CIFsmry (version 1.0.1.1)

CIFsm: Cumulative incidence function estimate and weighted summary statistics

Description

Estimate the cumulative incidence function for cause of interest in two-sample study. Provide the weighted summary statistics based on given method and weight.

Usage

CIFsm(ds,method="dif",pp = 0,qq = 0,conf.bd=T,n.sim=500)

Arguments

ds

ds is a dataset contains the time, cause of event and group. For cause, 0 means censoring, 1 is the cause of event, 2 is all other causes. Two groups need to be coded as 1 and 2.

method

method can be chosen from "dif"=risk difference, "rr"=risk ratio and "or"=odds ratio

pp

first parameter of weight function

qq

second parameter of weight function

conf.bd

logical; if TRUE, create confidence band cut point. Set to FALSE if the confidence band is not needed, which reduces the computational time

n.sim

number of simulations used in creating confidence band; will be ineffective if conf.bd is FALSE

Value

sample

Total sample size from both groups

used

Sample used in analysis. Subject with missing value in any of the three variables (time, cause or group) will be excluded from analysis

size

Sample size for each group

njp

Total number of unique event time points in two groups

tjp

Unique event time points from both groups

ny1

Number of subject at risk in group 1

f1

Estimate of cumulative incidence function for group 1

f1.se

Standard error of cumulative incidence function for group 1

ny2

Number of subject at risk in group 2

f2

Estimate of cumulative incidence function for group 2

f2.se

Standard error of cumulative incidence function for group 2

dif

Estimate of difference in cumulative incidence (risk difference) between the 2 groups

dif.se

Standard error of the risk difference

dif.pv

P-value of the risk difference

rr

Risk ratio of the cumulative incidence between the 2 groups

rr.se

Standard error of the risk ratio

rr.pv

P-value of the risk ratio

or

Odds ratio of the cumulative incidence between the 2 groups

or.se

Standard error of the odds ratio

or.pv

P-value of the odds ratio

cbcut

95% confidence band cut point of risk difference, risk ratio and odds ratio based on simulation method.

method

Method used for the weighted summary statistics

weight

Weight used for the weighted summary statistics

region

Time range of the data

nbd

Index of beginning and end of the data. For internal use only

ave

Time integrated weighted summary statistics

avese

Standard error of the time integrated weighted summary statistics

ci95

95% confidence interval of the time integrated weighted summary statistics

avepval

P-value of the time integrated weighted summary statistics

wt

Weight assigned at each unique event time point

Details

The estimates and summary statistics are described in Zhang and Fine (2008).

References

M.J. Zhang and J.P. Fine. Summarizing difference in cumulative incidence functions. Statistics in Medicine, 27:4939-4949, 2008.

J. Li, M.J. Zhang and J. Le-Rademacher. Weighted Summary of two Cumulative Incidence Functions with R-CIFsmry Package. Computer Methods and Programs in Biomedicine[Submitted].

Examples

library(CIFsmry)
data(sim.dat)
out <- CIFsm(sim.dat,pp=0,qq=0)
out$avepval
plot(out$tjp,out$f1,type="s",ylim=c(0,1),yaxt="n",xaxt="n",xlab="Time",ylab="CIF",
     lty=1,lwd=1)
points(out$tjp,out$f2,type="s",lty=1,lwd=3)
axis(1,at=seq(0,6,by=1),cex=0.6)
axis(2,at=seq(0,1,by=0.2),cex=0.6)
legend("bottomright",c("1","2"),title="group",lty=1,lwd=c(1,3))

out10 <- CIFsm(sim.dat,pp=1,qq=0)
out10$avepval