I am not a pharmacometrician but am attempting a preliminary data analysis for a grant submission, i.e., I do not yet have an expert involved so I am seeking advice here on the following:
I am attempting to use the 'nlme' package in R to generate preliminary estimates of Vd/F and Cl/F from multidose PK data of an orally administered drug. However, the package does not appear able to handle missingness, and I would like to avoid the added complexity of introducing a data imputation step.
The data are repeated measurements structured in long tidy format as c(id, time, dose, conc) and the issue is that almost all of the t>0 dosing times have no corresponding concentration data. Is there an alternative package to 'nlme', or other solution?
The raw data are formatted similar to NONMEM i.e. a column "amount" that contains either the dose amount or plasma concentration according to a "type" variable with values of "dose" or "conc".
I have also considered reducing the problem to single-dose PK but would prefer to use all of the data.
Thank you.