Metan code: How can I put the value of the average effect size in the x-axis?

[KHALID MEDANI]

I am using the metan to create this forest plot . what statement should I use to put 0.033 (the average effect size) in the x-axis (under the dotted red line, see attached image) ? Thank you all.

[Anne Dam]

Dear Khalid,
I recently did this but I used meta instead of metan, since metan gave me formatting issues. But I dont know if you have a version of Stata that can run meta?

For meta it would be something like

meta     esize nc meanc sdc nt meant sdt ,  esize(hedgesg) studylabel(Author)  //calculate your own mean, sd and n for control and treatment, use "author" as studylabel and whatever effectsize you are interested in

meta forestplot , ///
         esrefline(lp(dash)) /// <--- this is part that inserts the mean as a line, and here it is dashed
         title("{bf:Hemorrhagic events...}: ") ///
         subtitle("{it:Aspirin and Trigacelor}") ///
         name(p1, replace)

I hope this can be helpful

If you can run meta in your stata version, there is a good example video here from stataCorp LLC:
https://www.youtube.com/watch?v=8zzZojXnXJg
and

help meta

[David Fisher]

KHALID MEDANI
Here is one way of doing it. I've based the code on the first example from the metan help pages, but it should be fairly straightforward for you to modify for your own use case.
Here's how it works:
1) The first call to metan carries out the meta-analysis and prepares the forest plot but does not draw it; but it saves the plot text size in a local macro.
2) The second call to metan adds the extra tick-line and text to the plot, using the saved text size to make it consistent with the rest of the plot.
You may need to play around with the twoway options tlength() (and possibly also tlabgap()) to get it looking right. Let me know if you have any issues.

use http://fmwww.bc.edu/repec/bocode/m/metan_example_data, clear
metan tdeath tnodeath cdeath cnodeath, rd random label(namevar=id, yearvar=year) counts forestplot(nodraw savedims(myplotdims))
local textsize = myplotdims[1, "textsize'"]
qui metan tdeath tnodeath cdeath cnodeath, rd random label(namevar=id, yearvar=year) counts forestplot(xmlabel(`r(eff)', tlength(*6) labsize(`textsize') format("%05.3f")) xlabel(,format("%03.1f")))

Anne Dam
As co-author and maintainer of the metan package, I was interested to read your comment. Could you explain a little more what you mean by "metan gave me formatting issues"? In particular, are you using a version of metan released since the end of 2019? Thanks!

[Anne Dam]

Hi David Fisher
Yes of course, I just saw your message. I will have to dig up my old data and old plot to re-create the issue so it will take some more time. I will send you a message, since I'm never sure how to find this post again, once it has disappeared from the first page under the 'General' Forum..