Comparative Randomized Control Trial for Infants at-Risk of Autism Spectrum Disorder: Outcomes from a Standard Measure of Play

Background: There is increasing interest in interventions targeting the infant and early toddler years, when it is believed that the developmental course of ASDs can be substantially altered. Parent-mediated interventions are important at this age as they provide an opportunity to begin treatment early and at a high dose. Interventions that target joint attention, engagement and play are also important for this age.

Objectives: The goal of this study was to rigorously test parent-mediated Joint Attention, Symbolic Play, Engagement and Regulation (JASPER) for infants at-risk for ASD. JASPER augmented one of two early intervention classrooms that used a global infant curriculum, Assessment, Evaluation, and Programming System for Infants and Children (AEPS; Bricker, 2002). AEPS-only (referred to as Standard Baby) was compared to AEPS+ Baby JASPER (referred to as Baby JASPER). In addition, parents randomized to Baby JASPER received hands-on, individualized training in JASPER and parents in the Standard Baby condition received parent education classes.

Methods: In total, 105 at-risk toddlers, 12 to 22 months old, were referred and screened for study entry. Eighty qualified for the trial by scoring with at least mild to moderate concern on the ADOS-2 Toddler module. The children were predominately male (80%) and ethnically diverse (54%) with an average age of 17.66 months (SD = 3.08). Once enrolled, the children were randomized to Baby JASPER or Standard Baby, twice-weekly for two months. Assessments were completed at baseline and immediately following the 8-week treatment period, which included the Structured Play Assessment (SPA; Ungerer & Sigman, 1981), a standardized play assessment delivered by a clinician blind to treatment status and unfamiliar to the child. The SPA was coded for overall play level, play diversity, and frequency of play acts.

Results: A zero-inflated Poisson regression with a random intercept was used to model the change in symbolic play due to excessive zero counts. A majority of children, 88% in the Baby JASPER group and 83% in the Standard Baby group, had no acts of symbolic play at baseline. Both groups saw improvement in the percentage of children with at least some symbolic play at Week 8 Exit (80% versus 62%, respectively), however, children in the Baby JASPER group had a greater increase in the frequency of symbolic play compared to the Standard Baby group. On average, children in the Baby JASPER group had 4.2 times more symbolic play from baseline to Exit compared to the Standard Baby group, controlling for expressive language, receptive language, visual reception, severity, and chronological age with 95% CI (1.2, 15.2) and ANOVA Type III p-value = 0.03.

Conclusions: Both treatment groups noted improvement in play skills over the course of intervention. Of the children with at least some symbolic play, the children in the Baby JASPER treatment showed greater increases compared to the Standard Baby treatment group. Of note, play outcomes were assessed using a standard measure of play conducted by a novel administrator, blind to treatment status, suggesting that the treatment effect generalized outside of the parent-child dyad.