LOCALLY D-OPTIMAL DESIGNS BASED ON A CLASS OF MODELS COMBINING AN EMAX AND ONE-COMPARTMENT MODELS
A class of models combining a pharmacokinetic compartmental model and a pharmacodynamic Emax model is introduced. One 4-parameter and another 5-parameter models in this class are studied.These models allow the estimation of PK parameters without drawing blood samples and the drug effct to be predicted along the time. Locally D-optimal (LD) designs and two classes of robust designs are investigated for these two models. An LD design for the 4-parameter combined model is found to be a saturated 4-point uniform LD design with the two boundary points of the design space in the LD design support. For the 5-parameter combined model, a suffient condition for the LD design to require the minimum number of sampling points is given. An LD design is found to be equivalent to an LDk-1 design if the linear parameter in the two combined models is a nuisance parameter. Examples of LD designs are presented. Simulation results show that LD designs and robust LD designs can not be traded in terms of mean square error (MSE) if the nominal values are not close to the true values. Consequently, LD designs seem self-robust and are suggested to be used in practice.
D-optimal design, pharmacokinetic compartmental model pharmacodynamic Emax model non-linear model
X. Fang A.S. Hedayat
University of Illinois at Chicago
国际会议
2006 International Conference on Design of Experiments and Its Applications(2006实验设计及其应用国际会议)
天津
英文
2006-07-09（万方平台首次上网日期，不代表论文的发表时间）