ABSTRACT

Objectives.  Although screening guidelines exist to detect new cases of diabetes mellitus (DM), about one-third of cases are undiagnosed and complications may be present two to ten years prior to diagnosis.  This raises the question of whether the guidelines are adequate to detect new cases of DM.  This study evaluates the effectiveness of the 2002 Expert Committee on the Diagnosis and Classification of Diabetes Mellitus testing in asymptomatic, undiagnosed individuals.

Methods.  The Third National Health and Nutrition Examination Survey, NHANES III, (1988-1994) is a multi-stage probability sample of US adults with clinical and survey data.  Adults ³20 without a prior diagnosis of diabetes who had a morning fasting plasma glucose (FPG) were classified as non-diabetics (FPG<126, n=6,240) and diabetics (FPG³126, n=274).  The guideline risk factors were compared between the two groups.

Results.  Screening for one or more risk factors has 94.4% sensitivity and requires that 79.2% of the population be tested.  Screening when two or more risk factors are present has comparable sensitivity (92.8%) and requires that 56.2% of the population be tested.  Since DM occurs at younger ages for minorities, screening whites who are ³40 or minorities who are ³30 has 95.0% sensitivity and requires that 59.8% of the population be tested.   

Conclusions. Following screening recommendations would essentially eliminate undiagnosed cases of DM.  Screening for two risk factors or using a simple age-based decision-rule would obtain similar sensitivity, and would reduce the necessary testing by approximately 20%.  Efficacious guidelines exist, but implementation of the guidelines remains the greatest challenge.