摘要
目的对于多中心临床试验随机化采用竞争入组的情形,应该共用一个随机分配序列,还是各中心使用各自的随机分配序列,目前未见研究报道。本文将通过探讨整个分配过程中以及分配完成后整体和中心内部的均衡性行为,为多中心临床试验随机化竞争入组方式的选择提供参考和实用性指引。方法以随机化性能较好的大棒法(big stick design,BSD)作为随机化方法,通过设定不同的中心数、样本量以及最大容许不平衡例数(maximal tolerated imbalance,MTI),借助随机模拟的方法,分析两种不同竞争入组方式下整体和中心内部的均衡性行为。结果随着样本量的增加,各中心共用一个序列的过程均衡性变差,而选用各自序列时则保持稳定。随着中心数的增加,各中心选用各自序列的过程均衡性变差,而共用一个序列时变好;多中心共用一个序列时无法通过使用较小的MTI来减少中心内部的不均衡。中心数不是很多时,各中心使用各自序列最终出现严重不均衡的比例非常小。通过模拟可以找到各中心使用各自序列竞争入组比共用一个序列均衡性更优的最小样本量,结果表明最小样本量和中心数存在线性关系。结论如中心数非常多,分配到各中心的例数很少时,可考虑多中心共用一个随机分配序列,但需要持谨慎的态度;其他情况,一般推荐各中心使用各自的随机分配序列。
Objective There is no research reported on whether shared or respective randomly allocated sequence should be adopted for multi-center clinical trials implementing randomized competitive enrolment scenario.Herein,we provided references and practical guidance for the selection of randomized competitive enrolment mode in multi-center clinical trials by exploring the overall and central internal balance behavior.Methods Select the big stick design(BSD)with better randomization performance as the randomization method,by setting different number of centers,sample size and maximum tolerated imbalance(MTI),and the overall and central internal balance behaviors under two different competitive enrolment modes were subsequently analyzed using random simulation.Results The process balance of centers applying shared sequences deteriorated along with increasing the sample size,while the balance of centers with respective sequences remained stable.On the other hand,the process balance of centers with respective sequences degraded along with increasing the number of centers,yet the balance of centers with shared sequences improved.Furthermore,for centers applying respective sequences,it is unable to reduce the internal imbalance by adopting a smaller MTI.And it was less likely to obtain results with severe imbalance for centers applying respective sequences when a moderate number of centers was considered.Importantly,it was evidenced through simulation that a better minimum sample size of centers applying respective sequences could be identified than that of centers with shared sequence,and that minimum sample size had been linearly correlated with the number of centers.Conclusion If the number of centers is very large,and the number of cases assigned to each center is very small,it may be considered that the multi-center shares a random allocation sequence,but it needs to be cautious;in other cases,it is generally recommended to adopt respective random allocated sequences in multi-center clinical trials.
作者
袁阳丹
刘伟杰
刘玉秀
陈丽嫦
陆梦洁
刘雅琦
刘曼
Yuan Yangdan;Liu Weijie;Liu Yuxiu(Department of Biostatistics,School of Public Health,Southern Medical University(510515),Guangdong)
出处
《中国卫生统计》
CSCD
北大核心
2020年第3期335-339,共5页
Chinese Journal of Health Statistics
基金
国家自然科学基金面上项目(81473066)
国家自然科学基金青年科学基金项目(81803338)。
关键词
多中心临床试验
随机化
竞争入组
均衡性
Multi-center clinical trials
Randomization
Competitive enrollment
Balance
