OBJECTIVES: Outcomes in patients with type 2 diabetes may vary depending on the antidiabetic medication used. This study compared differences in total health care costs, medication adherence, and persistence in patients with type 2 diabetes enrolled in the North Carolina Medicaid Program that were newly started on thiazolidinedione (TZD) therapy with patients starting other oral antidiabetics (OAD) during the same period. METHODS: A total of 2660 patients newly starting TZD therapy between July 2001 and December 2003 were compared to 2050 patients starting other oral antidiabetic medication for health care costs and outcomes in the post-medication start year. Multivariate regression techniques incorporating health care utilization in the year prior to start of new therapy were utilized to determine the net cost impact of one therapy versus the other. RESULTS: The mean age of the TZD group was significantly higher as compared to OAD group (p<0.05). Total number of prescription refills was significantly higher with TZD than OAD (p<0.05). Other baseline characteristics were similar across the two groups. The TZD cohort had significantly higher persistence rates as well as adherence rates compared to OAD in the 2nd year (p<0.01). Multiple regression analyses found that patients starting TZD have better treatment persistence in the post-medication start year compared to patients starting other oral antidiabetic agents. (4 % increase in therapy persistence index, P<0.001). In addition, patients starting TZDs had 19% lower total annual health care costs (P<0.01) compared to patients starting other oral antidiabetic agents. CONCLUSIONS: Introduction of thiazolidinedione therapy in a Medicaid-enrolled type 2 diabetic population was associated with significantly greater treatment persistence and lower healthcare costs in the post-start year compared to patients starting other oral antidiabetic agents.