Community screening for diabetes. Low detection rate in a low-risk population.

OBJECTIVE: To evaluate glucose-based community screening for diabetes with regard to detection rate. RESEARCH DESIGN AND METHODS: A retrospective analysis of a community-screening questionnaire data base that included a screening for blood glucose. Referred subjects had fasting glucose levels > 6.4 mM (115 mg/dl) or postprandial levels > or = 8.9 mM (160 mg/dl). An attempt was made to contact referred subjects and to ascertain whether follow-up was undertaken and current status. A random sample of subjects not meeting the glucose criteria (nonreferred) also was contacted in an analogous fashion to referred subjects.
RESULTS: In 2,016 questionnaires, glucose-based referral criteria were exhibited by 148 (7.3%) individuals, and subsequent evaluation data were available for 111. Of those 111 individuals, 37 (33%) knew they had diabetes before the screening, and 39 (36%) did not seek further evaluation. Of the remaining 35 subjects, 6 (13%) were told of their new diagnosis of diabetes, and 29 were told they did not have diabetes. Three of 50 nonreferred subjects knew of their diabetes before screening. Thirty percent (14 out of 47) of nonreferred subjects underwent subsequent evaluation, although they were not told to do so. A single new case of diabetes occurred in the nonreferred group.
CONCLUSIONS: Community screening for diabetes that is based on measured glucose is of low yield. The known problems of glucose-based screening, coupled with its low yield, make a glucose-based approach difficult to justify. These results indicate that glucose-based community screening should be done only under the careful supervision of a health professional who is trained both in glucose measurement instrumentation and in screening.