Variability and singularity arising from poor compliance in a pharmacokinetic model. II: The multi-oral case (Q515832)

In the pharmacokinetic model considered here, the drug concentration level at time \(t~ (\geq 0)\) is given by \[ C(t) = \frac{F}{V_d}\frac{k_a}{k_a-k_e}\sum_i D_i \{(e^{-k_e(t-T_i)} - e^{-k_a(t-T_i)}\}I_{\{t\geq T_i\}}, \] where \(k_a\) is the fixed absorption rate, \(F\) is the bioavailability parameter, \(k_e\) is the elimination rate, \(V_d\) is the apparent volume of distribution of the drug in terms of concentration in plasma, and \(D_i\) and \(T_i\)'s are the consecutive dosage levels and dosing times. Distributional properties of \(C(t)\) are studied for four types of models that depend on the nature of \(F, k_e, V_d\), and the \(D_i\) and \(T_i\). The first three of these quantities are taken to be either deterministic or random, and in the latter case, they are assumed to be log-normally distributed. The dosage level \(D_i\) can either be deterministic, or have a two-point or uniform distribution. The \(T_i\) can either be deterministic, discrete random as determined by a Bernoulli mechanism, or form the arrival times of a homogeneous Poisson process. The paper derives distributional results for \(C(t)\) (such as expressions for the mean, variance, and characteristic functions, and bounds on tail probabilities) for finite \(t\), and as \(t \to \infty\), and discusses their practical consequences. Some simulation-based results are also presented to illustrate the effect of poor adherence (with sparse, random dosing schedule) on the properties of \(C(t)\). Similar results for a similar, simpler model were given in [the second author and \textit{P. Lévy-Véhel}, ``Variability and singularity arising from poor compliance in a pharmacodynamical model. I: The multi-IV case'', J. Pharmacokinet. Pharmacodyn. 40, No. 1, 15--39 (2013; \url{doi:10.1007/s10928-012-9284-y})].