What You Don’t Know Can Hurt You: Early Life Course Racial Health Disparities in Undiagnosed Diabetes

Population estimates of health conditions come from large national surveys. However, these estimates may vary across studies for groups with differing risk for diagnosable conditions like diabetes. The purpose of the current study is to investigate differences in prevalence estimates of diabetes risk for young adults between an age-matched subsample of the NHANES and the Add Health studies. This study seeks to determine whether differences in study designs alter estimates of diabetes risk for young adults. Focal analyses explore the impact of biomarker collection nonresponse in NHANES and Add Health on diabetes risk estimates. Results indicate that African Americans are disproportionately likely to be missing biomarker data despite completing the survey portion of both studies. When diabetes status is imputed for these individuals, increased odds of nonresponse do not change predicted risk of undiagnosed diabetes. However, the multivariate models predicting undiagnosed diabetes risk yield different conclusions across studies. This suggests that the NHANES estimates for undiagnosed diabetes may be less generalizable for young adults than the estimates from the Add Health data. Further study is needed to evaluate possible race-specific nonresponse bias of biomarker data collection across the two studies.