Hi, I'm using inverse probability treatment weighting for the propensity for vaccination, followed by Cox regression. My exposure is vaccination and my outcome is death in the first week of life. Can I add a random effect term for mothers who have had more than one birth (and subsequent death in the first week)?
thanks
Minda

logistic fluvac ///
i.matage5y i.mum_abor parity i.anctrimester i.year_season_birth ///
i.smoke i.seifa10 i.asthma i.hypertension i.diabetes i.gestdiabetes i.gesthypertension i.preeclampsia
predict p
gen ipw=.

replace ipw=1/p if fluvac_com==1
replace ipw=1/(1-p) if fluvac_com==0


** Unweighted
stset endrisk_firstweek, origin(origin_firstweek) enter(startrisk_firstweek) exit(time .) ///
fail(died==1) id(babyid) scale(365.25)
stcox i.fluvac, vce(robust)
stset, clear


** Weighted
stset endrisk_firstweek [pweight=ipw], origin(origin_firstweek) enter(startrisk_firstweek) exit(time .) ///
fail(died=1) id(babyid) scale(365.25)
stcox i.fluvac, vce(robust)
stset, clear