Predictive Value of Machine Learning for Poststroke Mortality Risk: Systematic Review and Meta-Analysis
Yujie Chen1, Zhujing Ou1, Yiting Deng1
1Department of Neurology, West China Hospital, Sichuan University, Chengdu, China.
Journal of Medical Internet Research
|April 2, 2026
Summary
Machine learning (ML) models show promise in predicting stroke mortality, aiding clinical decisions. However, study heterogeneity and bias necessitate caution in real-world use, with external validation recommended.
Area of Science:
- Medical Informatics
- Clinical Epidemiology
- Artificial Intelligence in Healthcare
Background:
- Stroke poses a significant mortality risk, underscoring the need for accurate prediction models.
- Machine learning (ML) is increasingly explored for stroke mortality prediction, but robust evidence is limited.
Conclusions:
- ML-based stroke mortality prediction is feasible and can serve as an auxiliary tool for risk stratification.
- Substantial heterogeneity and risk of bias in existing studies warrant caution for clinical implementation.
- External validation is crucial before widespread adoption of ML models in diverse clinical settings.
