On Evaluation of Individual Bioequivalence Based on Small Sample Confidence Inteval Approach

• Main Author: 李啟彰
• Other Authors: Liu, J.P.
• Format: Others
• Language: en_US
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/12755658135327694694

碩士 === 國立成功大學 === 統計學系 === 88 === For demonstrating the test and reference formulations of a drug have the similar bioavailability, Anderson and Hauck (1990) proposed a new concept called “individual bioequivalence”. Until recently, many researchers had proposed various methods to assess individual bioequivalence. In 1997 (Guidance for Industry, 1997), the U.S. Food and Drug Administration (FDA) recommended that use of the nonparametric percentile bootstrap confidence interval to claim individual bioequivalence. However, the method of bootstrap requires intensive and complex computations. Hyslop, Hsuan, and Holder (submitted to Statistics in Medicine) developed alternative confidence interval procedure to assess individual bioequivalence without utilizing bootstrap’s method. The procedure proposed by Hyslop, Hsuan, and Holder is a parametric method based on Normal distribution. And the procedure may be influenced by the underlying distribution. Thus, we adopted the simulation techniques to investigate the behavior of size and power under Normal, Uniform, Exponential and Cauchy distribution by using the method proposed by Hyslop, Hsuan, and Holder. Keywords: bioavailability, individual bioequivalence, bootstrap, confidence interval, size, power