OBJECTIVES: To compare overall and diabetes-related health care utilization and expenditures in Medicare beneficiaries with diabetes prior to and after implementation of Part D.  METHODS: Data were from the 2001-2010 Medical Expenditure Panel Survey.  Medicare beneficiaries aged ≥65 years with self-reported diabetes were identified, excluding individuals who were Medicare-Medicaid dual eligible, had TRICARE, or other public coverage. The outcomes included overall and diabetes-related prescriptions, medical services utilization, and expenditures as well as out-of-pocket costs. These outcomes were analyzed using generalized linear model regression models with a log-link and gamma (for costs) or Poisson (utilization) distribution.  All expenditures were inflated to 2010 dollars.  STATA survey commands were used to account for the complex survey design.  RESULTS: There were 21,864 eligible Medicare beneficiaries with diabetes that comprised the sample population.  After adjusting for sociodemographic characteristics and health status, implementation of Part D was associated with decreased out-of-pocket expenditures for insulin (Coeff.: -0.577, p<0.05) and oral antidiabetic agents (Coeff.:-1.292, p<0.001).  No statistically significant effect on total health care and prescription expenditures were found.  Implementation of Part D was associated with increased number of prescriptions filled (Coeff.: 0.145, p<0.001) without increase in emergency room visits or inpatient stays. CONCLUSIONS: The implementation of Medicare Part D increased medication use and reduced out-of-pocket costs for elderly diabetes patients.  This reduction would allow for enhanced access to necessary medications, and thereby may have a positive impact on adherence and health outcomes.