I plan to perform a meta-analysis with a primary outcome consisting in very rare events (sometimes not occurring in any group of some studies included), and a much more frequent secondary outcome.
To assess the risk of publication bias, I therefore decided to draw a funnel plot based on seconday outcome data.
Is that acceptable? If so, how would I conduct sensitivity analyses related to a possible publication bias?
$\begingroup$
$\endgroup$
3
-
1$\begingroup$ This Q&A might be helpful $\endgroup$– mdeweyCommented Jan 6, 2022 at 16:28
-
$\begingroup$ @mdewey unless I'm mistaken you forgot to include a link to the Q&A $\endgroup$– user47679Commented Jan 6, 2022 at 17:05
-
1$\begingroup$ stats.stackexchange.com/questions/193260/… $\endgroup$– mdeweyCommented Jan 6, 2022 at 17:19
Add a comment
|