To change the value of 'Nullrefline' in precomputed effect meta-analysis

[Adolfo Aramburu]

Good morning

I'm conducting a meta-analysis of precomputed effect sizes. When adding a null effect line in the forest plot, it automatically draws the line at 0. However, when dealing with relative risks, I would like the line to be drawn at 1. I could use the 'xline(1)' syntax, but I also want to employ the 'favorsleft' and 'favorsright' options of 'nullrefline'.

The issue I have could be represented in the following example:

webuse bcgrratio, clear
generate double log_rr = log(rr)
generate double log_rrll = log(rr_ll)
generate double log_rrul = log(rr_ul)
meta set log_rr log_rrll log_rrul
meta summarize
meta forestplot, nullrefline(favorsleft(Reduced risk) favorsright(Increased risk) lcolor(red))

Any suggestions will be welcome.

Thanks

[Houssein Assaad (StataCorp)]

Dear Adolfo,

You can use the eform() option (exponentiated form) with meta forestplot to report risk ratios (instead of log(rr)) and place the null reference line at the exp(0)=1 value. Try:

  
 meta forestplot, eform("Risk ratio") nullrefline(favorsleft(Reduced risk) favorsright(Increased risk) lcolor(red))

I hope this helps!

[Adolfo Aramburu]

Great, I tried it and I was able to solve it.
Thank you.