Christina Drost

Examination of therapist behaviours in therapist-assisted Internet-delivered cognitive behavioural therapy for generalized anxiety disorder

Doctoral Dissertation Results,Christina Drost 2016

Background: Therapist-Assisted Internet-Delivered Cognitive Behavioural Therapy (TAICBT) for the treatment of generalized anxiety disorder (GAD) is found to be efficacious. The nature of therapist-assistance, however, is not well understood. This research aimed to enhance understanding of the nature of therapist-assistance by examining: 1) therapist behaviours in emails sent to clients treated with TAICBT for GAD and the extent to which these qualitatively overlap with behaviours described in face-to-face therapy; and 2) if therapist behaviours in the current study (using a larger and more diverse therapist sample) compared to therapist behaviours reported in a past study of TAICBT for GAD (Paxling et al., 2013). The relationship between therapist behaviours and treatment outcome, treatment satisfaction, and therapeutic alliance was also examined.

Method: Data was collected as part of an open trial investigating a 12 module TAICBT program delivered to 107 clients presenting with symptoms of GAD. TAICBT was provided by 21 registered therapists or 31 graduate students; therapist-assistance was delivered through secure e-mails. Anxiety was assessed before and after treatment. 

Results: Nineteen therapist behaviours were identified in the secure messages sent from therapist to patients. These behaviours qualitatively overlapped with therapist behaviours that are commonly observed in face-to-face therapy. Alliance Bolstering was the most frequently coded therapist behaviour. Eight therapist behaviours were found that largely overlapped with behaviours reported in a past study of therapist behaviour in TAICBT (Paxling et al. 2013); however, the frequency of these behaviours varied substantially between studies. Correlational analyses suggested that therapists provided more frequent feedback on weekly mood ratings and self-report questionnaires when clients demonstrated less improvement on anxiety measures. Analyses also suggested that therapists were more likely to provide Psychoeducation and Task Prompting when clients demonstrated less improvement on anxiety. Of note, Repair of Potential Alliance Ruptures was significantly and positively associated with higher client ratings of therapeutic alliance. Therapist e-mails received during the first two modules by treatment completers contained significantly more instances of statements coded as Course Procedure and Goal Setting than e-mails received during the first two modules by treatment non-completers. 

Conclusions: Overall, preliminary evidence was found to suggest that: 1) there is both overlap and variability in terms of therapist-assistance provided in face-to-face therapy and across different TAICBT programs, and 2) the relationship between therapist behaviours, treatment outcome, therapeutic alliance, and treatment satisfaction is complex and needs to be examined further. The findings shed substantial light on the nature of therapist-assistance in TAICBT.