ArticlesLanguage bias in randomised controlled trials published in English and German
Introduction
Systematic review of evidence about the benefits and risks of medical interventions can influence decision-making in clinical practice and public-health medicine, identify areas in which further research is needed, and guide allocation of resources.1, 2 The dissemination of medical evidence, including the publication of results from randomised controlled trials (RCTs), is influenced by several factors, however, that modify the probability of whether a trial is included in a meta-analysis. Publication bias-selective publication of significant findings and the non-publication of those without such findings-has been documented repeatedly.3, 4, 5 (The word “significant” relates to statistical significance here and in the rest of the article.) Consequently, only biased samples of all the existing evidence are likely to be publicly available.
Several factors influence the probability of whether a published study is included in a systematic review.6, 7, 8 One factor that has received little attention is the language in which a paper is published.9 Investigators working in non-English-speaking countries publish some of their work in national journals. Authors may be more likely to report in an international, English-language journal results that are significant, whereas other findings are more likely to be published in local journals. English-language bias could, therefore, be introduced in systematic reviews and meta-analyses that are based exclusively on reports written in English. Although concern has been expressed about this type of bias,9, 10, 11 its significance in meta-analytic research is unclear at present.
In an attempt to identify all published controlled trials, the Cochrane Collaboration has embarked on an extensive manual search of many medical journals published in languages other than English.12 We manually searched through five leading German-language general-medicine journals13 and, at the same time, did a bibliographical study. Our objectives were to describe publication trends and quality features of RCTs done in German-speaking Europe and to assess whether trials with significant results are more likely to be published in an international English-language journal than trials without significant findings.
Section snippets
Methods
Five leading general-medicine journals published in German-speaking Europe were searched manually for RCTs. We searched thoroughly each issue of Deutsche Medizinische Wochenschrift, Schweizerische Medizinische Wochenschrift, Schweizerische Medizinische Rundschau (Praxis), Wiener Klinische Wochenschrift, and Wiener Medizinische Wochenschrift published between 1985 and 1994. A trial was defined as an RCT if assignment of participants to treatment and control groups was described as randomised by
Results
255 trials were identified through the manual search. 32 (13%) reports were excluded (23 in French, five in English, two in German from Hungary, and one in German from the UK and Croatia). Therefore, 223 German-language reports published by 529 key authors fulfilled the inclusion criteria. Of these, the numbers of trials published each year decreased from 40 in 1985 to eight in 1994. The Medline search on the 529 key authors found 570 English-language reports. The average number of
Discussion
English is the predominant language in contemporary medical research. Investigators outside the English-speaking world who want their work to be recognised have little choice but to attempt to publish in English. We found that clinical trials are more likely to be reported in an English-language journal if they contain significant results whereas other trials were published in national journals in German. In the USA19 and the UK,3 surveys have shown that investigators are reluctant to submit
References (32)
- et al.
Publication bias in clinical research
Lancet
(1991) - et al.
Publication bias in clinical trials
Control Clin Trials
(1987) - et al.
Redundancy, disaggregation, and the integrity of medical research
Lancet
(1996) - et al.
Selecting the language of the publications included in a meta-analysis: is there a Tower of Babel bias?
J Clin Epidemol
(1995) - et al.
Completeness of reporting of trials published in languages other than English: implications for conduct and reporting of systematic reviews
Lancet
(1996) - et al.
Randomised trials in German-language journals
Lancet
(1996) - et al.
Assessing the quality of reports of randomized clinical trials: is blinding necessary?
Control Clin Trials
(1996) - et al.
Assessing the quality of randomized controlled trials: an annotated bibliography of scales and checklists
Control Clin Trials
(1995) - et al.
Ginkgo biloba
Lancet
(1992) Cochrane's legacy
Lancet
(1992)