Acute cardiac care & resuscitation
Predicting Out-of-Hospital Cardiac Arrest in the General Population Using Electronic Health Records
Authors' conclusion
In a population-based case-control study, machine learning models incorporating readily available EHR data showed reasonable discrimination and risk enrichment for OHCA in the general population. Salient factors associated with OCHA risk were myriad across the cardiovascular and noncardiovascular spectrum.
Study context
- Population
- General population receiving health care at the University of Washington; individuals with OHCA and age- and sex-matched controls2,366 OHCA, 23,660 controls
- Comparison
- Machine learning models vs Baseline model
- Outcome
- OHCA
Methodological detailspopulation-based, age- and sex-matched, external validation
- Population-based
- Age- and sex-matched
- External validation
Main finding
As reported in the abstract
OHCA prediction (external validation)
| Group | AUC | PPV at 99% specificity |
|---|---|---|
| Machine learning models | 0.80-0.85 | 2.5% to 3.1% |
| Baseline model | 0.66 | 0.8% |
AUC
Machine learning models0.80-0.85
Baseline model0.66
PPV at 99% specificity
Machine learning models2.5% to 3.1%
Baseline model0.8%
Observational data: associations, not effects.
Secondary outcomes
Shockable OHCA
- Established CV risk factors retained predictive importance; demographic characteristics (minority race, single marital status) and noncardiovascular comorbidities (substance abuse disorder) also contributed
Nonshockable OHCA
- Range of salient predictors: comorbidities, habits, vital signs, demographic characteristics, ECG measures
Original publication
Predicting Out-of-Hospital Cardiac Arrest in the General Population Using Electronic Health Records. Circulation. Published 11 June 2024. doi:10.
Summarised from the published abstract; every number is checked against it