A heartbeat in your sleep: AI learns to read tomorrow's heart risk from tonight's ECG

A heart tracing taken while you sleep may already know something about your next ten years. In an NIH-supported study published in the journal Sleep, researchers trained a deep-learning model on single-lead ECGs recorded during sleep studies, combined with sleep-stage information, first on 15,809 patients at Massachusetts General Hospital and then tested it on 9,810 patients at Emory and 12,576 at Beth Israel Deaconess. The model estimated ten-year risk of atrial fibrillation, stroke, heart attack, heart failure and death — from data hospitals already collect every night. 'Using ECGs to predict which patients have a greater risk could help guide clinical decision-making,' said David Goff of the National Heart, Lung, and Blood Institute. It is prediction in service of prevention: finding the patients to help before the emergency, not after.

SDG 3 · Good health and well-being — Mechanism: tech innovation