Perioperative medicine & outcomes
Mortality prediction after major surgery in a mixed population through machine learning: a multi‐objective symbolic regression approach
Authors' conclusion
This study confirms the feasibility of developing a multi-objective symbolic regression-based model for predicting 1-year postoperative mortality in a mixed non-cardiac surgical population.
Study context
- Population
- Patients undergoing major elective non-cardiac surgery with previous cardiopulmonary exercise testingof 2145 in database
- Participants
- 1,190 patients
- Comparison
- Multi-objective symbolic regression model vs existing models
- Training datasetn = 952
- Test datasetn = 238
- Outcome
- Model performance for predicting 1-year mortality (F1 score)
- Follow-up
- 1 year
Methodological detailssingle-institution, machine learning (multi-objective symbolic regression), Median (IQR [range]) age 71 (61-79 [45-89]) years; 825 (69%) male
- Single-institution
- Machine learning (multi-objective symbolic regression)
- Median (IQR [range]) age 71 (61-79 [45-89]) years; 825 (69%) male
Performance
As reported in the abstract
Model performance (F1 score)
F1 score of the multi-objective symbolic regression model
0.712
Further findings
Major contributors to model performance (Shapley additive explanations)
- Ventilatory equivalents for carbon dioxide, oxygen at peak exercise and BMI influenced model performance most, surpassing surgery type and named comorbidities
Original publication
Mortality prediction after major surgery in a mixed population through machine learning: a multi‐objective symbolic regression approach. Anaesthesia. Published 8 January 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →