ICU sedation, delirium & recovery
ICURE: Intensive care unit (ICU) risk evaluation for 30‐day mortality. Developing and evaluating a multivariable machine learning prediction model for patients admitted to the general ICU in Sweden
Authors' conclusion
The ICURE prediction model predicts 30-day mortality rate at first-time ICU admission superiorly compared to the established SAPS 3 score.
Study context
- Population
- Patients ≥18 years admitted to general ICUs in Sweden 2008–2022 with SAPS 3 registered (Swedish Intensive Care Registry)
- Participants
- 296,344 first-time ICU admissions
- Comparison
- ICURE machine learning model vs SAPS 3 score
- Outcome
- 30-day mortality prediction (AUC-ROC)
- Follow-up
- 30 days
Methodological detailsregistry-based, machine learning, random 80/20 training/test split
- Registry-based
- Machine learning
- Random 80/20 training/test split
Performance
As reported in the abstract
Prediction of 30-day mortality, test data, all predictors
- AUC-ROC
- 0.884
- Specificity
- 95.2%
- Sensitivity
- 47.0%
- NPV
- 87.9%
- PPV
- 70.7%
Further findings
Most important predictors
- Age, GCS, creatinine, SBP and pH
Original publication
ICURE: Intensive care unit (ICU) risk evaluation for 30‐day mortality. Developing and evaluating a multivariable machine learning prediction model for patients admitted to the general ICU in Sweden. Acta Anaesthesiol Scand. Published 21 July 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →