simJLSDdata {QTOCen}

Function to generate simulation data from a sequentially randomized experiment designed in (Jiang et al. 2017)

Description

Function to generate simulation data from a sequentially randomized experiment designed in (Jiang et al. 2017)

Usage

simJLSDdata(n, case = "a", s_Diff_Time = 1, C_max = 5,
  Censored = TRUE, fix_x0_value = NULL)

Arguments

n

sample size

case

string. One of "a", "b", "c", corresponding to three models.

s_Diff_Time

Numeric. Default is 1. This is the length of time between two stages of treatment

C_max

Numeric. Default is 5. This the upper bound of the uniform distribution of the censoring time variable. Changing this value shifts the overall censoring rate easily.

Censored

Boolean. Default is TRUE. Whether the data has censoring or not. If TRUE, all survival time would not be censored at all in the returned data.

fix_x0_value

Numeric. Default is Null. If supplied, it will generate simulated data with a fixed value, fix_x0_value, of the univariate baseline covarate.

Details

This generative model is proposed in (Jiang et al. 2017), Section 5, the second example. It uniformly defined three sets of conditional distributions of the survival times given the observable covariates at each stage within the same framework.

All three models satisfy the independent censoring assumption.

Value

This function returns a data.frame with simulated subject trajectories.

• x0 the baseline covariate, always observable at relative time point 0;

• a0 the observed first-stage treatment level at relative time point 0;

• x1 an updated covariate observable to the relative time point s_Diff_Time, when the a second stage treatment is scheduled

• a1 the observed second-stage treatment level at relative time point s_Diff_Time.

References

Jiang R, Lu W, Song R, Davidian M (2017). “On estimation of optimal treatment regimes for maximizing t-year survival probability.” Journal of the Royal Statistical Society: Series B (Statistical Methodology), 79(4), 1165–1185.

Examples

dataA <- simJLSDdata(500,case="a")
dataB <- simJLSDdata(500,case="b")
dataC <- simJLSDdata(500,case="c")