如题,请问两家临床单位,因为一些原因,病例数不按1:1来完成,如2:1可以吗?临床统计上允许吗?
可以,统计上允许。
yes. but:
1 you'd precify this in the protocol. you cannot change the allocation ratio after the study initiated without an amendment
2 an allocation ratio of 1:1 has the biggest Power compared with other ratios (such as 2:1), so if you want to maintain the desired power, the sample size might be larger than that of a study with an allocation ratio of 1:1
不需要也不强求等比,但数目别差别比例过大组间不平衡,因为毕竟都会有分层的。
如果方案没有规定2:1,那凑2:1是不合适的,但如果比如肥胖的某疾病患者地区间存在差异,极端情况也是可以不得不接受的。
我的建议是一切按照方案执行,如果需要改变病例数的分配,需要咨询统计专家,修改方案后方可实施。
可以,统计上允许。
yes. but:
1 you'd precify this in the protocol. you cannot change the allocation ratio after the study initiated without an amendment
2 an allocation ratio of 1:1 has the biggest Power compared with other ratios (such as 2:1), so if you want to maintain the desired power, the sample size might be larger than that of a study with an allocation ratio of 1:1
不需要也不强求等比,但数目别差别比例过大组间不平衡,因为毕竟都会有分层的。
如果方案没有规定2:1,那凑2:1是不合适的,但如果比如肥胖的某疾病患者地区间存在差异,极端情况也是可以不得不接受的。
我的建议是一切按照方案执行,如果需要改变病例数的分配,需要咨询统计专家,修改方案后方可实施。