终检渐近正态法

摘要

目的：：基于经典渐近正态法，提出一种用于两生存率检验所需样本量的测定的终检渐近正态法。方法：该方法是终检样本的实际样本量随终检概率和时间衰减形成的有效样本量。结果：它具有还原性，摆脱了指数分布的假设，与两生存率检验相匹配；其观测功效和预定功效吻合；相比之下，终检非中心法略显保守， Lachin-Foulkes法对于生存率检验显得功效不足。结论：这种方法适于慢性病临床研究，以检出早、中、晚期疗效差别。%Objective An asymptotic normal procedure with censorship is proposed for determining the sample size required in the test on two survival rates. Methods The procedure is derived from its classical counterpart on the basis of that the real size of censored samples declines with the censoring probability and the time. Results It is re-ducible, free from the assumption of exponential distributions, and matched with the test on two survival rates. The observed power coincides with the prescribed power. By contrast, the non -central procedure with censorship is slightly conservative and the Lachin-Foulkes procedure has the power usually not adequate for the test on two sur-vival rates. Conclusion The procedure is appropriate with planning clinical studies of chronic diseases for detecting early, middle, or late treatment-control difference.