A Comparison of Clinical Versus Self-Report Measures: An Evaluation of the Broad Autism Phenotype in Parents of Individuals with Autism

Background: The Broad Autism Phenotype (BAP) refers to a constellation of subclinical behavioral features related to the core symptoms of autism, expressed among unaffected relatives and indexing genetic liability. Studying the BAP is critical for understanding the complex etiology of autism. While in-person objective assessments of the BAP have been well validated (e.g., Piven et al. 1997), they are often labor- and time-intensive. Questionnaires offer a time-efficient alternative, but may be subject to reporting bias. It is therefore important to identify valid yet efficient assessments that reliably capture the BAP. Recent work (Sasson et al., 2014) reported significant discrepancies on questionnaires of self- and informant-reports of the BAP in fathers of children with autism when they were positive for particular BAP features, suggesting that clinical judgment is required for a deeper characterization of the BAP, particularly in fathers. This study investigates agreement between objective clinical ratings of BAP features using gold standard objective, relative to questionnaire measures.  

Objectives: To compare agreement between questionnaire and direct assessment measures of the BAP in a large sample of parents of children with autism. 

Methods: Participants included 171 parents of children with autism (107 mothers; 64 fathers). All participants completed a self- and informant-report version of the Broad Autism Phenotype Questionnaire (BAPQ; Hurley et al., 2007), a well-validated questionnaire that quantifies BAP characteristics. Personality features of the BAP were assessed using the Modified Personality Assessment Scale (MPAS; Tyrer, 1988), in which subjects and informants are interviewed separately about personality styles, and ratings are conducted by blind, independent coders relying on concrete behavioral examples. Pragmatic language was assessed using the Pragmatic Rating Scale (PRS, Landa et al., 1992), in which a semi-structured conversational interview is conducted and objective coders rate the presence of clinically significant pragmatic language violations. BAP (+) status was defined by the presence of at least two of the directly assessed BAP characteristics (Aloof, Rigid, or Pragmatic Language deficit). On the BAPQ, parents were classified as BAP (+) based on published criteria.    

Results: A series of 2x2 chi-squared analyses were conducted to assess differences between questionnaire (BAPQ) and direct assessment (MPAS/PRS) scores. Overall, the BAPQ was less accurate in classifying fathers as BAP (+) (p < .05). This relationship was driven by discrepancies in Aloof ratings (p < .01).  Rigid traits and Pragmatic language ratings were consistent across questionnaire and direct assessment in both groups. 

Conclusions: Results mirror those of Sasson et al. 2014, suggesting that fathers may be less likely to endorse Aloof traits and overall BAP features in themselves. Alternatively, clinical evaluation (i.e., MPAS) may be more sensitive in capturing such traits in fathers. Data will be further explored to examine specific agreement between self- and informant-report measures on both the MPAS and BAPQ, assessing whether variables such as social cognition and gender may predict consistency (or lack thereof) between each measure. This study has implications for future phenotyping efforts relevant to molecular genetic, neurobiological studies of autism incorporating data from both affected individuals and unaffected family members.