临床随访研究中样本含量的估计  

Estimation of sample size needed in clinical follow-up study

作　　者：陈素领[1] 魏朝晖[2] 于浩[1] 陈峰[1] 

机构地区：[1]南京医科大学流行病与卫生统计学系,南京210029 [2]上海第二医科大学生物统计教研室,上海200025

出　　处：《中国新药杂志》2004年第8期732-735,共4页Chinese Journal of New Drugs

摘　　要：目的:介绍临床随访研究中生存分析资料的log-rank检验所需样本含量的估计法。方法:以离散性Markov链拟合生存过程,据此计算log-rank统计量的数学期望和方差,导出样本含量估计公式。结果:实例分析表明,该法能较好反映实际情况,应用灵活。结论:本法是一种有效、可行的样本含量估计法,值得推荐。Objective:To introduce method for estimation of sample size needed in log-rank test of survival data in clinical follow-up study. Methods: The survival process was modeled by discrete Markov chain and the asymptotic expectation and variance of log-rank statistics were calculated thereafter. The sample size formula was thus derived. Results: Practical case analysis indicated that actual situation may be appropriately reflected by the method and be conveniently used. Conclusion: The method for estimation of sample size described above is effective and practical and worth to be recommended.

关 键 词：样本含量 生存分析log-rank检验 MARKOV模型 

分 类 号：R969.4[医药卫生—药理学] R821.31[医药卫生—药学]