“Common heterogeneity estimates” in network meta-analysis in Stata

[Naz Islam]

Hi StataGurus!
Most papers on network meta-analysis say that they assume "common heterogeneity estimates" (referred to as tau-square, tau2) across different treatment loops. They also say that this tau2 is estimated from the network meta-analysis.

In Stata I tried

network meta c

In one of the Stata manuals it was said that

mvmeta

command saves the tau2 matrix as r(tau2)

Hence I tried to save this matrix with

matrix tau2 = r(tau2)

and to display the matrix with

matrix list tau2

, but it returned an empty matrix.

When we ask for

ifplot

, it yields all the loop-specific estimates of the heterogeneity (tau2). I was wondering if taking an average of those is a valid way (I suppose not, but putting it here).

How can I get an estimate of "common heterogeneity estimates" across the comparisons, in network meta-analysis, in Stata?

[Elen Xr]

Hi Naz,

I am dealing with the same problem. May ask you if you had managed to find a solution ??

thank you