ipdmeta (version 2.1)

make.clustered.ipd: Generates a Mixed Effects IPD meta dataset of Weibull event times

Description

Generates a random ipd datset for a covariate-treatment Weibull IPD mixed effects meta-analysis.

Usage

make.clustered.ipd(n,beta,D,censor.rate,tau2)

Arguments

n

vector of group sizes

beta

Fixed effects parameters (intercept, trt, x, interaction)

D

Matrix of bivariate frailty covariance-variance

censor.rate

Non-informative target censoring rate

tau2

Between-study variance for normal covariate

Value

• List of ipd, D, shape and censor.rate.

Details

The covariate x is normal with marginal distribution N(0,1+tau2).

The treatment indicator is Bernoulli(1/2).

The beta parameters are fixed and the study random effects are drawn from a MVN with varcov = D. If these components are zero than the corresponding study effect is set to zero. The shape of the Weibull is drawn from a Uniform(.5,3) and is the same for all studies.

Censoring is a non-informative exponential random variable. The rate is set to be the rate at which the probability of the censoring time being less than the median of the data set is censor.rate.

Examples

make.clustered.ipd(c(5,10,20),rep(0,4),D=diag(c(.1,.2)),.1,1/2)