ICU sedation, delirium & recovery
Predicting Intensive Care Delirium with Machine Learning: Model Development and External Validation
Authors' conclusion
Machine learning models trained with routinely collected electronic health record data accurately predict ICU delirium, supporting dynamic time-sensitive forecasting.
Study context
- Population
- Intensive care unit (ICU) patient stays with routinely collected electronic health record data
- Comparison
- 24-hour model and dynamic model vs modified reference model
- Outcome
- Delirium (positive CAM-ICU screen or Intensive Care Delirium Screening Checklist ≥4)
Methodological detailsmachine learning, multicentre development database, external validation on two single-centre databases, Development multicenter database 2014 to 2015; validation single-center databases 2001 to 2012 and 2008 to 2019
- Machine learning
- Multicentre development database
- External validation on two single-centre databases
- Development multicenter database 2014 to 2015; validation single-center databases 2001 to 2012 and 2008 to 2019
Performance
As reported in the abstract
Discrimination for ICU delirium (AUC)
| Test or model | Value | AUC |
|---|---|---|
| 24-hour model | mean 0.785 (95% CI 0.769–0.801) | 0.785 |
| Modified reference model | 0.730 (95% CI 0.704–0.757) | 0.730 |
| Dynamic model | mean 0.845 (95% CI 0.831–0.859) | 0.845 |
Value
24-hour modelmean 0.785 (95% CI 0.769–0.801)
Modified reference model0.730 (95% CI 0.704–0.757)
Dynamic modelmean 0.845 (95% CI 0.831–0.859)
AUC
24-hour model0.785
Modified reference model0.730
Dynamic model0.845
Further findings
Calibration (Brier score), 24-h and dynamic models
- Calibration similar in both models; mean Brier Score [95% CI]
- 0.102 [0.097 to 0.108] and 0.111 [0.106 to 0.116]
- Model discrimination and calibration maintained on the validation datasets
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Predicting Intensive Care Delirium with Machine Learning: Model Development and External Validation. Anesthesiology. Published 20 December 2022. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →