Fifty-six studies were included in the review (4,713 participants). The average attrition rate was 16.5% (range zero to 54.8%). Thirty studies used an intention-to-treat analysis.
Compared with waiting list or placebo, self-help treatment was associated with a moderate to large effect size (g=0.78, 95% CI 0.67 to 0.90; 54 studies) for reduction of anxiety symptoms. Self-help treatment compared with face-to-face treatment showed a small effect that favoured face-to-face treatment (g=−0.20, 95% CI -0.37 to -0.02; 28 studies). Significant heterogeneity was found for both outcomes (Ι²=48.32% and Ι²=66.85%).
When self-help treatment was compared to waiting list or placebo, subgroup analyses showed that self-help treatment format, different primary anxiety diagnosis and recruitment settings were significantly associated with treatment outcome in bivariate analyses. There was no significant relationship between treatment outcome and use of intention-to-treat analysis. Only the variable of recruitment settings remained significant in a meta-regression analysis.
The meta-regression analysis also indicated that type of control groups, treatment format and patient gender were significantly associated with treatment outcome when self-help treatment was compared to face-to-face treatment.
There was potential for publication bias. Results of other outcomes were reported.