Using RaaS to predict hospital readmission rates

Опубликовано 4 августа 2016, 19:58

What can happen when a team of machine learning researchers collaborate with cardiologists and clinicians to take advantage of cloud computing? It may actually transform healthcare. Medical institutions in the United States face increasing pressure to reduce readmission rates of patients with chronic conditions because of the financial penalties associated with these readmissions. A team of researchers from Center for Data Science at the University of Washington Tacoma partnered with Multicare Health System, with support from Edifecs, to build a Microsoft Azure-based toolset to predict risk of hospital readmission (RaaS). The RaaS service blends clinical and claims data including hundreds of attributes such as demographics, lab tests, vitals, comorbidities, and charges. This data is used to predict clinical risks for readmission of patients with congestive heart failure, to report the patients most at risk for readmission, and also to suggest meaningful insights and explanations behind each prediction.