Cardiac arrest & resuscitation
Prospective, multicenter validation of the deep learning-based cardiac arrest risk management system for predicting in-hospital cardiac arrest or unplanned intensive care unit transfer in patients admitted to general wards
Authors' conclusion
The DeepCARS™ predicts IHCA and UIT more accurately and efficiently than conventional methods.
Study context
- Population
- Adults admitted to general wards during a 3-month period, 4 teaching hospitals, South Korea
- Participants
- 55,083 adults
- Exposure
- DeepCARS™ vs modified early warning score (MEWS), national early warning Score (NEWS) and single-parameter track-and-trigger systems
- Outcome
- Predictive accuracy (AUROC) for IHCA or unplanned ICU transfer
- Follow-up
- Within 24 h of the alarm being triggered
Methodological detailsdiagnostic accuracy, multicentre
- Diagnostic accuracy
- Multicentre
Main finding
As reported in the abstract
IHCA or unplanned ICU transfer (composite) · within 24 h of alarm
| Group | AUROC |
|---|---|
| DeepCARS™ | 0.869 |
| MEWS | 0.756 |
| NEWS | 0.767 |
AUROC
DeepCARS™0.869
MEWS0.756
NEWS0.767
Observational data: associations, not effects.
Further findings
Incidence of in-hospital cardiac arrest (IHCA) and unplanned ICU transfer
- IHCA per 1,000 admissions
- 0.90
- Unplanned ICU transfer per 1,000 admissions
- 6.44
Alarm burden and appropriateness at same sensitivity
- Mean alarm counts per day per 1,000 beds significantly reduced with DeepCARS™; rate of appropriate alarms higher vs conventional systems
Original publication
Prospective, multicenter validation of the deep learning-based cardiac arrest risk management system for predicting in-hospital cardiac arrest or unplanned intensive care unit transfer in patients admitted to general wards. Crit Care. Published 5 September 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →