Prognostic machine learning models for predicting postoperative complications following general surgery in Bandar...

Area of Science:

• Surgical quality improvement
• Medical informatics
• Predictive analytics in healthcare

Background:

• Minimizing postoperative complications is crucial for enhancing surgical care quality.
• Understanding the occurrence and risk factors of postoperative complications is essential.
• Machine learning (ML) offers potential for developing predictive models in surgery.

Purpose of the Study:

• To develop risk factor prediction models for postoperative complications after general surgery.
• To assist surgeons in identifying high-risk patients.
• To improve patient outcomes through early risk identification.

Main Methods:

• Prospective analysis of general surgery patients (aged ≥18) at a tertiary referral center.
• Data collection from September 2025 to September 2026, excluding incomplete records.
• Application of four supervised ML techniques: logistic regression, decision trees, random forests, and extreme gradient boosting to predict 30-day postoperative complications.

Main Results:

• Models will be evaluated using accuracy, precision, recall, and F1 score.
• Identification of key patient-related, surgery-related, and postoperative factors influencing complication risk.
• Comparison of the performance of different ML algorithms in predicting complications.

Conclusions:

• ML models can effectively predict postoperative complications in general surgery.
• Identifying risk factors can guide targeted interventions to reduce complications.
• This approach has the potential to enhance surgical care quality and patient safety.