| Περίληψη : | Meta-analysis is an objective and quantitative methodology used for the synthesis of research studies conducted in the past on a particular issue, which leads to an overall conclusion. Unfortunately, it is often suspected that the available studies represent a biased subset of the evidence, possibly due to publication bias, statistical heterogeneity or other systematically different effects in small studies which is commonly recognized as a threat to the validity of the results of a meta-analysis. We use studies from a meta-analysis of treatment with serenoa repens versus placebo for men with benign prostatic hyperplasia. Fixed effects and random effects models provide an examination of the sampling behavior of the observed effect sizes .We study the differences in our conclusions using these two regression tools. To detect publication bias we use funnel plots along with forest plots, radial plots and qq plots. Asymmetry in a funnel plot may be a strong indication of publication bias but Egger test for funnel plot asymmetry is used as well to examine whether the association between estimated intervention effects and the standard error of the intervention effect is greater than might be expected to occur by chance. Then we represent two different methods to address publication bias; the trim-and-fill method and the Copas selection model. However, the trim-and-fill method drawback is that it is based on strong assumptions about the symmetry of the funnel plot. The Copas selection model shows how the random effects estimate of the overall log-odds ratio in favor of treatment, varies as the chances of small studies being published depends on their results. This approach is proved to work better than the trim and fill approach. We illustrate our results using the studies from a small meta-analysis of benign prostatic hyperplasia and programming in R language.
|
|---|
Copyright © 2013 Βιβλιοθήκη Ο.Π.Α. Με την επιφύλαξη παντός δικαιώματος.