針對藥物開發進行的第一期臨床試驗的藥物動力學研究，主要目的即是探求該藥物在血液或血漿中濃度隨時間變化的側寫，因為此一血中藥物濃度時間側寫經常提供該藥物藥效和/或毒性的有用資訊。 本人的2008-2010國科會研究計劃和2009年發表於 Statistics in Medicine的論文已經在一般的二期雙序列交叉設計之下，分別根據常用的雙區和開放一階單區藥物動力模式，提出具有廣義甘馬誤差分布的藥物濃度統計模式。事實上，口服藥、一般注射藥或是其他方式服用藥物的平均藥物濃度時間側寫未必適用上述特定的藥物動力模式。為了避免實務上模式辨識的困難，本人擬進行一個二年期的研究計畫，新創一個穩健的藥物濃度統計模式。第一年將提出具有廣泛右偏分布藥物濃度的半參數統計模式，使之適用於描述各種藥物的血中藥物濃度時間側寫。主要工作是估計此一穩健統計模式中的平均藥物濃度及相關參數，並且進行模式的評估。第二年則將專注於上述模式的應用。除討論如何針對此一平均藥物濃度時間側寫建立信賴帶，並且在二期雙序列交叉設計之下進行生物相等性的檢定。 The major purpose of the pharmacokinetic study in Phase I of a clinical trial for drug development is to investigate the profile of the concentration of the drug in plasma or blood over time. The reason is because that the drug concentration-time profile usually provides with valuable information about the efficacy and/or toxic effect of the drug. In my NSC research project 2008-2010 and the paper published in Statistics in Medicine in 2009, I have developed statistical models with a generalized gamma error distribution for the time-varying drug concentration under a two-period and two-sequence crossover design based on some particular pharmacokinetic models, including an ordinary two-compartment model and an open, first order, one-compartment model. However, in practice, these specific pharmacokinetic models may not well applicable to the situations involving medicines taken in oral, by injection or by any other methods. To avoid the difficulty on identifying the suitable pharmacokinetic model, we develop a robust statistical model for the drug concentration over time in this two-year research project. In the first year, we build up a semi-parametric model for right-skewed drug concentration so that the model can be applicable to a variety of pharmacokinetic studies in which drugs may be taken in different ways. The major work is the estimation of the mean drug concentration and the related parameters as well as the evaluation of the statistical model. In the second year, we focus on the application of the robust statistical model, including constructing a confidence band for the mean drug concentration over time and developing a bioequivalence test under a two-period, and two-sequence crossover design. 研究期間：10008 ~ 10107
