To determine whether clinical outcomes in patients with type 2 diabetes were improved by protocol-driven care in a Diabetes Centre compared with usual outpatient care. Descriptive analysis of a prospective cohort. During a median 7-year observational period, 91 patients with type 2 diabetes and no cardiovascular or renal complications were monitored by a nurse and a diabetologist in a clinical trial setting according to a structured protocol. Another 81 patients with comparable clinical characteristics were monitored by generalists at the medical clinic in the same hospital. Clinical end points, defined as death and cardiovascular and renal events, were evaluated in 1997 by review of case records. Patients receiving structured care had lower mortality (relative risk RR = 0.21; 95% confidence interval CI = 0.07, 0.65; P = .006) than the usual-care group, as well as a lower incidence of combined clinical end points (RR = 0.43; 95% CI = 0.22, 0.84; P = .01). In the usual-care group, patients who had no monitoring of glycosylated hemoglobin or plasma lipid levels during the entire observational period (8.6%) had a 14.6-fold (P < .01) and 15.7-fold (P < .01) increased risk of death and combined clinical end points, respectively, compared with those who had at least one measurement (60.5%). Management by protocol-driven care model improved survival and clinical outcomes in patients with type 2 diabetes. Definitive studies are required to confirm these findings and compare the cost effectiveness of these care models.