To determine the feasibility and efficacy of structured education in intensive insulin therapy (IIT) in patients with Type 1 diabetes mellitus commonly attended by primary care physicians, a prospective case-control study was carried out in co-operation with 26 general practitioners in rural Alpine region and the diabetes service at the University of Vienna, Medical School, Austria. From 89 rural Type 1 diabetic patients on conventional insulin therapy (CIT), those volunteering for better diabetes care (n = 52) were trained in IIT in the diabetes education centre and subsequently received their outpatient service by their general practitioners, as did those remaining on CIT (n = 37). Patients were matched as case-controls (n = 36 in each therapy group) for metabolic control at baseline (IIT/CIT: HbA1c 8.2 +/- 1.8 vs 8.1 +/- 2.0%, ns), age, duration of diabetes, incidence of retinopathy and nephropathy. Analysing an observation period of > 4.5 years, patients trained in IIT presented with improved metabolic control as compared to those on CIT (Mean HbA1c: IIT, 6.9 +/- 1.0%; CIT, 7.9 +/- 1.3%, p < 0.05, ANOVA). No difference between groups was, however, observed at the end-point of the study in HbA1c (IIT, 7.3 +/- 1.3%; CIT, 7.8 +/- 1.4%; IIT vs CIT, p = 0.14) and in the development of diabetic microangiopathy, frequency of reported severe hypoglycaemic episodes, and increase in body weight.(ABSTRACT TRUNCATED AT 250 WORDS)