Scripps Research develops AI to predict diabetes and kidney disease via ECG

Researchers at Scripps Research have developed a new artificial intelligence model called ECG-CLIP that can extract complex health information from standard electrocardiograms (ECG). While traditional ECG analysis focuses on identifying arrhythmias or signs of a heart attack, this AI model connects waveform patterns with clinical diagnoses to uncover deeper systemic insights.

According to a study published in ‘The Lancet Digital Health’, the model was trained on a massive dataset of over 1.7 million ECGs from 542,000 Scripps Health patients. Testing demonstrated that ECG-CLIP can reliably predict the likelihood of developing type 2 diabetes or chronic kidney disease within a three-year period, moving beyond simple cardiovascular risk assessment.