What happened
Technion Faculty of Biomedical Engineering researchers, led by Prof. Joachim Behar and Eran Zvuloni, developed DeepHHF, an AI model predicting heart failure risk up to five years before clinical onset. Published in npj Digital Medicine, DeepHHF analyses standard 24-hour Holter ECG recordings, identifying subtle abnormalities imperceptible to human clinicians. The model was trained on approximately 70,000 Holter examinations from Leumit Health Services.
Why it matters
Early heart failure detection enables preventive interventions, reducing hospitalisations and mortality for the 64 million people globally affected by the condition. For health insurers and healthcare providers, this five-year predictive window from routine, non-invasive ECGs offers a cost-effective mechanism to target high-risk patients. This follows other AI-ECG research, like UT Southwestern's work on heart failure precursor screening, underscoring a shift towards proactive cardiovascular health management.
