Moderating Effect of Temperament On the Relation Between Symptom Severity and Age of First Parental Concern: A Retrospective Study

Background: There is a great deal of variability in the age at which children are formally diagnosed with autism, likely due, in part, to the heterogeneity of ASD presentation. ASD specific behaviors (i.e., symptoms) and non-syndrome specific factors (i.e., temperament) may contribute to the age at which parents first become concerned about their child's development. This study is a retrospective investigation of how symptom severity and temperament relate to age of first concerns. Objectives: To identify the unique and combined effects of syndrome-specific and non-syndrome specific factors on parents' retrospective reports of the age when they first had concerns about their child's development. Methods: Preliminary analyses included 58 higher functioning children with autism (HFA; 8-16 years of age).  All children had a community diagnosis of autism, a verbal IQ greater than 70 on the WISC-IV, and met diagnostic criteria on 2 out of 3 diagnostic measures (ASSQ, SCQ, ADOS). Parents completed the Social Communication Questionnaire (SCQ), Autism Spectrum Screening Questionnaire (ASSQ), Early Adolescent Temperament Questionnaire (EATQ), and a subset of items from the Autism Diagnostic Interview-Revised (ADI-R). Three hierarchical regression models were conducted to examine the relations between (1) SCQ total symptom severity, (2) one of the three EATQ factors (Surgency, Effortful Control, or Negative Affect), and (3) the interaction of symptom scores and temperament on parent-reported age of first concern in months on the ADI-R (ADI-R #2). In each analysis, age of first spoken word (ADI-R #9) was entered in the first step as a control variable. Results: Across all three models, SCQ predicted first concerns in months, in which greater symptom severity related to earlier age of concern, B=-.76, t(56)=-2.13, p=.037. However, Negative Affect was the only temperament variable that significantly contributed, over and above the effects of symptom severity, R2change=.09, F(3,55)=3.48, p=.02. In addition, the interaction of SCQ and Negative Affect significantly predicted age of first concern, R2change=.01, F(4,54)=2.69, p=.04.  The significant interaction was probed using guidelines provided by Aiken and West (1991). The regression of Negative Affect on ADI-R was examined at different levels of SCQ (± 1 SD and the mean). Each slope value was tested to determine whether it was significantly different from zero. Results showed that only the slope of the regression at -1 SD of SCQ was significantly different from zero, t(54)=-2.29, p=.026. That is, Negative Affect was particularly important for indicating earlier concerns for children with low levels of symptom severity.  Conclusions: The results of this study corroborate recent prospective findings from Garon et al. (2009), which reported that diagnosed ASD-sibs displayed low levels of positive affect, high levels of negative affect, and difficulty controlling attention, compared to a control group and ASD-sibs whom did not receive a diagnosis. Together these findings suggest that high levels of Negative Affect may be an early marker of ASDs, particularly in children showing lower levels of classic ASD symptoms. This emphasizes the importance of examining non-syndrome specific factors, such as temperament, in the study of individual differences in the detection and presentation of ASD.