摘要
目的 :介绍成组序贯 L ogrank检验及 类错误率损耗函数期中分析方法在生存资料中的应用。方法 :利用 α损耗函数计算期中分析的界值 ,通过界值与标准化 L ogrank Wald统计量的关系作出统计结论。结果 :序贯 L ogrank Score检验统计量类似独立、正态随机变量部分和 ,近似独立增加 ,其方差与失效数成比例。结论 :标准化序贯 L ogrank Wald检验统计量分布结构类似立即应答变量 ,而针对后者已建立许多成熟的理论 。
Objective:To introduce the application of group sequential logrank test and type one error spending function methods in survival data.Method:To determine the interim boundaries by type one error spending function and draw conclusions according the relationship between the boundaries and standardized logrank wald statistics.Result:Group sequential logrank score statistics behaved like a partial sum of independent normal random varible,with variance growing proportional to the number of events observed.Conclusion:Standardized group sequential logrank wald statistics has a distribution structure which is identical to that for the instantaneous respones problems for which much theory has been developed and can be used in more complex clinical trials with survival data.
出处
《现代预防医学》
CAS
2002年第5期613-614,623,共3页
Modern Preventive Medicine
关键词
临床试验
期中分析
成组序贯
I类错误率损耗函数
生存资料
Clinical trials
Interim analysis
Group sequential
Type one error spending function
Survival data
