Acute cardiac care & resuscitation
Artificial Intelligence-Enhanced Electrocardiography and Health Records to Predict Cardiac Arrest
Authors' conclusion
In a large U.S. health care system, artificial intelligence-enhanced ECG and EHR data effectively discriminated individuals at risk of OHCA and identified those at clinically relevant risk of incident OHCA over a 2-year period.
Study context
- Population
- General population; individuals undergoing electrocardiography (ECG) in a health care system
- Comparison
- Discrimination of OHCA by the multimodal ECG + EHR model, the EHR-only model and the ECG-only model
Methodological detailstemporal validation, real-world cohort of individuals undergoing ECG
- Temporal validation
- Real-world cohort of individuals undergoing ECG
Main finding
As reported in the abstract
Discrimination of OHCA, temporal validation cohort
| Group | Area under the receiver operating characteristic curve | Area under the precision recall curve |
|---|---|---|
| ECG + EHR model | 0.83 | 0.44 |
ECG + EHR model
Area under the receiver operating characteristic curve0.83
Area under the precision recall curve0.44
- Highest for ECG + EHR model, followed by the EHR-only model and the ECG-only model
- Bonferroni adjusted P for all pairwise comparisons <0.05
Observational data: associations, not effects.
Further findings
| Outcome | High-risk | Low risk |
|---|---|---|
| 2-year cumulative incidence of OHCA by ECG + EHR model risk group · accounting for the competing risk of non-OHCA mortality | 2.4% (95% CI 2.0%–2.8%) | 0.5% (95% CI 0.3%–0.8%) |
Incident OHCA flagged as high-risk by the EHR + ECG model, real-world cohort · 2-year period
- Of those with incident OHCA flagged as high-risk
- two-thirds (153 of 228)
Original publication
Artificial Intelligence-Enhanced Electrocardiography and Health Records to Predict Cardiac Arrest. JACC Adv. Published 11 May 2026. doi:10.
Summarised from the published abstract; every number is checked against it