摘要
系统评价被认为是当前提供治疗性干预的最佳证据,但是由于无法全面获得相关的研究资料,仍然不可避免的存在偏倚。偏倚的类型主要包括文献发表性偏倚、文献查寻偏倚和文献筛选偏倚,其中最难克服的是文献发表性偏倚。偏倚评估的方法常用漏斗图,控制方法包括预先注册临床试验、发布研究信息、系统评价时全面收集所有发表和未发表的临床试验。
At present,the best evidences of pre vent and treatment can be supplied by systematic reviews,but the literat ure bias is avoided if all the related studies can not be collected as required.The type of bias mainly includes publica-tion bias,literature retrieval bia s and literature inclusion bias,amo ng which the publication bias is the hardest o vercome.A funnel plot is a useful tool to assess bias in systematic review.The methods of control the bias involve d in pre-registering the clinical trials,publishing research information and thor-oughly searching all the published a nd unpublished related studies.
出处
《中国临床康复》
CSCD
2003年第3期368-369,共2页
Chinese Journal of Clinical Rehabilitation
关键词
选择偏性
循证医学
系统分析
selection bias
evidence-based med icine
systems analysis
