Dear all, I would like to visualize in my presentations, that -given a limited number of subjects- the actual ct-profiles (or PK/PD in fact) of a small group will differ from the population predictions via VPC (which comes naturally to most of us but is advantegous in some presentations for expectation Management )
Therefore, I would like to perform a series of simulations with a limited number of subjects (n=10) to visualize that these predictions will differ from each other. I want the model to do repeated sampling of the PK Parameter with added IIV (one parameter set for each individual). Since my model has no covariates, I wondered if it matters to simulate 1 replicate for 10 subjects or 10 replicates for 1 subject. Given the example attached it seems to matter (the range of ct- profiles seems more narrow with the 1 subject).
Could you give advice on the best practise how to simulate repeated sets of limited subject number?