Twenty-four RCTs (n=1,275) were included.
Effect sizes.
Pre-/post-treatment: overall, there was a significantly higher effect size in the intervention group compared with the control group (d=0.66, 95% CI: 0.35, 0.96; t-test 6.92, d.f.=37, p=0.00). The effect size was 0.86 (95% CI: 0.69, 1.03; 24 trials) for treatment intervention, compared with 0.13 (95% CI: 0.03, 0.24; 16 trials) for waiting-list control and 0.58 (95% CI: -1.6, 1.3) for attention placebo control. There was no statistically significant heterogeneity between the trials.
The effect size using the RCMAS outcome measure was 0.83 (95% CI: 0.64, 1.02) in the intervention group compared with 0.22 (95% CI: 0.08, 0.35) in the control group. The effect size using the FSSC-R outcome measure was 0.85 (95% CI: 0.45, 1.24) in the intervention group compared with 0.32 (95% CI: 0.11, 0.54) in the control group. The effect size using the CDI outcome measure was 0.70 (95% CI: 0.54, 0.86) in the intervention group compared with 0.20 (95% CI: 0.07, 0.33) in the control group.
Follow-up effect size: the intervention effect size for studies reporting short term follow-up was 1.36 (95% CI: 0.78, 1.94; 16 trials).
Subgroups: child-focused therapy studies, family-focused therapy studies and individual therapy studies all showed a significant treatment effect size, both post-treatment and at follow-up. Group therapy studies showed a significant treatment effect size post-treatment, but a non significant effect size at follow-up.
Percentage recovery.
Significantly more participants in the intervention group no longer met the criterion for their principal pre-treatment anxiety disorder compared with participants in the control group (t-test 15.06, d.f.=34, p<0.00). Across all active treatments, 68.9% of participants in the intervention group (21 trials) who completed the therapy no longer met the criteria for their pre-treatment anxiety disorder at the end of treatment, compared with only 12.9% who were assigned to a waiting list. The recovery rate increased to 72% for participants in the intervention group who completed treatment at the end of follow-up. Recovery rates based on intention-to-treat were 55.4% after treatment.
In terms of the fail-safe N, 17 additional studies of zero effect would be needed to reduce the overall effect size from 0.86 to 0.50, and 79 studies to reduce it from 0.86 to 0.20.