Clinical Report: Interpretable Gradient Boosting Machine for In-Hospital Mortality in SIMI

Overview

This study developed an interpretable Gradient Boosting Machine (GBM) model to predict in-hospital mortality in patients with sepsis-induced myocardial injury (SIMI). The model demonstrated strong performance in both internal and external validations.

Background

Sepsis-induced myocardial injury (SIMI) is a severe complication of sepsis, significantly increasing in-hospital mortality rates. Traditional risk prediction models lack interpretability and multi-center validation, which limits their clinical utility. This study addresses the limitations of current models.

Data Highlights

Key Findings

• The GBM model outperformed other algorithms in predicting in-hospital mortality for SIMI patients.
• Key predictors identified included APS III Score, Hypertension, Albumin, Diabetes, SOFA Score, ALT, RBC, and Lactate.
• A simplified model with five variables achieved an AUC of 0.789.
• The model was deployed on a user-friendly online platform for real-time risk assessment.
• External validation confirmed the model's generalizability across different cohorts.

Clinical Implications

The development of an interpretable predictive model for SIMI patients may enhance risk assessment.

Conclusion

This multicenter study successfully created and validated an interpretable model for predicting in-hospital mortality in SIMI patients.

Related Resources & Content

1. JMIR Medical Informatics, 2026 -- Online Sepsis Prediction Using Vital Signs and Multiscale Temporal-Aware Contrastive Learning: Model Development and Validation Study
2. BMJ Health & Care Informatics, 2026 -- Early sepsis prediction using a hybrid LSTM-GAT model: a study on the PhysioNet 2019 dataset
3. Frontiers in Cardiovascular Medicine, 2026 -- A machine learning model for predicting short-term in-hospital mortality in acute myocardial infarction with coexisting chronic obstructive pulmonary disease
4. DIGITAL HEALTH, 2026 -- Development and deployment of an interpretable stacking ensemble model for predicting in-hospital mortality in ICU patients with chronic kidney disease and sepsis
5. Fourth Universal Definition of Myocardial Infarction (2018) | JACC
6. Sepsis mortality prediction using machine learning and deep learning - a systematic review | BMC Medical Informatics and Decision Making | Springer Nature Link
7. Fourth Universal Definition of Myocardial Infarction (2018) | JACC
8. https://academic.oup.com/eurheartj/article/46/34/3339/8152698
9. Sepsis mortality prediction using machine learning and deep learning - a systematic review | BMC Medical Informatics and Decision Making | Springer Nature Link