weighted_j2r: Create a weighted_j2r 'ImputeMechanism' object

In dejaVu: Multiple Imputation for Recurrent Events

Create a weighted_j2r ImputeMechanism object

Description

Missing counts for a subject in the active treatment arm will be imputed according to a point (determined by trt.weight) between the means of the placebo and treatment arms, conditioned on the number of events. Missing counts for subjects in the placebo arm will be imputed according to the mean of the placebo arm, conditioned on the subject's observed number of events.

Usage

weighted_j2r(trt.weight, delta = c(1, 1), proper = TRUE)

Arguments

trt.weight	See details
delta	If trt.weight=1 then delta is a vector of length 2 (control.delta,treatment.delta) and the mean number of expected events for the imputed missing data is multiplied by the appropriate delta
proper	If proper=TRUE then proper imputation is performed, in which each imputation is created based on parameters values drawn from the (approximate) posterior distribution of the imputation model. If proper=FALSE, improper imputation is performed. This means all imputed datasets are generated conditional on the maximum likelihood estimates of the parameters.

Details

If trt.weight = 0 then imputation using this mechanism will follow the jump to reference (j2r) model whereby missing counts for subjects in both arms will be imputed according to the mean of the placebo arm conditioned on the subject's observed number of events

If trt.weight = 1 then imputation using this mechanism will follow the MAR model whereby missing counts for subjects in each arm will be imputed according to the event rate of subjects in its treatment group conditioned on the subject's observed number of events

See the User guide vignette for further details

Value

An ImputeMechanism object

See Also

ImputeMechanism.object

Examples

sim <- SimulateComplete(study.time=365,number.subjects=50,
event.rates=c(0.01,0.005),dispersions=0.25)
sim.with.MCAR.dropout <- SimulateDropout(sim,
                     drop.mechanism = ConstantRateDrop(rate = 0.0025))
fit <- Simfit(sim.with.MCAR.dropout)
imps <- Impute(fit, weighted_j2r(trt.weight=0), 10)

dejaVu documentation built on Sept. 11, 2024, 8:24 p.m.