Does fear reduction predict treatment response to exposure for social anxiety disorder?

Fear activation and reduction have traditionally been considered important mechanisms of exposure therapy. Evidence to date is mixed and impeded by inadequate methodology. This study examined the extent to which fear activation and reduction within and across exposures predicted treatment outcomes for social anxiety disorder within a paradigm suitable for their measurement.

Sixty-eight adults with social anxiety disorder and fear of public speaking completed seven exposure sessions, each consisting of seven speeches conducted in virtual reality. Exposures were identical in duration, task requirements, and virtual public speaking situation. Fear was measured with skin conductance and subjective distress ratings. At baseline and post-treatment, participants completed a public speaking behavioral approach test with a panel of confederate judges; subjective fear was measured. A standardized questionnaire of anxiety symptoms was administered at baseline, post-treatment, and one-month follow-up.

No indices of within- or between-session fear reduction, measured by subjective distress and skin conductance response, predicted treatment outcome. One measure of fear activation was associated with outcomes such that less activation predicted greater symptom reduction; remaining indices did not predict outcomes.

Data were collected in the context of a randomized controlled trial of scopolamine; drug group was included in analytic models to account for drug influence. VR exposures elicited mild levels of distress that may underestimate levels of distress in clinical settings.

Findings failed to support fear activation or reduction within or across exposure sessions as significant predictors of treatment outcome for social anxiety. Treatment implications are discussed.