I posted a vignette that outlines the session in the StudyGroup issue tracker:
https://github.com/isop-phmx/studyGroup/issues/22
The direct link is here:
https://github.com/isop-phmx/studyGroup/files/331189/pts.pdf
We have “Probability of technical success” in the title and we’ll get to that. But the focus will be on running some larger scale population simulations with mrgsolve
. The workflow we do here will be common to many different M&S applications beyond PTS
.
We will:
- Give brief explanation of
PTS
- Build a simulation data set with covariates from
NHANES
data
- Look at the NONMEM model and how to get that translated into
mrgsolve
specification
- Get the NONMEM output into a format that
mrgsolve
can handle
- Run the simulation in
R
with a couple of alternatives for parallelizing that process
- Summarize the simulations a couple of ways to get at
PTS
- Give an example for simulating parameters from the covariance matrix (of the estimate) for those who aren’t using
METHOD=BAYES
I’m not sure if this is going to be 40 minutes too short or 40 minutes too long … but we’ll try to make it useful within the hour.