Early Childhood Predictors of Social Functioning in Adults with Autism Spectrum Disorder

Background: Low levels of adaptive social functioning in adults with Autism Spectrum Disorders (ASD) have been reported across a range of longitudinal studies (reviewed in Seltzer et. al, 2004). However, some adults with ASD achieve higher levels of social functioning (Kanner et al., 1972; Kobayashi et al., 1992). Identification of early predictors of differential adult social outcomes may inform intervention approaches. While early IQ scores and language skills (Howlin et. al, 2004, Lord & Bailey, 2002) are frequently indicated as the strongest predictors of adaptive outcomes in adolescents and adults with ASD, early joint attention skills predict subsequent language in middle childhood (Sigman & Ruskin, 1999) and adolescence (McGovern & Sigman, 2004). We hypothesized that joint attention, in addition to IQ and language skills, would predict adult social outcomes.

Objectives: This longitudinal study sought to identify developmental predictors of social functioning in adulthood and to relate our findings to prior studies (such as Howlin et al., 2004).

Methods: Participants included 20 adults (19-32 yrs) who were diagnosed with autism in early childhood. Social outcome ratings (based upon Howlin et. al, 2004) were used to differentiate between adults with good to fair social outcomes (some independence) and adults with poor outcomes (lack of independent work, living arrangements and friendships). The two groups were then compared on the following skills assessed in early childhood (3-6 yrs): initiation and response to joint attention, developmental intelligence quotient and developmental language quotient. Due to variability in age at first assessment, scores on the intelligence and language measures are developmental quotients (DQ) or age equivalent scores divided by chronological age.  

Results: Independent samples t-tests revealed that those with good to fair social functioning in adulthood (n=9) differed from those with poor outcomes (n=11) on intelligence scores [M=69.9 vs. 49.6, p < 0.05], language scores [M=52.4 vs. 33.7, p < 0.05], responsiveness to joint attention [M=16.8 vs. 6.3, p < 0.001], and initiation of joint attention [M= 13.4 vs. 5.3, p < 0.01]. However, when all of the variables were entered into a single regression model, only language DQ was a significant predictor of adult social outcome ( ß = .72, t (13) = 3.2, p < 0.01).

Conclusions: These findings suggest that in early childhood, autistic individuals who will achieve higher levels of adult social functioning differ from those with poorer social outcomes across a range of measures. However, by 3 to 6 years of age, language skills may exert more influence upon adult outcomes than either nonverbal social skills or intelligence scores. These findings encourage continued focus on language skills in treatment programs targeting social functioning in individuals with autism. Improvements in characterization and treatment of children with autism in the 30 years since this sample was first evaluated suggests the need for further longitudinal research in order to continue fine-tuning treatment for children with autism on the basis of early symptom profiles.