Creating Symptom Profiles to Anticipate Treatment Outcomes for Adolescents with ASD Following the UCLA PEERS® Intervention

Background:  Adolescents with autism spectrum disorder (ASD) present with unique symptom profiles and varying degrees of symptom severity with regard to social deficits. These deficits fueled the evolution of targeted social skills interventions addressing the core deficits seen in ASD (Rogers, 1998). One such example includes the Program for the Education and Enrichment of Relational Skills (PEERS^®), an evidence-based, parent-assisted social skills intervention for adolescents with ASD without intellectual disabilities. Traditionally, targeted interventions like PEERS^® have focused on classifying individuals with ASD according to IQ and/or diagnosis alone. Yet, few studies have examined the extent to which symptom profiles, rather than IQ or diagnosis, anticipate treatment outcome. Thus, classifying adolescents with ASD into symptom profiles may provide more useful insights into who is most likely to benefit from targeted interventions. 

Objectives: The purpose of this study is to (1) identify distinct symptom profiles related to social functioning among adolescents with ASD, and (2) distinguish unique treatment outcomes across these symptom profiles following the PEERS^® intervention. 

Methods:  Participants in this study included 139 adolescents with ASD from 11-18 years of age (M=14.17, SD=1.70) who presented for social skills treatment with their parents through the UCLA PEERS^® Clinic. In order to create symptom profiles of social functioning, a k-means Cluster Analysis was performed using baseline parent-reports of social functioning on the subscales of the Social Responsiveness Scale (SRS; Constantino, 2005). Treatment outcome was assessed for self-esteem (PHS-2: Piers et al., 2002), anxiety (SAS; La Greca, 1998), empathy (EQ; Baron-Cohen, 2004), social skills and problem behaviors (SSIS; Gresham & Elliot, 2008), friendship quality (FQS; Bukowski, 1994), peer engagement (QSQ; Frankel & Mintz, 2008), and social skills knowledge (TASSK; Laugeson & Frankel, 2010). One-way ANOVAs and independent T-tests were used to identify differences in treatment outcome across clusters (i.e., symptom profiles).

Results: Using a k-means Cluster Analysis, three distinct symptom profiles emerged across adolescents with ASD using the SRS: mild, moderate-to-severe, and severe impairments in social functioning. Significant differences across clusters were observed in relation to anxiety on the SAS [F(2,104) = 7.04, p=.001], social skills on the SSIS [F(2,74) = 5.82, p=.004], and overall empathy on the EQ [F(2,26) = 4.09, p=.028] in the areas of emotional reactivity [F(2,26) = 3.54, p=.044] and cognitive empathy [F(2,26) = 4.30, p=.024], according to parent reports following treatment. Significant differences in overall self-esteem on the PHS-2 were observed across clusters [F(2,107) = 2.77, p=.05], according to adolescent reports following treatment. While treatment gains were observed across all clusters in the following areas, there were no significant differences between clusters in empathic social abilities on the EQ, peer engagement on the QSQ, friendship quality on the FQS, or social skills knowledge on the TASSK.  

Conclusions: The marked differences in treatment outcomes across symptom profiles among adolescents with ASD highlight the multi-faceted phenotypes observed on the spectrum, and emphasize the necessity to “go beyond” IQ and diagnosis alone in order to better understand who is more likely to benefit from targeted treatment intervention in this highly diverse population.