A constructed binary-outcome network used throughout the examples and
tests. It is disconnected: sub-network 1 (Placebo, A, B) is
anchored on placebo, while sub-network 2 (A+B, A+B+C, A+B+D) shares
no treatment with sub-network 1. The shared components A and B bridge
the two, and components C and D are identified within sub-network 2.
Format
A data frame with 5 rows and 5 columns:
- studlab
Study label.
- treat1, treat2
Treatments compared (components joined by
"+").- TE
Log odds ratio of
treat1versustreat2.- seTE
Standard error of
TE.
The attribute "truth" holds the data-generating component log-odds
ratios.
Details
Studies S3 and S4 also have individual patient data
(cpaic_bin_ipd); their rows here are the unadjusted contrasts, which
cmaic() / cstc() replace with population-adjusted versions.