中文摘要：

目的 探讨瓮模型在选择性交叉设计中的应用,提出一种新的基于加速瓮模型的选择性交叉设计（accelerated urn model-selective crossover design,AUM-SC）方法,并探索调速因子函数中参数k的合理取值范围。方法 AUM-SC设计是将改良后的瓮模型引入选择性交叉设计的一种新的临床试验设计方法,采用Monte Carlo方法模拟对不同疗效假定下AUM-SC设计在抗肿瘤新药临床试验中的随机化转组过程进行探讨。结果 AUM-SC设计中调速因子函数为加速因子函数,且加速因子函数中参数k在（1,3）范围时,安慰剂组病人能够以合理的转组可能性转到试验组继续接受治疗。结论在抗肿瘤新药临床试验中,如果试验药物确实有效,AUM-SC设计方法可以使病情发生进展的安慰剂组受试者以较高的转组可能性转到试验组接受治疗,从而最大程度地保护受试对象的利益。

英文摘要：

Objective To discuss the application of Um Model in Selective Crossover Design, we put forward a new Selective Crossover Design method based on Accelerated Urn Model（AUM-SC） and explore the reasonable range of parameter k in the control factor function. Methods AUM-SC design is a new method of clinical trial design, which employs the improved urn model in the Selective Crossover design. Under the assumptions of different efficacy effect, we use the Monte Carlo method to discuss the treatment switch process in oncology clinical trials. Results The simulation suggests that when the control factor function in the AUM-SC model is accelerated factor function and the parameter k varies through 1 to 3, the subject in the placebo group is able to switch to the treatment group with a reasonable ratio. Conclusion The simulation shows that the subjects in the placebo group of an oncology trial are able to switch to the treatment group with a higher ratio when the test drug has better efficacy so as to satisfy the ethical consideration.