I have a longitudinal dataset with 2 matched cohorts (Cohort==1 & Cohort==0) in a ratio of 1:3 that I am following longitudinally until a failure/event occurs.
I want to compare failure functions (cumulative incidence) at a time _t between the 2 cohorts
To set my data I use:
stset timeelapsed, failure(failure) id(id)
then I use
sts list, failure by(Cohort) saving(KM)
to get the failure function result for each cohort at a time _t
I want to sample from these cohorts 500 matched individuals and do 500 bootstraps of the difference in failure functions to get an estimate of the difference in cumulative incidence and related 95% confidence interval at each time point _t.
Any advice?
0 Response to Bootstrapping to obtain confidence intervals around differences in failure functions
Post a Comment