Abstract

Diabetes screening in urban Emergency Departments (EDs) can identify previously undiagnosed diabetes cases and link newly diagnosed patients to outpatient care. NYU Langone has recently implemented a system-wide initiative to screen ED patients for diabetes. However, only 23% of ED patients with newly diagnosed diabetes have been to a follow-up outpatient visit at NYU Langone. Thus, we wish to develop a machine learning-based predictive model, predicting the likelihood that an ED patient will follow up with outpatient care after a new diagnosis of diabetes. This model will inform a targeted, telemedicine-based intervention to help bridge these gaps in care.