Hilltop’s Analytics & Research team—Leigh Goetschius, PhD; Danielle Barefoot, MEd; Fei Han, PhD; Ruichen Sun, MS; and Morgan Henderson, PhD—co-authored a recently released article in the American Journal of Managed Care titled Predicting Severe Diabetes Complications Using Administrative Claims Data in Maryland.
The researchers document the operations and performance of Hilltop’s Pre-DC model in the Maryland Primary Care Program (MDPCP) population. This is a large-scale predictive model developed by Hilltop in 2022 and designed to identify beneficiaries at highest risk for severe type 2 diabetes complications. Using more than 12 million person‑months of data, the model performed very strongly, with the top 10% of predicted‑risk individuals accounting for nearly 57% of severe events in the following month. This model is one of several predictive models that Hilltop developed, maintains, and uses to create predictive risk scores; moreover, while this study focuses on the MDPCP population, the Pre-DC model is currently live for the entire Medicare fee-for-service and Medicaid populations (covering approximately 2 million Marylanders).
These findings are an example of how Hilltop’s advanced analytics can support proactive, data‑driven care for Maryland’s publicly insured populations.
|
