无对照二分类资料的Meta分析方法及Stata实现  

Meta-Analysis of Non-Comparative Binary Outcomes and Its Solution by Stata

作　　者：王佩鑫[1,2] 李宏田[2,3] 刘建蒙[2,3] 

机构地区：[1]北京大学公共卫生学院,北京100191 [2]北京大学生育健康研究所,北京100191 [3]北京大学卫生部生育健康重点实验室,北京100191

出　　处：《循证医学》2012年第1期52-55,64,共5页The Journal of Evidence-Based Medicine

基　　金：国家自然科学基金资助项目(81072372)

摘　　要：目的介绍无对照二分类资料Meta分析方法及在Stata软件中的操作步骤。方法首先介绍3种数据类型无对照二分类资料Meta分析的原理及方法,再用Stata软件对3个实例数据进行Meta分析。结果无对照二分类资料Meta分析的关键是选择服从正态分布或可转化为正态分布的指标。3个实例数据经正态转换后进行Meta分析,结果与原文一致。结论 Stata软件可实现无对照二分类资料(含患病率、发病密度和比值)的Meta分析,操作简单,实用性强。Objective To introduce the method of meta-analysis for non-comparative binary outcomes and its realization in Stata. Methods We first introduced principles and methods of meta-analysis for three types of non- comparative binary outcomes, and then replicated results of three published recta-analyses in Stata. Results The key point of doing these meta-analyses was to choose the effect size indices which were of normal distribution or could be transformed into normal distribution. The replicated results were consistent with the original literatures. Conclusions Meta-analyses for three types of binary outcomes, including prevalence, incidence density, and odds, could be done in Stata conveniently.

关 键 词：二分类变量 无对照 STATA META分析 

分 类 号：R195.1[医药卫生—卫生统计学]