Workshop Proposal: Meta-Analysis with R

[wviechtb]

I would be happy to give an entire workshop on how to do meta-analyses with R as part of ESMARConf2022, covering the following topics:

• computation of effect size / outcomes,
• equal- and random-effects models,
• moderator analyses / meta-regression,
• methods related to publication bias,
• some advanced methods like multilevel/multivariate models and robust variance estimation.

This could be just a few hours, but for this to be sufficiently comprehensive, I would really need more time (like a full day). I don't want to 'hog the time' though, so this could also be a 'satellite event'. Happy to discuss this further!

[mcguinlu]

Hi @wviechtb,

This would be amazing - thank you for suggesting it and for volunteering your time!

I'll have to confirm with the other organisers, but I think a full-day satellite event/workshop just prior to the conference would be the way to go. It would mean that anyone interested could have a primer/refresher in meta-analyses with R, and so (hopefully) get more out of the conference as a result.

(Though one important detail is conspicuously absent from your workshop description - will there be smooth jazz??)

[wviechtb]

Sounds like a plan, but let's wait and see what the other organizers think. And yes, there 🎷 shall 🎷 be 🎷 smooth 🎷 jazz 🎷.

[shiduopili]

would you also talk about "network meta-analysis" as well? should be useful in health care settings to compare different interventions.

[wviechtb]

Covering everything mentioned above is a lot even if the workshop runs for a full day. Usually, it takes multiple days to cover these topics in proper detail. But I can try to add some pointers when we get to multilevel/multivariate models since network meta-analysis is a relatively straightforward extension of these models.