一般反向方差法在Meta分析的应用  

Application of Generic Inverse Variance Method in Meta-analysis

作　　者：向安莉[1] 张铃[1] 叶方立[1] 

机构地区：[1]武汉科技大学医学院,武汉430065

出　　处：《数理医药学杂志》2012年第1期9-11,共3页Journal of Mathematical Medicine

摘　　要：以探讨肥胖为前列腺癌病因可能性的Meta分析为例,检索近10年公开发表的有关肥胖与前列腺癌的研究文献,纳入文献6篇,累计病例3328例,对照535724例,分别采用RevMan 5.0软件中的分类数据和一般反向方差法进行Meta分析,结果显示,两方法的OR值及95%CI等统计量相似,OR=1.11,95%CI为1.01～1.23。In the Meta analysis of dichotomous data, it often happens that literatures do not directly provide original data but only the effect value OR and 95% CI. By adopting general reverse variance method to take the logarithm of the effect value and transferring 95 % CI lower limit into standard error bY logarithmic, this problem is hopefully solved. This paper adopts Meta analysis in an attempt to explore the possibility that obesity causes prostate cancer. The paper retrieves research literature of the past ten years on obesity and prostate cancer, including 6 articles, a total of 3328 cases of illness compared with 535724 control cases. With the dichotomous data in RevMan 5.0 software and general reverse variance method of Meta analysis, the paper demonstrates that the results obtained from the two methods, OR value and 95% CI statistical quantities are similar, with OR being 1.11, 95% CI somewhere between 1.01 and 1.23.

关 键 词：META分析 一般反向方差法 肥胖 前列腺癌 

分 类 号：R195.1[医药卫生—卫生统计学]