The generalisability of pharmacoeconomic studies: issues and challenges ahead.

Developing from a previous review, this article revisits the generalisability theme to summarise recent advances in methodology and provide an update of challenges faced by producers and users of pharmacoeconomic data. Our original evaluative criteria encompassed technical issues, applicability and transferability. The technical elements of best practice are comparatively uncontroversial: choosing relevant alternatives; transparent reporting of methods and findings; accessing and applying the best-quality evidence; using best methods to synthesise data; and using deterministic sensitivity analysis to explore potential systematic bias whilst employing probabilistic sensitivity analysis to explore the influence of random error at the whole model level. The applicability of economic findings within their original policy context (e.g. national analyses based on generalisable within-country data) can be determined, provided that best practice guidelines for economic modelling are adhered to. The transferability of economic findings (from one policy setting to another, e.g. country, region, clinical setting or patient population) requires careful exploration of changes in resource implications, unit prices and outcomes, a process facilitated again by transparent reporting of methods, adjustment for baseline risk and potentially by recent statistical developments intended to deal with hierarchically structured data. Although there is considerable consensus in the published literature about these key issues, limitations remain for economic analysis as implemented because of its opaqueness of method, failure to reflect the opportunity cost of decisions and lack of societal mandate. If the primary purpose of health economic evaluation is to help society to obtain the best value from limited resources, then, at a time when most technologically advanced societies need to engage with the realities of limited healthcare funding, technocratic solutions alone appear insufficient. Making health economic findings accessible to patients, clinicians and society, in the form of relevant narratives, will help this essential debate and expose assumptions underpinning economic analysis to broader critical inspection.