Using RaaS to predict hospital readmission rates

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.

Date:
Speakers:
Ankur Teredesai, David Hazel, Florence Chang, George Wu, Senjuti Basu Roy, and Tony Kim
Affiliation:
MultiCare Health System, Clinical AVP Edifecs, University of Washington Tacoma,