05-19-2017 02:40 AM
I'm trying to run a Cox model with time-dependent covariates, furthermore some of these covariates are predictable in time, for example macroeconomic covariates.
We are using PHREG, here is the code when we estimate the model:
proc phreg data=union_surv_macro PLOTS(overlay)= (SURVIVAL); baseline covariates=baseline_scenario; model (time1,time2)*cens_1(0)=unemployment rscore; run;
So, in our code, unemployment is a time dependent covariate and rscore is a fixed time covariate. When we want to use a survival curve in order to predict, we have to input a single unemployment value to the model, but what we want instead is to input a whole scenario of unemployment, with several values over time. So we would have here several survival curves for each operation and we have to combine them somehow (how?)
Is there any automatic implementation in PHREG to do this?
Thank you very much for your time.