Predicting Improvement In Social/Communication Symptoms of ASD Using Retrospective Treatment Data

Background:  

Studies examining change over time in symptoms of autism spectrum disorders (ASD) have reported mixed findings depending on sample and design.  Some have shown improvement in social/communication abilities, while others have shown worsening social impairment.  Although early language and IQ have emerged as predictors of improvement in some studies, these models have not examined the role of previous treatment.   Among smaller samples of young children, treatment-outcome studies have also found IQ and early language to be predictors of response to therapy; however, these findings have not been replicated among large community-based samples or among older children and adolescents.

Objectives:  

Primary objectives were to:  1) examine the developmental course of social/communication symptoms across ages among a large sample of children/adolescents with ASD, and 2) examine potential predictors of symptom improvement, particularly the roles of community-based treatment type and intensity.

Methods:  

The sample included 1433 children/ adolescents with ASD (ages 6-17) who participated in the Simons Simplex Collection (SSC), a North American multi-site study involving families with only one child with ASD.  Measures for the current study included the Autism Diagnostic Interview-Revised (ADI-R), IQ test, demographic information, and parent-reported treatment history (including occupational, speech, and behavioral therapy).  Baseline and current social/communication symptom severity scores were calculated using sums of “current” and “age 4-5” codes for 15 relevant ADI-R items.

Results:  

The majority (95%) of the sample showed at least some improvement in social/communication symptom severity.  Analysis of variance (ANOVA) revealed cohort effects based on age, with older children showing greater current and historical severity, lower likelihood of having received any type of treatment, and, for those who did receive treatment, a later age at first treatment (p<.01).  With regard to improvement in social/communication symptoms, ANOVA revealed no gender, race, or age differences.  Those who received speech, behavioral, or occupational therapy demonstrated greater improvement (p<001); and age at first treatment was correlated with improvement (p<.01).  Linear regression examined potential predictors of current severity, controlling for initial severity and age. Gender, race and verbal status at age 5 were not significant predictors.  Significant interactions between nonverbal IQ and treatment were found (p<.01), with all 3 treatment types associated with greater improvement for those with higher IQ.  Separate regression models examined the effect of treatment intensity on outcome.  Significant IQ by treatment interactions were found for all 3 treatment types, with age at first treatment predicting improvement for speech therapy and OT.

Conclusions:  

Among a large, well-characterized sample of children and adolescents with ASD, we found improvement over time in social/communication symptom severity.  Treatment was associated with improvement, but effects were most pronounced for individuals with higher IQ.  Earlier age at first treatment (for speech and occupational therapy) also predicted improvement.  Despite the limitations of the retrospective design, this study provides important information about the course and outcomes of ASD symptoms. Future large-scale studies with prospective designs will be essential in further elucidating these issues.