Do Clinicians Operationalize the Broader Autism Phenotype the Same Way Across Sites?

Background: Many researchers and clinicians have started to utilize the term and concept of a ‘Broader Autism Phenotype (BAP)’ in an attempt to better understand the complex etiology, genetic/familial contributions, developmental trajectories, and range of outcomes associated with siblings of children with  Autism Spectrum Disorders (ASD). To use the term effectively, however, we must establish shared operational definitions and methods for assessing the BAP at various points in development. At this point, it is unclear whether clinicians across sites identify BAP concerns based on the same symptoms and profiles.

Objectives: To examine variables associated with clinician-rated BAP in siblings of children with ASD and typically-developing controls and investigate the relative emphasis placed on these variables across three clinical research sites.

Methods:  N=88 toddlers (N=45 sib-ASD, N=43 control) aged 36 months (Mullen ELC M=108.7, SD=16.9) were assessed at three clinical research sites collaborating on a prospective study of infant sibling development. No children receiving a diagnosis of ASD (N=12) were included in analyses.  As part of a best-estimate diagnostic procedure, clinicians blind to group status rated whether they had BAP concerns (yes/no).  This was done for both siblings of children with ASD and controls.  If “yes” to BAP concern, clinicians specified their concern (yes/no) on the following domains: Social, Language/Communication, Atypical Behaviors, Emotional/Behavioral , or Discrepancy between observed and reported behavior.  Data also were available from the Autism Diagnostic Interview-Revised (ADI-R; Rutter, Couteur & Lord 2005), the Autism Diagnostic Observation Schedule (ADOS; Lord et al., 2000), Vineland Adaptive Behavior Scales, 2^nd edition (Vineland II; Sparrow et al., 2005), and the Mullen Scales of Early Learning (MSEL; Mullen, 1994). Logistic regression models were used to examine which continuous scale scores and binary clinician-rated symptom domain variables were associated with BAP concerns, with and without controlling for site.

Results:  Across all sites, the presence of BAP concern was significantly associated with the ADOS Social Affect and Restricted Repetitive Behavior domain scores, the Mullen Early Learning Composite, and the Vineland-II Socialization and Communication domain scores. Out of 34 children with identified BAP, clinicians were concerned about language and communication functioning for 56%, and about social and atypical behaviors for 41% respectively, with 12% eliciting concern over other emotional/behavior symptoms. In 6% of these children, a discrepancy was noted between observed and reported behavior. No predictor-by-site interactions were significant for either these binary BAP domain ratings or the continuous variables.  Some 20% (N=8) of these children rated as having BAP concerns were from the non-sib control group.

Conclusions: Binary ratings of BAP may represent ratings of meaningful subclinical concerns, as they are correlated with commonly used phenotyping measures across sites.  Our findings suggest that these ratings were not different across three clinical sites although clinicians were given no specific training in overall BAP appraisal.  However, the content and meaning of these subclinical concerns warrants substantial consideration.  A majority of concerns related to language, and children with no family history were often rated as evidencing BAP concerns.