Heart HealthAI Uses Overnight Sleep Study Heart Recordings to Flag Future Cardiovascular Risk
Heart recordings collected during sleep studies may reveal more than sleep problems. Researchers trained an artificial intelligence model using overnight heart recordings and sleep stages from 15,809 patients, then tested it in two independent hospital groups totaling 22,386 patients. Higher model scores were linked to greater risks of irregular heart rhythm, stroke, heart failure, and death, even after accounting for common cardiovascular risk factors and sleep measurements. The strongest associations involved atrial fibrillation and heart failure. Results for heart attacks were less consistent across hospitals. The findings suggest that existing sleep study data could eventually help guide cardiovascular monitoring, but they do not show that using this model improves patient outcomes. This summary is based on the abstract only, which does not report several measures needed to evaluate clinical usefulness.
