msm.predict {qgcomp}

Secondary prediction method for the (non-survival) qgcomp MSM.

Description

this is an internal function called by qgcomp.glm.boot, but is documented here for clarity. Generally, users will not need to call this function directly.

Get predicted values from a qgcompfit object from qgcomp.glm.boot.

Usage

msm.predict(object, newdata = NULL)

Arguments

object

"qgcompfit" object from qgcomp.glm.boot function

newdata

(optional) new set of data (data frame) with a variable called psi representing the joint exposure level of all exposures under consideration

Details

(Not usually called by user) Makes predictions from the MSM (rather than the conditional g-computation fit) from a "qgcompfit" object. Generally, this should not be used in favor of the default predict.qgcompfit function. This function can only be used following the qgcomp.glm.boot function. For the qgcomp.glm.noboot function, predict.qgcompfit gives identical inference to predicting from an MSM.

Examples

set.seed(50)
dat <- data.frame(y=runif(50), x1=runif(50), x2=runif(50), z=runif(50))
obj <- qgcomp.glm.boot(y ~ z + x1 + x2 + I(z*x1), expnms = c('x1', 'x2'), 
                       data=dat, q=4, B=10, seed=125)
dat2 <- data.frame(psi=seq(1,4, by=0.1))
summary(msm.predict(obj))
summary(msm.predict(obj, newdata=dat2))