Nonlinear Mixed Effects Models: Practical Issues

Abstract

This chapter provides practical advice in the development of nonlinear mixed effects models. Topics that are discussed include how to choose an estimation method, how to incorporate various covariates in the model (concomitant medications, weight, age, smoking, pregnancy, pharmacogenomics, food, formulation, race, renal function, and laboratory values), and when and how to collect samples for analysis. Missing data and censoring within the context of a repeated measures analysis is also discussed. The chapter concludes with material unique to NONMEM, the leading software for population pharmacokinetic–pharmacodynamic analysis. NONMEM topics that are covered include NONMEM errors and how to resolve them, mu modeling (also called mu-referencing), and the consistency of NONMEM parameter estimates across various platform–compiler combinations.