张天嵩, 熊茜. 有序数据的Meta分析方法及SAS实现[J]. 循证医学, 2012, 12(2): 125-128. DOI: 10.3969/j.issn.1671-5144.2012.02.019
    引用本文: 张天嵩, 熊茜. 有序数据的Meta分析方法及SAS实现[J]. 循证医学, 2012, 12(2): 125-128. DOI: 10.3969/j.issn.1671-5144.2012.02.019
    ZHANG Tian-song, XIONG Qian. Meta-Analysis of Ordinal Data and Its Solution by SAS[J]. Journal of Evidence-Based Medicine, 2012, 12(2): 125-128. DOI: 10.3969/j.issn.1671-5144.2012.02.019
    Citation: ZHANG Tian-song, XIONG Qian. Meta-Analysis of Ordinal Data and Its Solution by SAS[J]. Journal of Evidence-Based Medicine, 2012, 12(2): 125-128. DOI: 10.3969/j.issn.1671-5144.2012.02.019

    有序数据的Meta分析方法及SAS实现

    Meta-Analysis of Ordinal Data and Its Solution by SAS

    • 摘要: 目的 介绍有序数据的Meta分析方法及其在SAS软件的实现。 方法 以实例说明,采用两步法: 第一步,基于累积比数模型,采用SAS中的GENMOD过程计算产生每个研究的效应量及其标准误;第二步,采用固定效应模型和随机效应模型以SAS中的MIXED过程进行经典Meta分析。 结果 固定效应和随机效应模型Meta分析结果显示,汇总比值比及95%可信区间均为2.853 9 (2.106 4,3.864 8)。 结论 对于有序数据,可以应用SAS中的GENMOD过程和MIXED过程进行Meta分析。

       

      Abstract: Objective To introduce methodology for undertaking a meta-analysis on ordinal data and its realization in SAS. Methods It was illustrated on an example data by using two-step method. In the first step, it was calculated treatment effect and its standard error for an individual study by using PROC GENMOD in SAS based on cumulative odds models; in the second step, a classical meta-analysis was proposed for fixed and random effect model by using PROC MIXED in SAS. Results Results from the traditional meta-analysis of summary odds ratio and its 95%CI were same 2.853 9(2.106 4,3.864 8) for fixed and random effect model. Conclusions The procedure SAS PROC GENMOD and MIXED were utilized to undertake a meta-analysis on ordinal data.

       

    /

    返回文章
    返回