App-based cognitive therapy for social anxiety disorder: A randomized-controlled comparison of guided and unguided app-based treatment versus waitlist

Internet-based cognitive behavioral therapy has demonstrated encouraging outcomes for patients diagnosed with social anxiety disorder (SAD). However, the impact of additional therapist-guidance remains debatable. The present study evaluates the effectiveness of the cognitive therapy app “Mindable: Social Phobia” comparing therapist-guided (vCT) and unguided self-help app (iCT) against a waitlist control group (WLC). We assigned 164 patients diagnosed with SAD randomly to one of the three groups, treatment lasting 12 weeks with 8 manualized video-conference delivered sessions in vCT. Assessments took place pre-treatment, midtreatment, posttreatment and follow-up 6 months after treatment (follow-up). The primary outcome was clinician-rated symptom severity (Liebowitz Social Anxiety Scale; LSAS), secondary outcomes included symptoms of depression and interpersonal pleasure. Drop-out rates were 9.3% in WLC, 42.6 % in iCT and 15% in vCT ( χ 2 = 19.5, p < .01). Both treatments reached significant stronger reduction of LSAS scores compared to WLC in all analyses (d [iCT vs. WLC] = -1.01, d [vCT vs. WLC] = -1.1). Using multiple imputation analyses, no significant difference was found in the LSAS between the two treatment groups ( t = -0.22, p = .830, d = 0.04). Patterns of results were similar for secondary outcomes. At follow-up, both treatment groups maintained treatment gains. Overall, the app demonstrates promising results, with patients clearly benefiting from the intervention. The add-on effects of the therapist-guidance favor provision of blended care as it prevents drop-out. • Cognitive therapy, delivered via self-help app, is effective with or without guidance • Therapist guidance is helpful to reduce the rate of attrition and enhance client satisfaction • Patients may also benefit regarding secondary outcomes such as depressive symptoms • The app's usage patterns significantly predict attrition