[R-meta] Issue with netmeta package

Chrysostomos Kalyvas CKalyvas at mapigroup.com
Tue Aug 8 18:13:54 CEST 2017

Dear All,

We are using the netmeta package in R to conduct a frequentist network meta-analysis. However, we have a trial with 4 arms of which 1 is the reference treatment (Topiramate 100mg) and the 3 are the same active treatment (they only differ in the way were administered). The outcome of interest is binary and the data is as follows:

r1            r2            r3            r4            n1           n2           n3           n4           t1            t2            t3            t4            na           Study
0.5          0.5          0.5          0.5          11           11           11           11           Topiramate 50mg             Topiramate 50mg                Topiramate 50mg             Topiramate 100mg          4              Lo 2014

The error that we get from R is "Error in pairwise(treat=list(t1, t2, t3, t4), event=list(r1, r2, r3,:Identical treatments for the following studies:'Lo 2014' - 'NA' - 'NA' - 'NA' - 'NA'"

Our preference is to not use an indicator for the treatments that are the same as we only need to test the efficacy and not the way that were administered.

Do you might know a potential solution to that issue?

Many thanks in advance.

Best regards,

	[[alternative HTML version deleted]]

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