[R-meta] Network Meta-analysis metafor

Caroline Woehl caroline.woehl at pg.hs-fulda.de
Wed Feb 14 08:45:29 CET 2018

Dear all,

I've only calucated network meta-analyses with the package netmeta yet.
Due to the possibility to calculate subgroup analyses, I'd like to have
a try at a network meta-analysis with metafor.

The analysis is based on 20 randomized controlled trials. They compare
physical activity or cognitive activity with usual care. One of them is
a multi-arm trial comparing physical and cognitive activity with usual
care. I could extract sample size, mean und standard deviation
(posttreatment) from each study arm. Thus, the file contains the study,
treat1, treat2, n1i, m1i, sd1i, n2i, m2i and sd2i.

Study                                                treat1       treat2
        n1i          m1i      sd1i        n2i          m2i       sd2i
       1      physical activity    usual care     37   17,16    4,33   
36   15,17    4,48
       2                    physical activity    usual care     39  
19,40    8,68    35   18,48    8,40
       2                    cognitive activity   usual care    36  
19,98    8,16    35   18,48    8,40
       3      cognitive activity usual care    35    42,34   9,30    33 
  44,49   9,53

So far, I’ve calculated the standardized mean differences (yi) and
variances (vi) for each pairwise comparison.
nma <- escalc(measure="SMD", m1i=m1i, sd1i=sd1i, n1i=n1i, m2i=m2i,
sd2i=sd2i, n2i=n2i, data=nma)

How can I define usual care as the reference group in the next step?

Many thanks for your help.

Best regards,

More information about the R-sig-meta-analysis mailing list