Leave-one-out sensitivity meta-analysis

[Crystal Marruganti]

Goodmorning everyone,

I am performing a meta-analysis with a continuous outcome and 2 groups (test vs control). I have STATA version 16.1.

I would like to perform a leave-one-out sensitivity meta-analysis. I performed it with the command metaninf, but I cannot edit the final graph.

Are there any other commands to obtain the same graph and calculations?

Best regards,

Crystal

[David Fisher]

Hi Crystal,
You can perform this analysis with the user-written command metan , available from the SSC archive (type ssc describe metan or ssc install metan). The relevant option is called influence. The forest plot is fully customizable.
Best regards,
David.