What happened
UT Southwestern Medical Center researchers developed an AI algorithm to interpret electrocardiograms (ECGs), accurately screening nearly 6,000 patients in Kenya for left ventricular systolic dysfunction (LVSD), a precursor to heart failure. The AI-augmented ECG (AI-ECG) demonstrated a 99.1% negative predictive value and 95.6% sensitivity for LVSD detection in a subset of 1,444 patients verified by echocardiograms. These findings, published in JAMA Cardiology, establish AI-ECG as a low-cost, scalable method for early heart failure detection in resource-limited settings.
Why it matters
Access to early heart failure detection will expand significantly in underserved regions. For public health officials and medical device procurement teams, this AI-ECG method reduces diagnostic costs by replacing expensive echocardiograms with AI-interpreted standard ECGs, which are widely available. The 99.1% negative predictive value and 95.6% sensitivity prove its efficacy as a scalable screening tool, addressing a critical gap in global cardiovascular care where resources are constrained.
