National patterns in diabetes screening: data from the National Health and Nutrition Examination Survey (NHANES) 2005-2012.

backgroundThere are few current population-based estimates of the patterns of diabetes screening in the United States. The American Diabetes Association (ADA) recommends universal screening of adults ≥ 45 years, and high-risk adults < 45 years, but there is no current assessment of ADA guideline performance in detecting diabetes and prediabetes. Furthermore, data on racial/ethnic patterns of screening are limited.

objectiveOur aim was to estimate diabetes screening prevalence for the US adult population and specifically for those who meet ADA criteria; to report the prevalence of prediabetes and diabetes among these groups; and to determine if high-risk race/ethnicity was associated with reported screening.

designThis was a Cross-sectional survey.

participantsNon-pregnant adults (≥ 21 years) without diabetes or prediabetes who participated in the National Health and Nutrition Examination Survey (NHANES) in 2005-2012 (n = 17,572) were included in the study. "Screening-recommended" participants, classified by ADA criteria, included (1) adults ≥ 45 years and (2) "high-risk" adults < 45 years. "Screening-not-recommended" participants were adults < 45 years who did not meet criteria. MAIN MEASURES: Diabetes screening status was obtained by self-report. We used calibrated HbA1c and/or fasting glucose levels to define undiagnosed diabetes and prediabetes. KEY

resultsSeventy-six percent of the study population (approximately 136 million US adults) met ADA criteria. Among them, less than half (46.2%) reported screening; undiagnosed diabetes affected 3.7% (5 million individuals), and undiagnosed prediabetes affected 36.3% (49 million people.) African Americans were more likely to report screening, both among adults ≥ 45 years and among "high risk" younger adults (OR 1.27 and 1.36, respectively.) Hispanic participants were also more likely to report screening (OR 1.31 for older adults, 1.42 for younger adults.) The screening rate among "screening-not-recommended" adults was 29.6%; the prevalence of diabetes and prediabetes were 0.4 and 10.2%, respectively.

conclusionsIn a nationally representative sample, 76% of adults met ADA screening criteria, of whom fewer than half reported screening. Limitations include cross-sectional design and screening self-report.