of Science:
- Medical Informatics
- Machine Learning in Healthcare
- Surgical Outcomes Research
Background:
- Postoperative delirium (POD) is a frequent and severe complication following surgery.
- Early identification of patients at high risk for POD is crucial for implementing timely interventions.
Purpose of the Study:
- To develop and validate a machine learning (ML) model for predicting POD.
- To compare the performance of eight different ML algorithms for POD prediction.
Main Methods:
- A multicenter retrospective cohort of 3000 surgical patients was analyzed.
- Eight ML algorithms were evaluated, with Random Forest selected as the optimal model based on AUC.
- SHapley Additive exPlanations (SHAP) were used for model interpretability.
Main Results:
- The Random Forest model achieved high predictive accuracy (AUC: 0.823 in external validation).
- Key predictors identified include Chinese Mini Mental Status (CMMS) score, Prognostic Nutritional Index, ASA classification, age, dementia, and ICU admission.
- The model demonstrated good calibration and discrimination.
Conclusions:
- A validated, interpretable ML framework for POD risk stratification is presented.
- The findings highlight the importance of nutritional status as a modifiable target for perioperative interventions to reduce POD.
