OBJECTIVES: The objective of this research project was to evaluate the effect of health insurance plan type on anti-diabetic medication adherence as measured by medication possession ratios (MPR>0.80: good MPR) for oral medications.  METHODS: The data source for this research project was the 2000-2001 MarketScan dataset, which is comprised of administrative claims data for over 2.5 million privately insured individuals in the United States.  MPR was defined as the sum of days supply each anti-diabetic medication divided by the number of days between the first and last prescription fill dates plus the number of days for the last refill.  RESULTS: The odds of having a good MPR for oral anti-diabetic medications were evaluated using multiple logistic regression models controlled for insurance plan type plus demographic and clinical characteristics of the target patient population.  A propensity score analysis was done using a stratification approach to further control for selection bias.  The proportion of patients with MPR ≥80% was 68.2%.  CONCLUSIONS: Patients in capitated plans were more likely to have good MPR compared to FFS, PPO and POS plan types (odds ratios: 0.75, p<0.001; 0.81, p<0.001; 0.81, p<0.001, respectively).  This relationship remained in all quintiles of the propensity score analysis.