Ires Ghielen

102 Chapter 6 and corresponding standard deviations measures were used to calculate Hedges’ g . Means and standard deviations from anxiety, depression, and general mental health outcome measures within each study were pooled within the CMA program so that one ‘psychological distress’ measure for each study was included in the meta- analyses. Two separate main meta-analyses were conducted: the first to investigate psychological interventions that were compared with waitlist or TAU conditions, the second to investigate psychological interventions that were compared with other active interventions (such as supportive listening, relaxation, and psycho- education). Within the first main meta-analysis, besides the combined psychological distress measure, the individual effect sizes on anxiety, depression, and general mental health outcomes were investigated using separate smaller meta-analyses. Subgroup analyses were conducted for disease type, control condition, and high vs low risk of bias. In addition, the relationship between risk of bias and effect size was investigated with a regression analysis. Within the second main meta-analysis, the different types of interventions of interest (CBTs and MBTs) were investigated by performing two separate meta-analyses. There were too few studies to perform further subgroup analyses. As considerable heterogeneity was expected, all analyses were conducted using the random effects model. The I² statistic was calculated as an indicator of heterogeneity. We calculated the 95% confidence intervals (95% CI) around I² [25] using the non-central chi-squared based approach within the heterogi module for Stata [26]. When the I² estimate reached 40%, this was classified as considerable heterogeneity [27]. Subgroup analyses were conducted according to the mixed-effects model [28], and the meta-regression analysis was conducted according to the procedures developed by Borenstein et al. [28]. Publication bias was examined with Duval and Tweedie’s trim and fill procedure which estimates how many studies are missing in the meta-analyses and then imputes these [29], as well as Egger’s test for the asymmetry of the funnel plot. The protocol of this meta-analysis was not pre-registered.