Your ECG just got smarter…Narayana Health’s AI is India’s first to detect heart failure in a flash
Trained on over 100,000 ECG images paired with echocardiogram reports, the model has demonstrated strong predictive performance
Bengaluru: Narayana Health Clinical Research team and Medha AI - the Advanced Analytics & AI division of Narayana Health, on Wednesday announced the development of a breakthrough artificial intelligence (AI) model capable of predicting heart function—specifically, left ventricular ejection fraction (EF)—from a standard ECG image in near
real-time. Designed for seamless integration into clinical workflows, the model aims to support earlier identification of heart failure and to improve diagnostic access in resource-limited settings.
This innovation addresses a critical gap in the diagnostic pathway for millions of patients. Heart failure affects nearly 10 million people across India and accounts for 1.8 million hospital admissions annually. Yet access to echocardiography (ECHO)—the gold standard for evaluating heart function—remains uneven, especially in rural areas. Narayana Health’s AI model offers a novel, scalable approach to screen patients using widely available ECG machines and cloud-based analytics.
Trained on over 100,000 ECG images paired with echocardiogram reports, the model has demonstrated strong predictive performance.
In external validation across 14 tertiary centres involving more than 57,000 patients, it identified 97% of individuals with severely reduced EF (≤35%).
On average, these patients were flagged 58 days earlier than when echocardiography was eventually ordered—offering valuable lead time for clinical intervention.
