Abstract
Although client-perceived therapist empathy relates to positive therapy outcomes, including in cognitive behavioral therapy (CBT), little is known about how empathy exerts its ameliorative effect. One possible way is by promoting clients’ subsequent homework compliance, a variable that also predicts positive outcomes in CBT. The present study sought to investigate simultaneously, in the context of 43 therapist–client dyads receiving 15 sessions of CBT for generalized anxiety disorder, (1) the association of early client-perceived therapist empathy (averaged over sessions 1, 3, 5) with mid-treatment client homework compliance (averaged over sessions 6, 8, 10); (2) the association of mid-treatment homework compliance on client posttreatment worry severity; and (3) the indirect effect of early perceived therapist empathy on posttreatment worry through mid-treatment homework compliance. Given that clients were nested within therapists, we examined both within- and between-therapist differences in clients’ ratings of therapist empathy and homework compliance, and tested both of these indices as predictors of the relevant dependent variables in a multilevel model. At the within-therapist level (i.e., differences between clients within a given therapist’s caseload), greater early empathy was associated with greater mid-treatment homework compliance. At the between-therapist level (i.e., differences between therapists across all of their cases), greater between-therapist homework compliance was related to lower posttreatment worry. Finally, homework compliance was not found to mediate the relationship between empathy and posttreatment outcome. The results underscore the importance of parsing client and therapist effects, and are discussed with regard to their training and research implications.
Notes
1. Because therapists in the MI-CBT condition received explicit training on empathy as an intervention, studying the effects of empathy in this group represents a different empirical focus than our interest here in examining how natural variability in therapist empathy relates to the well-known variability in client homework compliance, and how both of these variables relate to outcome in traditional CBT.
2. For all clients, criteria were met retrospectively for DSM-5 (American Psychiatric Association, Citation2013), which was under development at the time of this study (the core symptoms of GAD did not change from DSM-IV to DSM-5).
3. A score of 65/80 has been shown to differentiate a diagnosis of GAD on the PSWQ (Behar, Alcaine, Zuellig, & Borkovec, Citation2003). A score of 68/80 was required for study enrollment in the present trial based on moderator analyses from a previous RCT for GAD (Westra, Arkowitz, & Dozois, Citation2009), demonstrating the high severity cutoff to be a score of 68/80.
4. The low internal consistency of the PSWQ at baseline (i.e., .57) is likely due to the restricted range of the intake baseline PSWQ scores in the current trial. That is, a score of at least 68/80 on this measure was required by all clients in order to enroll in the study. Thus, there was very limited variability in clients’ baseline PSWQ scores.