Hi Statalist,

I would be grateful for some help predicting IPT-weighted 5-year (60 months) survival and 95% confidence intervals by treatment group (treated vs not treated) after multiple imputation. No other variables require adjustment.

I have an imputed dataset and I have created propensity score IPT weights for each individual.

I have worked through the following thread for help in predicting 5 year survival:
https://www.statalist.org/forums/for...d-mi-predictnl

I'm just not sure how to get the confidence limits??

Code:
mi stset date_death [pweight=ipt_wt], failure(death_censor==0) origin(date_diagnosis) scale(30.4)

mi estimate, dots cmdok sav(mi_stpm2, replace): stpm2 treatment, df(5) scale(hazard) eform

gen t60

mi predictnl surv1 = predict(survival at(treatment 0) timevar(t60)) using mi_stpm2, ci(surv1_lb surv1_ub) force
mi xeq: summarize surv1

mi predictnl surv2 = predict(survival at(treatment 1) timevar(t60)) using mi_stpm2, ci(surv2_lb surv2_ub) force
mi xeq: summarize surv2