中文摘要：

目的解决新药非劣效性临床试验评价中两组率差值的协变量调整问题。方法利用ROC曲线与两组率差值的对应关系,通过ROC分析间接得到调整协变量之后两组率之差的估计值及标准误。结果给出了一种新的非参数分层分析方法 ,如果协变量对有效率有影响,可以通过调整分析得到更准确的检验结果。结论本文提供的方法能够更直观、有效地对新药临床试验的数据进行非劣效性评价。

英文摘要：

Objective To address the covariate adjustment for rate difference in non-inferiority clinical trials. MethodsThe relationship between rate difference and ROC curve was used to indirectly obtain the estimator for rate difference and its standardized error. Results We present a novel non-parametric stratified analysis method. ConclusionThis method can visually and effectively assesses the non-inferiority of clinical trials.