Audit of diabetes in general practice.

To complete a first audit cycle of diabetes care in a general practice and to develop a simple method for continuing the audit cycle. Retrospective examination of the medical records of all diabetic patients in a general practice in 1990. A group general practice in a Bristol health centre with roughly 13,200 patients, which since 1983 had had a protocol for care of its diabetic patients. 223 known diabetic patients in the practice. MAIN AUDITED MEASURES: Comparison against previously agreed standards of process and outcome of diabetes care in the practice, including number of patients whose care had been reviewed in accordance with the practice protocol, serum fructosamine and blood glucose concentrations in patients aged under 70, and number of newly diagnosed patients given explicit education and referred for diatetic advice. Defined standards were not met for several criteria--for example, percentages of patients aged below 70 (n = 149) with serum fructosamine concentrations less than 3.5 mmol/l (62% v 90% defined value) and less than 2.8 mmol/l (35% v 70%) and last recorded blood glucose concentrations less than 10 mmol/l in insulin dependent patients (n = 48) (23% v 90%) and less than 8 mmol/l in non-insulin dependent patients (n = 101) (17% v 90%). Of newly identified diabetic patients (n = 32), 59% and 28% respectively were referred to dietitians and given educational material compared with the 100% standard. The practice has a high prevalence of diabetes (1.7%) but has the resources for their care. The format and implementation of the agreed systematic process of care for diabetic patients needs improvement. A simple audit suitable for most general practices might record two measures of the process of care--a disease register of all diabetic patients in a practice and an attendance register to determine whether they have regular check ups--and one measure of the outcome of care, such as serum fructosamine concentration (or local equivalent). A practice could establish its own standards for these measures and monitor its performance against them.