Internet-Delivered Cognitive Behavioral Therapy for Social Anxiety Disorder: Nationwide Retrospective Cohort Observational Study

J Med Internet Res. 2026 Sep 10;28:e85676. doi: 10.2196/85676.

ABSTRACT

BACKGROUND: Internet-delivered cognitive behavioral therapies (iCBTs) can address the accessibility and affordability limitations of conventional therapy. Although iCBTs for social anxiety disorder (SAD) are efficacious, their effectiveness in routine care remains less well established. Helsinki University Hospital (HUS) provides a novel nationwide iCBT program for SAD.

OBJECTIVE: We aimed to assess whether this 7-session, flexible time-scheduled, therapist-supported HUS-iCBT for SAD, delivered by a specialized clinic, (1) is effective in routine care and (2) can enhance adherence, compared to previously published programs with ≥9 sessions.

METHODS: This retrospective, registry-based, observational cohort study used data from the HUS-iCBT register from 2019 to 2022. For treatment, the inclusion criteria were a diagnosis of SAD and possession of an email address and an official e-identification. Severe personality disorders, suicidal behavior, acute psychosis or mania, and disorders with cognitive decline were exclusion criteria. Altogether, 2360 physician-referred patients were accepted for therapy. For the study, we applied an additional criterion of ≥21 baseline score on the Social Phobia Inventory (SPIN, primary outcome) to verify the exclusion of subclinical conditions. Patients aged less than 18 years, with missing baseline information, who did not log-in, or reentered treatment were excluded. The remaining patients (n=1654) formed the modified intent-to-treat (mITT) population, provided they completed ≥1 on-therapy sessions. Secondary measures were Overall Anxiety Severity and Impairment Scale and Patient Health Questionnaire-9.

RESULTS: Linear mixed effects modeling revealed a large SPIN-measured effect size for mITT and completer populations (Cohen d=1.09, 95% CI 1.03-1.16 and d=1.15, 95% CI 1.06-1.23, respectively). Late and early dropouts also benefited from the treatment, with medium and small effect sizes, respectively (d=0.41, 95% CI 0.32-0.50 and d=0.72, 95% CI 0.58-0.86). Reliable Change Index analysis showed improvement in 679 out of 1654 (41.1%, 95% CI 38.7-43.5) patients in the mITT population and 516 out of 893 (57.8%, 95% CI 54.5-61.0) patients in the completer population. Completers exhibited improvement also on secondary measures. Patients completed a mean of 5.74 (SD 1.99; 82%) of all 7 sessions (vs 59% and 68% in earlier programs with ≥9 sessions and 85.3% in one 6-session program). In the mITT population, 893 out of 1654 (54%, 95% CI 51.6-56.4) patients were completers. Independent sample t test was applied for comparisons between groups. Missing data were analyzed with the Little MCAR (missing completely at random) test. Multiple testing was compensated using the Benjamini-Hochberg false discovery rate. Cronbach α was calculated at baseline and end point for outcomes.

CONCLUSIONS: The innovative 7-session HUS-iCBT for SAD (vs ≥9 sessions in most previous programs) may improve adherence without compromising effectiveness. Task shifting from highly specialized to less specialized professionals may further improve cost-effectiveness, although formal health-economic evaluation is needed. These findings may inform the development of future iCBT programs for mental disorders.

PMID:42721423 | DOI:10.2196/85676