Schwarzer Guido, Rücker Gerta
Institut für Medizinische Biometrie und Medizinische Informatik, Universitätsklinikum Freiburg.
Z Evid Fortbild Qual Gesundhwes. 2010;104(4):306-13. doi: 10.1016/j.zefq.2010.03.016.
Publication bias in meta-analysis occurs if the probability of publication depends on the estimated treatment effect and the precision/size of a study. In this article we describe the funnel plot as a graphical method to evaluate the presence of publication bias. Furthermore, tests of publication bias based on funnel plots are explained, and methods are presented that estimate treatment effects adjusted for publication bias. The article closes with a critical discussion of the various statistical methods.
如果发表的可能性取决于估计的治疗效果以及研究的精度/规模,那么在荟萃分析中就会出现发表偏倚。在本文中,我们将漏斗图描述为一种评估发表偏倚是否存在的图形方法。此外,还解释了基于漏斗图的发表偏倚检验,并介绍了估计针对发表偏倚进行调整后的治疗效果的方法。本文最后对各种统计方法进行了批判性讨论。