2000 年 31 巻 5 号 p. 659-665
Nonlinear mixed effects models are often used for the analysis of repeated measure ments in clinical trials. A typical example is population pharmacokinetic analysis. Due to the statistical complexity, careful attention must be paid to some statistical issues in order to obtain reliable and robust results. This paper provides a description of the notes on the use of nonlinear mixed effects models for the analysis of repeated measurements in clinical trials. Three main issues are discussed in this paper: 1) a first-order lineariza tion, 2) a missing data, and 3) a measurement error.