Best Practices in Technology and Analytics

Optum Predicts Risk of Atrial Fibrillation Using Artificial Intelligence

An estimated 3-6 million people in the United States have (AFib), an irregular heartbeat that increases the risk of stroke and heart disease. Due to the aging population, the numbers are expected to increase and many patients are unaware of their condition. If AFib is detected early on, the risk of stroke and heart disease is reduced by >75% with proper care and medications. The department of Data Science and Transformation of Optum Enterprise Analytics is exploring the ability of machine learning and artificial intelligence to identify patients who might have an abnormal heart rhythm. In a recent study, Optum trained a tree-based model using a de-identified dataset and identified patterns of data that are associated with the presence of ICD coding for atrial fibrillation. If patients are determined to be at risk for AFib, they can be referred for further testing and participation in anticoagulant drug therapy. Identifying AFib and preventing potential heart disease and stroke presents health organizations with opportunities for medical cost savings. The end goal of Optum’s research is to improve the health of AFib patients, decrease medical spending, and possibly detect other chronic diseases and complications through machine learning, artificial intelligence, and predictive analytics.

This best practice was discussed as part of eHealth Initiative’s July 2018 Technology & Analytics Workgroup.