Meta forestplot Hartung-Knapp variance correction sub-groups

[Kim van Daalen]

Hello!

I am currently conducting a meta-analysis of a small study number (n=3) with different subgroups (3 disease outcomes). I am using a random effects meta-analysis with a Hartung-Knapp variance correction applied to the DerSimonian-Laird model.

I am trying to create a single forestplot depicting the results of the different sub-groups with the Hartung-Knapp variance correction applied to the overall subgroup effectsizes. However, I am unable to figure this out. When using the code below I get an error message as se(khartung) is applied to the overall effect size obtained not the subgroup overall effectsizes.

option subgroup() may not be combined with option se().

meta set lnrr selnrr, studylabel(author_year_pub)
meta forestplot, eform random(dlaird) se(khartung) subgroup(outcome)

I have also tried to use admetan, but when using hksj there I also receive an error message:

option hksj not allowed
Error in admetan.BuildResultsSet
(Note: meta-analysis model was fitted successfully)

I hope I have explained my problem clearly! Is there anyone that might know a solution to this?

Best wishes,

[David Fisher]

Dear Kim,

I can't speak for meta forestplot, other than to agree with you -- it looks like this combination of options is currently invalid. Possibly this will be corrected in a later version.

But admetan should work (and indeed does work with my own data). Would it be possible to supply a sample of your data and code using dataex, so that I can try to diagnose the issue?

Best wishes,

David.