Prior information for population pharmacokinetic and pharmacokinetic/pharmacodynamic analysis: overview and guidance with a focus on the NONMEM PRIOR subroutine

Fig. 1. Plot of individual η clearances (black line) on top of theoretical η-distribution N(0, ω2) (dotted line). The model resulting in the top plot is to be preferred over the one resulting in the bottom plot. ETACL: η clearances. Adapted from [43]

Fig. 2. Illustration of parameters of the random-effects-model, from [38]. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\widehat{\theta }$$\end{document}θ^ k = μ + ϵk + ζk (1),\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\widehat{\theta }k$$\end{document}θ^k: typical value in the study k, μ: typical value in the « universe» of population, ϵk: deviation from the typical value because of sampling errors in the study k, ζk: deviation from the typical value because of over-arching distribution of true effect sizes with the mean, μ, ζk ~ N(μ, τ2)

Fig. 3. Example of codes of a NONMEM control file for implementation of PRIOR, NWPRI subroutine, informative and non-informative priors; as defined by Bauer [1] and Gisleskog et al. [3]

Fig. 4. Case example of the plot proposed by Tsamandouras et al. [11], data from [43]. The estimate of THETA(CL) in the model built with prior was 12.7 L.h−1 (black line). THETA(CL) in the reference model was 9.89 L.h−1 and its standard deviation 3.71 L.h−1 (dotted lines)