Fantastic news - huge congratulations!!! 🎉🍰🍾

I can't say I totally understand everything but the maps are beautiful. Congratulations! 🎉

I'm sure there's more than one regression model people have strong opinions about 😅

Cox regression vs. pooled logistic regression? I'd need another coffee before getting into this...

Cardamom bun availability is arguably the most important aspect to assess here!

Karolinska? Bristol? I guess we're not meant to self-nominate 😂 Honestly though I think there's excellent people spread across so many different institutions - you might be better of finding one or two people whose work you'd like to emulate and just see if you can work with them...

Ultimately it was arbitrary for us, but the review isn't a clinical meta-analysis but an overview of methodological approaches to a specific problem... Good point about the double screening - unclear!

I've only used it informally - so for reviews going into background sections/funding applications etc. We've got a PhD student writing up a paper where we've used it though so I'll report back! I think it's probably more suited for some kinds of questions than others...

Not an ML expert so not sure! I never got beyond this cartoon level of understanding: www.youtube.com/watch?v=k-a2.... The stopping rule is the tricky part, I've only used it for informal reviews. Seems there are some papers though... systematicreviewsjournal.biomedcentral.com/articles/10....

Yes - I've used ASreview and I love it (asreview.nl). It uses ML to present you with more relevant papers earlier on, and then you only screen until a stopping role. You train the model yourself. Free! Open Source!

Thanks! Just goes to show how important a deep understanding of context is for Epi studies, doesn't it! We're slowly starting to see more use of centre and/or GP practice as IVs in pharmacoepi, the idea being that it's basically a proxy for clinician preference. Validity of that likely varies..!

This is interesting to me although I don't work in this area of epi specifically - does centre affect outcome other than through exposure in this context? I've been worried about centre adjustment in my field (RxEpi) in case it's more of an IV - as in: www.sciencedirect.com/science/arti...