Estimate the tau process at specified time points. The estimated variances at the last time point under complete randomization design and random allocation rule (urn model) are provided.
tau.fit(data, t = numeric())an object of class "tauFit" with components
N0 | number of individuals with arm=0 |
N1 | number of individuals with arm=1 |
t | the specified truncation time |
tau | the estimated value of tau measure |
var.r | the estimated variance under random grouping design (complete randomization design) |
var.f | the estimated variance under fixed grouping design (random allocation rule / urn model) |
a data.frame consisting of arm, surv.time, event.
a sequence of specified times. If the user do not specify the sequence, the default is an equally-spaced sequence from 0 to the last identified time.
The estimation and inference procedure are proposed by Yi-Cheng Tai, Weijing Wang and Martin T. Wells. The value of tau measure serves as a clinically meaningful measure of treatment effect. It supplements the traditional hazard ratio (HR) under nonproportional hazard scenario.