ARTIFICIAL intelligence (AI)-enhanced electrocardiography (ECG) combined with electronic health record (EHR) data successfully identified individuals at increased risk of out-of-hospital cardiac arrest (OHCA), according to a major new U.S. study.
The findings suggest that AI-driven cardiac screening could help clinicians detect high-risk patients before a life-threatening event occurs.
AI-Enhanced ECG Improves Cardiac Arrest Prediction
Out-of-hospital cardiac arrest remains one of the leading causes of sudden death worldwide, with many cases occurring in individuals without previously recognised high-risk cardiovascular disease. Despite advances in emergency response and defibrillation strategies, predicting who will experience cardiac arrest has remained extremely difficult in the general population.
Researchers developed and validated AI-enhanced ECG models using both ECG waveforms and routinely collected EHR data to determine whether these tools could better stratify OHCA risk. The case-control study matched participants according to age and sex before assessing model performance in temporal validation cohorts.
The multimodal AI-enhanced ECG and EHR model achieved the strongest predictive performance, with an area under the receiver operating characteristic curve of 0.83 and an area under the precision recall curve of 0.44. These results significantly outperformed models based solely on ECG or EHR data.
Real-World Performance of AI-Enhanced ECG Screening
To evaluate clinical relevance, investigators tested the AI-enhanced ECG approach in a real-world healthcare population undergoing ECG assessment. Over a 2-year follow-up period, the combined ECG and EHR model identified approximately two-thirds of individuals who later experienced incident OHCA.
Among patients classified as high-risk by the AI-enhanced ECG model, the 2-year cumulative incidence of OHCA reached 2.4% (95% CI: 2.0%–2.8%). By comparison, individuals designated low-risk demonstrated a substantially lower cumulative incidence of 0.5% (95% CI: 0.3%–0.8%).
These findings indicate that integrating AI-enhanced ECG analysis with clinical health records may provide a practical method for identifying patients who could benefit from closer cardiovascular monitoring or preventive interventions.
Implications for Preventive Cardiology
The researchers noted that current prevention strategies for sudden cardiac arrest primarily focus on patients with known structural heart disease or severely reduced cardiac function. However, many OHCA cases occur outside these traditional high-risk groups.
AI-enhanced ECG screening may therefore offer a scalable population-level approach for earlier cardiac arrest risk identification. Further prospective studies are needed to determine how these models should be integrated into routine clinical practice and whether earlier intervention could reduce mortality.
The study also highlights the growing role of AI in cardiology, particularly in improving prediction models using routinely available healthcare data.
Reference
Sharma S et al. Artificial intelligence-enhanced electrocardiography and health records to predict cardiac arrest. JACC Adv. 2026;DOI:10.1016/j.jacadv.2026.102787.
Featured image: bilanol on Adobe Stock
