在存活分析上，常見到的檢定為兩存活函數是否相等，有分成在固定時間點下存活機率是否相等與整體時間下存活函數是否相等兩種；但是在醫學應用上，則是比較關心檢定於兩存活函數在一個所能容忍的誤差範圍內是否相等，稱之為對等性，若不具對等性，再進一步去探討其較優性與較劣性。Goeman et al. (2010)提出了三邊假設檢定(Three-sided hypothesis testing)，控制在型Ⅰ錯誤下，同時檢定較優性、對等性與較劣性，免去重複做兩次假設檢定導致型Ⅰ錯誤的增加。本文將三邊假設檢定的想法用在存活函數上，利用統計模擬計算此假設檢定的型I錯誤、型II錯誤與型III錯誤機率之大小，並應用在兩種不同治療狀況下的存活機率或函數的實例上，檢定他們是否具有對等性、較優性與較劣性。

In survival analysis, it is common to test equality of two different survival functions. It contains two cases: at fixed time point and for all time in survival function. In medical applications, it is considered the equivalent test of the two survival functions in the error range which one can accept. It is called the equivalent test. If two survival functions are not equivalent, the superiority test or the inferiority test will be considered. Goeman et al. (2012) had proposed the three-sided hypothesis testing. It can test the equivalent, superiority and inferiority at the same time when the type I error rate was fixed. It can avoid to increase type I error rate because multiple hypotheses are tested. In this paper, we use the three-sided hypothesis testing on survival analysis. The statistical simulation is conducted to compute type I error rate, type II error rate and type III error rate in this hypothesis testing. Finally, this method is used on real example to compare equivalent, superiority and inferiority of two survival functions.

1. Fleming T.R. and Harrington D.P. (2005), Counting Processes and Survival Analysis, Wiley-InterScience.
2. Fleming T.R. and Harrington D.P. (1981), “A class of hypothesis tests for one and two samples of censored survival data”, Communications In Statistics, 10:763-794.
3. Gehan E.A. (1965), “A Generalized Wilcoxon test for comparing arbitrarily singly censored samples.”, Biometrika, 52:203-223.
4. Gisela Tunes da Silva, Logan B.R. and Klein J.P. (2009), “Methods for Equivalence and Noninferiority Testing”, Biology of Blood and Marrow Transplantation, 15(1 Suppl) : 120-127.
5. Goeman J.J. , Solari A. and Stijnen T. (2010),“Three-sided hypothesis testing : Simultaneous testing of superiority, equivalence end inferiority”, Statistics in Medicine, 29 : 2117-2125.
6. Hirotsu C. (2007), “A unifying approach to non-inferiority, equivalence and superiority tests via multiple decision processes”, Pharmaceutical Statistics, 6 : 193-203.
7. Hosmer D.W. Jr. and Lemeshow S. (1999), Applied Survival Analysis: Regression Modeling of Time to Event Data, New York, Wiley.
8. Kalbfleisch J.D. and Prentice R.L. (2002), The Statistic Analysis of Failure Time Data (Second Edition), Wiley-Interscience.
9. Kaplan E.L. and Meier P. (1958), “Nonparametric estimation form incomplete observation”, Journal of the American Statistical Association, 53:457-481.
10. Klein J.P. and Moeschberger M.L. (2003), Survival Analysis Techniques for Censored ans Truncated Data (Second Edition), Springer, Now York, USA.
11. Klein J.P. , Logan B. , Harhoff M. and Andersen P.K. (2007),” Analyzing survival curves at a fixed point in time”, Statistics in Medicine, 26:4505–4519.
12. Mantel N. (1966), “Evaluation of survival data and two new rank order statistics arising in its consideration.”, Cancer Chemotherapy Reports, 50 (3): 163–70.
13. Pepe M.S. and Fleming T.R. (1989), “Weighted Kaplan-Meier Statistic: A Class of Distance Test for Censored Survival data”, Biometrics, 45 :497-507.
14. Peto R. and Peto J. (1972), “Asymptotically efficient rank invariant test procedures (with discussion)”, Journal of the Royal Statistical Society A, 135:185-206.
15. Prentice R.L. and Marek P. (1979), “A qualitative discrepancy between censored data rank tests”, Biometrics, 35:861-867.
16. Tarone R.E. and Ware J.H. (1977), “On distribution-free test for equality for survival distributions”, Biometrika, 64:156-160.