Examining Predictors of Treatment Success In a Parent-Assisted Social Skills Training Program for Teens with Autism Spectrum Disorders: The UCLA PEERS Program

Background: Social skills training has become a widely used treatment in assisting individuals diagnosed with Autism Spectrum Disorders (ASD) effectively fit into their social environment. However, most social skills treatment research has focused on younger and lower functioning children with ASD, with very little research targeting higher functioning older adolescents on the spectrum. Among the few intervention studies conducted with this population, very little emphasis has been placed on examining the factors that predict treatment success for this highly underserved population. Objectives: This study examines the predictors of treatment success for the UCLA PEERS Program, a manualized evidence-based parent-assisted social skills intervention aimed at improving overall social skills and friendship quality among adolescents 13-17 years of age with high-functioning autism or Asperger’s Disorder.
Methods: 61 adolescent participants and their parents were randomly assigned to a treatment condition or a delayed treatment control condition.  Participants attended weekly 90-minute group treatment sessions over a 12-14 week period. Targeted skills included: conversational skills; electronic communication; choosing appropriate friends; appropriate use of humor; peer entry and exiting skills; good host behavior during get-togethers; good sportsmanship; strategies for changing bad reputations; and strategies for handling rejection including teasing, bullying, arguments, and rumors/gossip. Skills were taught through didactic instruction using concrete rules and steps of social etiquette in conjunction with role-playing exercises.  Teen participants practiced newly learned skills during behavioral rehearsal exercises within the group, and parent-assisted weekly socialization homework assignments outside of the group.  Factors associated with treatment success were examined across both groups. 
Results: Participants were divided into two groups, based upon whether they hosted more get-togethers with peers by the end of treatment. Of the 43 participants (70.4%) who hosted more get-togethers after treatment, significantly better quality of friendships were also reported (p <.05), as well as significantly more invitations for get-togethers from other teens (p <.025). Adolescents who improved, reported an average of 3.4 more hosted get-togethers per month following treatment (p <.025) and an average of 2 more invited get-togethers at the end of treatment (p <.025). Baseline socialization variables that appear to predict treatment success included lower scores on adolescent knowledge of social skills (p <.025), number of hosted get-togethers prior to the intervention (p <.05), and number of invited get-togethers prior to the intervention (p <.005).  Statistical analysis further revealed that higher scores on autistic symptom traits at baseline predicted treatment success (p <.05). Other demographic variables including age, grade, gender, ethnicity, school placement, IQ, and adaptive functioning were not correlated with treatment outcome.
Conclusions: These findings suggest that predictors of treatment success in the UCLA PEERS Program are correlated with greater autistic symptomatology at baseline and poorer social skills.  Participants who were more socially impaired at baseline and were experiencing greater autistic symptomatology were more likely to improve their frequency of hosted get-togethers following treatment.  Additionally, participants who had better quality of friendships were more likely to host get-togethers and to be invited for get-togethers at a peer’s house following the intervention.