Socioeconomic Disparities in ASD Screening Outcomes Using the M-CHAT(-R)

Background: With its goals to reduce costs and increase accessibility, healthcare reform has become a popular topic of discussion among US citizens. Lower maternal education, a marker for socioeconomic status (SES), has been linked to decreased access to and utilization of healthcare, due to factors such as reduced awareness of health risk factors and prevention initiatives, limited resources, and poor provider-patient communication.

Objectives: As part of a large autism screening study using the M-CHAT(-R), informal observations have raised concerns regarding differences in screen positive rates across families of different SES.  In order to better understand healthcare disparities, this study examines whether socioeconomic variables differentially affect outcome within this multistep screening study.

Methods: The M-CHAT(-R) is a questionnaire used to screen toddlers for ASDs; the Follow-Up Interview (FUI) clarifies at-risk responses.  Parents in metro-Atlanta completed the M-CHAT(-R) at their child’s 18- and/or 24-month pediatric visits to identify children at risk for autism spectrum disorders (n=11,918). 1,025 (8.6%) screened positive on the questionnaire, 758 of whom were successfully contacted to complete the FUI.  A total of 232 were offered a free evaluation after demonstrating continued risk on the FUI. Diagnostic evaluations were completed by 145 families with 82 (57%) resulting in an ASD diagnosis.  An additional 6,421 toddlers were excluded from the sample due to missing demographic data (i.e., sex, maternal education, ethnicity), language barrier, significant motor delays that would preclude standardized assessment, or because they were part of a developmentally typical control sample.

Results: Level of maternal education ranged from less than 8^th grade to graduate-level education (median was a bachelor’s degree, 33% of sample).  Among parents who completed the M-CHAT(-R), screening results were significantly related to level of maternal education (i.e., < 8^th grade, high school, some college, bachelor’s, graduate), χ²(4, 11,918)=213.8, p<.001.  That is, as maternal education decreased, children were more likely to initially screen positive.  Among those who completed an FUI, interview results were also significantly related to maternal education, χ²(4, 758)=14.7, p=.005. Children of families with the highest level of maternal education (graduate level) were more likely to continue to screen positive on the FUI.  This suggests that the greatest consistency across questionnaire and interview level of screening is seen in those with the highest level of education.  Finally, at evaluation, maternal education was not significantly related to diagnostic outcome, χ²(4, 145)=6.1, p=.194.

Conclusions: Results suggest that children of parents with lower SES are more likely to initially screen positive on the M-CHAT(-R) than those with higher SES.  Screening outcome remains most consistent from questionnaire to interview for those with graduate level degrees, possibly due to increased awareness of normative childhood development.  This also suggests that FUI plays a significant role in reducing the inflated screen positive rate for families of lower SES.  At the diagnostic level, SES disparities are no longer apparent, supporting previous findings of consistent ASD rates across diverse backgrounds.  Additional research on reasons for these differences is needed in order to eliminate disparities in identification and referrals for intervention.