Objective: To examine factors influencing the identification of kidney impairment in patients with type 2 diabetes in Australian primary care.
Design, setting and participants: 348 general practitioner investigators were asked to estimate kidney function and its severity in 10-15 consecutively presenting patients with type 2 diabetes (n = 3893). They were then asked, for each patient, whether they routinely estimated kidney function. No instruction was provided on how kidney function should be estimated or categorised. Data were collected between April and September 2005.
Main outcome measures: Kidney function estimated by the Cockcroft-Gault equation using clinical and laboratory data provided by the GP; estimates of kidney function made by the GP.
Results: In 24% of the patients with type 2 diabetes, their GP routinely estimated kidney function. However, few of these patients had impaired kidney function or risk factors for kidney disease. There was a good statistical correlation between the estimates made by GPs and the data-derived estimates (R2 = 0.72). GPs identified patients with data-derived estimates of kidney function < 60 mL/min in over 83% of cases, with a specificity of 90%. Impaired kidney function was reported by GPs in 34.4% of men and 36.4% of women. These figures were discordant with function categorisation using both GP estimates and data-derived values, overlapping in half of the patients. Despite GPs' ability to assess creatinine clearance, "raw" (unstandardised) serum creatinine levels inappropriately influenced the perception of impairment of kidney function.
Conclusion: GPs can accurately assess kidney function, without reporting of estimated glomerular filtration rate (eGFR). However, even in patients at increased risk of chronic kidney disease, routine estimates are seldom made. Our findings underline the value of the recent initiative recommending automatic reporting of eGFR in Australia.