Wrapped in Cotton Wool: Using Implementation Science to Examine Unexpected Results in a Recess-Based Intervention Study for Children With Autism Spectrum Disorder (ASD)

PURPOSE: Through the occupation of play, children experience wellbeing and develop social-emotional and communication skills (Parham & Fazio, 2008). Children with disabilities such as autism spectrum disorder (ASD) and intellectual disabilities (ID) may benefit from opportunities to practice these skills. However, children in substantially separate special education programs may face barriers to unstructured, child-directed play (Sterman et al., 2016). Occupational therapists are well-equipped to advocate for free play during the school day for these children. However, implementing interventions in schools is notoriously difficult. There is a paucity of research exploring the challenges to implementing playground interventions for children with disabilities. In this study, we examine the factors that supported or inhibited implementation of a novel playground intervention designed to promote recess play for children with autism spectrum disorder (ASD) and intellectual disabilities (ID). DESIGN: We asked: what were the facilitators and barriers to intervention implementation at two schools that participated in a playground intervention? In a larger study, we implemented a two-pronged playground intervention in five substantially separate education programs (four specialized schools and one mainstream school with a support classroom). For this study, we merged an implementation science framework (Domitrovich et al., 2008) with a qualitative comparative case study design to examine two schools that participated in the playground intervention. Each school served as an instrumental, bounded case. We selected these two schools because each had surprising quantitative results: one school had higher-than-expected outcomes, while the other did not improve at all on targeted outcomes. METHOD: Semi-structured evaluative interviews conducted with school staff at each school served as the primary data source for this study (n = 5 at School A and n = 7 at School B). To supplement these data, we conducted interviews with members of the research team and reviewed artifacts, including school policy documents. We employed cross-case analysis techniques, deductively coding each interview according to the Domitrovich et al. (2008) model. This model proposes factors that may impact intervention implementation in schools (e.g., personnel expertise). We compared the schools on each element of this model. RESULTS: Facilitators included strong and supportive leadership and a stren