Effectiveness of Screening Tools in a Community-Based Sample: Which Children Are Missed and Why?

Friday, May 12, 2017: 12:00 PM-1:40 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
S. Richardson1, M. Reid2, C. Beacham3 and C. Klaiman4, (1)Marcus Autism Center, Atlanta, GA, (2)Catholic University of America, Washington, DC, (3)Marcus Autism Center, Children's Healthcare of Atlanta, Atlanta, GA, (4)Emory University, Atlanta, GA

Early screening of Autism Spectrum Disorder (ASD) is essential for referring children for evaluations and early intervention, with benefits found in starting intervention as young as possible (Zwaigenbaum et al., 2015). With pediatricians on the front lines, providers are in need of measures that efficiently and effectively identify concerns. The Ages and Stages Questionnaire-3 (ASQ-3) is a common developmental screener for general developmental concerns and the Modified Checklist for Autism in Toddlers, Revised (M-CHAT-R) is a widely used autism specific screening tool. Our previous work has shown that parent concern on the ASQ accurately captures developmental delays in children with ASD, but that only the communication scale was associated with autism symptomatology on an ADOS, suggesting the need for an ASD specific screener as well (Hamner et al., 2015).


To build upon our previous work by evaluating two commonly used screening tools and their effectiveness in identifying children in need of further evaluation.


78 children (64 boys) between the ages of 16-43 months (mean = 28.80 months, SD = 6.93) referred based on parent concerns and/or provider recommendations who received an ASD diagnosis. Approximately 85 additional participants are expected before May 2017. Parents completed the ASQ-3 and M-CHAT and children were administered the Mullen Scales of Early Learning and the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2).


On the ASQ, 63 out of 78 children failed the communication scale (81%), 58 (74%) failed the problem-solving scale, and 56 (72%) failed the personal-social scale. Of the 15 children who passed the communication scale, 6 more were identified by either the problem-solving or personal-social scales. On the M-CHAT, 66 (85%) screened with concerns (total score 3 or higher), of which 61 also failed at least one ASQ scale. When combining the ASQ communication scale and the M-CHAT, 72 (92%) of children were identified and combining an ASQ fail on any scale and the M-CHAT increases that number to 96%. Children who passed the ASQ communication scale performed significantly higher on Mullen visual reception (t(77) =4.56, p <.01), receptive language (t(77) = 5.44, p < .01), and expressive language (t(77) = 4.37, p < .01) scales. A similar, significant pattern was seen for problem-solving and personal-social scales. Additionally, those who passed the communication scale were significantly older (t(77) = 2.80, p <.01); this pattern was not seen on other scales. Children missed on the M-CHAT had significantly higher visual reception scores (t(76) = 2.04, p <.05) and a trend for higher expressive language scores (t(76) = 2.01, p = .06).


In our community sample, both the ASQ and M-CHAT flagged the majority of children with ASD, supporting the utility of these measures in pediatric practices. The most children were identified with the combination of ASQ scales and the M-CHAT, indicating a need for both measures. Higher functioning children are more likely to be missed when using only one of the measures. Pediatricians are encouraged to monitor parent concerns in children without developmental delays even if passing screening measures.