Sixty-six studies reporting data on 85 samples (n=approximately 7,291, range eight to 927) were included. The median drop-out rate was 10% across studies.
Studies that included intervention and comparator groups reported significant beneficial effects of the intervention (summary ES of 0.11, 95% CI 0.05 to 0.17; 42 samples). There was modest heterogeneity (p<0.10). Studies that reported pre-post data in treatment groups also reported a beneficial effect of treatment (ES 0.27, 95% CI 0.22 to 0.33). There was significant heterogeneity (p<0.001). There was no difference in pre-post data in studies that provided this information for control groups.
Unpublished studies reported larger mean effect sizes (ES 0.39) compared with published studies (ES 0.09), as did those without external funding (ES 0.31) compared to those with external funding (ES 0.08). Studies that did not use educational/motivational sessions reported larger effect sizes (ES 0.24) than those that did (ES 0.02). Other factors were not found to be significantly related to QoL effect size. Funnel plots suggested no evidence of publication bias.