Abstract: Healthcare costs are extremely high in the U.S, and only half the population accrues 97% of the costs. To improve care for diabetics, we are developing a model that predicts attributes correlated with high utilization. Using an integrated medical records database, we identified a cohort of diabetes patients that were high utilizers. These patients were likely to have pain, addiction, depression, retinopathy, and ketoacidosis. We hope these predictors can improve care and reduce costs.

Learning Objective 1: Develop an understanding on how to improve care in diabetic patients in order to reduce healthcare costs and improve health outcomes using electronic health records.

Learning Objective 3 (Optional): Understanding why a diabetic patient is becoming a high utilizer of health care and how to prevent it using predictive modeling


Anika Ghosh (Presenter)
Northwestern University

Presentation Materials: