Meta-analysis: how to use fixed-effect and random-effect model for different subgroups

[Jonnathan Andrew]

Hi all, I am new to STATA. I was previously told that subgroup analysis by sex can be performed using one of these commands:

• metan _ES _seES, fixed eform label (namevar=Author) by(sex)
• metan _ES _seES, random eform label (namevar=Author) by(sex)

However, I now encounter a problem. The heterogeneity in the male subgroup is low, so I would like to use the fixed-effect model for male. On the contrary, the heterogeneity in the female subgroup is high, so I would like to use the random-effect model for female. So, my question is, what is the correct command that I can use in this situation?

Thank you in advance.

[Marc Jarczok]

Hi Jonnathan,
up to my knowledge u can't mix the random und fixed effect option by group. Instead, u have to run two meta commands - one for each sex - and specify random or fixed.
hope this helps, Marc