Aims  To describe the IQED, a quality‐assurance system started in 2001 in Belgian hospital‐based multidisciplinary diabetes centres, and its effects on the quality of care. Methods  The study was conducted through four data collections (in 2001, 2002, 2004 and 2006). Approximately 120 diabetes centres provided data on a systematic random sample of 10% of their adult diabetic patients on at least two daily insulin injections. Data on patient characteristics, glycaemic control, cardiovascular risk, diabetes complications, follow‐up procedures and treatment were obtained. Local quality promotion was encouraged by returning comprehensive feedback (benchmarks) and during information meetings. Results  Nearly all diabetes centres (98–100%) participated. The pooled sample consisted of 9194 (32%) Type 1 and 19 828 (68%) Type 2 diabetes patients, with mean diabetes duration of 17 years and 14 years, prevalence of microvascular complications of 23% and 38% and prevalence of macrovascular complications of 9% and 26%, respectively. At the start, the quality of care was good in terms of risk‐factor testing rates and moderate in terms of patients meeting goals for risk‐factor management. At least 50% of the centres initiated quality‐promoting initiatives. After 5 years, significant improvements were seen in risk‐factor testing rates, apart from renal screening. Improvements in intermediate outcomes were less obvious, apart from an increase in patients reaching the targets for blood pressure and LDL cholesterol. Conclusions  It is feasible to implement a continuous quality‐improvement project on a nationwide scale, with improvements particularly in process indicators.