wlr.Stat: Weighted Logrank Test Statistic
In keaven/nphsim: Non proportional hazards sample size and simulation

Description Usage Arguments Details Value Examples

Weighted logrank adopted from survMisc package

1	wlr.Stat(survival, cnsr, trt, stra = NULL, fparam)

`survival`	time-to-event variable
`cnsr`	censoring variable: 1=censoring, 0=event
`trt`	treatment varaible. Accepted values are either "experimental" or "control"
`stra`	stratification variable for stratified weighted logrank test. Default is `NULL`.
`fparam`	a list input to request additional FH test and time of delayed separation in the simplified APPLE method `fparam$stdset` if not `NULL`, the standard set of tests are added `fparam$rho` a vector of `rho` to be used in Fleming-Harrington FH(rho, gamma) `fparam$gamma` a vector of `gamma` to be used in Fleming-Harrington `FH(rho, gamma)`. The length of `gamma` needs to be the same as `rho` `fparam$wlr` test name of which the p-Value will be stored in the variable `pval`. Accepted values are "1", "n", "sqrtN", "S1", "S2", "APPLE", or "FH(rho,gamma)" where `rho` and `gamma` is one of the pair specified by user in the `fparam` `fparam$APPLE` If not left `NULL`, it is the time of separation used in the simplified piecewise exponential method.

The one-sided p-Values from the following tests can be obtained based on input from fparam$stdset and fparam$APPLE

1: log-rank (part of standard set)
n: Gehan-Breslow generalized Wilcoxon (part of standard set)
sqrtN: Tarone-Ware (part of standard set)
S1: Peto-Peto's modified survival estimate (part of standard set)
S2: modified Peto-Peto (by Andersen) (part of standard set)
APPLE: Simplified APPLE method with user specified late separation time
FH(rho,gamma): Fleming-Harrington with user specified pairs of rho and gamma

The function return a list with the follow components

pval: One-sided p-Value from user specified test
pval_[XXX]: One-sided test from various weighted tests

# weighted logrank on the simulated data
library(survMisc)
medC = 6 
hr <- c(1, 0.6)
intervals <- 3 
gamma <- c(2.5, 5,  7.5,  10) ## a ramp-up enrollment
R     <- c(2  , 2,  2  ,  6 ) ## enrollment period: total of 12 months
eta <- -log(0.99) ## 1% monthly dropout rate
sim1 <- nphsim(nsim=2,lambdaC=log(2)/medC,lambdaE=log(2)/medC*hr, ssC=300,ssE=300,
               intervals=intervals,gamma=gamma, R=R,eta=eta)
test1 <- simtest(x=sim1, anaD=c(250,300), method=wlr.Stat,fparam=list(rho=c(0,0,1), 
                 gamma=c(0,1,1), wlr='FH(0,1)', APPLE=3))
test1$result[]

# direct function call (without cutoff)
wlr.Stat(surv=sim1$simd$survival, cnsr=sim1$simd$cnsr, trt=sim1$simd$treatment,
         fparam=list(rho=c(0,0,1), gamma=c(0,1,1), wlr='FH(0,1)', APPLE=3))