种族敏感性临床试验的样本量计算方法探讨

Sample Size Calculation Method of Ethnic Sensitivity Studies

目的考虑种族敏感性临床试验在各地区间招募人数不同的情况,介绍基于药代动力学的样本量计算方法,并探讨在多剂量情况下效能计算过程中存在的多重性问题。方法通过Monte Carlo模拟和数值积分的方法拟合效能函数,比较Bonferroni,Dunnet和Sheffe方法在多重性问题中的使用优劣。结果在不同偏态情况下,计算为达到相应的效能所需入组的受试者人数。在多剂量种族敏感性临床试验的设计中,为达到相同的整体置信水平,按照Dunnet方法给出的置信区间最窄。结论在相同的效能水平要求下,地区间入组人数不同时相比入组人数相同时需要额外增加样本量。在多重性问题中,Dunnet方法相比于其他两个方法优势明显。

Objective Introduce a sample size and power calculation method when numbers of subjects differs among regions for ethnic sensitivity studies based on pharmacokinetics（PK） parameter and solves the multiplicity issue when simultane- ously comparing ethnic effect over several doses. Methods Statistical model is constructed to calculate sample size and power with Monte Carlo simulation and numerical integration. Compare the performances of Bonferroni, Dunnet and Sheffe method in solving the multiplicity issue. Results Calculate the required sample size under different recruitment situation. Construction of confidence interval by Dunnet method is illustrated. Conclusion It requires minimal sample size when subject numbers are i- dentical among regions in order to achieve the same power~ Dunnet method is better than other two methods （Bonferroni and Dunnet method）.