Preventable hospitalizations are common among Medicaid-enrolled adults with type 2 diabetes, yet health plans lack evidence on what reduces them.

Type 2 diabetes affects millions of Medicaid enrollees and drives substantial health care costs, particularly through preventable hospitalizations for complications like diabetic ketoacidosis, severe hypoglycemia, and infections. While clinical guidelines outline standards of care, health plans have limited evidence about which services—from medication to behavioral supports—reduce these costly acute events in real-world Medicaid populations. Additionally, Medicaid enrollees receive care through either managed care organizations or traditional fee-for-service arrangements, yet little research compares outcomes across the two.

The Blue Cross Blue Shield Association (BCBSA) needed an analysis of national Medicaid data to identify which service combinations are associated with fewer hospitalizations and how outcomes vary across delivery systems.

NORC analyzed national Medicaid claims to identify which services were associated with fewer preventable hospitalizations and how delivery systems compared.

NORC conducted analyses of Transformed Medicaid Statistical Information System (T-MSIS) data from calendar years 2021-2023 for all states and the District of Columbia. The study examined adult Medicaid enrollees aged 18-64 with type 2 diabetes, comparing service use and outcomes across managed care and fee-for-service delivery systems.

The methodological challenge was identifying which services are associated with better outcomes while accounting for factors that influence both service access and those outcomes. NORC grouped services into categories aligned with the American Diabetes Association’s Standards of Care, including blood glucose, foot, and eye screenings; traditional antidiabetics; GLP-1s; behavioral and lifestyle supports such as diabetes self-management education; and specialist visits. The team tracked these services over 12 months in a cohort of Medicaid enrollees with diabetes.

We then measured outcomes over the following six months, including all-cause inpatient admissions, emergency department visits, and diabetes-related potentially avoidable inpatient admissions (Prevention Quality Indicators, or PQIs). By measuring services first and then measuring outcomes, we could assess whether earlier services related to later health events. In a subsequent six-month period, to determine if prior utilization of different services and treatments influenced the outcomes of interest. To ensure reliable findings across state Medicaid programs, NORC applied risk adjustment to control for age, demographics, geography, comorbidities, and delivery system type.