rfdist in metan does not work

[Alex Lehr]

Hello,

I'm currently on a set of data and conducting a meta-analysis with metan. For the results I do need the prediction intervals but I always get the following error:

Error in predictive interval data
Error in metan.BuildResultsSet
(Note: meta-analysis model was fitted successfully)
r(9);

This is the simplified code I use:

gen logRR=log(RR)
gen loglc=log(lc)
gen loguc=log(uc)

metan logRR loglc loguc if analysis==1, lcols(author year cases_total) random eform rfdist

I seriously do not know what's the problem, really would appreciate it if someone could give me a hint.
Thanks in advance and all the best,

Alex

[carlo custodero]

Hi all,

I have the same issue with my Stata version 14.1.

Hope someone can help us!

Best,

Carlo

[Giorgia Sulis]

Hi,

I am experiencing the same problem on Stata 17 and was wondering if anyone had found a solution over the past few months.
Many thanks

Giorgia

[David Fisher]

Dear all,
Firstly, apologies for the late response!
Unfortunately, I am unable to recreate this error (under Stata v17.0). I am aware that older versions of Stata sometimes encountered issues with metan in the past, but I thought this had been fixed.
Could you check what version of metan you are using, by typing into Stata which metan ? It should say "version 4.04 16aug2021". If not, please update (e.g. by typing adoupdate metan, update ).
If you are definitely running the latest version but are still seeing the error, could you provide some example data e.g. using dataex ?
Many thanks,
David.

[Mark Ebell]

Didn't work for me either, using latest version of you package and Stata 18:
*! version 4.08 David Fisher 17jun2024
*! version 4.08.1 David Fisher 12jul2024

Get same error. Prefer not to post our data publicly but happy to share .dta file with you via email.

Best, Mark

[David Fisher]

Hi Mark Ebell ,
Apologies, the focus of my work has changed in recent months and I haven't been checking the Forums as much as I'd like.

As I say, I'm unable to recreate this error, although I see that multiple users have reported it.

If you are happy to send me a data file, my email can be found within the information shown after typing: ssc describe metan
Alternatively, if you (or another reader!) could isolate a small sample of data which still triggers the error, which you could post publicly (modified if necessary), that would be great.

Many thanks,
David.