AI-Powered ECG Technology Shows Promise in Early Heart Failure Detection in Kenya

An innovative artificial intelligence (AI) electrocardiogram (ECG) algorithm has demonstrated remarkable effectiveness in the early identification of heart failure among patients seeking healthcare in Kenya, according to recent research presented at the Heart Failure 2025 conference.

Heart failure is a significant health concern in Sub-Saharan Africa, where the condition is notably prevalent among younger populations, often leading to poorer health outcomes compared to those observed in high-income countries. Dr. Ambarish Pandey, a researcher from the University of Texas Southwestern Medical Center, emphasized the importance of early detection of left ventricular systolic dysfunction (LVSD) as a critical strategy for identifying individuals at risk of developing heart failure. He noted that access to echocardiography, the conventional method for diagnosing LVSD, is limited in resource-constrained environments.

In light of this challenge, a study was conducted in Kenya to explore the feasibility of assessing LVSD through ECG using validated AI software, presenting a potentially scalable solution for screening larger populations.

The study involved a prospective cross-sectional multicenter screening of adult patients attending eight healthcare facilities throughout Kenya. Researchers assessed cardiovascular risk factors by classifying individuals into high-risk groups based on previous cardiovascular disease (CVD) or a Framingham Risk Score (FRS) exceeding 10%.

Every participant underwent a 12-lead ECG, and the AI-ECG algorithm (AiTiALVSD, developed by Medical AI Co, Seoul, South Korea) evaluated the probability of LVSD based on a pre-established risk threshold. A subset of participants also received echocardiography assessments to compare the performance of the AI-ECG model.

The evaluable cohort comprised 5,992 participants, with an average age of 55 years, of which 66% were female. Remarkably, 65% of the participants were classified as having a high cardiovascular risk. The AI-ECG algorithm identified a prevalence of LVSD at 18.3%, with higher rates observed among individuals with elevated Framingham risk scores (22.9%) or prior CVD (32.0%), compared to those with low FRS (9.9%).

Among 1,444 participants who received both AI-ECG and echocardiography assessments, LVSD was confirmed in 14.1% of cases. The AI-ECG algorithm exhibited impressive performance metrics in comparison to echocardiography, achieving a sensitivity of 95.6%, a specificity of 79.4%, and a negative predictive value of 99.1%.

Dr. Bernard Samia, the senior author and President of the Kenya Cardiac Society, remarked on the potential of AI-ECG algorithms as a cost-effective and scalable screening tool for heart disease, particularly heart failure, in at-risk populations within resource-limited settings.

The findings highlight the significant number of individuals identified with LVSD, prompting further investigations. Dr. Pandey expressed a desire to conduct larger, more extensive screening studies across multiple African countries to ascertain whether the identification of LVSD leads to an increased adoption of evidence-based treatments.