[R] Plot survival curves after coxph() with frailty() random effects terms

[Andreu Ferrero]

I would like to plot a survival curves of a group with different categories
after running a Cox model with frailty() random effects terms.

I just could display a survival plot of the covariable’s mean.



Here an example:



library(survival)

fit<-coxph(Surv(time, status) ~ sex+  frailty(litter, dist='gamma',
method='em'), rats)

summary(fit )

suf<-survfit(fit)

plot(suf,  xscale=1, xlab = "Days", ylab="Survival", conf.int=FALSE)

lines(suf[1], lwd=2)

Warning message:

In `[.survfit`(suf, 1) : survfit object has only a single survival curve



#But if I use the next code I get the 2 groups.

surv_group <- survfit(Surv(time, status) ~ sex, rats)

lines(surv_group[1:2], lwd=2, conf.int=FALSE)




However, I am not sure about these 2 curves are well done, appropriate.

If any of you could help…



-- 
Andreu Ferrero Gregori

	[[alternative HTML version deleted]]