研究期間：10108~10207;Statistical models for the drug concentrations over time in a pharmacokinetic (PK) study and their applications were investigated in my previous NSC research projects. In the statistical models, the mean drug concentration-time profile is described either by the compartment PK model or by the spline regression model, and the generalized gamma distribution is used for the drug concentration. In practice, however, the involved healthy volunteers are usually different in their physical conditions. Moreover, the drug concentrations that are repeatedly measured from the same subject after drug administration are often correlated. Therefore, to take into account of the real-world situations, we plan to conduct a two-year NSC research project to develop a more practical statistical model for the oral-administered drug concentration-time profiles in the PK study. The newly developed model employs mixed effect model for the trend of the drug concentration-time profiles which includes the between-subject variation. Moreover, the repeated measured drug concentrations are described by a multivariate generalized gamma distribution which is constructed by using the Gaussian copula function. The estimation of the PK statistical model is discussed in the first year and several tests based on the proposed PK mixed effect model are considered for detecting possible outlying profiles in the second year.
