Mediation of Treatment Effect in Teacher-Implemented Social Communication Intervention for Young Children with Autism

Background:

Few studies have examined mediators of treatment outcomes for young children with autism (e.g., Gulsrud et al., 2015; Pickles et al., 2014), and none for teacher-implemented interventions. Previous studies have examined specific strategies, or active ingredients, within interventions; however, it is important to consider that interventions are often packaged for optimal effect. Thus, we have identified and grouped strategies within the intervention into three tiers of support (Basic, Foundational, and Advanced) to examine how these different sets of strategies mediate the outcome measure of joint engagement.

Objectives:

The current study is a secondary data analysis from two previously published teacher implemented intervention studies (Chang et al., 2016; Shire et al., 2016). The aim of this study is to examine whether sets of strategies mediate joint engagement.

Methods:

Participants. 179 children age 2-5 years (mean age=38 months; SD = 9.35) from two independent community intervention deployment studies (Chang et al., 2016 and Shire et al., 2016) were included in the study. Children in both studies were randomized to immediate JASPER (Joint Attention, Symbolic Play, Engagement, and Regulation) treatment (IT) or school as usual control (WL) for 3 months.

Intervention. Teachers and paraprofessionals received training and in-vivo coaching with the research team to learn the JASPER intervention. In the IT group, teachers implemented JASPER 30 minutes a day with their students for 3 months, while WL teachers continued with standard school curriculum.

Ten-minute Teacher-child play interactions (TCX). Teachers’ accurate implementation of JASPER was coded at entry and exit. Seven main strategy subscales were coded. Each subscale was summarized in a percentage score where higher score indicates greater implementation accuracy. Three tiers of strategies were created: Tier 1 (Basic) included supporting regulation/engagement, environmental arrangement, and imitation/modeling. Tier 2 (Foundational) included play routines and language strategies. Tier 3 (Advanced) included play expansion and programing for social communication gestures. Children’s time jointly engaged was coded as proportion of the play interaction that children were jointly engaged with their teachers (1 minute intervals).

Results:

Basic and Advanced strategies partially mediated the effect of treatment (% mediated: 36% and 28% respectively; p-values<0.01) while Foundational strategies had close to full mediation (50% mediation; p<0.01).

Conclusions:

All three sets of strategies mediated the effect of treatment on joint engagement. Joint engagement was most influenced by Foundational strategies. These results highlight the importance of Foundational strategies, which include establishing developmentally appropriate play routines and responding to and expanding children’s communication. It may be that Basic strategies, such as setting up the environment, may not be enough support to engage children with autism, and Advanced strategies, such as expanding routines, may be more effective after Foundational strategies, such as having a solid play routine, are established to create a context for social communication.