(PS3-81) Predictors of Icbt Treatment Outcomes for Anxiety Among Autistic Youth

Introduction:  

Digitally-delivered CBT (iCBT) that automates CBT content is one method that has been proposed to reduce therapist burden without sacrificing efficacy. While iCBT is highly efficacious, its impact varies among individuals. Pinpointing patient traits that forecast iCBT treatment efficacy will enable clinicians to personalize treatments to align their patients with the most appropriate interventions. Studies on predictors of CBT outcomes among autistic and non-autistic youth with anxiety disorders have produced relatively inconsistent findings. This points to the need to further explore the predictors of parent-led iCBT treatment outcomes among autistic youth with anxiety disorders. 

Methods:   

Using data from a randomized clinical trial evaluating two versions of parent-led, therapist-supported iCBT for autistic youth with anxiety-related disorders, one with weekly email contact from a therapist (iCBT-Email), and one with alternating bi-weekly email contact and 30-minute telehealth coaching calls with a therapist (iCBT-Video), this study examined predictors of treatment response in a sample of 57 children. The majority of youth in this study identified as boys (72%) and White (70%). 28% of the participants were of Hispanic or Latino ethnicity. Potential predictors were examined using validated assessments, including repetitive behaviors and restrictive interests (Social Responsiveness Scale 2-Restrictive Interests/Repetitive Behaviors [SRS-RRBs]), pretreatment anxiety severity (Pediatric Anxiety Rating Scale [PARS]), parent-rated externalizing symptoms (Child Behavior Checklist-Externalizing [CBCL-E]), and verbal intellectual ability (Kaufman Brief Intelligence Test-II-Verbal IQ [VIQ]). Hispanic/Latino ethnicity was also evaluated as an exploratory aim. All predictors were examined in hierarchical linear modeling, using PARS scores across the trial (pre-, post-, 1-month follow-up). as the primary treatment outcome measure. There were no differences found between the two groups in the primary study, so groups were combined in this analysis. 

Results:  

Higher baseline anxiety severity (PARS) was significantly associated with more improvement in PARS across the trial, B = -.16, p = .009. Other variables were not significantly associated with differential rates of change in PARS, including CBCL-E, B = .029, p = .070, SRS-RRBs, B = -.023, p = .44, VIQ, B = .001, p = .95, or Hispanic/Latino identity, B = .42, p = .23. 

Discussion:   

Youth who initiated treatment with more severe anxiety experienced more rapid improvement across the trial. This is likely an artifact of regression to the mean, as participants with more severe anxiety before therapy had more statistical opportunity for a decline relative to participants who initiated therapy without as severe anxiety. No other tested pre-treatment characteristics were associated with differential outcome; more externalizing behaviors trended towards predicting poorer outcomes, and it may be that follow-up work with larger samples would identify this factor as predictive.