Methods of improving the fitting of Double Peak Phenomenon data in pharmacokinetics

Two methods of using compartmental modelling to describe the Double Peak Phenomenon in pharmacokinetics have recently been compared using two sets of data provided by AstraZeneca and a further set from the literature. The first method has the advantage of clear physiological interpretation, but the model has a comparatively large number of parameters to be estimated. The second approach has considerably fewer parameters to be estimated, but it lacks the direct relationship to physiology. Almost equally good fits to the data sets were obtained from the two approaches. In this paper, the two methods are used to fit a further set of data from the literature. For all the data sets, it is found that, for some of the responses, the tail end is not fitted very well. Methods of improving this fit by weighting the tail end of the data before fitting and by adding an extra compartment to the second modelling approach are investigated.