A model-based validation study of postoperative complications with considerations on operative timing

Area of Science:

• Medical Informatics
• Surgical Outcomes Research
• Machine Learning in Healthcare

Background:

• Surgical procedures demand high surgeon acuity and dexterity.
• The influence of surgeon fatigue and circadian rhythms on surgical outcomes remains inconsistently studied.

Purpose of the Study:

• To investigate the significance of operative timing in classifying surgical complications.
• To apply an interpretable machine learning approach to surgical records.

Main Methods:

• Retrospective study utilizing surgical record data from a Chinese tertiary teaching hospital.
• Trained and optimized various machine learning models (linear, generative, tree-based) using grid search cross-validation for F1 score optimization.

Main Results:

• XGBoost demonstrated superior performance in classifying complications.
• Surgery duration was the most significant predictor of complications.
• Nighttime surgery (9 pm - 7 am) showed a moderate association, less impactful than patient age, gender, or surgery type.

Conclusions:

• Surgical records can indicate a mild effect of operative timing on postoperative complications.
• Operative timing's influence is potentially entangled with multiple other patient and procedural factors.