BACKGROUND: Although specialty care has been shown to improve short-term outcomes in patients hospitalised with acute medical conditions, its effect on patients with chronic conditions treated in the ambulatory care setting is less clear. OBJECTIVE: To examine whether specialty care (ie, consultative care provided by an endocrinologist or a general internist in concert with a patient's primary care doctor) within the first year of diagnosis is associated with improved outcomes after the first year for adults with diabetes mellitus treated as outpatients. DESIGN: Population-based cohort study using linked administrative data. SETTING: The province of Saskatchewan, Canada. Sample: 24 232 adults newly diagnosed with diabetes mellitus between 1991 and 2001. Method: The primary outcome was all-cause mortality. Analyses used multivariate Cox proportional hazards models with time-dependent covariates, propensity scores and case mix variables (demographic, disease severity and comorbidities). In addition, restriction analyses examined the effect of specialist care in low-risk subgroups. RESULTS: The median age of patients was 61 years, and over a mean follow-up of 4.9 years 2932 (12%) died. Patients receiving specialty care were younger, had a greater burden of comorbidities, and visited doctors more often before and after their diabetes diagnosis (all p< or =0.001). Compared with patients seen by primary care doctors alone, patients seen by specialists and primary care doctors were more likely to receive recommended treatments (all p< or =0.001), but were more likely to die (13.1% v 11.7%, adjusted hazard ratio (HR) 1.17, 95% confidence interval (CI) 1.08 to 1.27). This association persisted even in patients without comorbidities or target organ damage (adjusted HR 1.16, 95% CI 1.01 to 1.34). CONCLUSION: Specialty care was associated with better disease-specific process measures but not improved survival in adults with diabetes cared for in ambulatory care settings.