I have computed the difference between the two arms of the log(-log survival ) vs log(time) plot above. This is the code and the result… so?
survi<-summary(survfit(sur~arm, data = kn061), time=seq(1,28,0.1 ) )
survi$surv=log(-log (survi$surv))
survi$time=log(survi$time)
survi2<-data.frame(strata=survi$strata, time=survi$time, surv=survi$surv)
s0<-survi2[survi2$strata=="arm=1",]
s1<-survi2[survi2$strata=="arm=0",]
survi2$diff= s0$surv- s1$surv
ggplot(data=survi2, aes(x=time, y=diff)) +
geom_line()
