25772
Investigating the Causes of Informant Discrepancies in the Assessment of Autism Spectrum Conditions

Friday, May 12, 2017: 12:00 PM-1:40 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
J. J. Finnemann1,2,3 and K. Barnes1, (1)Department of Psychology, University of Cambridge, Cambridge, United Kingdom, (2)Department of Psychiatry, University of Cambridge, Cambridge, United Kingdom, (3)Department of Music and Performing Arts, Anglia Ruskin University, Cambridge, United Kingdom
Background:  
“Informant discrepancies are ubiquitous in research on the assessment, development and treatment of psychopathology” (De Los Reyes, 2011) and pose a major challenge to the diagnostic process of neurodevelopmental conditions such as autism. The current ‘gold standard’ for diagnosing autism includes an observational assessment (ADOS) and a caregiver interview (ADI), but little is known about the causes of discrepancies between these.
Past research has shown that parent report of child behaviour is influenced by factors including parental mental health (Treutler, 2003), parental gender (Hay, 1999), the age of the child (Bitsika, 2016) and parental income (Johnston, 2010).

Objectives:  
We were interested in investigating whether there are systematic differences in evaluations of children with autism and whether specific contributions to these discrepancies can be identified.

Methods:
The sample consisted of 364 children (who took part in a randomised controlled trial of music therapy) between the ages of 4 and 8. The baseline assessment included the ADOS, the ADI, the Social Responsiveness Scale (SRS) and a life satisfaction questionnaire (QoL) for the child and the family. All children were followed up with ADOS assessments after 2, 5 and 12 months and parents filled in the questionnaires and attended counselling sessions. The latter were evaluated by conducting sentiment analyses on the transcripts of the sessions.
Growth curve modelling was used to examine the longitudinal data: The final model was a parallel linear growth model for informant discrepancies (z-scores of the ADOS and SRS scores were used to derive a signed difference score across time) and QoL. The predictors tested included the age of the child, the linguistic ability of the child, parental education and the gender of the parent.

Results:
As data collection finishes in November 2016 only preliminary results are currently available.
However the findings replicate previous research indicating stronger correlations between the parent report measures (ADI and SRS) than between parent and clinician measures (ADI & ADOS and SRS & ADOS).
In addition both the quality of life ratings as well as obtained sentiment scores correlated negatively with autism severity reported in the SRS, whereas none of the measures correlated with the total ADOS scores.
Only language ability emerged as a significant predictor in the growth curve analysis.

Conclusions:
Two factors were identified which are associated with informant discrepancies: the language ability of the child and parental wellbeing.
Language delay is often the first concern noted by parents (De Giacomo & Fombonne, 1998) and thus could play a role in the perceived severity of the condition by parents whereas behavioural observations should be sensitive enough to assess other diagnostic domains which are sometimes missed by parents of young children (Chawarska et al. 2006).
Furthermore the impact of the family's quality of life on the cognitive appraisal of the child's condition underlines the need to investigate the importance of post-diagnostic support on parental wellbeing.