Screening for Autism and Developmental Disorders in Toddlers from a Lower Socio-Economic Strata Using the MCHAT and ASQ; Fail Rates by Demographic Characteristics

Background:

The American Academy of Pediatrics recommends early screening for developmental disabilities (DDs), including autism spectrum disorders (ASD). Surveillance of ASD has consistently shown lower prevalence rates among Hispanics than Whites. To investigate possible reasons for this, including limited health care access, we initiated a screening program in a county health system serving a primarily Hispanic population.

Objectives:

Screen all children born in 2006 attending well-child pediatric visits between 16-30 months of age at two clinics and compare fail rates to other populations and by demographic characteristics.

Methods:

The Modified-Checklist for Autism in Toddlers (MCHAT) and the Ages and Stages Questionnaire (ASQ) were given to parents of children who fit our criteria, with a choice of English or a Spanish translation. These were scored on-site by our staff, and results provided to pediatricians to inform their exams, discussion with parents, and decision to refer or continue monitoring.  We linked our screened population to the clinic data base and the California birth certificate file to obtain demographic data including maternal race, age and education, and child’s race and age at screening. We calculated MCHAT fail rates for first screens only and for any screen and then compared them by ASQ fail rates and by demographic characteristics, using the chi-square test (p-value <0.05 indicating statistical significance); multi-variate modeling will also be conducted

Results:

We linked 92% of the screening records to birth certificates, yielding data on 2,010 screens, representing 1,518 individual children. Overall, 21% of screens resulted in a failed MCHAT and 32% in a failed ASQ (on any domain). Examining the first screens only, 21% of children failed the MCHAT, of which, half also failed the ASQ; 32% failed the ASQ communication domain and 18% the social domain. Children of Hispanic race/ethnicity (81% of the sample) were significantly more likely to fail the MCHAT than Whites; 22.6% vs. 13.3 % respectively. Also, the 60 percent of children with Spanish language screens had a significantly higher fail rate (25.1%) than those with English (16.0%). MCHAT fail rates were also higher among children with less educated mothers (p=0.07) and somewhat higher among younger mothers, but not significantly.

Conclusions:

This sample’s MCHAT fail rate is about twice that found in other studies of children of a similar age. This may be due to a higher prevalence of DDs in this lower socio-economic population, or to differences in interpretation of the instruments by these parents. The higher fail rate observed among Hispanic children and less educated mothers is opposite the pattern typically seen for ASD rates. However, we do not yet know which children will be diagnosed with an ASD or DD. Our results indicate that screening a low SES, Hispanic population can be conducted but may require more clinical resources to facilitate administration and to conduct follow-up on a higher proportion. The surveillance we ultimately complete in this county will determine if screening helped identify more children with ASD than in the prior year.