Social Communication Screening and Parent Concern at 9-21 Months of Age: Comparison of a Large Primary Care Sample and Children Later Diagnosed with ASD

Background: The American Academy of Pediatrics recommends screening all children for ASD at 18 and 24 months. However, there is limited evidence of well-validated autism-specific screening tools in primary care settings. Unlike families referred for suspected ASD or high-risk siblings who may have heightened concern, families in the primary-care population may receive positive screening results before raising concerns about their child. In a review of current evidence for ASD screening, examining how broadband and autism-specific screening tools can be used together to improve accuracy was a priority for future research (Zwaigenbaum et al., 2015). 

Objectives:  To compare the results of a broadband screen for social communication (SC) delay and parent concern from a large sample screened in primary care by the FIRST WORDS^®Project with a subgroup of children later diagnosed with ASD.

Methods:  Children were first screened for SC delay based on parent-report with the Infant-Toddler Checklist (ITC; Wetherby & Prizant, 2002) through primary care providers at 9-21 months of age and negative screens were invited for re-screening. Two autism-specific screening tools were used for children with a positive SC screen— the Early Screening for Autism and Communication Disorders (ESAC) based on parent-report and the Systematic Observation of Red Flags of ASD(SORF) based on video-recorded observation. Children with a positive autism screen were invited for a diagnostic evaluation to confirm or rule out ASD between 24-36 months of age. Parent concern reported on the ITC was coded into 16 types of concern. This sample represented the region with 57.6% white, 30.6% black, 8.4% multiracial; 7.6% Hispanic; 47.2% first-born, 32.1% second-born, 20.5% later-born; 15.4% bilingual. The results were grouped into 3-month age intervals from 9-21 months and analyzed.

Results:  Based on screening of 8,161 children, the percent of positive SC screen ranged from 15-20%, indicating selection bias of the sample based on the tenth-percentile screening cutoff. There were 194 children diagnosed with ASD by 3 years. The percent of parents with concern in the full sample increased from 6% at 9-11 months to 15% at 18-21 months. For the ASD subgroup, the percent of positive screen ranged from 53% at 9-11 months to 78% at 18-21 months. The percent of parent concern in the ASD subgroup was 22% at 9-11 months, 33% at 12-14 months, 50% at 15-17 months, and 67% at 18-21 months. Most concerns were initially expressive language and motor skills, and predominantly expressive language by 18-21 months. Patterns of screening outcome and parent concern will be examined by age and race/ethnicity.

Conclusions:  These findings document the effectiveness of a developmental surveillance system to screen for ASD in a general population sample of toddlers using a broadband SC screener followed by two autism-specific screeners. There was an increase in parent concern from 9-21 months, but the broadband SC screening outcome was more sensitive than parent concern across this age range. These findings illustrate that parents are fairly accurate at reporting social communication milestones but less accurate at knowing whether to be concerned about delays in these milestones.