Development of a machine learning algorithm model to predict intraoperative hypotension in elderly patients...

Area of Science:

• Anesthesiology
• Medical Informatics
• Geriatric Medicine

Background:

• Intraoperative hypotension (IOH) poses risks in elderly patients undergoing general anesthesia.
• Predicting IOH is crucial for patient safety and effective anesthetic management.

Purpose of the Study:

• To develop and validate machine learning (ML) models for predicting IOH risk in elderly patients.
• Identify key predictors of IOH in this demographic.

Main Methods:

• Secondary analysis of 1,720 elderly patients from a randomized controlled trial.
• Eight ML models were developed and evaluated using cross-validation and various performance metrics.
• SHapley Additive exPlanations (SHAP) were used for model interpretability.

Main Results:

• Anesthesia protocol, comorbidities, preoperative labs, intraoperative drugs, and surgical factors were key predictors.
• The random forest model demonstrated exceptional performance (accuracy=0.9917, AUC-ROC=0.9998).

Conclusions:

• A robust ML model can accurately predict IOH in elderly surgical patients.
• Findings support individualized anesthesia management and targeted preventive strategies to mitigate IOH incidence.