OBJECTIVES: Although the relationship between diabetes and risk of pancreatic cancer are well-documented, limited research has examined whether racial/ethnicity differences accounted for the association between diabetes and pancreatic cancer. The aims of this study were to 1) assess whether diabetes is associated with pancreatic cancer in the elderly Medicare population, and 2) identify if any racial/ethnicity disparities exist in the association between diabetes and pancreatic cancer. METHODS: In a retrospective cross-sectional study, community-dwelling and institutionalized Medicare beneficiaries aged 65 or over without health maintenance organization enrollment from the Medicare Current Beneficiary Survey (MCBS) from 2001 to 2010 were included. The outcome of the study was pancreatic cancer; and the key explanatory variables were diabetes and race/ethnicity. Potential confounders considered included age, gender, education, marital status, income, residence area, body mass index (BMI), smoking status, Charlson comorbidity index (CCI), and use of preventive services. The International Classification of Diseases-9 codes were used to identify the outcome and explanatory variables from the Medicare Part A and B claims. Logistic regressions were performed to estimate the association between diabetes and pancreatic cancer. RESULTS: A total of 98,966 person-years of elderly Medicare beneficiaries were included in the analysis. After controlling for potential confounders, the association between diabetes and pancreatic cancer was statistically significant (odds ratio [OR], 3.15; 95% confidence interval [CI], 2.07-4.77). Compared to non-Hispanic whites, non-Hispanic blacks (OR, 0.43; 95% CI, 0.15-1.20) and Hispanics (OR, 0.78; 95% CI, 0.33-1.85) had similar risks of pancreatic cancer. CONCLUSIONS: This study found that diabetes is associated with pancreatic cancer in the elderly Medicare population. However, racial/ethnicity differences do not explain the association between diabetes and pancreatic cancer.