Skip to content

[R-meta] Meta-analysis with R, use treatment effect as input instead of 2 groups?

3 messages · Felix M, Guido Schwarzer

#
Dear All,


I?m fairly new to R and would like to meta-analyze data where I only have
available the differences between the outcomes for the experimental groups
and the control groups, as well as the SEMs and SDs of those delta values.

However it appears that metafor, as well as other meta-analysis packages
I?ve looked at, expect the mean effect sizes of the two groups and their
SDs as input.



Is it possible to use the delta values mentioned above as input for a
meta-analysis using the metafor package? And if not, does anyone know of an
R package with which this can be done?



Best regards

Felix
#
Am 17.10.17 um 17:57 schrieb Felix M:
I assume that you have the following information for each study:
delta = mean difference between experimental and control group
SEM = standard error of delta
SD = standard deviation of delta

In order to conduct a meta-analysis you only need delta and SEM.

You can use, for example, metagen() in R package meta in the following way:
metagen(delta, SEM)

Similarly, you can use rma() in R package metafor:
rma(yi = delta, sei = SEM)

See the corresponding help pages for more information, e.g. help(metagen).

Best wishes,
Guido
#
Dear Guido,

thank you very much for the swift reply. Yes, that is indeed the
information I have available for each study.

I will start testing this right away! I may have more detailed questions
later.

I'll study the help pages as well of course. I just didn't know where to
start on this particular issue, and I figured it would be best to just ask
people who already have lots of experience with the software. I'm glad this
mailing list exists!

Best regards, Felix

2017-10-17 20:18 GMT+02:00 Guido Schwarzer <sc at imbi.uni-freiburg.de>: