- 1.1 Using the relationships and , we obtain and .
- 1.2 Using the formula , we find
- 1.3 The treatment network is given in Figure 1.1c. Notice that an indirect comparison can be estimated when there is no single common comparator, although the standard error obtained is quite large:
- 1.4 Analysis 1 comes from a network meta-analysis, and analysis 2 from pairwise comparisons. We do not see a coherence relationship between the posterior means because the odds ratios have skewed posterior distributions. However, we observe coherence in the posterior medians as these will reflect the approximately normal log odds ratios. For example, .
- 1.5 This can ...