A Example Data Set"
In bpcp: Beta Product Confidence Procedure for Right Censored Data

knitr::opts_chunk$set(echo = TRUE)


DIR<-"C:/GitHub/bpcp/R/"
source(paste0(DIR,"bpcp2sample.R"))
source(paste0(DIR,"fixtdelta.R"))
source(paste0(DIR,"kmciFunctions.R"))

Chemotherapy for Stage B/C Colon Cancer

Here is a data example. I got the data from the survival R package.

See ?survival::colon

Because we are dealing with small sample issues. Consider only those with a perforation of the colon.

library(survival)
data(cancer,package="survival")
# use time to death only (etype==2)
d<- subset(colon,perfor==1 & rx!="Obs" & etype==2)
#d<- subset(colon,obstruct==1 & rx!="Obs" & etype==2)
# change time to years
d$time<- d$time/365.25

s<-survfit(Surv(time,status)~rx,data=d)
plot(s,lty=c(1,2),mark.time=TRUE)
legend("bottomleft",levels(d$rx)[2:3],lty=1:2)

Consider the tests at different years.

bpcp2samp(d$time,d$status,d$rx,testtime=2)

fixtdelta(d$time,d$status,d$rx,testtime=2)



bpcp2samp(d$time,d$status,d$rx,testtime=3)

fixtdelta(d$time,d$status,d$rx,testtime=3)



bpcp2samp(d$time,d$status,d$rx,testtime=4)

fixtdelta(d$time,d$status,d$rx,testtime=4)



bpcp2samp(d$time,d$status,d$rx,testtime=5)

fixtdelta(d$time,d$status,d$rx,testtime=5)


bpcp2samp(d$time,d$status,d$rx,testtime=6)

fixtdelta(d$time,d$status,d$rx,testtime=6)



bpcp2samp(d$time,d$status,d$rx,testtime=7)

fixtdelta(d$time,d$status,d$rx,testtime=7)

We could talk about how the delta method does not change between years 4 and 5 nor does it change between years 6 and 7.

It would also be nice to have some simulated data sets that look similar to the Kaplan-Meier curves, but also have a simulated scenario that looks similar to the Kaplan-Meier curves except that at year 7 the curves are equal. And hopefully when we simulate the tests at year 7, only the bpcp2samp maintains the type I error rate.