Description Usage Arguments Details Value Author(s) References See Also Examples

Estimate the information based on the data, which is the numerator of the information fraction.

1 2 3 |

`data_ref` |
Input reference dataset which provides the survival curves for the estimation. It could be some dataset entirely external to |

`data_check` |
Input dataset to check the estimated information. It should follow the sample format as |

`rho` |
First power parameter for the Fleming-Harrington weight which weighs on the early departures: |

`gamma` |
Second power parameter for the Fleming-Harrington weight which weighs on the late departures: |

The `I_t`

function estimates the information up to the maximum follow-up time in the data of `data_check`

, which is identical to the numerator of the information fraction proposed by Hasegawa (2016):* \hat{P}_1(t)\hat{P}_0(t)\int_0^t W(t,s)^2N(t,ds)*. Note that the datasets `data_check`

and `data_ref`

input here are output data from `data.trim`

functions, or any datasets including `survival`

as time to event or censoring, `delta`

as event indicators, and `trt`

denotes treatment assignment (1 is treatment, 0 is control). Note that `I_t.2`

is another option which is slightly different from the one proposed in Hasegawa(2016), but is identical to the estimate of variance of the weighted log-rank test, which considers the total at-risk set *R(t)* and treatment arm *R_1(t)*: *\int_0^t\frac{R_1(s)R_0(s)}{R(s)^2}W(t,s)^2N(t,ds)*.

The returned value is the calculated information estimated from the input dataset `data_check`

using the survival function estimated from `data_ref`

.

Lili Wang

Hasegawa, T. (2016). Group sequential monitoring based on the weighted log‐rank test statistic with the Fleming–Harrington class of weights in cancer vaccine studies. Pharmaceutical statistics, 15(5), 412-419.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | ```
# install.packages("devtools")
# library(devtools)
# install_github("keaven/nphsim")
library(nphsim)
eps<-2 # delayed effect
p<-0.5 #treatment assignment
b<-30 # an intrinsic parameter to decide the number of intervals per time unit
tau<- 18 # end of the study
R<-14 # accrual period [0,R]
omega<- tau-R
lambda<-log(2)/6 # control group risk hazard
theta<-0.7 # hazard ratio
lambda.trt<- lambda*theta #hazard after the change point for the treatment arm
rho<- 0 # parameter for the weights
gamma<-1 #parameter for the weights
alpha<-0.025 #type 1 error
beta<-0.1 #type 2 error
# First we decide the sample size:
size_FH <- sample.size_FH(eps,p,b,tau,omega,lambda,lambda.trt,rho, gamma,alpha,beta)
n_FH <-size_FH$n
n_event_FH<-size_FH$n_event
accrual.rt<-n_FH/R # the needed arrual rate
#Generate data accordingly, use eta=1e-5 to inhibit censoring
data_temp <- nphsim(nsim=1,lambdaC=lambda, lambdaE = c(lambda,lambda.trt),
ssC=ceiling(n_FH*(1-p)),intervals = c(eps),ssE=ceiling(n_FH*p),
gamma=accrual.rt, R=R, eta=1e-5, fixEnrollTime = TRUE)$simd
#Obtain the full information at the final stage based on the generated data
#Trim the data up to the final stage when n_event_FH events have been observed
data_temp1 <-data.trim.d(n_event_FH,data_temp)[[1]]
I_t(data_temp1,data_temp1,rho,gamma) # the estimated information at the final stage
#Trim the data up to certain event numbers at the interim stage when 60% of the events have been observed. Have been trimmed once to get data_temp1, no need to add additional variables, thus set the third argument to be F.
I_t.2(data_temp1,data_temp1,rho,gamma) # If we consider the change of the at-risk set, which is not necessary to be a fixed probability.
data_temp2 <- data.trim.d(ceiling(0.6*n_event_FH),data_temp1,F)[[1]]
I_t(data_temp1,data_temp2,rho,gamma) # Use the full dataset data_temp to provide the survival function, and check the estimated information for the trimmed data set data_temp2 with only 60% of the planned events have been observed.
I_t.2(data_temp1,data_temp2,rho,gamma) # If we consider the change of the at-risk set, which is not necessary to be a fixed probability.
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.