Table 3 Advantages and limitations of three methods of analysis of multi-domain scores
Method | Main advantage | Main limitation |
|---|---|---|
GPC with non-prioritized outcomes [13] | Can be applied to outcomes of all types | Statistically significant treatment effects may be due to small differences |
GPC with prioritized outcomes [3] | Prioritizes outcomes and allows thresholds of clinical relevance to be specified | Net treatment benefit may be difficult to understand (in probabilistic terms) |
Multi-domain responder index [19] | Has better statistical power than a single responder analysis | Requires ability to define “responder” status for all outcomes |