Development and validation of interpretable machine learning models to predict 30-day mortality in patients with intracerebral hemorrhage.
Machine learning models explainable to doctors predict which brain hemorrhage patients face highest 30-day mortality risk, supporting earlier intensive interventions.
This study develops and validates interpretable ML models (with SHAP value-based explainability) for predicting 30-day mortality in intracerebral hemorrhage patients using clinical and laboratory data, with performance superior to conventional scoring approaches. Feature importance analysis identifies GCS, hematoma volume, and select laboratory values as top predictors, providing insight into model decision-making and potential for clinical deployment.
What the study was
- Study design
- retrospective_cohort
- Population
- Intracerebral hemorrhage patients
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- Sci Rep
Why it surfaced
AI/ML diagnostics watchlist; interpretable ML (SHAP) for ICH mortality prediction; open-access Sci Rep; limited novelty in a crowded space but solid methodology; corrected pipeline_priority from MEDIUM to STANDARD; classification_confidence corrected from moderate to medium.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.