Psychotherapy and Pharmacotherapy for Social Anxiety Disorder: A Comprehensive Meta-Analysis

Abstract

Social anxiety disorder is the fourth most prevalent mental disorder in the US. Over the past several decades, psychotherapeutic, specifically cognitive behavioral, and pharmacologic approaches have been found efficacious for social anxiety disorder. A number of meta-analyses have been conducted since 1995 examining the efficacy of cognitive behavioral therapy (CBT) and/or pharmacotherapy for social anxiety disorder. Though there have been numerous trials in the past decade, no meta-analysis examining both psychotherapy and pharmacotherapy for social anxiety disorder has been published since 2001. For the present study, a comprehensive literature search produced 93 publications featuring 94 controlled trials (N = 11,503), which were included in the final analyses. We found a moderate to large effect size for all active treatments compared to control conditions. Significant heterogeneity among treatment effects was evident, largely accounted for by true variation between effects, versus standard error. Examination of potential study characteristic moderators indicated that treatment type (CBT, medication, combination), analysis type (intent-to-treat vs. completer), funding source, type of screening interview, type of treatment clinic (academic or private), version of diagnostic criteria, type of social anxiety sample (generalized social anxiety disorder only vs. mixed sample of generalized and specific social anxiety disorder) and type of inclusion/exclusion criteria related to other anxiety disorders were significant moderators. Publication type, inclusion/exclusion criteria related to depression and substance abuse/dependence, and full sample comorbidity with another disorder were not. Treatment type was no longer a significant moderator once control condition was accounted for. In psychotherapy trials, self-exposure (as compared to all other types of CBT) and psychotherapist training were significant moderators, whereas variables corresponding to treatment modality and delivery were not. Medication class and specific drug type were significant moderators for pharmacotherapy studies comparing an active treatment to a control condition. Head-to-head comparisons, which included trials comparing active treatments, indicated no differences between psychotherapy, medication, and the combination of the two. Further, social anxiety treatment had moderate to large effects on depression and quality of life.