Sample Size Determination for Two-Stage Equivalence Test

兩階段相等性檢定樣本數的決定

馬瀰嘉(Mi-Chia Ma) ； 葉尚柏(Shang-Bo Ye)

《智慧科技與應用統計學報》 9卷1期 (2011/06) Pp. 25-41

https://doi.org/10.29807/JTITAS.201106.0003

摘要

在臨床研究上，兩階段設計通常是在第一階段從數個治療方法當中找出最佳的治療方法，並在第二階段與標準治療方法做比較。本研究提出一兩階段設計在非劣性檢定和相等性檢定上，第一階段仍從數個治療方法當中找出最佳的治療方法，在第二階段觀察此最佳的治療方法是否不比標準治療方法差，或是等價於標準治療方法。兩階段設計法的優點在於，當虛無假設被接受時，這個試驗會被提早終止。本研究在固定的顯著水準以及檢定力之下，藉由近乎最小化期望總樣本數，可以得到最佳的樣本數以及拒絕域臨界點。最後利用統計模擬，去比較傳統一階段和兩階段非劣性檢定和相等性檢定的設計在樣本數上的差異，得到兩階段非劣性檢定和相等性檢定的設計分別有較小總樣本數。

關鍵字

兩階段設計 ； 非劣性檢定 ； 相等性檢定

並列摘要

In clinical research, a two-stage design usually indicated that the best of several experimental treatments was selected at the first stage and confirmed whether it was better than the standard treatment at the second stage. This paper proposed a two-stage design for non-inferiority test or equivalent test. We select the best of several experimental treatments and test whether it is not inferior to a standard treatment or is equivalent to a standard treatment. The sample size is parsimonious only for the best treatment and the standard treatment at the second stage. The design allows early termination when the null hypothesis does not reject. By minimizing expected total sample size for fixed significance level and power, an adaptive optimal sample size and cut-off parameters are obtained. Finally, a simulation study was conducted to compare the difference of sample size between one-stage design and two-stage design for non-inferiority test or equivalent test. The two-stage design has small total sample size for non-inferiority test or equivalent test.

並列關鍵字

two-stage design ； non-inferiority test ； equivalent test