Screening for Autism Spectrum Disorders in Young Children Referred for Developmental Assessment: Guiding Efficient Assessment Practices within a Tertiary Clinic Setting

Background: While there is an increased awareness and demand for effective screening tools for young children at risk for ASD, there is not yet consensus on which measures are most effective. The M-CHAT and CSBS-ITC show tremendous promise for use in primary care settings to identify toddlers at risk for ASD, but these have not been investigated as closely among children who are referred for evaluation due to known or suspected developmental delays. Given high rates of referral and few clinicians skilled at diagnosing ASD, it is essential that clinics identify children in need of ASD-specific evaluation as efficiently as possible.

Objectives: To evaluate the utility of two screening measures for ASD in young children in a tertiary care clinic setting, including:

- utility in a clinic-referred sample of children at risk for developmental disabilities

- identification of children without a referral question of ASD who were in need of ASD-specific evaluation

- utility in a more expanded age range (up to 48 months)

- how each measure individually and together predicted ASD diagnosis.

Methods: As part of the intake process for referred children under the age of 4 years, caregivers completed the CSBS-ITC and M-CHAT in addition to a general intake form requesting information about development and referral question. Screeners are used by clinic staff to inform assessment procedures, including whether the child will receive an ASD-specific evaluation in addition to other interdisciplinary assessments. Based on referral question and/or scores on screeners, children were routed to either ASD-specific or more general developmental evaluations. Final diagnoses utilized ADI-R/ADOS and were concluded by members of the interdisciplinary team.

Results: 138 children have been screened using these procedures; diagnostic assessments have been completed on a subset (n = 29). Within the available data, 69% of the children who had a question of ASD were confirmed to have an ASD diagnosis. Among the children with confirmed ASD, 85% were detected with the M-CHAT and 92% were detected with the CSBS-ITC. Within this sample, the positive predictive values for the M-CHAT and CSBS-ITC were 58% and 50%, respectively. Negative predictive values were 80% for the M-CHAT and 75% for the CSBS-ITC. Of the children who screened positive on one or both of the screeners but did not have a referral question of ASD, none received an ASD diagnosis upon specific evaluation. Evaluations are in progress or will be scheduled for an additional 60 children.

Conclusions: Preliminary data indicates that children at-risk for ASD are already being referred for an ASD-specific evaluation. The positive predictive values are lower in the current sample than has been previously reported in community-based samples, warranting additional analyses. The rates of false positives observed on the CSBS-ITC in this sample is likely related to the nature of the non-ASD referrals to our clinics, as most children referred in this age range have concerns regarding language and/or other developmental delays. Subsequent analyses will explore the utility of these measures both within and beyond the typical age ranges for these measures.