Using Temperament Traits to Identify Subgroups of School-Aged Children with Autism Spectrum Disorder

Background:   Temperament is defined as a set of individual differences in behavioural tendencies. An extensive body of literature establishes the association between temperament and childhood psychiatric disorders and more recent studies have documented specific temperament profiles in infants with Autism Spectrum Disorder (ASD). While previous studies have been informative in distinguishing ASD from non-ASD cases in infancy and early childhood, our understanding of how temperament varies within the ASD population and how that variability is associated with outcome in children with ASD later in life is limited.

Objectives: The primary objective of this study was to identify subgroups of school-aged children with ASD with distinct temperament profiles. A secondary objective was to examine whether children from different temperament subgroups have different outcomes – adaptive functioning skills, social communication deficits, and emotional-behavioural problems – later in life.  

Methods: Data came from the Pathways in ASD study, a large Canadian longitudinal study of children with ASD. The sample was composed of 185 school-aged children with ASD (mean age: 7.8 years; 83% males). Temperament was indexed using the Temperament in Middle Childhood Questionnaire (TMCQ). A total of 17 TMCQ dimension scores were used in hierarchical cluster analysis to derive subgroups of children with ASD at baseline (T₁). Analysis of variance (ANOVA) was used to explore subgroup differences on: (1) T₁ (baseline) TMCQ dimension scores; and (2) T₂ (12 months later) outcome on variables of interest –adaptive functioning (indexed by the Vineland Scales of Adaptive Behaviour; VABS II); social communication deficits (indexed by the Social Responsiveness Scale; SRS), and emotional behavioural-problems (indexed by the Child Behaviour Checklist; CBCL 6-18).

Results: A 2-cluster solution was selected as the best fit to the data. At T₁ the two clusters were characterized by distinct temperamental profiles (p<.001 for most TMCQ dimensions). Specifically, compared to children in Cluster 2 (60.5% of sample), children in Cluster 1 (39.5% of sample) had higher scores on activation control, affiliation, assertiveness, attentional focusing, fantasy/openness, and soothability and lower scores on anger/frustration, impulsivity, discomfort, fear, sadness, and shyness. Moreover, at T₂ (12 months later) children in Cluster 1 had better adaptive functioning skills, fewer social communication deficits, and fewer emotional behavioural problems (p<.001 for all).      

Conclusions: Study findings show that in school-aged children with ASD distinct temperament profiles are associated with differential outcomes regarding adaptive functioning, social communication, and emotional-behavioural problems over a 12-month period. These findings support the use of temperament traits as a potentially informative approach in disentangling phenotypic heterogeneity in ASD. Since temperament is believed to be linked to genetic factors, such an approach could prove fruitful for the development of nosological models that help bridge the gap between clinical/behavioural symptoms and the biological underpinnings of the disorder.