Decisional Conflict and Regret | 45 2 When at least four sufficiently homogeneous quantitative studies were available, we calculated pooled prevalence estimates for DC and DR, using the GLMM model because of high variance in events and sample sizes26. Since considerable heterogeneity was expected, considering different tumor locations and treatments, as well as differences between countries and hospitals, we adopted a random-effects model. The betweenstudy component of variance τ2 was estimated by use of the DerSimonian–Laird method. Subgroup analyses were performed on type of used instrument. All analysis were done in R (version 4.2.2) and Rstudio, using packages ‘meta’27 and ‘metafor’28.
RkJQdWJsaXNoZXIy MTk4NDMw