46 | Chapter 2 Results Study selection The search yielded 27.258 records. Following deduplication, 16.009 unique articles remained. Of the manually screened first 50% of records, 178 publications were retained for full-text screening. Rayyan ratings for the remaining records (n=7180) ranged from 4.5 stars (n=1), 3.5 stars (n=66) to 2.5 stars and lower (n= 7113). After screening the 2000 highest ranked records, the 4.5-stars and thirteen of the 3.5-stars records were retained for full-text screening. After score recalculation, the 660 records with a 2.5 star rating were screened, and none included. After a second recalculation, the remaining 4520 articles received ≤ 1.5 stars. Of these, 500 random records (>10%) were screened, without any inclusions, after which the remaining records were discarded. Thus, 192 studies were retained for full-text review, of which 164 were subsequently excluded, mostly because of ineligible outcomes (n=105), leaving a total of 28 included studies for data extraction. See figure 1. Description of included studies Of the 28 included studies, five had a qualitative design (DC=3, DR=1, DC and DR=1) and 23 were quantitative (DC=3, DR=19, DC and DR=1). Publication dates ranged from 1980 to 2022, and twelve studies were < 5 years old. Two quantitative studies reported on the same study using different but overlapping subsamples (Windon et al29,30); only the data from the larger study was used for metaanalysis29. For ten quantitative studies, we contacted the authors for additional data regarding the prevalence of DR. For two, the raw data were no longer available (Gill et al.31, Shuman et al.32). One study by Shaverdian et al.33 considered DR of a de-escalation treatment instead of regular treatment, and we excluded this study from the meta-analysis on grounds of clinical heterogeneity. Hence, meta-analyses were performed on sixteen articles, using three different instruments (Appendix B): study-specific questionnaires, the Shame and Stigma Scale (SSS) and the Decision Regret Scale (DRS). Qualitative studies (N=5) Risk of Bias Three of the five qualitative studies had a low RoB30,34,35. All had clear aims and a thorough method. One article36 did not comment on the relation between researcher and study participants. Another article37 did not have dedicated results and discussion sections and was therefore rated as high RoB (Table 1).
RkJQdWJsaXNoZXIy MTk4NDMw