[R] survexp with weights

Mike Harwood harwood262 at gmail.com
Wed Apr 20 18:03:47 CEST 2011


Hello,

I probably have a syntax error in trying to generate an expected
survival curve from a weighted cox model, but I can't see it.  I used
the help sample code to generate a weighted model, with the addition
of a "weights=albumin" argument (I only chose albumin because it had
no missing values, not because of any real relevance).  Below are my
code with the resulting error messages.  Thanks in advance!

> pfit <- coxph(Surv(time,status>0) ~ trt + log(bili) + log(protime) + age +
+     + platelet,  data=pbc
+     )
>
> pfit
Call:
coxph(formula = Surv(time, status > 0) ~ trt + log(bili) +
log(protime) +
    age + +platelet, data = pbc)


                  coef exp(coef) se(coef)        z      p
trt          -0.000624     0.999  0.17304 -0.00360 1.0000
log(bili)     0.985497     2.679  0.08949 11.01262 0.0000
log(protime)  2.794001    16.346  0.95289  2.93215 0.0034
age           0.020590     1.021  0.00805  2.55666 0.0110
platelet     -0.001699     0.998  0.00085 -2.00130 0.0450

Likelihood ratio test=164  on 5 df, p=0  n= 308, number of events=
143
   (110 observations deleted due to missingness)
>
> plot(survfit(Surv(time, status>0) ~ trt, data=pbc))
> lines(survexp( ~ trt, ratetable=pfit, data=pbc), col='purple')
>
> pfit.wtd <- coxph(Surv(time,status>0) ~ trt + log(bili) + log(protime) + age +
+     + platelet,  weights=albumin, data=pbc
+     )
>
> pfit.wtd
Call:
coxph(formula = Surv(time, status > 0) ~ trt + log(bili) +
log(protime) +
    age + +platelet, data = pbc, weights = albumin)


                 coef exp(coef) se(coef)      z       p
trt          -0.01354     0.987 0.094204 -0.144 8.9e-01
log(bili)     0.99282     2.699 0.048690 20.391 0.0e+00
log(protime)  2.54136    12.697 0.525797  4.833 1.3e-06
age           0.01913     1.019 0.004398  4.350 1.4e-05
platelet     -0.00165     0.998 0.000462 -3.578 3.5e-04

Likelihood ratio test=535  on 5 df, p=0  n= 308, number of events=
143
   (110 observations deleted due to missingness)
> plot(survfit(Surv(time, status>0) ~ trt, data=pbc))
> lines(survexp( ~ trt, ratetable=pfit.wtd, data=pbc), col='purple')
Error in eval(expr, envir, enclos) : object 'albumin' not found
> lines(survexp( ~ trt, ratetable=pfit.wtd, weights=albumin, data=pbc), col='purple')
Error in eval(expr, envir, enclos) : object 'albumin' not found
In addition: Warning message:
In survexp(~trt, ratetable = pfit.wtd, weights = albumin, data =
pbc) :
  Weights ignored
> lines(survexp( ~ trt, ratetable=pfit.wtd, weights=pbc$albumin, data=pbc), col='purple')
Error in eval(expr, envir, enclos) : object 'albumin' not found
In addition: Warning message:
In survexp(~trt, ratetable = pfit.wtd, weights = pbc$albumin, data =
pbc) :
  Weights ignored
>



More information about the R-help mailing list