Acute cardiac care & resuscitation
Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning
Authors' conclusion
The novel CShock score has the potential to provide automated detection and early warning for cardiogenic shock and improve the outcomes for millions of patients who suffer from myocardial infarction and heart failure.
Study context
- Population
- Patients admitted to the cardiac intensive care unit (ICU) with acute decompensated HF and/or MI
- Participants
- 1,500 vs 131
- Comparison
- Development vs New York University (NYU)
- Developmentn = 1,500
- NYUn = 131
- Outcome
- Onset of cardiogenic shock (CShock discrimination, AUROC)
Methodological detailsdeep learning, physician-adjudicated outcomes
- Deep learning
- Physician-adjudicated outcomes
Performance
As reported in the abstract
Prediction of cardiogenic shock onset (AUROC)
| Test or model | Value | AUROC |
|---|---|---|
| Development | 204 events | 0.821 (95% CI 0.792-0.850) |
Development
Value204 events
AUROC0.821 (95% CI 0.792-0.850)
Further findings
Top predictors (Shapley values)
- Elevated heart rate most predictive; others: admission diagnosis of STEMI, admission diagnosis of acute decompensated HF, Braden Scale, Glasgow Coma Scale, blood urea nitrogen, SBP, serum chloride, serum sodium, arterial blood pH
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning. Eur Heart J Acute Cardiovasc Care. Published 22 March 2024. doi:10.
Summarised from the published abstract; every number is checked against it