Are Children Severely Affected By ASD Underrepresented in Treatment Studies?: An Analysis of the Literature

Background: Despite significant advances in the field of autism research, many experts have noted that those severely affected by Autism Spectrum Disorder (ASD) appear to have been understudied, particularly as the ASD phenotype has broadened over time. The extent to which this is true, and how severity is represented in the literature, has not been quantified. This question may exert its most proximal effects in the arena of treatment studies, which inform current clinical practices.

Objectives: We aimed to evaluate the degree to which the more severely affected population of children with ASD has been included in treatment studies, how severity is measured and represented, and if there have been changes in these areas over the past two decades in the autism literature.

Methods: A comprehensive literature search was conducted encompassing all autism research articles published from January 1991 to March 2013. Inclusion criteria for our analysis were peer-reviewed treatment studies with a sample size greater than one of pediatric subjects with ASD. Review of the articles based on these criteria yielded a final sample of 367 publications. The following was extracted from the full text of each article: publication year, sample size, age and sex of subjects, study design, type of treatment, and treatment target. To assess severity of participants in each study, information was extracted regarding communication level, IQ/level of intellectual disability (ID), and adaptive behavior, noting whether each was reported and with what measures. To determine overall presence of severely affected subjects, we evaluated each study based on severity cut-off scores on measures used in three domains (communication, IQ/ID, adaptive behavior). If an article included severely affected subjects in at least one domain, it was categorized as having evidence of inclusion of the severely affected population.

Results: Of the three severity domains, information regarding IQ/level of ID was most commonly reported (in 65.7% of papers, n=241), followed by communication (42.2%, n=155) and adaptive behavior (21.8%, n=80). Forty-nine percent of the studies (n=180) demonstrated evidence of inclusion of the severely affected population by our criteria. Notably, when examining the inclusion of the severely affected population over time, there was a significant decrease in the proportion of studies that included severely affected subjects (Χ² (8, n=367) = 36.7, p<0.001). Of the assessed papers published 1990-1994, 90% (n=18) included the severely affected population, whereas only 36.2% (n=42) of papers published 2010-2013 included the severely affected.

Conclusions: In reviewing treatment studies of children with ASD, it is clear that even by our liberal criteria of needing just one reported domain (IQ, communication, adaptive) to fall within the severe range, children severely affected by ASD were increasingly underrepresented in the literature over time. Particularly notable, in a disorder with a core feature of communication deficits and typically lowered adaptive functioning, was the underreporting of both domains. Future research could examine possible factors related to the underrepresentation of those severely affected by ASD, such as relative lack of gold-standard measures for communication and challenges in accessing more severely affected samples for outpatient studies.