Neuroanesthesia & neurocritical care
Machine learning-based prognostic analysis of patients with status epilepticus in the neurological intensive care unit
Authors' conclusion
Compared with the other models and scoring systems, XGBoost demonstrated superior predictive performance
Study context
- Population
- Patients with SE with multisystem complications in the neurological ICU
- Participants
- 169 patients
- Comparison
- Six machine learning (ML) algorithms vs Status Epilepticus Severity Score (STESS), ENDIT and EMSE scores
- Outcome
- Prognosis (predictive performance, AUC)
Methodological detailsmachine learning, comparison with clinical scores
- Machine learning
- Comparison with clinical scores
Performance
As reported in the abstract
AUC in test dataset
| Test or model | AUC |
|---|---|
| Decision tree (DT) | 0.660 |
| Random forest (RF) | 0.644 |
| Support vector machine (SVM) | 0.663 |
| K-nearest neighbors (KNN) | 0.689 |
| XGBoost | 0.825 |
| Least absolute shrinkage and selection operator (LASSO) logistic regression | 0.610 |
AUC
Decision tree (DT)0.660
Random forest (RF)0.644
Support vector machine (SVM)0.663
K-nearest neighbors (KNN)0.689
XGBoost0.825
Least absolute shrinkage and selection operator (LASSO) logistic regression0.610
Further findings
Top ten predictors of XGBoost model (SHAP analysis)
- Hypoalbuminemia, nutritional risk score, age, ventilation duration, NCSE, GCS score, duration of impaired consciousness, creatinine level, APACHE II score, and CCI
Original publication
Machine learning-based prognostic analysis of patients with status epilepticus in the neurological intensive care unit. J Crit Care. Published 9 March 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →