K层四格表OR值的一致性检验方法模拟比较研究  

A Monte Carlo Investigation of Homogeneity Tests of the Odds Ratio in K×2×2 Tables

作　　者：潘建红[1] 何春拉[1] 陈靖[1] 陈平雁[1] 

机构地区：[1]南方医科大学公共卫生与热带医学学院生物统计学系,510515

出　　处：《中国卫生统计》2013年第5期626-629,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助(30972554);广东省自然科学基金资助(915180200400001)

摘　　要：目的比较11种K层四格表OR值的一致性检验方法的优劣,并为其应用提供依据。方法采用Monte Carlo模拟方法,考虑第i层样本量N i、OR i值、对照组暴露概率π0i和层数K等四种参数的不同组合,使用SAS9.2软件编程,检验水准取值为0.05,比较各个检验方法的Ⅰ型错误以及检验效能。结果与结论 Zelen精确概率法过于保守;Mantel-Haenszel方法由于计算的便捷性,在样本量以及层数较大时,可以视为最佳的选择;Breslow-Day-Tarone法比BreslowDay法更为保守,因此本研究不建议对Breslow-Day法进行校正;近似非条件方法与近似条件方法相比,前者优于后者;加权最小二乘方法相对于Mantel-Haenszel以及近似非条件方法来说偏于保守,但是当样本量很大时,由于计算相对简单,因此,被广泛使用于Meta分析中;稀疏数据检验方法不提倡使用。Objective To compare the 11 homogeneity test methods of the odds ratio in K× 2 × 2 tables. Methods Data with differ- ent setting of parameters N （ sample size）, OR, π0 （ exposure rate in con- trol） of the ith stratum and K （the stratum number） were generated using Monte Carlo in SAS9.2OR. The type one error and power of the 11 meth- ods were compared. Results ＆Conclusion Zelen exact test is too conservative;Mantel-Haenszel methods should be viewed as the best choice when the sample size and K are large because of its facility in computation; In the study Breslow-Day test should not be corrected because Breslow- Day-Tarone test is more conservative than Breslow-Day test;The asymptot- ic unconditional methods perform better than the asymptotic conditional methods;Although the weighted least square method is more conservative than Mantel-Haenszel methods and asymptotic unconditional methods, it is widely used in meta-analysis because of its facility in computation when the sample size is large;The sparse data test methods should be discarded.

关 键 词：一致性检验 OR值 MONTE Carlo 

分 类 号：R195[医药卫生—卫生统计学]