Assessment of racial and ethnic bias in autism spectrum disorder prevalence estimates from a US surveillance system.

Abstract

This study assessed potential under-ascertainment of autism spectrum disorder due to missing administrative information for Hispanic and non-Hispanic Black children within the Autism and Developmental Disabilities Monitoring Network. We analyzed data from two Network sites (Colorado and Wisconsin) for surveillance years 2012 and 2014 to determine whether children excluded from autism spectrum disorder prevalence estimates due to missing residency and other information differed from those included by race and ethnicity. We used multiple approaches to impute missing information to evaluate impacts on racial and ethnic disparities in autism spectrum disorder prevalence. Compared with confirmed autism spectrum disorder cases, those excluded due to missing residency were more than twice as likely to be Hispanic (19% vs 44%; p < 0.002), yet the number of cases excluded due to missing residency information was too small to account for prevalence differences. Confirmation of autism spectrum disorder case status was more likely for children with relevant health records than for those with school records only. Moreover, relevant health records were more likely to be missing for Black and Hispanic children than for White children. Observed disparities in autism spectrum disorder prevalence were not accounted for by missing demographic data, but may reflect disparities in healthcare access for developmental evaluations.