ABSTRACT
Background and objective: Motivational interviewing (MI) was originally developed to treat problematic drinking but is increasingly integrated into treatment for anxiety disorders. A causal model has been proposed which suggests technical and relational factors may account for the efficacy of MI. The technical hypothesis suggests that therapist MI-consistent behaviours are related to client change talk, and change talk is linked to treatment outcome. Research examining the technical hypothesis has typically been conducted in MI for substance use; therefore, the current study aimed to explore the technical hypothesis in MI for social anxiety disorder (SAD). Method: Participants diagnosed with SAD (n = 85) each received MI prior to receiving group cognitive-behavioural therapy (CBT). MI sessions were coded for behaviours relevant to the MI technical hypothesis. Results: The proportion of MI-consistent therapist behaviours and reflections of change language significantly predicted the proportion of change talk by the client during MI sessions; however, therapist and client behaviours did not predict treatment outcome. Conclusion: The findings support one path of the MI causal model in the context of social anxiety, though indicate that the occurrence of these behaviours during an MI pre-treatment may not extend to predict treatment outcome following CBT.
Supplemental data
Supplemental data for this article can be accessed here at https://doi.org/10.1080/10503307.2020.1751892.
Correction Statement
This article has been republished with minor changes. These changes do not impact the academic content of the article.
Notes
1 There were four outlying MI sessions that contributed to the wide range, two particularly short (23.36 and 24.44 min), and two particularly long (73.04 and 89.21 min). The discrepancy in length of sessions was due to two participants arriving late to the session (short sessions) and study administration occurring within the session (i.e., administering questionnaires rather than session content; long sessions).
2 The Optimal Design computer program was utilized to calculate power for multilevel model analyses in the current study (Raudenbush, Citation2011). Given that the number of treatment groups in the study was fixed, the MDES approach was used. This approach computes the minimum effect size that can be detected at a particular level of power for a pre-specified sample size. With power set at .80, and the predicted intra-class correlation at the group level set at 0.10, a moderate effect size could be achieved with 24 groups.
3 MITI standards for beginning proficiency: Percent MI-Adherent = 90%; Percent open questions = 50%; Percent complex reflections = 40%; Reflection-to-question ratio = 1.0; Global score ratings of Empathy, Direction and MI Spirit = an average of 3.5.
4 In order to examine whether or not each type of motivational language made distinct contributions to client outcome, we conducted additional analyses for which we entered the percentage of change talk (change talk/total utterances) and the percentage of counter-change talk (counter-change talk/total utterances) in separate models. Neither the percentage of change talk (t (91.49) = −0.18, p = .859) or the percentage of counter-change talk (t (90.94) = 0.20, p = .845) predicted CBT treatment outcome.
5 In order to examine whether outcome analyses were affected by missing data due to participant dropout, we utilized the pattern-mixture approach detailed by Atkins (Citation2005). A dummy variable for participant drop-out was added to each multi-level model as a predictor, along with the interaction term between drop-out and predictors. In both the therapist behaviour and client language models, the drop-out variable and it’s interactions were non-significant.