科研动态期刊摘要

凯·迪克森:循证医学的先驱与Cochrane协作网的创始人

The BMJ · 2026-11-12

摘要

bmj;395/nov12_7/e101440/F1F1f1 1987年,凯·迪克森及其同事发表了一项简短研究,其结果令人不安。1 他们询问了318名已发表随机试验的作者,是否也参与过从未发表的试验。在已完成但未发表的试验中,14%支持新疗法;而在已发表的试验中,55%支持新疗法。编辑并非主要责任方:大多数未发表的试验只是从未被撰写和投稿,通常是因为结果为阴性——即未能拒绝零假设——或者结果缺乏趣味。如果显示获益的试验比显示无效或有害的试验更有可能发表,那么综述、荟萃分析、指南以及依赖它们的临床医生将高估获益并低估危害。

English
Kay Dickersin: pioneer of evidence based medicine and a founder of the Cochrane Collaboration

bmj;395/nov12_7/e101440/F1F1f1In 1987 Kay Dickersin and her colleagues published a short study with an uncomfortable finding.1 They had asked 318 authors of published randomised trials whether they had also taken part in trials that were never published. Among the completed but unpublished trials, 14% favoured the new treatment; among the published ones, 55% did. Editors were not chiefly to blame: most unpublished trials had simply never been written up and submitted, usually because the results were negative—meaning that they failed to reject the null—or were uninteresting.If trials showing benefit are more likely to reach print than those showing no effect or harm, then reviews, meta-analyses, guidelines, and the clinicians who rely on them will overestimate benefits and underestimate ha

内容仅供参考,不构成诊疗建议,不能替代执业医师的诊断。

← 返回元界