Evaluating Social Communicative Behaviors Across Treatment Settings for Children with Autism

Background: Many intervention options for children with autism spectrum disorders (ASD) are widely available. The majority of research focuses on the effectiveness of interventions delivered one-on-one by a trained therapist in the home and in specialized inclusive preschools. However, there has been a lack of research conducted on the direct comparison of different treatment delivery models.

Objectives: This study compares the acquisition of social skills in children at-risk for ASD enrolled in either an in-home one-on-one intervention program or an inclusive preschool program. 

Methods: A total of 17 children, (mean age = 25 months) identified as at-risk for ASD were recruited from two early intervention programs. Nine children were recruited from a one-on-one in-home intervention program and the remaining eight children were recruited from an inclusive preschool. Both programs are public options for children under age three with ASD and staff from both programs were trained using similar behavioral and developmental strategies, reducing differences due to intervention types and highlighting the difference in intervention delivery; inclusive versus in-home. At intake chronological age, Early Learning Composite Score on the Mullen Scales of Early Learning, and score on the Social Domain of the Vineland Adaptive Behavior Scales were not significantly different between groups. Children were assessed within eight weeks of entering intervention and then again six months later. At each assessment period, children and parents participated in a battery of social skill assessments including the Early Social Communication Scales (ESCS), a structured play observation with a parent and another with a peer, and a video-referenced rating of early reciprocal social behavior (vrRSB) developed in the Constantino Laboratory at Washington University in St. Louis.

Results: A repeated measures ANOVA revealed significant decreases in parental ratings of social deficits on the vrRSB (M₁=51.24, M₂=41.53, F=10.444, p=0.007), and an increase in rate of initiating joint attention with the experimenter on the ESCS (M₁=0.64, M₂=.91, F=4.92, p=.042) across assessment periods for both groups of children. There were no differences over time or between groups in peer play or play with parents.

Conclusions: Generally, improvements in social skills across children were seen over time, although child performance was variable. Children showed improvements in joint attention skills, and were rated as having fewer social deficits by their parents over time in both groups. These data indicate that early intervention delivered in one-on-one or inclusive settings can be effective in supporting social skill development in children at-risk for ASD. However, the lack of changes in play with peers and parents over time warrant further investigation. The lack of differences between programs highlights the need for the analysis of resources consumed by each program type and child characteristics that may predict differing outcomes in one service delivery setting over another.