The Singapore public healthcare system has increasingly used the term "right-siting" to describe the principle that stable chronic disease patients should be managed in primary care rather than specialist settings. The majority of primary healthcare providers in Singapore are general practitioners (GPs). The aims of this paper were to measure the quality of diabetes care in specialist and GP settings, and assess right-siting efforts in a tertiary centre in Singapore. Three hundred eighty-three consecutive patients with type 2 diabetes referred to the Singapore General Hospital Diabetes Centre (SGH DBC) between January and March 2005 were analysed. At the first visit, 51 patients (13.3%) were classified as inappropriate referrals and discharged back to the referral source or to primary care. After 12 months, 136 patients (group A = 35.5%) remained on follow-up at SGH DBC. In these patients, significant improvements were seen in mean HbA1c but not blood pressure (BP) or low density lipoprotein-cholesterol (LDL-C). One hundred twenty-eight (group B = 33.4%) patients were discharged from DBC within the 12 months of the study period. Mean follow-up duration in group B was 5.5 months and HbA1c, blood pressure and LDL-cholesterol had improved significantly in these patients. Glycaemic control of group B patients at the time of discharge was significantly better than group A at 12 months (mean HbA1c = 7.15% vs 8.16%; P <0.001). More than half (55.6%) of group B patients achieved HbA1c targets compared to 32.4% from group A (P <0.001). Although mean BP and LDL-C levels fell in group B patients, the percentage of patients achieving BP and LDL-C targets did not improve significantly in both groups. From August 2005 to January 2008, GPs participating in SingHealth's Delivering on Target (DOT) programme enrolled 579 patients under their care for additional diabetic counselling by community nurse educators. Pre- and post-programme HbA1c results were submitted for 370 patients (64%). Mean HbA1c levels of these patients decreased from 8.23% to 7.32% (P <0.001). The proportion of patients who achieved HbA1c <7% increased from 26% to 51% (P <0.01). However, BP and LDL-C levels did not improve. It is difficult to base referral or discharge decisions solely on these indicators. Our studies show that both in the specialist and GP settings, significant improvements in HbA1c are seen. Results for BP and LDL-C, however, showed little improvement. Some degree of rightsiting was seen at SGH DBC with discharged patients showing greater improvements than patients who were retained. However, >30% of patients remained in SGH DBC despite achieving HbA1C targets. Our results indicate the need for better strategies to address the underlying obstacles to right-siting. Of greater concern, the lack of improvement in BP and LDL-C indicates a high degree of clinical inertia to these issues among specialists and GPs treating diabetes in Singapore.