利用线性模型的置换检验实现Meta分析:基于SAS宏的实现

Meta-Analysis Using The Permutation Test of The Linear Model:Implementation Based on The SAS Macro

目的介绍利用线性模型的置换检验实现小样本研究的Meta分析方法。方法通过Fleiss93cont实例介绍美国南佛罗里达大学Kromrey等共同研发的一款基于线性模型的置换检验实现协变量分析的SAS宏命令(%METAPERM2)。该数据集中的小样本研究并不能满足正态性、独立性、方差齐性等前提假设。结果采用广义线性模型的回归系数为:X_1(年龄)=0.125,X_2(地区)=0.291。五种回归权重检验方法的结果为:传统加权最小二乘法β_1=0.000,β_2=0.338;Freedman Lane模型β_1=0.228,β_2=0.180;Kennedy模型β_1=0.472,β_2=0.557;Manly模型β_1=0.064,β_2=0.040;Ter Braak模型β_1=0.075,β_2=0.142。结论在正态性、独立性、方差齐性理论假设条件下,传统最小二乘法系数检验的显著性比任何置换检验都要大;小样本研究的Meta分析采用置换检验可能是一种更为合适的统计学方法。

Objective To introduce a meta-analysis method based on permutation test of linear models for small-sample meta-analysis.Methods A Fleiss93cont example was used to introduce a SAS macro(%METAPERM2)for covariate analysis developed by Kromrey of the Southern University of Florida in USA.This small-sample dataset did not satisfy the assumptions such as normality,independence and homogeneity of variance.Results The regression coefficient of generalized linear model was X1(age)=0.125,X2(area)=0.291.The results of the five regression weight test methods were:The traditional weighted least squares(WLS)β1=0.000,β2=0.338,Freedman Lane modelβ1=0.228,β2=0.180,Kennedy modelβ1=0.472,β2=0.557,Manly modelβ1=0.064,β2=0.040,Ter Braak modelβ1=0.075,β2=0.142.Conclusions Based on the hypothesis of normality,independence and homogeneity of variance,the significance of the traditional WLS coefficient test was larger than that of any permutation test,and permutation test may be a more suitable statistical method for small-sample meta-analysis.