Machine learning for predicting mortality in adult critically ill patients with Sepsis: A systematic review
Authors' conclusion
ML models demonstrate a modest improvement in predicting sepsis-associated mortality.
Study context
- Population
- Critically ill adults with sepsis
- Participants
- 1,477,200 adults · 31 studies
- Comparison
- Machine Learning (ML) models vs conventional scoring systems for predicting mortality
- Outcome
- Prediction of mortality (AUROC)
Methodological detailsprotocol registered at PROSPERO, Databases searched from inception to February 2024; 1822 articles screened
- Protocol registered at PROSPERO
- Databases searched from inception to February 2024; 1822 articles screened
Pooled result
As reported in the abstract
Mortality prediction performance (AUROC range)
ML models
0.62-0.90
AUROC range
Conventional scoring systems
0.47-0.86
AUROC range
Further pooled results
Study methodology
- Studies with high risk of bias
- 19
- Studies with internal and external validation
- 9
Original publication
Machine learning for predicting mortality in adult critically ill patients with Sepsis: A systematic review. J Crit Care. Published 25 July 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →