i am perform a meta-analysis, some published studies provide median and 95%CI of hospital stays, according to some search, i found two ways to pools median data:
one is use excel or web-based (http://www.comp.hkbu.edu.hk/~xwan/median2mean.html, it can not be opened) tools to convert median and IQR/range to mean and SD, it looks dosen't suit me, because my data is median and 95% CI
another is using R package metamedian pool medians data or the difference of medians data,
pool.med() function need vector of medians with sample size(optional), it seems also dosen't suit the situation i faced
how to perform a meta-analysis of median and 95% confidence interval?
You can try the following spreadsheet http://www.iun.edu/~mathiho/medmath/Estimating.xls
The related article describes the underlying assumptions
Article Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and vari...
A more recent article (also with a spreadsheet) is the following
Article Estimating the sample mean and standard deviation from the s...
What is the definition for a confidence interval around a median? Producing a confidence interval traditionally requires calculating a standard error, which would be based on a mean, not a median.
You may look in this thread https://www.researchgate.net/post/how_to_perform_a_meta-analysis_of_median_and_95_confidence_interval?isAnswerFieldFocused=true and of course the Cochrane handbook. However, this assumption is met only when the result is based in data from normal distribution (interpreted as mean, 95% CI. I believe that for abnormal distributions this is not possible. The formula of Hozo et al may be used for median (range), however it does not apply to 95% CI in abnormal distributions. The problem is that in abnormal distribution the 95% CI do not reflect the actual range of data.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size