Funnel plot usually gives you an idea of publication bias via visual inspection; however I came across Begg's or Egger's regression tests. Do I need to do both visual and mathematical assessments? Or, visual inspection through funnel plot would be enough?
In general, if you have enough studies, it is preferable to move towards formal statistical tests. You can find concise and sound explanation in the Cochrane Handbook.
However, visual inspection is always a good habit.
http://handbook.cochrane.org/chapter_10/10_4_3_tests_for_funnel_plot_asymmetry.htm
Thanks very much Dr. Paolo for your answer. As I have only 9 included studies; shall I then proceed even funnel plot? According to the Cochrane handbook "As a rule of thumb, tests for funnel plot asymmetry should be used only when there are at least 10 studies included in the meta-analysis, because when there are fewer studies the power of the tests is too low to distinguish chance from real asymmetry".
The performance of the tests depend also on the type of outcome and the between-trial heterogeneity. I suggest you to perform the test and see what happens, if no evidence of publication bias is found look at funnel plot.
See also the attached references:
Sterne JA1, Gavaghan D, Egger M. Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature. J Clin Epidemiol. 2000 Nov;53(11):1119-29.
Harbord RM, Egger M, Sterne JA. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints.Stat Med. 2006 Oct 30;25(20):3443-57.
Macaskill P1, Walter SD, Irwig L. A comparison of methods to detect publication bias in meta-analysis. Stat Med. 2001 Feb 28;20(4):641-54.
you should consider a trade off between all issues. In one hand it depends to the number of accessible papers. On the other hand there is issue of the validity of data. Funnel plot gives you only the idea or a general judgment. Calculations are also estimations but are more accurate.
Dear Prof. Khoshdel,
Thanks very much for your response. Could you please suggest me how can I proceed to calculations of Begg's or Egger's regression? Please recommend any software by which I can assess this.
I think we have a text on how to interpret the results of a meta-analysis that is of interest of you since it also covers several publication bias analyses.
Also, we have developed a simple and free tool to perform a meta-analysis in Excel. It covers all the common publication bias analyses, including the Egger Regression and the Begg and Mazumdar Rank Correlation.
Everything is freely available from http://www.meta-essentials.com
Good luck with your meta-analysis!
Article How to Interpret Results of Meta-Analysis
Deleted research item The research item mentioned here has been deleted
Great! Thanks very much Dr. Rhee. I am sure its going to help me a lot.
Inspection of the funnel plot may be somewhat problematic, as different people come to different conclusions (Bax et al., 2009). Therefore, it may seem appropriate to use a regression test. However, funnel plot asymmetry doesn't necessarily mean that there is publication bias (Field & Gillett, 2010).
If you want to decide whether your results are robust then you could consider other established methods:
- trim and fill analysis
- sensitivity analysis (gray literature vs published studies)
- fail-safe N
With regard to the statistical software: All of the procedures are implemented in R (i.e. metafor package)
http://www.metafor-project.org/doku.php
Article More Than Numbers: The Power of Graphs in Meta-Analysis
Article How to do a meta-analysis
Please go through the 'recommendations on testing for funnel plot asymmetry.'
Cochrane Handbook for Systematic Reviews of Interventions, Version 5.1.0
Editors: Julian PT Higgins and Sally Green
Good Luck
http://handbook.cochrane.org/chapter_10/10_4_3_1_recommendations_on_testing_for_funnel_plot_asymmetry.htm
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis