Prevention & lipidsCohort study
Artificial Intelligence–Enabled Serial Electrocardiograms for Prediction of All-Cause Mortality in Secondary Care Settings
- retrospective cohort study
- deep-learning framework
- training, validation, and test sets
Authors' conclusion
A serial AI-ECG model provides accurate, clinically interpretable prediction of 1-year mortality and may support risk stratification in secondary care.
Study context
- Population
- Adults aged over 50 years treated in emergency department or inpatient settings between 2014 and 2019patients included, 4,372 (34%) died within 1 year
- Participants
- 13,417 adults
- Exposure
- AI-ECG model for predicting 1-year all-cause mortality; incremental value of serial ECGs
Main finding
As reported in the abstract
Discrimination of the serial AI-ECG model for 1-year mortality · 1 year
serial ECGs
0.84
area under the curve · 95% CI 0.824–0.862
Observational data: associations, not effects.
Further findings
| Outcome | single index ECG | serial ECGs |
|---|---|---|
| Performance at threshold of 0.24 | ||
| sensitivity | Not reported in the abstract | 70% |
| specificity | Not reported in the abstract | 80% |
| positive predictive value | Not reported in the abstract | 49% |
- Performance at threshold of 0.24: threshold selected on the validation set by maximizing the F1 score and applied to the test set
Calibration
- Brier score
- 0.123
- good agreement between predicted and observed risks
Decision-curve analysis
- greater net benefit than treat-all and treat-none strategies
Addition of clinical variables beyond age and sex
- did not improve performance
Original publication
Artificial Intelligence–Enabled Serial Electrocardiograms for Prediction of All-Cause Mortality in Secondary Care Settings. JACC Adv. Published 17 June 2026. doi:10.1016/j.jacadv.2026.102875
Summarised from the published abstract; every number is checked against it