A randomized controlled trial was conducted to compare three forms of diabetes follow-up: (1) general practitioner care, (2) a system of care shared between the general practitioner (GP) and clinic and (3) conventional clinic care. Two hundred and six diabetic patients without significant diabetes-related or other medical complications were randomized to one of these follow-up systems. Metabolic control and blood pressure improved significantly and equally in all three groups (p < 0.05). The shared care group performed as well as or better than either of the other two groups in all other outcome measures. In particular, final attendance rates were 72% for shared care compared with only 35% for GP care and 53% for clinic care. Data collection rates for shared care were comparable with the clinic group for random blood glucose (88.9% vs 95.1%), weight (93.5% vs 98.3%), and blood pressure (94.8% vs 92.7%). Only in the case of glycosylated haemoglobin did shared care have poorer data collection (66.0% vs 98.4%). In all these parameters, except blood pressure, shared care out-performed the GP group. We conclude that with adequate support from and communication with hospital-based diabetes services, GPs are capable of providing care appropriate to the needs of uncomplicated diabetic patients.