OBJECTIVE: To investigate the longitudinal relationship between psychiatric diagnostic variables and metabolic control among youths with IDDM. RESEARCH DESIGN AND METHODS: A group of 88 youths, 8 to 13 years old at onset of 1DDM, were evaluated repeatedly during a 9-year follow-up period, on average, using a standardized psychiatric protocol. Levels of HbA1 were also assessed repeatedly. Psychiatric diagnoses were derived independently of HbA1 values. RESULTS: In univariate longitudinal analyses, the psychiatric diagnosis of noncompliance with medical treatment was significantly related to HbA1 level. There was a trend of an association between any major psychiatric disorder, as well as nondepressive disorder, and HbA1 Interaction terms between IDDM duration (or age) and psychiatric variables were also significantly related to metabolic control. According to the final multivariate model of repeatedly assessed HbA1, noncompliance with medical treatment (irrespective of IDDM duration) and the interaction between nondepressive psychiatric disorder and IDDM duration contributed to worse metabolic control. CONCLUSIONS: We found some support for the hypothesis that psychiatric morbidity negatively affects blood glucose regulation and that its consequences are more marked the longer young patients have had IDDM. We did not confirm the hypothesis that depressive illness has particularly deleterious consequences on metabolic control. Noncompliance with medical treatment and having had nondepressive psychiatric illness in interaction with IDDM duration account for a statistically significant but clinically modest amount of variability in HbA1 over time. The weak relationship among these variables may explain the inconsistent findings in the literature regarding psychiatric morbidity and metabolic control.