23 replies to this topic

[serge guzy]

You are mostly welcome. Yes it is indeed a very nice example to estimate the doses when not known accurately or even if for some reason there is a relative bioavailability assessment to be made. Then now you can answer to your real problem. Do not forget that not fixing the other model parameters will lead to potential multiple or even infinite optimal solutions and the optimizer will randomly generate one. If f1=f2, then it is definitevely a must to fix V at least.

Let me know if you need further assistance. Sprry for the typo errors, there is no spelling check I think here in the forum.

best Regards;

Serge

[Angus McLean]

Serge: my experience is in developing products: I seek to use model work assist the process.

 

 When developing modified release formulations often companies have a series of formulations with different rates of release {from slow to fast}. Consequently cross over comparative bioavailability studies are often done on these similar formulations. Often the plasma concentrations fit the same PK model and the absorption is complete (>95%}.

When one looks at the model parameters the Volume of distribution changes quite a bit between these formulations. The K01 for sure will be different between the formulations, but the drug elimination rate would be similar. So one would think in terms of fixing K10 {as a parameter}, but for me the fixing of V could be a problem. I realize that is what you have to do in the specific example above considered. Even if we consider the case of a two component formulation {e.g. 2 types of beads in a capsule} absorbed at different parts of the intestine then the K01 will be different between the formulations.

 

Angus

[serge guzy]

Dear Angus

This is definitively a modeling issue. My point is that if you define f1 and f2, then estimating v will create some sort of identifiability issue because the change in v you are talking about is apparenlty due to this f effect. If you estimate both, who knows what you are going to get.

Best

Serge

[Angus McLean]

Serge: I do follow your thinking very well.

 

Your point is very sound.I like the option of fixing the paramaters. One would have to deal with situations like this by seeing what you get with the different options.

 

Angus