Objective: The use of self-reported data to assess diabetes risk in the general population is a key component of the National Diabetes Prevention Program. The American Diabetes Association Diabetes Risk Test (ADADRT) is a popular tool used for this purpose, yet there is paucity of data on its validity in real-world clinical settings. We assessed the performance of ADADRT in identifying undiagnosed prediabetes or diabetes at a Federally Qualified Health Center serving predominantly African Americans.
Methods: The ADADRT was administered to participants without pre-existing diagnosis of prediabetes or diabetes and standard validation measures were calculated using HbA1c as the gold standard. Findings are reported for participants who self-identified as being Black or African American (n = 151).
Results: HbA1c identified 25.8% of the sample as having prediabetes or diabetes. At the recommended cutoff for ADADRT (i.e., score ≥5) the following measures were observed: proportion of high-risk persons 47.7%, sensitivity 74.4%, specificity 61.6%, positive predictive value 40.3%, negative predictive value 87.3%, Youden Index 36, and area under the receiver-operating characteristic curve 0.68.
Conclusion: ADADRT reasonably identifies low income African Americans with high diabetes risk. Findings are comparable to the precursor to ADADRT as validated with nationally representative data.
C.C. Chima: None. N. Anikpezie: None. B.C. Wade: None. L.S. Pongetti: None. T. Powell: None. B. Beech: None.