在传统临床试验中，常事先确定受试对象数及受试时间，试验结束后进行期终分析；或在α＝０．０５的水准上反复进行显著性检验，因而加大第一类错误概率，甚至出现错误结论。本文介绍两组计量资料的期中分析及Ｐｏｃｏｃｋ、ＯＢｒｉｅｎ和Ｐｅｔｏ名义检验水准方法，并就其意义、公式及基本思想进行了讨论。实例分析表明，日本产Ｍｅｖｏｌｏｌｉｎ（美百乐镇）在第４周时，可使高脂血症病人ＴＣ显著下降，并持续下降；到第８周时，ＴＧ也有显著下降。 In traditional clinical trials, the number of subjects and the duration of the test are often determined in advance, and end-of-period analysis is performed after the end of the trial; or the significance test is repeated at the level of α=0.05, thereby increasing the probability of the first type of error. , even wrong conclusions. This article describes the mid-term analysis of two sets of measurement data and Pocock, OBrien and Peto’s nominal test level methods, and discusses its significance, formula and basic ideas. Case studies have shown that Mevololin produced in Japan can significantly reduce the TC of patients with hyperlipidemia and continue to decline at the 4th week; by the 8th week, TG also decreased significantly.