Parents' Concerns Predict a Later Autism Spectrum Disorder Outcome: A Prospective Study of High-Risk Siblings from 6 to 36 Months

Background:   Studies of ‘high-risk’ (HR) infants  (infant siblings of children diagnosed with Autism Spectrum Disorder (ASD)) provide a unique opportunity to prospectively track parents’ developmental concerns in relation to subsequent ASD diagnosis. Using this method, parents of children with ASD have reported concerns as early as 12 to 14 months of age, the most frequent concerns relate to delayed communication (Ozonoff et al., 2009; Hess & Landa, 2012). Although promising, the findings of these studies are limited in scope, in respect to range of developmental domains covered, number of time points, and lack of low-risk (LR) community controls. The present study addresses these limitations.  

Objectives: To examine which parental concerns in the first 24 months distinguished HR infants who were diagnosed with ASD at 36 months from other HR and LR infants.

Methods:  Participants: Three groups of children: (1) HR siblings who did not receive an ASD diagnosis at 36 months (HR-N), (2) HR siblings who did receive a diagnosis of ASD at 36 months (HR-ASD), and (3) infants without a family history of ASD (low-risk; LR).  At 36 months, an independent, diagnostic assessment for ASD using the ADOS and ADI-R was conducted for all participants, blind to risk status and prior assessments, including parent concerns.

Parent Concerns: Parents of LR and HR infants were interviewed about current concerns regarding their infants’ development at ages 6, 9, 12, 15, 18, and 24 months. The interview queried 10 domains: sleep, diet, sensory interests or unusual reactions, gross and fine motor development, repetitive movements and restrictive behaviors, communication development (both verbal and non-verbal), regression in communication skills, social skills, play behavior, and behavioral problems.  A coder blind to group membership coded the concerns using a binary system, with “0” representing the absence and “1” the presence of a concern. A second coder, also blind, coded 30% of all forms to assess reliability using Cohen’s kappa (overall 0.71).

Statistical Analyses: Total number of concerns and each individual domain were analyzed using mixed model analyses, with Group and Age as independent variables. To determine when group differences emerged, group x age interactions were explored using Benjamini & Hochberg (1995) corrections. 

Results:  Total number of concerns differentiated HR-ASD from the HR-N and LR groups by 12 months of age (qs < .003). Domain-specific concerns revealed that sensory (qs < .032) and motor (qs < .028) concerns at 6 months differentiated HR-ASD from the other two groups, whereas other domain-specific concerns did not differentiate HR-ASD until 12 months or later; social (12 months; qs < .025), communication (15 months; qs < .028), repetitive behaviors (18 months; qs < .03), and behavioral problems (24 months; qs< .017).

Conclusions:   The results suggest that the presence or absence of parental concerns provides valuable predictive information to aid in differentiating between HR infants who will and will not receive an ASD diagnosis. The sequencing of parent concerns reflects the developmental course of ASD, with early concerns mirroring the ‘prodrome’ of ASD, and later concerns reflecting the ‘core’ domains of the disorder.