A Machine Learning Approach in Predicting Mortality Following Emergency General Surgery.

Area of Science:

• Surgical outcomes research
• Artificial intelligence in medicine
• Health informatics

Background:

• Emergency general surgery (EGS) procedures are associated with significant mortality.
• Predicting mortality in EGS is crucial for patient management and resource allocation.

Purpose of the Study:

• To develop and validate a machine learning (ML) algorithm for predicting mortality after EGS.
• To compare the ML algorithm's performance against established risk prediction models.

Main Methods:

• Utilized the American College of Surgeons National Surgical Quality Improvement Program database (2012-2017).
• Developed an ML algorithm to predict EGS mortality.
• Compared ML model performance against American Society of Anesthesiologists (ASA) classification, American College of Surgeon Surgical Risk Calculator (ACS-SRC), and modified frailty index (mFI) using AUC, sensitivity, specificity, PPV, and NPV.

Main Results:

• The developed ML algorithm demonstrated very high performance in predicting EGS mortality.
• The ML algorithm exhibited superior predictive performance compared to ASA classification, ACS-SRC, and mFI.
• Key performance metrics including AUC, sensitivity, specificity, PPV, and NPV favored the ML approach.

Conclusions:

• Machine learning presents a promising tool for predicting outcomes in EGS.
• ML can aid clinicians in surgical decision-making, patient counseling, and identifying modifiable risk factors.
• Improved prediction may lead to better clinical outcomes, optimized resource allocation, and reduced treatment costs.