clinDR (version 2.4.1)

predict.fitEmaxB: Estimated mean and posterior intervals derived from a Bayesian hyperbolic or sigmiodial Emax model.

Description

The mean/proportion response for different doses estimated from a Bayesian Emax model is computed along with corresponding posterior intervals. The results are computed for a vector of input dose levels. The estimates are posterior means or medians of the MCMC generated means/proportions. For binary outcomes, the estimated response rates are computed on the logit scale and then back-transformed before forming the estimates and posterior intervals.

Usage

# S3 method for fitEmaxB
predict(object, dosevec, clev = 0.9,
	int = 1, dref = 0, xvec=NULL, ...)

Value

A list with estimated mean/proportion (pred, predMed), lower bound, upper bound, posterior SD, and corresponding values for differences with the reference dose. One value for each dose in dosevec. The MCMC response means (proportions for binary data) are in simResp, and the residual SD for continuous data are in sigsim.

Arguments

object

Output of fitEmax with class "fitEmaxB".

dosevec

Vector of doses to be evaluated.

clev

Level for the posterior intervals about the mean/proportion at each dosevec.

int

The index for the protocol (intercept) to use for the predictions

dref

Differences in response between doselev and dref are computed.

xvec

The vector of centered baseline values for the prediction model when xbase was specified in the model fit. Centering must be done using the protocol-specific means consistent with int. See details for the default calculations when xvec is not specified.

...

No additonal parameters will be utilized.

Author

Neal Thomas

Details

Results computed from simple tabulations of the MCMC parameters evaluated in the Emax function.

If baseline covariates were included in the fit and xvec is not specified, then the predicted value is the mean of the predictions for all patients in the specified protocol. Note that the protocol must be specified, or the prediction defaults to patients from the first protocol. Note that for binary data, the distinction between the mean of the predicted values and the predicted value at the mean of the covariates can be important.

See Also

fitEmaxB

Examples

if (FALSE) {
data("metaData")
exdat<-metaData[metaData$taid==6 & metaData$poptype==1,]

prior<-emaxPrior.control(epmu=0,epsca=10,difTargetmu=0,difTargetsca=10,dTarget=80.0,
        p50=3.75,sigmalow=0.01,sigmaup=20)
mcmc<-mcmc.control(chains=3)

msSat<-sum((exdat$sampsize-1)*(exdat$sd)^2)/(sum(exdat$sampsize)-length(exdat$sampsize))
fitout<-fitEmaxB(exdat$rslt,exdat$dose,prior,modType=4,
				count=exdat$sampsize,msSat=msSat,mcmc=mcmc)

predout<-predict(fitout,dosevec=sort(unique(exdat$dose)))
}