Individual Bayesian posterior predictions at short intervals — makePost • Pmetrics

Returns the Bayesian posterior predictions at short intervals specified during the NPAG run, up to 12 minutes. These results are contained separately from the main output of NPAG, in the PRTBxxxx file.

Usage

makePost(run, NPdata)

Arguments

run: The number of the folder that contains the relevant run. If missing will be set to current working directory.
NPdata: Optional name of NPdata object if run is missing.

Value

A dataframe of class PMpost with columns:

id: Subject id
time: Time of predictions in decimal hours
icen: Prediction based on mean or median of Bayesian posterior parameter distribution
pred: Predicted output for each outeq
outeq: Output equation number
block: Observation blocks within subjects as defined by EVID=4 dosing events

Details

Predictions are calculated using the median (default) or mean of the Bayesian posterior parameter values and each subject's dosing, sampling, and covariate (if relevant) information.

Author

Michael Neely

Examples

if (FALSE) {
post <- makePost(run = 1)
# would work if there is a complete run in folder 1
}