To assess whether the programs of diabetologic education (PDE) are effective to improve metabolic control in diabetes mellitus, depending on their inclusion in an overall therapeutic program and on the setting where they are imparted (hospital versus primary care center, PCC), we evaluated the changes in the HbA1 values in 230 insulin-treated diabetic patients, after receiving a PDE in a PCC (group A; n = 88) or as hospital inpatients (group B; n = 72) or outpatients (group C; n = 50). The remaining 20 patients (group D) received the same PDE, but their treatment did not depend on our Unit. The initial mean HbA1 levels were significantly reduced 3, 6 and 9 months after the PDE in group A patients (10.6% vs 9.6%, p less than 0.05; vs 9.4%, p less than 0.02; vs 9.1%, p less than 0.01), in group B patients (10.0% vs 9.5%, p less than 0.05; vs 9.1%, p less than 0.05; vs 9.0%, p less than 0.05) and group C patients (10.1% vs 9.5%, p less than 0.05; vs 9.1%, p less than 0.02; vs 8.9, p less than 0.01); however, they did not significantly change in group D patients (10.3% vs 10.7%, vs 10.3%, and vs 10.0%). Mean HbA1 in months 3, 6 and 9 were not different between groups A, B and C, but they were lower (p less than 0.05) than in group D. Our results confirm that PDE and treatment are effective to improve metabolic control in diabetic patients, whereas diabetological education alone is not useful. The place where PDE is given is not determinant for its effectiveness.