Indigenous Australians have high rates of diabetes and its complications, including blindness[1], kidney disease[2], and cardiovascular mortality[3]. The reasons for this racial disparity are multifactorial, involving a complex interplay of behavioural, biological, economic and societal factors that vary across health indicators and healthcare settings. One confounding factor is the quality of diabetes care delivered in populations that are culturally and socio-economically disadvantaged [4, 5]. Indeed, the fact that minority patients, such as Indigenous Australians, are more likely to receive care at lower-performing facilities in remote and regional sites has biased many comparative studies. Moreover, when controlled for in managed-care settings, ethnicity is not consistently associated with worse processes or outcomes in individuals with type 2 diabetes [4].

The National Evaluation of the Frequency of Renal Impairment cO-existing with NIDDM (NEFRON) study was an large national survey of patients with type 2 diabetes in the Australian primary care setting [6]. The principal aim of this clinic-based study was to determine the frequency of any consultation between any patient with type 2 diabetes and their general practitioner to be complicated by chronic kidney disease, and the factors that influence it's management. In this paper, we examine the management of patients identified by their GP as having an Indigenous ethnic background, and contrast it with the care of non-Indigenous patients consecutively presenting to the same practice, and over the total NEFRON cohort.

Despite advances in diabetes care, the clinical outcomes of Indigenous Australians with type 2 diabetes in general practice remain poor. While ethnic differences in the quality of diabetes care partly contributes to disparity in diabetic complications, even when compared to non-Indigenous patients consecutively attending the same practitioner and receiving a comparable pattern of care, Aboriginal patients have worse glycemic control, and more micro- and macrovascular complications. These findings are consistent with smaller community-based studies[11] and collected regional data from remote and rural settings [12–15], where HbA_1c levels are typically 1% higher in Indigenous patients, and high rates of complications are observed. However, these data represent the first national survey including large numbers of Indigenous patients and adjusting for centre effects, to characterise this disparity.

The National Evaluation of the Frequency of Renal Impairment cO-existing with NIDDM (NEFRON) study was an clinic-based, cluster-stratified survey of patients with type 2 diabetes from across the Australian primary care setting. It was not designed to be representative of all patients with type 2 diabetes, and clearly excludes patients outside of the mainstream of primary care (e.g. Aboriginal Medical Services, Tertiary Care etc). Data collected as part of NEFRON also relied on the GP's assessment of each patient, which may be inaccurate with regards to assigned ethnicity or other factors associated with diabetes management. As a clinic-based incident-driven study it has a number of limitations, being inherently (and deliberately) biased towards the kind of patients that regularly attend their practitioner, and the non-standardised laboratory results on which they base their daily management. Nonetheless, from the practitioner's perspective, the likelihood that any routine GP-patient encounter will be accompanied by specific complications of diabetes, and the opportunities for improvement within the context of established practice may be better estimated by a clinic-based survey, when compared to cross-sectional population studies including individuals who are never seen by practitioners, and data not available to practitioners. However, as Indigenous patients are less likely to seek care from a General Practitioner (Family Physician), there could be a bias towards more unwell subjects in that group, including those with a history of cardiovascular or kidney complications. Equally non-compliant individuals who carry a greater burden of complications, may be less likely to be represented in an encounter-driven survey. Nonetheless, the percentage of NEFRON patients with an Indigenous background (3.7%) was similar to that observed in other recent cross-sectional studies, including the DEMAND and the AusDiab surveys (3%) where ethnicity was self-identified (Robert Atkins, personal communication).

Although every effort was undertaken to ensure a representative distribution of practices through random selection of investigators[6, 7], initial recruitment in the NEFRON study was also based on interest in undertaking the study. It is also possible that these clinics might therefore not be an accurate representation of primary care services accessed by Indigenous Australians with diabetes. Consequently some selection bias in relation to participating investigators and subsequently enrolled diabetic patients, cannot be ruled out. Nonetheless, the practice and practitioner characteristics of the NEFRON study investigators, were similar to that recorded for Australian general practice as a whole, and on the whole represent the 10,000 encounters between patients with type 2 diabetes and their GP that occur every working day in Australia.

While the NEFRON study shows that ethnic disparities persist even within the same practice, this survey did not include individual data on socio-economic, educational, cultural and other factors known to impact diabetes management. Poverty, lack of education and other factors associated with socio-economic disadvantage are strongly linked to adverse health outcomes, independent to ethnicity. However, Indigenous Australian's carry a much greater burden. As seen in other minority populations worldwide[16], socio-economic disadvantage is strongly correlated with the higher prevalence of chronic disease seen in Aboriginal peoples [17, 18]. Data from the community-based Fremantle Diabetes study suggest that major disparities in income, housing and eduction may underlie many of the ethnic differences in diabetes care[11]. Similarly, issues regarding acceptability, uptake and compliance have not been addressed in the NEFRON study. However, recent data suggest that uniform access to and quality of care can result reduce process or outcome disparities in diabetes care[4]. While such data have not been confirmed in Indigenous Australians, even after adjusting for practice differences by analysing consecutive patients, key areas of unmet need in Indigenous patients with type 2 diabetes attending their GP may be clearly demonstrated.

Albuminuria was more common in Indigenous patients in this survey, where 23% of patients with type 2 diabetes had an ACR in the macroalbuminuric range, compared to 6% of non-indigenous individuals. However, the NEFRON study was limited by restricting classification of the UAE to the most recent ACR for any individual patient, Since there is known to be significant day-to-day variability in albumin excretion, ideally, at least two of three collections in a 3- to 6-month period should show elevated levels before a patient is classified as having an abnormal albumin excretion. Nevertheless, the NEFRON data are consistent with findings from the AusDiab study, which employed three collections and found that 34% of individuals with diabetes had an elevated UAE[19]. Although there is strong familial clustering of CKD in Indigenous patients, some of the observed excess in CKD may be attributable to modifiable differences in health care delivery, including glycaemic control and smoking cessation [20, 21]. There were no clear differences in mean blood pressure control or the initiation or use of antihypertensive therapy. Indeed, blood pressure targets were more often achieved in Indigenous patients, consistent with previous reports [11]. Similarly, the measures of obesity were similar in Indigenous and non-Indigenous cohorts, although such parameters have not been validated in Indigenous Australians.

It has been argued that micro- and macro-albuminuria in Aboriginal people is common in the absence of diabetes. Indeed, it has been suggested that the 'normal' range for microalbuminuria and other variables may be different in Indigenous and non-indigenous groups. Nonetheless, from a practitioner's perspective, the finding of CKD in Indigenous patient with diabetes is a potent risk factor for adverse outcomes, and should not be dismissed as a racial difference. While a date of diagnosis of diabetes is not the same as the onset of diabetes (which is likely to have been at least 5–10 years earlier), our data suggests that similar numbers of Indigenous and non-Indigenous patients with recently diagnosed diabetes had an abnormal ACR, defined according to conventional guidelines. However, at later time-points, albuminuria was more common, potentially reflecting accelerated microvascular disease in Indigenous patients.

Although microvascular disease was more common in Indigenous patients, the frequency of patients with macrovascular disease or impaired kidney function was overall very similar to that observed in non-Indigenous patients. Such data may suggest a survival bias, whereby Indigenous patients who have died or attended tertiary hospitals because of more severe complications (e.g. ESRD, CHF), would not have been included in the NEFRON study. Indeed, Indigenous patients with type 2 diabetes were on average 11 years younger than non-Indigenous probands, making cardiovascular events and ESRD appear less common, though at the same time, after age adjustment, inordinately high (figure 4). Moreover, 43% of Indigenous patients with type 2 diabetes had a first degree relative with cardiovascular disease diagnosed before the age of fifty.

There are many potential contributors to impaired glycaemic control in Indigenous patients. Although from the NEFRON survey it is apparently that patients are receiving the same combinations of the same antidiabetic drugs, and the targets for optimal glycaemic control are the same, few Indigenous patients are reaching these targets. Compliance with prescribed antidiabetic therapies, lifestyle and dietary modifications are likely contributors to this disparity, possibly reflecting issues surrounding the frequency, accessibility, affordability, and cultural sensitivity of diabetic care. Genetic predisposition, dyslipidemia, chronic inflammation and reduced 'endowment' of pancreatic islets (much as nephrons) in low-birth weight babies, may also make diabetes more difficult to treat. It is interesting to note that racial differences in metabolic control diminished with the duration of diabetes. Whether this is a lead-time effect of late diagnosis of diabetes, or the loss of recalcitrant patients to co-morbid disease remains to be established by longitudinal studies. However, it is plausible that at the time of presentation Indigenous patients have more advanced β-cell depletion and higher levels of insulin resistance, making them less sensitive to standard oral antidiabetic therapies, in a similar way described in Caucasian populations with advanced disease[22]. This has led to the suggestion that both groups would potentially benefit from insulin therapy[11], a treatment approach which remains underutilised in the general practice management of type 2 diabetes.

The NEFRON study was designed to examine factors that influence the management of diabetes in the context of GP-patient encounter. As such, it does not address the pivotal socio-economic determinates of health, including income and social status; social support networks; education and literacy; employment and working conditions; social environments; physical environments; and housing; which also contribute to high rates of diabetes and its complications. It is clear that Indigenous Australians Islanders will not achieve equal health outcomes until their economic, educational and social disadvantages have been eliminated[23]. Nevertheless, while social disadvantage continues, the NEFRON study also demonstrates key ethnic disparities in Australian diabetes care. The aim of the national health care funding system is to give universal access to health care. Yet while seeing the same doctors and receiving the same medications, treatment targets remain unfulfilled in Indigenous patients. More importantly, intensification of management, which should follow the identification of risk, is seldom applied. This cross-sectional study confirms Aboriginal ethnicity as a powerful risk factor for microvascular and macrovascular disease, which should be used by practitioners to identify candidates for more intensive multifactorial intervention.