The inclusion of reports of randomised trials published in languages other than English in systematic reviews.

摘要: Objective To assemble a large dataset of language restricted and inclusive systematic reviews, including both conventional medicinal (CM) complementary alternative medicine (CAM) interventions. then assess the quality these reports by considering comparing different types reviews their associated RCTs; CM CAM interventions; effect restrictions compared with inclusions, whether results are influenced other issues, statistical heterogeneity publication bias, in review process. Data sources MEDLINE, EMBASE, Cochrane Database Systematic Reviews Centralised Information Service for Complementary Medicine. Review methods Three were included: restricted; inclusive/English (EL) that searched RCTs languages than English (LOE) but did not find any and, hence, could include any, quantitative data synthesis; LOE included them synthesis. Fisher's exact test was applied to compare three respect reporting characteristics assessment tool. The odds ratio trials versus EL computed each this information pooled across examine influence type intervention (CM, CAM) have on estimates effect. Several sensitivity analyses performed. Results predominantly French German. Language inclusive/LOE highest reviews. higher There only minor differences eight considered. However, there inconsistent depending type. results, those reported previously, suggest excluding from analytical part is reasonable. Because present research previous efforts every RCT resulting possibility uncertainty as when bias will be LOE, it always prudent perform comprehensive search all evidence. This result applies investigating benefits does imply reviewers should neglect LOE. We recommend regardless language. may merit some aspects process although decision likely depend several factors, fiscal resources being available. significantly shift an intervention's effectiveness CAM. Here, inclusion, resulted reduced do appear bias. Conclusions With exception recently published less optimal. marker better review. effectiveness. substantial if excluded it.