Objective - To develop a simple, economically viable, and effective means of population screening for diabetes mellitus. Design - A postal request system for self testing for glycosuria with foil wrapped dipsticks. Preprandial and postprandial tests were compared with a single postprandial test. The subjects were instructed how to test, and a result card was supplied on which to record and return the result. All those recording a positive test result and 50 people recording a negative result were invited for an oral glucose tolerance test. Setting - General practice in east Suffolk, list size 11 534. Patients - All subjects aged 45-70 years registered with the practice were identified by Suffolk Family Health Services Authority (n = 3057). The 73 subjects known to have diabetes from the practice's register were excluded, leaving 2984 subjects, 2363 (79.2%) of whom responded. 1167 subjects completed the single test and 1196 the two tests. Main outcome measures - Response rate and number of patients with glycosuria. Sensitivity, specificity, and positive predictive value of a single postprandial test and preprandial and postprandial tests. Number of new cases of diabetes identified and cost of screening. Results - Of the patients completing the single postprandial test, 29 had a positive result; an oral glucose tolerance test showed that eight (28%) had diabetes, six (21%) impaired glucose tolerance, and 14 (48%) normal glucose tolerance. 44 of the group who tested before and after eating had a positive result; nine (20%) had diabetes, five (11%) impaired tolerance, and 26 (11%) normal tolerance. Screening cost 59p per subject and 81 pounds per case detected. Of the 17 people with previously undiagnosed diabetes, eight were asymptomatic and 11 had not visited their general practitioner in the past three months. Conclusions - A postal request system for self testing for postprandial glycosuria in people aged 45-70 is a simple and effective method of population screening for diabetes mellitus.
