Pediatric anesthesia & critical care
Development and external validation of a machine learning model for brain injury in pediatric patients on extracorporeal membrane oxygenation
Authors' conclusion
The Random Forest model based on machine learning demonstrates high accuracy and robustness in predicting brain injury in pediatric patients supported by ECMO, with strong generalization capabilities and promising clinical applicability.
Study context
- Population
- Pediatric patients supported by extracorporeal membrane oxygenation (ECMO) (Chinese Society of Extracorporeal Life Support registry and local hospitals)15 features selected
- Participants
- 1,633 vs 154
- Comparison
- Ten machine learning (ML) methods compared for predicting brain injury; random forest best
- Developmentn = 1,633
- External validationn = 154
- Outcome
- Brain injury prediction (accuracy and AUC)
Methodological detailsmachine learning, retrospective cohort, internal and external validation, multicentre registry
- Machine learning
- Retrospective cohort
- Internal and external validation
- Multicentre registry
Performance
As reported in the abstract
Random forest model performance for brain injury
| Test or model | Value | AUC (internal validation) | AUC (external validation) |
|---|---|---|---|
| Development | 181 events | 0.912 | |
| External validation | 30 events | 0.807 |
Value
Development181 events
External validation30 events
AUC (internal validation)
Development0.912
External validation— not stated in the abstract
AUC (external validation)
Development— not stated in the abstract
External validation0.807
Original publication
Development and external validation of a machine learning model for brain injury in pediatric patients on extracorporeal membrane oxygenation. Crit Care. Published 9 January 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →