Hi everyone, I recently published a paper on cochlear implant outcomes among patients with congential malformations and was working with RevMan, a software provided by the Cochrane Collaboration. I was wondering if anyone had thoughts on RevMan software and it's particular functionality in calculating the I^2 ("i-squared) statistic.
Thanks.
I wouldn't study R if the only need is I^2 calculation. Actually, you can calculate it even in MS Excel (just for fun) - it is not difficult, really. Here is the paper on it: http://www.biomedcentral.com/1756-0500/5/52 All formulas are also nicely explained by Dr. M.Borenstein in his book "Introduction to Meta-Analysis".
You could also use some freeware like MetaXL: http://www.epigear.com/index_files/metaxl.html
I thought that RevMan gives you the I-squared estimate, but not the confidence interval. It's also easy to work out I-squared based on Q and the df, both of which it gives you as well. The Higgins paper cited by Chris should have the simple formula, and a helpful guide to interpretation. What aspect of functionality were you concerned about?
Hi Michael. I'd suggest additional tools for I^2. I'd avoid RevMan unleast you are in a Cochrane team. I prefer working with both CMA (Comprehensive Meta-Analysis) and MIX. Try them. I^2 calcultation is really very easy with them.
To add to Mikhail's comment, another useful and freely available software is Meta-Analyst. I have used it for a few meta-analyses and compared the results using SAS. The results were comprable. Please see the website and publication for more detail:
http://www.cebm.brown.edu/open_meta
Reference: BMC Med Res Methodol. 2009 Dec 4;9:80. doi: 10.1186/1471-2288-9-80
The I2 is provided with the forest plots as an useful estimates of heterogeneity. I further recommend metan from Stata version 12.0. I've recently finished two metaaanalyses using this software and I got a good experience. Besides, Higgins's article is very useful.
Regards
Another (free) option that I have used extensively for meta-analysis is the metafor package for R. This takes a bit of learning, but is well documented and much more versatile than Revman. Amongst all of the other things that it does, it will give the I^2 statistic.
Simon R.K is right. R is a very useful and useful resource for conducting meta-analysis.
I would ike to share these tools:
Chapmann & Hall. Applied Meta-Analysis with R. 2013
Brian S. Everitt and Torsten Hothorn. A Handbook of Statistical Analyses Using R — 2nd Edition
I hope these can be useful for you all
Sincerely
Jose F. Meneses-Echavez
I wouldn't study R if the only need is I^2 calculation. Actually, you can calculate it even in MS Excel (just for fun) - it is not difficult, really. Here is the paper on it: http://www.biomedcentral.com/1756-0500/5/52 All formulas are also nicely explained by Dr. M.Borenstein in his book "Introduction to Meta-Analysis".
You could also use some freeware like MetaXL: http://www.epigear.com/index_files/metaxl.html
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Computer Software