Expectancy updating predicts anxiety symptom reduction from exposure therapy: Predictive analyses from a randomized clinical trial.

Our objective was to assess predictors of exposure therapy outcome for social anxiety disorder and panic disorder. Specifically, we investigate whether expectancy updating or fear change during exposures are predictive of long-term anxiety reduction. We conducted a randomized controlled trial of 84 participants evaluating two forms of 9-session exposure therapy. One form of exposure therapy was based on the inhibitory retrieval model (i.e., expectancy updating), and the other was based on habituation principles (i.e., fear reduction). Our treatment outcome measures of anxiety symptomatology included the Liebowitz Social Anxiety Scale and the Panic Disorder Severity Scale; these were analyzed at pre-treatment, post-treatment, and 3-month follow-up. Our putative predictive measures were collected during exposures and included expectancy for the feared outcome and fear/anxiety. From these measures, we calculated expectancy change (post-exposure minus pre-exposure, averaged across exposures), learning rate derived from theory-based computational modeling of expectancy data, and fear change (post-exposure minus pre-exposure, averaged across exposures). We conducted structural equation modeling to assess whether these putative predictive measures predicted treatment outcome. Results showed that expectancy change significantly predicted post-treatment and 3-month follow-up anxiety; similarly, expectancy-based learning rate marginally predicted post-treatment anxiety and significantly predicted 3-month follow-up anxiety. Conversely, fear change did not predict anxiety symptoms. Lastly, expectancy change and expectancy-based learning rate predicted long-term anxiety even when controlling for fear change. In conclusion, reducing expectancy for feared outcomes during exposure predicts anxiety symptom reduction. Exposure therapy could benefit from exposures designed to update expectancies about feared outcomes rather than reduce within-exposure fear.