论文部分内容阅读
本文以模拟试验展示以前提出的两生存率检验所需样本容量测定方法的行为。取α=0.05(单侧),β=0.1,按不同数值的最小临床承认疗效差量或两生存率和终检率交叉组合得出32种方案结合Weibull生存分布形状参数1/3—3形成160个试验集各重复1000次。在几乎全部指数生存分布试验集,该检验的观测功效符合预定功效。在如此大范围形状参数Weibull分布下,变异甚微。这提示,这种方法对于指数生存分布十分精确,对于不同形态的实际分布,仍然适用。
In this paper, we simulate the behavior of the proposed method for determining the sample volume required for the two survival tests. Twenty-two regimens were combined with the Weibull survival distribution shape parameters 1 / 3-3 by taking a = 0.05 (unilateral), β = 0.1, and the minimum clinically recognized efficacy difference or survival rate and final survival rate of different values were cross-combined Each of the 160 trial sets was repeated 1000 times. In almost all of the exponential survival distribution trial sets, the test’s observed efficacy was in line with the intended efficacy. In such a wide range of shape parameters Weibull distribution, the variation is minimal. This suggests that this method is very accurate for exponential survival distributions and still applies to the actual distribution of different morphologies.