WVU Researchers Create AI Model for Heart Disease in Rural Areas

West Virginia University (WVU) researchers are developing artificial intelligence (AI) models designed to enhance the diagnosis and prediction of heart disease specifically for rural populations. This initiative is particularly significant as many existing healthcare AI systems tend to rely heavily on data from urban environments, which may not accurately reflect the needs of rural patients.

Prashnna Gyawali, an assistant professor in the Benjamin M. Statler College of Engineering and Mineral Resources, highlighted a critical gap in current AI applications in healthcare. Most AI models have been trained on data from urban settings, where patients often have different biological and socioeconomic characteristics compared to those in rural areas. This discrepancy can hinder the effectiveness of AI in diagnosing health conditions in less populated regions.

To address this issue, Gyawali and his team have begun training a new AI model using exclusively rural patient data collected from various locations across West Virginia. He emphasized, “You have to ensure your algorithms have seen the populations where you want them applied.” This focus aims to create models that accurately reflect the unique characteristics of rural health populations, ensuring the AI can be effectively utilized for diagnosing heart disease and other conditions.

The team has been gathering anonymous patient datasets from multiple regions in West Virginia to evaluate how different AI models perform in diagnosing heart disease based on test results. Gyawali believes that if implemented correctly, AI could significantly relieve the burden on healthcare professionals in rural areas. He noted, “Health care problems are growing, and we have manpower shortages.” A patient in West Virginia may need to travel several hours just to receive an initial diagnosis, making the potential for AI-assisted diagnosis even more crucial.

Gyawali envisions a future where clinics equipped with affordable scanning devices and AI systems can facilitate early detection of ailments like heart disease. “If we have more clinics with inexpensive scanning devices with an AI system attached, we can have an early detection system flagging certain patients,” he stated. However, he cautioned that achieving this vision requires rigorous training and validation of AI models to ensure they are reliable and unbiased.

While the team is optimistic about the early results from their testing, Gyawali stressed that the AI models have only interacted with historical rural datasets to this point and have yet to assist with real-world patient diagnoses. “Whenever we talk about safety-critical applications like health care, we need to make sure they’re reliable,” he said. The priority is to refine the model until both medical and computer science professionals are confident in its safety and reliability for human use.

The ongoing work involves adding layers of complexity to enhance the model’s performance. Gyawali is also exploring the possibility of validating these algorithms in clinics outside of their initial study to determine if the AI’s effectiveness extends beyond West Virginia. “Can we find a clinic that’s not been involved with this study and test it out on their dataset?” he questioned, indicating a desire to expand the research.

In addition, Gyawali pointed to the necessity of policy-level interventions that could facilitate real-world clinical trials for these algorithms. “That’s the roadmap toward adopting these tools in clinics,” he remarked, underscoring the importance of both technological development and supportive policy frameworks in advancing rural healthcare.

As this project continues, Gyawali and his team remain committed to ensuring their AI model is as reliable as possible, with an emphasis on its potential to transform the future of healthcare access in rural communities.