Evidence of Increasing Socioeconomic Disparity in the Prevalence of Autism Spectrum Disorder Among U.S. Children

Background:  Previous studies have shown significant positive associations between autism spectrum disorder (ASD) prevalence among children in the U.S. and indicators of socioeconomic status (SES), raising the possibility that there is under-ascertainment of ASD in less socioeconomically advantaged children.  An analysis of data from 12 Autism and Developmental Disabilities Monitoring (ADDM) Network sites for the combined years of 2002 and 2004, for example, found ASD prevalence to increase in a stepwise manner from 4.8, to 6.5 and 8.1 per 1,000 children in the lowest, middle and highest SES tertiles, respectively.

Objectives:  To replicate the previous ADDM Network study using comparable data for the year 2006, and evaluate the hypothesis that the SES disparity in ASD prevalence would be reduced in 2006 relative to 2002.  Support for this hypothesis would be consistent with the possibility that awareness and identification of ASD among children of low SES have improved over time, and have contributed to ASD prevalence increases.

Methods:  A cross-sectional study was implemented combining data from all 10 ADDM Network sites that participated in both the 2002 and 2006 study years of the ADDM Network.  The ADDM Network is a multiple source surveillance system that incorporates abstracted data from health care providers and school records to determine the number of 8-year-old children meeting DSM-IV TR criteria for autistic disorder or other ASD.  In 2006, the population base for this study included 287,606 8-year-old children. This population was defined by geographic boundaries that were common across the two study years, 2002 and 2006.  Within this population, 2,446 children were found to meet diagnostic criteria for ASD in 2006. Census block group-level SES indicators were used to compute ASD prevalence by SES tertiles of the population.

Results:  The prevalence of ASD in the combined study area in 2006 was 8.5 per 1,000 children, a 51% increase relative to the same study area in 2002. In addition, similar to 2002, the prevalence per 1,000 children increased stepwise with increasing SES, from a low of 5.5 in the lowest SES tertile to 9.3 in the middle SES tertile and 11.7 in the highest SES tertile. Although prevalence increased over time in all three SES tertiles, the magnitude of the increase was greatest for high SES children, resulting in an increase rather than a decrease over time in the magnitude of the SES disparity.  The prevalence ratios for the highest versus lowest SES tertiles increased significantly (p<0.05) between the two study years, from 1.74 (95% confidence interval 1.54, 1.96) in 2002 to 2.13 (95% confidence interval 1.92, 2.36) in 2006.

Conclusions:  The SES disparity in ASD prevalence has increased during a time period of increasing overall prevalence of ASD. Because surveillance in this study was based on review of health and educational records, these results point to possible under-ascertainment and lack of access to services for children with autism who are socioeconomically disadvantaged. Further research is needed to monitor and understand the sources of the SES disparity in ASD prevalence.