Using R for meta-analysis

[lytong22]

I would love to learn more about using R for meta-analysis, and more specifically the metafor package.

[ekothe]

@lytong22 is there something specific you want to know more about?

[lytong22]

Thanks for the reply. I am conducting a meta-analysis for a systematic review, and the included studies report a mix of both numerical and categorical data. Here are a few things I would be interested in:

• If the included studies report both numerical and categorical data, is it possible to pool the effect sizes together? (I know that I can convert all of the effect sizes into standardised mean difference, but can't seem to locate that feature yet in available R packages).
• How to store data extracted from the included studies so that future updates of the review are possible?
  I am pretty new to using R for meta-analysis in general so would love to hear other ideas/people's experiences.

[kristyrobledo]

@lytong22 :
The Cochrane group have a lot of resources on meta analysis principles. This would be were to get a good grounding first on the methods.
?numerical and categorical data - do you mean RR and their SE versus events and denominators? You need clearly defined outcomes and can work from there.

There are a stack of tutorials and vignettes online for performing metaanalysis in R (assuming simple pairwise and not a network metanalysis), but i think its really important to have an understanding of the principles prior to jumping in.

I think the only way to put together the trials is really to use excel... I'm not sure of any other way to manage this easily moving forward.

[ekothe]

I've got a bunch of resources for meta-analysis in R using different packages/approaches - happy to have a chat over lunch if that would help