I am conducting a systematic review and have an observational study. I want to perform some meta-analyses using the already calculated odds ratios. Can I use RevMan for this purpose? If not, do you know of any other website where this is possible?
Bonjour Clara,
You can either backtransform the OR in raw number using a previously reported method (Veroniki AA, Pavlides M, Patsopoulos NA, Salanti G. Reconstructing 2 x 2 contingency tables from odds ratios using the Di Pietrantonj method: difficulties, constraints and impact in meta-analysis results. Res Synth Methods. 2013 Mar;4(1):78-94. doi: 10.1002/jrsm.1061. Epub 2012 Nov 12. PMID: 26053541.).
The other solution (maybe easier) would be to convert raw numbers in OR from other studies and using a statistical software. In R, this can be easily done by using metafor package (see this example using log odds ratio (yi in the example). Assembling Data for a Meta-Analysis of (Log) Odds Ratios [The metafor Package] (metafor-project.org)
En espérant que cela aide.
Dear Clara,
This is not an uncommon problem when we do not have the raw data from some of the studies to be included in the MA. However, it would not be possible to include OR directly in the individual studies, we need exact number, mean, sd etc.for that particular outcome. Possible solutions to this problem are
1. Contact study authors and ask for the raw data
2. If you are not able to get data from authors, then (if you have a good number of studies ( with relevant raw data) qualifying for MA part, then you can exclude the studies without raw data and can keep them till SR part by summarizing the study characteristics and findings in the included studies. However, it would not be possible to carry those studies till MA part.
I think it's still required to have the SE of the OR. So, there are two ways: 1. contact the authors and ask for the count data and 2. impute. R, STATA, Comprehensive Meta-Analysis Software, and Python can solve the problem for you. I suggest Comprehensive Meta-Analysis Software and STATA for your question since they are more user-friendly.
- Badges
- Science topic