单组临床试验目标值法的精确样本含量估计及统计推断  

Exact method of sample size calculation and statistical inference in single-arm clinical trials with objective performance criteria

作　　者：成琪[1] 刘玉秀[1] 陈林[1] 刘丽霞[1] 

机构地区：[1]南方医科大学南京临床医学院(南京军区南京总医院)医务部

出　　处：《中国临床药理学与治疗学》2011年第5期517-522,共6页Chinese Journal of Clinical Pharmacology and Therapeutics

基　　金：江苏省"六大人才高峰"第三批资助项目(06-C-031)

摘　　要：目的:针对结局为二分类变量以率作为终点评价指标的单组临床试验,探讨基于二项分布原理进行单组目标值法样本含量估计及统计推断的精确方法。方法:系统综合和全面论述了单组临床试验目标值法所涉及到的样本含量计算、可信区间估计和假设检验的精确方法。结果:按目标值起点为75%,间隔1%增加,给出预期事件发生率起点为76%,间隔1%增加,与目标值对应的精确样本含量结果,并列表给出β在0.20条件下,α分别取单侧和双侧0.05水平下的两套结果,可供直接查用。提供了可信区间估计及假设检验的精确计算公式。结论:为单组临床试验目标值法的样本含量计算、可信区间估计及假设检验提供了系统、实用的方法学支持。AIM： To explore the exact method of sample size estimation and statistical inference in single-arm clinical trials with Objective Performance Criteria（OPC） based on binomial distribution principle. METHODS： This article will systematically and comprehensively discuss the exact method of sample size calculation,confidence interval estimation and hypothesis testing which the single-arm clinical trial method involves in.RESULTS：With the starting point of OPC at 75% by an interval of 1% increasing and the starting point of expected event rate at 76% by an interval of 1% increasing,corresponding exact sample sizes are calculated.And two sets of results under the condition of β of 0.20 and α of 0.05（one sided and two sided） are tabulated respectively.The exact calculating formulas of confidence interval estimation and hypothesis testing are provided.CONCLUSION： Systematic methodology support is provided for OPC method in single-arm clinical trials in calculating sample size,estimating confidence interval and hypothesis testing.

关 键 词：单组临床试验 目标值 精确法 样本含量 可信区间 假设检验 

分 类 号：R311[医药卫生—基础医学]