在药物的第二、三阶段的临床试验时,药物的安全性和有效性是测验的主要指标。Hsu和Berger在药物各个剂量组的方差2σ都相等这一假设下,提出一种识别最小有效剂量的逐步置信区间方法。而方差齐性假设条件通常是不合理的,本文主要利用Stein两阶段抽样方法,给出方差无任何限制条件下识别最小有效剂量(MED)的逐步置信区间方法。 In the second and third phase of the drug clinical trials, the safety and efficacy of the drug are the main indicators of the test. Hsu and Berger proposed a stepwise confidence interval method to identify the minimum effective dose under the assumption that the variance 2σ for each dose group of the drug was equal. However, the assumptions of homogeneity of variance are usually unreasonable. In this paper, Stein’s two-stage sampling method is used to obtain a stepwise confidence interval method to identify the minimum effective dose (MED) under the condition of no variance.