Pharmacokinetics with uncertainties in rate constants- III: the inverse problem

The identification of a random compartmental model from kinetic data in pharmacokinetics is considered, where the randomness arises due to uncertainties in our knowledge of the rate constants. An identification scheme together with computational algorithms for estimating the properties of the random rate constants is presented. A two-compartment model is used as an example, and the problem of identification based upon incompletely observed data is commented upon. The results also show that, if the underlying model is actually random, serious degradation in results can be expected if it is incorrectly modeled as deterministic.