stmixed for recurrent time-to-event data

[Carole Khairallah]

Hi all,
I am doing an individual participant data (IPD) meta-analysis on recurrent time-to-event data, i.e. multiple failure-time survival data, as described here

https://www.stata.com/support/faqs/statistics/multiple-failure-time-data/#cond1/

.
The failures are ordered, and a variable identifies the failure risk group for each time span.

Here is my syntax using Stata v17.0 : 1) stset 2) stcox with the option "strata()" which identifies the order of an event and "robust"

stset tstop, fail(fail) exit(time .) id(p_id) enter(tstart)
stcox treatment age i.study i.site, robust strata(event) nolog

I would like to add random effects in my model, so I looked at the command stmixed from SSC, but it doesn't allow the "strata" variable. Would it be similar to add the strata variable as a random effect in the model?

Also, I tried to run stmixed, but only got error messages.





Thank you for your help,
Carole