Bioequivalence and Bioavailability Forum

Hi ElMaestro,

❝ One thing that occasionally crossed my mind was that I would love to see someone (no, I am of course not naming anyone and of course I do not have any particular person in mind, Helmut ) writing a function in R which takes some arguments of relevance and spits out a list of proposed sampling points. [etc., etc., …]

Fantastic idea!
In any software for PK modeling – with some efforts even in Base  nls() – you could obtain the VIF of the model based on different sampling schedules. Plot them for the primary parameters vs time. Around the extrema you find time points which are the most important ones. Others are luxury.
I did that for many years in WinNonlin and sometimes still today.
In the early 1990s I wanted to go even further. Given a guesstimate of the between-subject variability of \(\small{f}\), \(\small{C_\textrm{max}}\), and \(\small{k_\textrm{el}}\) establish the goalposts of the bioanalytical method (LLOQ, ULOQ). Questions: Where to place the calibrators, how many calibrators + replicates are needed to give the most accurate/precise back-calculated concentrations? Tried a lot till I realized that bioanalytics is the least important part in this game.*

---

• Gaffney M. Variance Components in Comparative Bioavailability. J Pharm Sci. 1992; 1(4): 315–7. doi:10.1002/jps.2600810402.