摘要
目的推荐并改进一种用于新药临床试验中生存分析三臂非劣效性设计的半参数统计推断方法。方法采用Cox比例风险模型建立用于生存时间数据的三臂非劣效检验方法,通过模拟试验验证该方法的有效性,并给出样本含量估计的R语言程序。结果模拟试验结果证实,在样本删失风险率维持在一定水平时,三臂Cox模型能有效地控制I类错误发生概率,并保证较高的检验效能,实际应用简单有效。结论本文推荐的方法,可以有效地应用于三臂生存时间数据的非劣效性研究。
Objective To introduce a new semi-parametric method for non-inferiority analyses in three-arm trials w ith censored timeto-event data.Methods Three-arm non-inferiority trials w ith time-to-event endpoints could be analyzed based on the Cox proportional hazard model.Simulation tests w ere conducted to investigate the validity of the proposed method,and the R-language program of calculating required sample size w as given.Results When the hazard of censoring remained w ithin a certain value,the new method can control the type I error and has great pow er.Conclusion The recommended method in this paper could be used in the three-arm non-inferiority trials w ith censored time-to-event data.
出处
《中国卫生统计》
CSCD
北大核心
2013年第4期476-479,483,共5页
Chinese Journal of Health Statistics
基金
国家青年自然科学基金资助(项目编号:81102201)
