Combining health records and heart tests predicts heart failure
In the UK Biobank, combining three risk tools scored 0.811 for distinguishing future heart failure, versus 0.788 for clinical risk factors alone.
Edited by Boris Djordjevic · AgingResearch.News
In an observational study, researchers evaluated 51,237 UK Biobank participants, averaging 64.7 years old (51.6% women), and 14,800 Mass General Brigham Biobank participants, averaging 55.7 years old (53.8% women). They tested three ways to predict future heart failure: a clinical risk score, a genetic risk score based on many inherited variants, and an artificial intelligence model analyzing electrocardiograms, which record the heart’s electrical signals.
The three-part model distinguished people who developed heart failure from those who did not better than any single approach. Its discrimination score was 0.811 in each group. Combining only the clinical score and electrocardiogram model performed nearly identically, scoring 0.811 in the UK group and 0.810 in the Mass General Brigham group. Adding the genetic score did not meaningfully improve prediction.
Why it matters
Heart failure is a progressive condition associated with illness and death. The researchers called for studies testing preventive interventions in people with high predicted risk, before clinically apparent disease develops.
Caveats
The observational study assessed prediction, not whether using these models to guide care improves patient outcomes. The findings were reported in an ESC Congress 2026 conference abstract.
The paper
Broad Institute of MIT and Harvard
ESC Congress 2026 Abstract, Oct 26, 2026
Presented at ESC Congress 2026
