Effect of selective screening for gestational diabetes.

OBJECTIVE: To estimate the percentage of pregnant women who would not be screened and the percentage of women with gestational diabetes mellitus (GDM) who would possibly remain undiagnosed if the American Diabetes Association's (ADA's) new selective screening recommendations are used rather than universal screening for GDM. RESEARCH DESIGN AND METHODS: Since 1987, the University of Michigan Health System has performed universal screening for GDM. In 1997, the ADA recommended that women having all four of the following characteristics need not be screened: age < 25 years, not members of an ethnic/racial group with a high prevalence of diabetes, normal body weight, and no family history of diabetes. We studied a random sample of the 25,118 deliveries at the University of Michigan between 1987 and 1997 to determine the prevalence of these four characteristics in our obstetric population. We also studied the prevalence of these four characteristics in 200 women who were diagnosed with GDM in the Endocrine Testing Unit and delivered at the University of Michigan between 1987 and 1997. RESULTS: Approximately 10-11% of women who delivered possessed all four low-risk characteristics and would not have been screened for GDM according to the new ADA recommendations. Only 4% of women (5 of 141) with GDM who delivered and for whom data on all four characteristics were reported possessed all four low-risk characteristics and would not have been screened. CONCLUSIONS: If the new ADA selective screening recommendations are used, few women with GDM will be missed (4%) but approximately 90% of pregnant women will still need to be screened for GDM.