本文研究在兩期雙序列交叉設計之下，測試藥及專利藥的生物對等性檢定。檢定的對象為生體可用率參數，例如：血液中的藥物總濃度 (AUC) ，最高藥物濃度 (Cmax) 及達到最高藥物濃度的時間 (Tmax) 。美國食品藥物管理局 (U.S. FDA) 建議的兩個單尾檢定是在上述參數估計式為對數常態分布假設下建立的，但是，實務中對數常態分布假設可能不合理。本文在對數轉換之參數估計為偏斜常態分布假設下，建立一個隨機效應模式，然後應用期望值最大化演算法估計模式中的參數，藉此建立較為穩健的兩個單尾檢定，進行研究兩種藥物的生物對等性檢定。本文以模擬研究上述兩種生物對等性檢定的型I誤差率及檢定力，結果顯示傳統及本文所提的兩個單尾檢定在型I誤差率表現相近，但是本文所提的檢定在偏斜常態分布下有較佳的檢定力表現。本文最後分析一筆資料，說明所提隨機效應模式及其生物對等性檢定的應用。;The thesis considers the bioequivalence (BE) test between the test drug and patent drug under a 2×2 crossover design. The bioavailability (BA) parameters under study are, for example, the overall drug concentration in blood (AUC) , the maximum drug concentration (Cmax) , and the time to reach the maximum drug concentration. Note that the two one-sided tests (TOST) suggested by U.S. FDA that are constructed under the assumption of lognormal distribution for the estimated BA parameters may not be practical. Therefore, this article constructs a random effect model for the estimated BA parameters by assuming that the logarithm of the estimated BA is distributed according to a skew-normal distribution. The parameters of the random effect model are estimated by using the EM algorithm, a robust TOST is then conducted for the BE between the two drugs. A simulation study is implemented to investigate the type I error rate and power of the two different BE tests. The result show that the two BE tests are similar on holding their type I error rates, but the robust BE test has a better power performance than the conventional TOST. Finally, analyze a real data set is illustrated to demonstrate the application of the proposed random effect model and the bioequivalence test.
