Key result
The Criticality Index-Mortality model tracks dynamic hospital mortality risk in PICU patients with ~0.85 AUC.
Why the study?
Accurate and dynamic mortality risk prediction models for children in ICUs that update serially during care are needed to monitor clinical improvement and deterioration in real time.
Population
27,354 ICU admissions of children from 2009 to 2018
Comparison
None (development and validation of machine learning mortality risk model)
Design
Retrospective analysis of a national database
Follow-up
Up to 180 hours
Loading...
May support dynamic pediatric ICU monitoring; leaves open prospective outcome trials before practice change.
Observational (n=27,354)
Yes
Effect estimate: AUC 0.852 (95% CI 0.843-0.861)
A machine learning model incorporating physiology, therapy, and care intensity can accurately track dynamic changes in hospital mortality risk for children in the ICU.
A 2022 study conducted an observational in Children in ICUs (n=27,354). Criticality Index-Mortality (CI-M) machine learning model was evaluated on Hospital mortality risk estimates determined at 6-hour time periods (AUC 0.852, 95% CI 0.843-0.861). The Criticality Index-Mortality machine learning model accurately tracked dynamic hospital mortality risk in pediatric ICU patients, with an overall AUC of 0.852 (95% CI, 0.843-0.861).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: