DeepCOPD: Development and Implementation of Deep Learning to Predict and Prevent COPD Health Care Encounters
Chronic obstructive pulmonary disease (COPD) affects about 24 million people in the United States and is responsible for nearly 700,000 hospitalizations each year, costing approximately $50 billion per year. Since many of these hospitalizations are preventable, the Brown Lab is developing a clinical decision support system to help providers and patients better understand and anticipate their risk for outcomes that could lead to hospitalizations. DeepCOPD seeks to leverage electronic health record data and clinical notes to develop deep learning models for predicting emergent healthcare encounters (e.g., hospitalizations, emergency room visits) and COPD disease progression (e.g., exacerbations, progression to home oxygen use). This project combines biomedical informatics, advanced predictive analytics, and implementation science to develop, validate, and evaluate the DeepCOPD clinical decision support tool within Dartmouth Health and Vanderbilt University Medical Center.
Study Status: Active