conf.MTL: Confidence interval of mean total lifetime

Description

This function is used to calculate confidence intervals of mean total lifetime using jackknife resampling.

Usage

conf.MTL(obj, state = NULL, nsim = 1000, L = 120)

Arguments

obj

An object returned by optim.fit, which contains the transition probabilities and other information used to simulate mean total lifetime.

state

A numeric vector indicating from which state the mean total lifetime is simulated. Default is NULL, where no mean total life for a specific state is output. If obj is returned by optim.fit with treatment=NULL, there is no need to set this argument.

nsim

The times of simulation for mean total life. The default is 1000.

The prespecified threshold for blocking the increase of residual lifetime. The default is 120.

Value

If the input object comes from optim.fit with treatment=NULL, a list object with elements:

conf.state.MTL

A data frame containing states, corresponding mean total lifetime, standard error and 95% confidence interval. If state=NULL, this element does not exist.

state.table

The correspondence of state number and state label.

If the input object comes from optim.fit with treatment is not NULL, a list object with elements:

conf.strategies

Mean total lifetime for different strategies, along with standard error and 95% confidence interval

Details

This function systematically leaves out each subject from the original datset and simulates mean total lifetimes for each n-1-sized subsample. The jackknife mean and variance are calculated by aggregating n simulated mean total lifetimes. For each jackknife dataset, mean total lifetime is simulated using the algorithm described in sim.MTL.

Examples

Run this code

# NOT RUN {
# }
# NOT RUN {
library(OptimalTiming)

##################################
## Example 1: This example shows how to calculate confidence
## intervals for different treatment strategies

## read data
data(SimCml)

## fit multistate model with treatment not equals NULL
fit=optim.fit(data=SimCml,
       transM=matrix(c(0,1,0,0,0,1,0,0,0,1,0,1,1,1,0,0,0,1,1,1,1,
       0,0,0,0,1,1,1,0,0,0,0,0,1,1,0,0,0,0,0,0,1,0,0,0,0,0,0,0),7,byrow=TRUE),
       nstate=7,state_label=c("diagnose","cp1","ap","cp2","bc","sct","death"),
       event_label=c("cp1.s","ap.s","cp2.s","bc.s","sct.s","death.s"),
       treatment=c("sct","sct.s"),absorb=c("death","death.s"),
       cov=c("age"),cov_value=c(0))

## compare different treatment strategies
conf.MTL(obj=fit,nsim=1000,L=120)

##################################
## Example 2: This example shows how to calculate confidence
## intervals for a given state

## read data
data(SimCml)

## delete the information of transplant time
data=SimCml[SimCml$sct.s==0,]
del=which(names(SimCml)%in%c("sct","sct.s"))
data=data[,-del]

## fit multistate model with treatment equals NULL
fit=optim.fit(data=data,
        transM=matrix(c(0,1,0,0,0,0,0,0,1,0,1,1,0,0,0,
        1,1,1,0,0,0,0,1,1,0,0,0,0,0,1,0,0,0,0,0,0),6,byrow=TRUE),
        nstate=6,state_label=c("diagnose","cp1","ap","cp2","bc","death"),
        absorb=c("death","death.s"),event_label=c("cp1.s","ap.s","cp2.s","bc.s","death.s"),
        cov=c("age"),cov_value=c(0))

## calculate mean total lifetime and confidence intervals
## for state 1,2,3,4
conf.MTL(obj=fit,state=c(1,2,3,4),nsim=1000,L=120)
# }
# NOT RUN {


# }