Perioperative medicine & outcomes
Development and preliminary assessment of a machine learning model to predict myocardial infarction and cardiac arrest after major operations
Authors' conclusion
The present ML model outperformed the Gupta score in the prognostication of MI/CA across a heterogenous range of operations.
Study context
- Population
- Patients undergoing major operations, 2016–2020 ACS-NSQIP25,177 (1.0%) MI/CA
- Participants
- 2,473,487 patients
- Exposure
- Machine learning (ML) model vs Gupta score for predicting MI/CA
- Outcome
- Discrimination (C-statistic) and calibration (Brier score) for intra/post-operative MI/CA
Methodological detailsmachine learning prediction model development, ACS-NSQIP registry, training/testing split
- Machine learning prediction model development
- ACS-NSQIP registry
- Training/testing split
Main finding
As reported in the abstract
Discrimination and calibration for MI/CA
| Group | C-statistic | Brier score |
|---|---|---|
| ML model | 0.84 | 0.0097 |
| Gupta score | 0.80 | 0.0100 |
C-statistic
ML model0.84
Gupta score0.80
Brier score
ML model0.0097
Gupta score0.0100
- ML model vs Gupta score: P < 0.001 — testing dataset
Observational data: associations, not effects.
Further findings
Model performance by operative subgroup (subgroup)
- Performance markedly improved among thoracic, aortic, peripheral vascular and foregut surgery
Original publication
Development and preliminary assessment of a machine learning model to predict myocardial infarction and cardiac arrest after major operations. Resuscitation. Published 15 May 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →