UCLA

The past thirty years have seen a proliferation in intervention development and testing for autism spectrum disorders (ASD). As a result, there are evidence-based approaches effective at improving the lives of individuals with ASD and their families. However, these approaches frequently have small effects, largely due to the wide heterogeneity that is characteristic of ASD. As those with ASD have a range of cognitive, social, and language characteristics, one-size-fits-all intervention approaches may not best meet the varied needs of this population. Additionally, little intervention research has included or focused on those most in need of effective interventions: those with little to no language. The present examination consists of three studies with the overall aim to characterize heterogeneity in minimally verbal children with ASD and determine whether response to naturalistic versus structured interventions are different depending on child profile. This aim was achieved via secondary data analysis of three separate minimally verbal intervention samples. All included participants had 30 different words or less but in one sample were school-age and received intervention (naturalistic v. structured) for 6 weeks and in another study were preschool-age and received intervention (naturalistic v. structured) for 6 months. Subgroup profiles identified in Study 1 were used as predictors of intervention response in Studies 2 and 3. Results suggest that minimally verbal children with ASD are heterogeneous and can be characterized based on nonverbal versus verbal measures into subgroups, with patterns that emerge in preschool and persist through early childhood. Furthermore, profile subgroups are suggested to account for much of the variability in response to intervention. In particular, having one word on a naturalistic language sample at entry made a difference in subgroup by intervention interactions. For some of the language outcomes, subgroup by intervention interactions suggest that, at least initially, a more structured intervention approach may be more beneficial for those participants with at least one word and the lowest nonverbal and verbal scores across measures while those children with higher scores across measures may benefit from a more naturalistic approach. Implications for future interventions and details are discussed.