Avoiding Frankenstein’s Monster and Partial Analysis Problems: Robustly Synthesising, Translating and Extrapolating Evidence

摘要: Chapter 2 highlighted how satisfying decision analytic principles of coverage and comparability with joint consideration cost effects are fundamental to allowing unbiased trial-based effectiveness analysis for two-strategy comparisons. In this chapter, we extend these identify methods robust relevant making in any given jurisdiction interest (e.g. Australia a PBAC remit, the UK NICE remit), while highlighting common biases, pitfalls fallacies that should be avoided. Satisfying is shown usually require evidence synthesis translation potentially extrapolation. particular, required which avoid inferential selection biases associated partialisation (O’Brien 1996; Briggs O’Brien 2001) alternative framing binary outcomes (Eckermann Willan 2009, 2011).