Within-person changes in aversive reactivity predict session-to-session reductions in anxiety and depression in the unified protocol

ABSTRACT

The Unified Protocol (UP) theoretically leads to reductions in emotional disorder symptoms by reducing aversive reactions to emotions. However, aversive reactions can take many forms (e.g., non-acceptance, behavioral avoidance). We examined if (1) multiple aspects of aversive reactivity predicted session-to-session changes in anxiety and depression in the UP, (2) these aspects reflected a single latent construct, and (3) changes in this latent construct predicted changes in anxiety and depression. Participants (N = 70, M_age = 33.74, 67.1% female, 74.3% white) completed six sessions of UP modules and measures of aversive reactivity, anxiety, and depression before each session. We used hierarchical linear modeling and random-intercept cross-lagged panel models to test aspects of aversive reactivity and a latent factor of aversive reactivity, respectively, as predictors of session-to-session changes in anxiety and depression. Within-person improvements in four of five aspects of aversive reactivity predicted decreases in anxiety, and improvements in two aspects predicted decreases in depression. However, within-person improvements in latent aversive reactivity predicted decreases in anxiety at five sessions and in depression across all sessions. These results add to the growing literature highlighting the role of aversive reactivity as a potential transdiagnostic process involved in improvements in emotional disorder symptoms during treatment.

Keywords:

• within-person
• transdiagnostic
• mechanisms
• aversive reactivity
• anxiety
• depression

Acknowledgements

The authors would like to thank Patrick Goh and Caitlyn Hood for conducting therapy sessions. The authors would also like to thank Destiney MacLean, Anna Garlock, and Alex Urs for their project assistance throughout the study. The authors received no funding from an external source.

Disclosure Statement

Shannon Sauer-Zavala receives royalties from Oxford University Press in her role as an author of the Unified Protocol. The other authors declare no conflicts of interest.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10503307.2023.2254467.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Notes

1 Krippendorff’s αs ≥ .80 indicate reliable variables; αs between .67 and .80 indicate tentative reliability (Krippendorff, 2004).

2 Although half of participants were randomized to discontinue treatment after session 6, they continued to complete the same pre-session measures on a weekly basis until week 12. We thus included data from session/week 7 to capture effects of session 6 content for all participants.

3 We included the non-target symptom as a covariate to better specify the effects of aversive reactivity on the target symptoms. We included session number to account for time in line with Wang and Maxwell’s (2015) recommendations. We included the sequencing condition indicator variable, using the standard sequencing condition as the reference category, to account for effects of module sequences. We included a therapist indicator variable to account for therapist effects rather than a random effect of therapists because, with only four therapists, the models with random effects of therapists did not converge.

4 To test if these results held when including all available sessions, we repeated the above analyses using data from all available sessions from all participants (Tables S11–S20). We found three notable differences: in contrast to results from the first 7 sessions, within-person improvements in negative beliefs about emotions, B = .04, SE = .02, p = .04, 95% CI [.001, .09], R² = .01; behavioral avoidance, B = .07, SE = .03, p < .01, 95% CI [.02, .12], R² = .03.; and cognitive flexibility, B = −.13, SE = .04, p < .01, 95% CI [−.20, −.05], R² = .03, significantly predicted session-to-session reductions in depression. The number of available sessions differs between our analyses of the first 6 sessions and all available sessions, but the total sample size does not, because participants completed all measures at every session.

5 We repeated the above analyses using data from all available sessions. The results did not substantively change with two exceptions: within-person reductions in anxiety significantly predicted session-to-session reductions in behavioral avoidance, B = .21, SE = .08, p = .01, 95% CI [.05, .37], R² = .07, but within-person reductions in depression did not significantly predict session-to-session reductions in behavioral avoidance, B = .13, SE = .08, p = .09, 95% CI [−.02, .27], R² = .04.

6 Because it was unclear at what lag different constructs exert different effects, we also performed analyses of these data using an RI-CLPM with aversive reactivity at session t–1 predicting anxiety/depression at session t, controlling for anxiety/depression at session t–1. Although in the expected direction, aversive reactivity was not a significant predictor of next-session changes in anxiety, B = .02, p = .62, or depression, B = .10, p = .13.