【摘要】 目的探讨非劣效试验研究中率指标非劣效性界值的设定问题。方法从一个实例入手,通过理论推导与软件模拟相结合,分析不同样本量及样本率条件下非劣效性界值的变化。结果非劣效性界值△主要由两部分组成:△≈△E+△0,即非劣效性界值的最低要求限△0和期望检验的最大率差△E。当样本量相同时,样本率越接近0.5,则△值越大。当样本率相同时,随着样本量的增大,△值逐渐减小。结论非劣效性试验中率指标非劣效性界值的设定应同时考虑最低要求限△和期望检验的最大率差△,这将为实际工作提供指导和参考。更多还原

【Abstract】 Objective To explore the limits to no-inferiority margin of binary endpoint in no-inferiority trial.Methods Beginning with an example,theoretical derivation and simulation was used to explore the fluctuation of no-inferiority margin under different sample sizes and proportions.Results No-inferiority margin △ mainly consists of two parts,△≈△E+△0,namely,the limit △0 of theminimum requirement for the margin plus the expectation maximum rate difference △E to be tested.Under the same sample size,the more a proportion gets close to 0.5,the larger the margin required.Similarly,under the same proportion,the margin decreases as the sample size increases.Conclusion The no-inferiority margin of binary endpoint in no-inferiority trial should consider both theminimum limit △0 and the expectation maximum rate difference △E.This will be a guidance and reference for determining the margin in a trial.更多还原