Machine learning-based triage model for elderly traumatic brain injury patients in Chinese emergency department

1

An XGBoost-based triage model was developed for elderly TBI patients to optimize ICU allocation in emergency settings.
2

The model achieved an AUC of 0.93, indicating high accuracy in predicting ICU triage disposition within 48 hours.
3

Key predictors for the model included symptoms, CT hematoma density, contusion severity, age, and anticoagulant use.
4

Decision curve analysis suggested the model offers a higher net benefit compared to traditional triage strategies.
5

Further prospective validation is needed to confirm the model's effectiveness in reducing unnecessary ICU admissions.

1