Description
Bring-Your-Own-Device mHealth data provides opportunities for the application of machine learning (ML) methods to determine what features may be the most predictive of outcomes for patients with inflammatory bowel diseases (IBDs). We evaluated multiple ML models and found that among the numerous mHealth features, the amount of moderate-to-vigorous physical activity performed was the most predictive of IBD disease activity. Step count was only predictive of disease activity for patients with Crohn’s disease.
Learning Objective: To understand the performance of various machine learning models when applied to a bring-your-own-device mHealth data to predict patient outcomes.
Authors:
Tim Coleman (Presenter)
University of Pittsburgh
Lucas Mentch, University of Pittsburgh
Kimberly Glass, Brigham & Women's Hospital
David Gotz, University of North Carolina at Chapel Hill
Nils Gehlenborg, Harvard Medical School
Arlene Chung, University of North Carolina at Chapel Hill
Presentation Materials: