Hello Stata community,
I want to perform survival analysis regarding factors associated with developping an event (kidney graft failure).
Two of the main covariates I consider to add in the model are time-varying events (rejection and infectious episode after transplantation), I'm thus using a long-type format database and Cox for this analysis.
I would like to test for proportionality hazard assomption for these variables, but I'm not sure what is the appropriate technic there.
I assume that the comparison of predicted curve to the observed one cannot be used there since KM comparator cannot be applied there.
Is it possible to use Schoenfeld residuals analysis (Stata function estat phtest) and if yes are there specific items to check in that particular situation ?
Thanks in advance for your help.
Romain
I want to perform survival analysis regarding factors associated with developping an event (kidney graft failure).
Two of the main covariates I consider to add in the model are time-varying events (rejection and infectious episode after transplantation), I'm thus using a long-type format database and Cox for this analysis.
I would like to test for proportionality hazard assomption for these variables, but I'm not sure what is the appropriate technic there.
I assume that the comparison of predicted curve to the observed one cannot be used there since KM comparator cannot be applied there.
Is it possible to use Schoenfeld residuals analysis (Stata function estat phtest) and if yes are there specific items to check in that particular situation ?
Thanks in advance for your help.
Romain
