The effects of a new integrated system of diabetes care with an enhanced role of the diabetes specialist nurse based in a purposed design diabetes centre, on diabetes control, attendance and cancellation rates, and admission for diabetic emergencies have been reviewed. Glycaemic control was examined in: (a) a cohort of 163 insulin-treated and 47 non-insulin treated diabetic subjects (age < 65 years) studied prospectively before and 3 years following the introduction of a new system of care; (b) a second cohort of more elderly patients aged greater than 65 years studied for the 3 years after the change over; (c) a cross-sectional study of the clinic population (n = 700) the year before and 3 years after the changeover; (d) a group of patients attending standard unaltered clinics in the same district (n = 157). Significant and sustained falls in HbA1 were observed in all groups of subjects attending the centre, with the means for those aged less than 65 falling from 11.9 +/- 2.3% to 9.9 +/- 1.9% and for those aged over 65 from a mean of 11.7 +/- 2.0% to 10.3 +/- 2.3%, 3 years later. The cross-sectional study provided similar results with a mean HbA1 of 12.2 +/- 3.0% prior to changeover and 10.4 +/- 4.4%, 3 years later. Smaller but significant changes were observed in patients continuing to attend the routine clinic (from 12.2 +/- 2.3% to 11.3 +/- 2.6%) over a similar period. Yearly admission rates for ketoacidosis and hypoglycaemia fell from 44 and 23, to 33 and 5 per annum, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)