Sergei Leonov (1), Alexander Aliev (2)
(1) Vertex Pharmaceuticals, Inc., Cambridge, MA, U.S.A. (2) Institute for Systems Analysis, Russian Academy of Sciences, Moscow, Russia
Introduction: Optimal experimental design of population PK/PD studies has received considerable attention in the literature and software development in recent years. Since 2006, the theory of optimal design for nonlinear mixed effects models and its applications in drug development were discussed at the annual PODE workshop (Population Optimum Design of Experiments); see http://www.maths.qmul.ac.uk/~bb/PODE/PODE.html. Discussions included the comparison of different software tools for population optimal design, and results of such comparison were presented at PAGE 2007 and 2011 meetings; see Mentré et al. [1, 2].
Methods and Objectives: The key object in estimation and optimal design for PK/PD models is the Fisher information matrix (FIM) of a particular sampling scheme. For nonlinear mixed models, the FIM does not have a closed-form expression and, therefore, approximate formulae have to be used. As reported in Mentré et al. [2], under the same assumptions all software tools produce identical FIMs. Simple approximations, e.g. linearization of the response, are relatively straightforward to implement and often give quite accurate results. However, there are instances when such simplified approaches lead to a substantial distortion of the variability estimates of model parameters. To improve the quality of the FIM approximation, we propose to calculate the mean response vector and its covariance matrix via Monte Carlo simulations.
Results: The alternative approximation of the FIM is implemented in PkStaMp library which is intended for the construction of D-optimal sampling schemes for compartmental PK and combined PK/PD models; see Aliev et al. [3]. We present several examples of the calculation of the FIM via the new approach, and compare the new outputs with those obtained via previously considered options.
References:
[1] Mentré F, Duffull S, Gueorguieva I, Hooker A, Leonov S, Ogungbenro K, Retout S (2007). Software for optimal design in population pharmacokinetics and pharmacodynamics: a comparison In: Abstracts of the Annual Meeting of the Population Approach Group in Europe (PAGE). ISSN 1871-6032, http://www.page-meeting.org/?abstract=1179
[2] Mentré F, Nyberg J., Ogungbenro K, Leonov S, Aliev A, Duffull S, Bazzoli C, Hooker A (2011). Comparison of results of the different software for design evaluation in population pharmacokinetics and pharmacodynamics. In: Abstracts of the Annual Meeting of the Population Approach Group in Europe (PAGE). ISSN 1871-6032. www.page-meeting.org/?abstract=2066
[3] Aliev A, Fedorov V, Leonov S, McHugh B, Magee M (2012). PkStaMp library for constructing optimal population designs for PK/PD studies. Comm. Statist. Simul. Comp., 41 (6), 717-729.
Reference: PAGE 21 () Abstr 2479 [www.page-meeting.org/?abstract=2479]
Poster: Study Design