改進均數比值假設檢定-應用於2×2交錯設計之生體藥效相等性試驗
Improving the Test of Mean Ratio Hypotheses- Application to Bioequivalence Trial in 2×2 Crossover Design

劉惠美，劉相妮

摘要/Abstract

生體藥效相等性試驗(bioequivalence trials)旨在顯示兩個處理(treatments)間之差異沒有超過事先預定的臨床值。本文考慮觀察之資料服從常態分配具有未知之均數和未知之變方之情況下，檢定均數比值(ratios of mean)是否通過生體藥效相等性試驗。依據樣本數和檢定尺度之大小，吾人利用概似比檢定給定充分統計量構建出齊一檢力較強於概似比檢定，新檢定是以學生隨機變數和交集-聯集檢定方法構建而成。將新檢定應用2 2交錯設計模型，以電腦模擬呈現新檢定與概似比檢定之檢力表現。

關鍵字：均數比值假設，齊一較強檢力檢定，交集-聯集檢定，生體相等性試驗

In this article, we consider bioequivalence trial with ratios of mean parameters and the data are normally distributed with unknown variance, . Based on the likelihood ratio test (LRT) given sufficient statistic and sample size, the new tests are size- tests and only involve univariate student t random variable. We apply the new tests to bioequivalence trials in 2 2 cross design.

Keywords：intersection - union test , likelihood ratio test , unknown variance.

關閉視窗