sim.bssr.gee.1subgroup {spass}

Simulation of a longitudinal one subgroup design with internal pilot Study

Description

Given estimates of the treatment effects to be proven, the variances, and the prevalence, sim.bssr.gee.1subgroup calculates an initial sample size and performs a blinded sample size recalculation after a pre-specified number of subjects have been enrolled. Each observation is simulated and a final analysis executed. Several variations are included, such as different approximations or sample size allocation.

Usage

sim.bssr.gee.1subgroup(
  nsim = 1000,
  alpha = 0.05,
  tail = "both",
  beta = 0.2,
  delta = c(0.1, 0.1),
  vdelta = c(0.1, 0.1),
  sigma_pop = c(3, 3),
  vsigma_pop = c(3, 3),
  tau = 0.5,
  rho = 0.25,
  vrho = 0.25,
  theta = 1,
  vtheta = 1,
  Time = 0:5,
  rec.at = 0.5,
  k = 1,
  model = 1,
  V = diag(rep(1, length(Time))),
  OD = 0,
  vdropout = rep(0, length(Time)),
  missingtype = "none",
  vmissingtype = "none",
  seed = 2015
)

Arguments

nsim

number of simulation runs.

alpha

level (type I error) to which the hypothesis is tested.

tail

which type of test is used, e.g. which quartile und H0 is calculated

beta

type II error (power=1-beta) to which an alternative should be proven.

delta

vector of true treatment effects, c(overall population, inside subgroup).

vdelta

vector of treatment effects to be proven, c(overall population, inside subgroup).

sigma_pop

vector of true standard deviations of the treatment effects, c(overall population, subgroup).

vsigma_pop

vector of assumed standard deviations, c(overall population, inside subgroup).

tau

subgroup prevalence.

rho

true correlation coefficient between two adjacent timepoints

vrho

initial expectation of the correlation coefficient between two adjacent timepoints

theta

true correlation absorption coefficient if timepoints are farther apart

vtheta

expected correlation absorption coefficient if timepoints are farther apart

Time

vector of measured timepoints

rec.at

blinded sample size review is performed after rec.at*100\% subjects of the initial sample size calculation.

k

sample size allocation factor between groups: see 'Details'.

model

which of the two often revered statistical models should be used?: see 'Details'.

V

working covariance matrix.

OD

overall dropout measured at last timepoint

vdropout

vector of expected dropouts per timepoint if missingness is to be expected

missingtype

true missingtype underlying the missingness

vmissingtype

initial assumptions about the missingtype underlying the missingness

seed

set seed value for the simulations to compare results.

Details

This function combines sample size estimation, blinded sample size re-estimation and analysis in a design with a subgroup within a full population where we want to test for treatment effects between a control and a treatment group. The required sample size for the control and treatment group to prove an existing alternative delta with a specified power 1-beta when testing the global null hypothesis H_0: \Delta_F=\Delta_S=0 to level alpha is calculated prior to the study and then recalculated in an internal pilot study.

For sample sizes n_C and n_T of the control and treatment group, respectively, the argument k is the sample size allocation factor, i.e. k = n_T/n_C.

Value

sim.bssr.1subgroup returns a data.frame containing the mean and variance of recalculated sample sizes within the control group and treatment group respectively and the achieved simulated power along with all relevant parameters.

Source

sim.bssr.gee.1subgroup uses code contributed by Roland Gerard Gera.

See Also

sim.bssr.gee.1subgroup makes use of n.gee.1subgroup, bssr.gee.1subgroup, and r.gee.1subgroup.

Examples

sim.bssr.gee.1subgroup(nsim = 5,missingtype = "intermittened")