To a person new to PK modeling, understanding the differences between population PK modeling (top down) and physiologically based PK (PBPK) (bottom up) modeling might be confusing. Hopefully, this analogy can help you out.

Let’s say you wanted a new cabinet in your house. How would you go about doing so?

The most apparent way would be to measure out the dimensions of the room you want the cabinet in and hire a carpenter to tailor make one for you. This is like (top down) population PK modeling, where we fit the model directly to the data.

If for some reason, however, you cannot measure your room dimensions, how then can you go about getting the right cabinet? One possibility would be to ask your neighbour, whose house floor plan is similar, what model their cabinet is. You can then buy the same model from the furniture store. This is akin to using a popPK model from literature instead of collecting your own data, if you can assume that the population used to develop the model is similar enough for your purposes. I do use these models as the basis to build PK-PD models where appropriate.

However, if your neighbor does not allow you to reference their house (no popPK model from literature available), and the carpenter is unavailable (cannot collect PK data to build the popPK model), we can try to build it ourselves, using prior knowledge of how a cabinet should be built. You must gather wood, screws, and hinges. Based on your knowledge of how cabinet doors and drawers work, and the standard sizing of a cabinet in a house model like yours, you can build a cabinet with reasonable knowledge that it should fit into your room. This is (bottom up) PBPK modeling, where the PK model is built based on the drug’s physicochemical properties, as well as our knowledge of human physiology.

There are other options too, that include a blend of the methods, but I will save that for another time.

Hope this post was helpful!