Visual predictive check (VPC)
A VPC compares observed summaries with their distribution under repeated simulation of the study design.
A VPC compares observed summaries with their distribution under repeated simulation of the study design (FDA. Population Pharmacokinetics. 2022.).
Simulate complete replicate datasets at the observed doses and sampling times, including variability and residual error. In each time bin, compare observed percentiles with uncertainty intervals for the corresponding simulated percentiles. Stratifying by regimen can reveal lack of fit hidden by pooling. A VPC under the original design does not test a new regimen directly.