Innovative AI Models Use Low-Tech ECGs to Detect Heart Failure in Rural Populations

Researchers at West Virginia University have developed advanced artificial intelligence (AI) models designed to diagnose heart failure among rural residents by analyzing simple electrocardiogram (ECG) data. This initiative addresses a critical gap in healthcare, where traditional diagnostic tools like echocardiograms are costly and often unavailable in underserved areas. The AI models are trained on data from over 55,000 West Virginia patients, focusing on the electrical signals captured by basic ECG devices, which are widely accessible and easy to operate.

Heart failure, a chronic condition where the heart cannot pump blood effectively, is a significant health concern, especially in rural regions where access to specialized care is limited. Dr. Prashnna Gyawali emphasized that existing AI diagnostic tools are primarily based on urban datasets, which may not accurately reflect the socioeconomic and environmental factors influencing rural populations.

To bridge this gap, the scientists trained machine learning models, particularly deep learning frameworks like ResNet, to interpret electrocardiograms and estimate ejection fraction—the measurement of the heart's pumping efficiency. Unlike echocardiography, which uses sound waves to produce detailed images and requires expensive equipment, ECGs are inexpensive and straightforward, making them ideal for resource-limited settings.

The study found that the deep learning models, especially when trained with specific electrode data, performed accurately in predicting ejection fractions. These findings could soon help clinicians identify heart failure more quickly and accurately in rural patients, enabling earlier intervention.

While not yet in widespread clinical use due to ongoing reliability validations, this research marks a promising step toward improving cardiac care in underserved communities. As heart failure affects millions, with a higher prevalence in rural America, advancing diagnostic methods like these could significantly impact public health outcomes.

This breakthrough was published in Scientific Reports, with contributions from doctoral student Alina Devkota and other WVU researchers. The study underscores the potential of combining accessible diagnostic tools with AI technology to reduce healthcare disparities.

Source: https://medicalxpress.com/news/2025-08-ai-heart-failure-rural-patients.html