UVA Health researchers have developed a new risk assessment tool called CARNA that uses machine learning and artificial intelligence to predict outcomes in heart failure patients. This tool is made available for free to clinicians and aims to improve care for the growing number of Americans living with heart failure by identifying individual patient risks for adverse outcomes. The researchers used anonymized data from heart failure clinical trials funded by the National Institutes of Health to develop the model, which outperformed existing predictors for determining the need for heart surgery or transplant, rehospitalization risk, and risk of death. The incorporation of hemodynamic clinical data and the ability to make decisions with missing or conflicting factors sets this model apart from others, enabling clinicians to personalize care and make treatment decisions more efficiently.
Heart failure is a progressive condition that can lead to fatigue, weakness, and ultimately death, making it crucial for clinicians to identify patients at risk of adverse outcomes. With over 6 million Americans currently living with heart failure and an expected increase to over 8 million by 2030, developing new tools like CARNA to improve care and outcomes for these patients is essential. The innovative use of AI and machine learning, along with collaboration between experts in heart failure, computer science, data science, and statistics, allowed the researchers to create a model that can help doctors personalize care for individual patients, potentially leading to longer, healthier lives.
The CARNA model is a significant advancement in risk assessment for heart failure patients, as its ability to analyze complex sets of data and intelligently present risk factors reduces decision burden for clinicians. By incorporating hemodynamic data, which describes how blood circulates through the heart, lungs, and body, this tool provides a more comprehensive view of patient risk and can assist doctors in making informed treatment decisions. The researchers hope that by using this model, clinicians will be better equipped to personalize care and improve outcomes for heart failure patients, ultimately benefiting their overall health and well-being.
The collaborative research environment at the University of Virginia played a critical role in the development of the CARNA model, bringing together experts in various fields to combine their knowledge and skills. Multidisciplinary biomedical research that integrates computer science with clinical medicine will be essential in leveraging AI and machine learning to benefit patients in the future. The researchers have published their findings in the American Heart Journal and have made the CARNA tool available online for free, with no financial interest in the work. By sharing their results and making the tool accessible to clinicians, the researchers aim to support advancements in heart failure care and research while contributing to the growing field of AI applications in healthcare.
The research project was supported by grants from the National Science Foundation and the NHLBI, based on a winning submission to the National Heart, Lung and Blood Institute’s Big Data Analysis Challenge. The researchers involved in the project, including Josephine Lamp, Yuxin Wu, Kenneth Bilchick, and others, worked together to develop and evaluate the CARNA model, highlighting the importance of collaboration in advancing medical research. By sharing their work and making the tool available for free, the researchers hope to contribute to the improvement of care for heart failure patients and inspire further exploration of AI applications in healthcare. For the latest updates on medical research from UVA, readers can subscribe to the Making of Medicine blog.
