OBJECTIVETo identify factors that predict medication adherence and to examine relationships among adherence, glycemic control, and indices of insulin action in TODAY (Treatment Options for Type 2 Diabetes in Adolescents and Youth).RESEARCH DESIGN AND METHODSA total of 699 youth 10–17 years old with recent-onset type 2 diabetes and ≥80% adherence to metformin therapy for ≥8 weeks during a run-in period were randomized to receive one of three treatments. Participants took two study pills twice daily. Adherence was calculated by pill count from blister packs returned at visits. High adherence was defined as taking ≥80% of medication; low adherence was defined as taking <80% of medication. Depressive symptoms, insulin sensitivity (1/fasting insulin), insulinogenic index, and oral disposition index (oDI) were measured. Survival analysis examined the relationship between medication adherence and loss of glycemic control. Generalized linear mixed models analyzed trends in adherence over time.RESULTSIn this low socioeconomic cohort, high and low adherence did not differ by sex, age, family income, parental education, or treatment group. Adherence declined over time (72% high adherence at 2 months, 56% adherence at 48 months, P < 0.0001). A greater percentage of participants with low adherence had clinically significant depressive symptoms at baseline (18% vs. 12%, P = 0.0415). No adherence threshold predicted the loss of glycemic control. Longitudinally, participants with high adherence had significantly greater insulin sensitivity and oDI than those with low adherence.CONCLUSIONSIn the cohort, the presence of baseline clinically significant depressive symptoms was associated with subsequent lower adherence. Medication adherence was positively associated with insulin sensitivity and oDI, but, because of disease progression, adherence did not predict long-term treatment success.