Simultaneous Administration of an Autism-Specific and General Developmental Screener in an Urban Pediatric Population

Background:  Little data are available on the performance of simultaneous administration of autism-specific and general developmental screeners in pediatric practice.

Objectives:  To determine if the addition of an autism specific screener, the Modified Checklist for Autism in Toddlers (M-CHAT), to a pediatric primary care visit where a general developmental screener, the Ages and Stages Questionnaire version II (ASQ II), identifies additional toddlers at risk for neurodevelopmental issues.  In addition, the influence of positive screening on referral to EI services was also examined. 

Methods:   Existing data from a randomized trial of developmental screening conducted in 4 urban pediatric clinics from December 2008 to June 2010 were reanalyzed.  Of 2,092 toddlers <30mos of age participating in the trial, 658 were screened with both the ASQ II and the M-CHAT in the same primary care visit and were included in analyses here.  Subjects were categorized as failing the M-CHAT only, failing the ASQ-II only, failing both, or failing neither.  In addition, for those failing at least one screener, data were assembled on referral to Early Intervention (EI) services as documented in the medical record. The influence of number and type of screener failure on frequency and timing of EI referrals was assessed via Kaplan Meier analysis. 

Results:  170 (25.8%) toddlers failed at least one screener with 46 of these (7.0% overall) failing the M-CHAT.  The proportion of children who failed the M-CHAT but who did not fail the ASQ-II was 32.61% (95% CI: 19.06% - 46.16%). A total of 91 (53.5%) of children failing at least one screener had a subsequent EI referral. For those failing the M-CHAT only, 30% were referred to EI, similar to the rate for those failing the ASQ II only (46%).  The EI referral rate for children failing both screeners was 77%.  The χ² statistic for the difference in these proportions is 10.1 (p-value=0.006) and the p-value for the log-rank test from a Kaplan Meier analysis of the difference in the time to EI referral across these groups was 0.014.

Conclusions:  This study suggests that ASD-specific screening might identify a substantial number of additional children at-risk for neurodevelopmental disorders who are missed by general developmental screens.   Further, children who screen positive on both ASD and general developmental screeners are more likely to be quickly referred to EI services than are children failing just one of these screeners. However, the final diagnostic status of children in this study is not known.  Also, the M-CHAT used here did not include the newly recommended follow-up interview and it is likely that fewer children would have screened positive were this interview used.  While the ASD positive predictive value of the M-CHAT with follow-up questions has most recently been estimated at just under 50%, the positive predictive value for any neurodevelopmental concern was >95%, so it is reasonable to expect that some children identified with ASD specific screeners who not confirmed to have an ASD still would benefit from EI services.