Since I would like to compare the survival time between treated and untreated groups for an observational data set, I used propensity score weighting method to get IPTW. After this, I want to get cumulative incidence curves (CIFs) for competing risk. I know how to get a non-weigted CIF; however, I could not find a R function for weighted CIF. Thanks for everyone.
non-weighted CIF crr(ftime, fstatus, cov1, cov2, tf, cengroup, failcode=1, cencode=0, subset, na.action=na.omit, gtol=1e-06, maxiter=10, init, variance=TRUE)
Regards, Stanley
There is no R-Package which implements this method so far. I am currently workin on an R-Package that does this sort of stuff (including confidence intervals, hypothesis tests etc.), but it's still under development so I can't share my code yet.
What you could do is the following: 1.) Estimate a parametric competing risks model (Fine & Gray / Cause-Specific-Cox ...) 2.) Use the model to predict the CIF at all event times for every individual 3.) Take a weighted mean (with your IPTW) at each time point inside each of your groups of interest
This procedure is decribed in some papers. See for example Neumann (2016)