Dear R-help,
I am using R 2.14.1 on Windows 7 with the 'gfcure' package (cure rate
model).
I have included the treatment variable in the cure part of the model as shown
below:
Ø ref_treat <-
gfcure(Surv(rem.Remtime,rem.Rcens)~1,~1+strata(drpa)+factor(treat(delcure)),data=delcure,dist="loglogistic")
>From that I can obtain the coefficients, standard errors etc as per
alternative models (with covariates only fitted to the survival part of the
model say).
> summary(ref_treat)
However, only one standard error is output:
Log-logistic mixture model
The maximum loglikelihood is -927.0449
Terms in the accelerated failure time model:
Coefficients Std.err z-score p-value
Log(scale) -0.894528 0.0236 -37.8324 0.0000000
(Intercept) 6.929351 0.0151 460.4157 0.0000000
Terms in the logistic model:
Coefficients Std.err z-score p-value
(Intercept) 2.542726
strata(drpa)drpa=2 18.767777
factor(treat(delcure))2 0.184192
factor(treat(delcure))3 0.472809
factor(treat(delcure))4 0.255565 953.6876 0.0003 0.9997862
factor(treat(delcure))5 0.401713
Warning message:
In sqrt(diag(solve(object$infomat))) : NaNs produced
Can anyone explain why this is the case?
Very many thanks,
Laura
[[alternative HTML version deleted]]
