Researchers at the University of Virginia have created a new risk assessment tool that can be used to predict outcomes when patients have heart failure. The researchers have made the tool publicly available for free to clinicians. The study has been published in the American Heart Journal.
Heart failure occurs when the heart is unable to pump enough blood. Symptoms can include fatigue, weakness, swollen legs and feet and, ultimately, death. Image Credit: University of Virginia
The new tool enhances current heart failure risk assessment tools by utilizing machine learning (ML) and artificial intelligence (AI) to identify patient-specific risks of adverse heart failure outcomes.
Heart failure is a progressive condition that affects not only quality of life but quantity as well. All heart failure patients are not the same. Each patient is on a spectrum along the continuum of risk of suffering adverse outcomes. Identifying the degree of risk for each patient promises to help clinicians tailor therapies to improve outcomes.
Sula Mazimba MD, Researcher and Heart Failure Expert, University of Virginia
About Heart Failure
Heart failure is a serious condition where the heart is unable to pump enough blood to meet the body’s needs, leading to symptoms such as fatigue, weakness, and swelling of the legs and feet, and it can ultimately be fatal. As heart failure progresses, it is crucial for clinicians to identify patients at risk of adverse outcomes.
Heart failure is an escalating issue in the US, with over 6 million individuals currently affected, and projections suggest this number will rise to more than 8 million by 2030. To enhance patient care, researchers at the University of Virginia developed a new model named CARNA. This initiative aligns with UVA Health’s first-ever 10-year strategic plan, which aims to improve patient care across Virginia and beyond.
The CARNA model was crafted using anonymized data from thousands of patients who participated in heart failure clinical trials previously supported by the National Institutes of Health’s National Heart, Lung, and Blood Institute.
In testing, CARNA surpassed existing prediction models in assessing a broad range of patient outcomes, including the necessity for heart surgery or transplantation, the risk of rehospitalization, and mortality rates.
The success of the CARNA model is attributed to its utilization of machine learning/artificial intelligence and the integration of hemodynamic clinical data, which detail the circulation of blood through the heart, lungs, and the rest of the body.
This model presents a breakthrough because it ingests complex sets of data and can make decisions even among missing and conflicting factors. It is really exciting because the model intelligently presents and summarizes risk factors reducing decision burden so clinicians can quickly make treatment decisions.
Josephine Lamp, Researcher, Department of Computer Science, University of Virginia School of Engineering
The researchers hope that by using the model, physicians will be better able to tailor care to specific patients, enabling them to live longer, healthier lives.
The collaborative research environment at the University of Virginia made this work possible by bringing together experts in heart failure, computer science, data science, and statistics. Multidisciplinary biomedical research that integrates talented computer scientists like Josephine Lamp with experts in clinical medicine will be critical to helping our patients benefit from AI in the coming years and decades.
Kenneth Bilchick MD, Cardiologist and Researcher, University of Virginia Health
Journal Reference:
Lamp, J. B. S., et al. (2024) Characterizing advanced heart failure risk and hemodynamic phenotypes using interpretable machine learning. American Heart Journal. doi.org/10.1016/j.ahj.2024.02.001