Since 2006, the American Academy of Pediatrics (AAP) has recommended developmental screening using a validated screening tool at all well-child visits. The AAP additionally recommended autism-specific screening at 18- and 24-months. Pediatricians note a lack of time and resources as barriers to autism-specific screening. Although some preliminary research has indicated that broad-band screeners may be able to detect autism, at this time there is insufficient data to support using broadband developmental screeners in lieu of autism-specific screeners.
Objectives:
To extend previous study results testing the sensitivity and specificity of both the broadband Ages and Stages Questionnaire (ASQ), and the autism-specific Modified Checklist for Autism in Toddlers- Revised (M-CHAT-R) in detecting autism spectrum disorders (ASDs) amongst children screened at 18- and 24- months.
Methods:
1846 children were screened with the M-CHAT-R and the ASQ. Of those children, 157 failed the M-CHAT-R. The parents whose children failed the M-CHAT-R were contacted for a follow-up phone interview. Forty-six children failed the phone-interview and were invited to the University of Connecticut for a developmental and diagnostic evaluation. Thus far, 33 of these 46 children have received evaluations. The sensitivity and specificity of the ASQ for detecting children who failed the M-CHAT-R, children who failed the follow-up interview, and children found to have an ASD upon evaluation, was assessed. Data is presented on the specific ASQ domains (Communication, Gross Motor, Fine Motor, Problem Solving and Personal-Social) that children with an ASD diagnosis failed.
Results:
In the following analyses, to fail the ASQ, the child had to fail one or more domain. Among children who failed the M-CHAT-R (n= 157), the ASQ demonstrated 52% sensitivity and 89% specificity with the M-CHAT-R screener. Among the children who failed the follow-up phone interview (n = 46), the ASQ demonstrated 85% sensitivity and 61% specificity ( χ2 = 27.63, p < .001) at detecting the children who failed the follow-up phone interview. In terms of children who received an ASD diagnosis (n=11), the ASQ demonstrated 82% sensitivity, and 14% specificity ( χ2 = 0.098, p = .75).
Thirty-three children were evaluated and received the following diagnoses: 11 ASD, 11 Developmental Delay, 4 Developmental Language Disorder and 7 No Diagnosis. Eight failed the Communication domain, 5 Personal-Social, 4 Fine Motor, and 3 Problem Solving.
Conclusions:
The ASQ identified 9 of the 11 children who received an ASD diagnosis. Therefore the ASQ demonstrates good sensitivity in identifying children at risk for an ASD. However, the ASQ’s specificity is less than that demonstrated by the M-CHAT-R. Therefore, the data corroborate the suggestion to use a broad-based screener in conjunction with an autism-specific screening measure; specifically those children who fail the ASQ, should be given an autism-specific screener. Further analyses will be run on the specific predictive power of each ASQ domain, and total ASQ scores.
See more of: Epidemiology
See more of: Prevalence, Risk factors & Intervention