Yixue xinzhi zazhi (Aug 2020)

Meta-analysis and R software-based implementation based on median as effect size

  • Qiao HUANG,
  • Joey S.W. KWONG,
  • Tie-Jun TONG,
  • Zhi MAO,
  • Liang ZHENG,
  • Hao CHEN,
  • Xiao-Feng HE,
  • Sheng-Jie DONG,
  • Chao ZHANG,
  • Cheng-Feng QIU,
  • Hong WENG,
  • Yu-Liang GUI,
  • Xue-Qun REN,
  • Xian-Tao ZENG

DOI
https://doi.org/10.12173/j.issn.1004-5511.2020.04.10
Journal volume & issue
Vol. 30, no. 4
pp. 265 – 271

Abstract

Read online

The median is a descriptive statistic for continuous variable following abnormal distribution; hence, we aimed to explore the feasibility of meta-analysis based on median as effect size. Statistical basis and common implementation methods of meta-analysis using median as effect size are described and a metamedian package in R statistical software was applied to conduct meta-analysis and display of results based on median, heterogeneity in meta-analysis was also measured if suitable. This article provides another feasible solution to deal with median as effect size for meta-analysis.

Keywords