ICU sedation, delirium & recovery
Machine Learning for In-hospital Mortality Prediction in Critically Ill Patients With Acute Heart Failure: A Retrospective Analysis Based on the MIMIC-IV Database
Authors' conclusion
The XGBoost model can effectively predict the in-hospital mortality of AHF patients admitted to the ICU
Study context
- Population
- Critically ill intensive care unit (ICU) patients with acute heart failure (AHF), MIMIC-IV database
- Participants
- 5,114 patients
- Exposure
- XGBoost vs other machine learning (ML) models and common scoring systems
- Training setn = 3,580
- Validation setn = 1,534
- Outcome
- In-hospital mortality prediction
Methodological detailsMIMIC-IV database, machine learning prediction model development and validation, least absolute shrinkage and selection operator (LASSO) variable selection
- MIMIC-IV database
- Machine learning prediction model development and validation
- LASSO variable selection
Main finding
As reported in the abstract
Predictive ability for in-hospital mortality
- XGBoost had the best predictive ability (AUC) among all ML models and was chosen as final model for higher net benefit
- XGBoost predictive ability superior to common scoring systems
Observational data: associations, not effects.
Further findings
- In-hospital mortality
- 12.5%
Original publication
Machine Learning for In-hospital Mortality Prediction in Critically Ill Patients With Acute Heart Failure: A Retrospective Analysis Based on the MIMIC-IV Database. J Cardiothorac Vasc Anesth. Published 16 December 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →