I am currently doing a meta-analysis on outcome of arthroplasty and come across different outcome measure across the literature. After discussion with a generic statistician, the suggestion is either perform Standard Mean Difference (SMD) or Ratio of Mean (RoM) for pooling the data of different outcome.
Cochrane has described the use of RoM; however, I am not able to find a way of performing the meta-analysis on Review Manager Softward.
Any idea how do we perform RoM meta-analysis?
I believe you mean "Response Ratio" when you say "Ratio of Mean". It is discussed in Introduction to Meta-Analysis by Michael Borenstein et al (2nd edition, 2021; pages 30-31). The overall effect size is calculated using the log of the response ratio for each study. I do not use RevMan so I do not know if it can calculate the overall effect size from study response ratios. It can be done using the Comprehensive Meta-Analysis software. There is an article available in ResearchGate that discusses the calculation; it is attached.
This may help. https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/1471-2288-8-32
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis