Nineteen RCTs (n=454 patients) were included in the review. Sample sizes ranged from 10 to 84. The mean Jadad score was 2.11 (SD=0.66). No evidence of publication bias was found.
Psychotherapy control conditions were associated with a statistically significant reduction in anxiety severity across a range of anxiety disorders (mean effect size 0.45, 95% CI 0.35 to 0.46, p<0.001). Significant heterogeneity was detected (Q-statistic 42.78, p=0.001). Mean weighted response and attrition rates were 25.0% and 14.2%.
Subgroup (moderator) analyses: There was no significant difference in the treatment effect across diagnostic groups. Similarly, there was no significant difference in response rates across diagnostic groups, except for between OCD and PTSD (p=0.009). The attrition rate for generalised anxiety disorder was greater compared with the other disorders (OCD, acute stress disorders, panic disorder), but there was no significant difference for PTSD (p=0.06) and social anxiety disorder (p=0.05). There was no significant difference in effect sizes across types of outcome measures (clinician-rated versus self-reported), study year (earlier versus recent) (p=0.12) and number of psychotherapy treatment sessions (p=0.42).