Correspondence Between Parent Report and Clinician Observation at 12 Months in Infants at High Risk for ASD

Background:  Given the current emphasis on early screening for autism spectrum disorders (ASD), it is crucial to examine the efficacy of screening tools and understand the potential vulnerabilities associated with utilizing parent reports. Although parents have an optimal perspective from which to observe and engage with their infants across diverse contexts, utility of parent ratings in a high-risk population is unclear, given contrast effects based on an older sibling with ASD along with other limitations of informant reporting. To further the development of ASD screening tools for infants, it is critical to evaluate concordance between parent report and clinical observation.

Objectives:  To investigate agreement across parental and clinician ratings, we compared similar items on the First Year Inventory (FYI; Baranek et al., 2003) a parent questionnaire designed to screen for ASD at 12 months, and the ADOS-Toddler (ADOS-T; Lord et al., 2013), a standardized ASD diagnostic tool, administered concurrently to 12-month-olds at high and low risk for ASD.

Methods: 161 (high-risk(HR)=95, low-risk(LR)=66) infants were evaluated with the ADOS-T at 12 months by clinicians blind to their risk status. Prior to the ADOS-T assessment, parents completed the FYI. Nine items describing comparable behaviors were identified across the instruments, including social, communication, and repetitive behavior. Each measure rates behaviors on a 0 to 3 scale. Items were compared using paired-samples t-tests with Bonferroni correction for multiple comparisons.

Results: Regardless of risk status, parents rated three speech and communication-related items (amount of babbling, directed vocalizations, and gestures) and object imitation as more typical compared to the clinical assessment (p<.005). Clinicians and parents in both risk groups agreed on items assessing response to name, response to joint attention, showing, and repetitive behaviors. Interestingly, response to name appeared twice on the FYI in slightly different formats. When the question was phrased globally (i.e., “Does your child answer to his name?”), both HR and LR parents rated the infant’s behavior as more typical than the clinician (p<.001); however, when the same question was presented in a multiple-choice format (“What do you typically have to do to get your baby to turn towards you?”), parent and clinician ratings were more similar in both risk groups.

Conclusions: Parents of high- and low-risk infants rated their children’s behaviors similarly to clinicians on a number of items tapping into social and repetitive behavior, but discrepancies between expert clinicians and parent ratings on speech and communication items were common in both risk groups and may reflect the effect of unfamiliar context on child’s behavior during direct assessment.  Wording of questions appeared to affect the degree of discrepancy between clinician and parent report. These findings suggest a further need to examine the sources of disagreement between parental report and clinician rating of key behaviors as well as the impact of design features on screening instruments’ ability to capture behavior most accurately. In addition to these issues, our discussion will also include preliminary results of the FYI’s utility in identifying those with a later ASD diagnosis in a high-risk sample.