2018/01/29 by Andrea Gabrio, Gabrio, Andrea, Alexina J. Mason +3
Decision Sciences · Economics, Econometrics and Finance · #Applications (stat.AP) #Capital Investment and Risk Analysis #Economic and Environmental Valuation #FOS: Computer and information sciences #Innovation Diffusion and Forecasting
paper · pdf · doi:10.48550/arxiv.1801.09541
openalex publication_date 2018/01/29 · openalex created_date 2023/02/16 · openalex updated_date 2026/07/28
Economic evaluations from individual-level data are an important component of\nthe process of technology appraisal, with a view to informing resource\nallocation decisions. A critical problem in these analyses is that both\neffectiveness and cost data typically present some complexity (e.g. non\nnormality, spikes and missingness) that should be addressed using appropriate\nmethods. However, in routine analyses, simple standardised approaches are\ntypically used, possibly leading to biased inferences. We present a general\nBayesian framework that can handle the complexity. We show the benefits of\nusing our approach with a motivating example, the MenSS trial, for which there\nare spikes at one in the effectiveness and missingness in both outcomes. We\ncontrast a set of increasingly complex models and perform sensitivity analysis\nto assess the robustness of the conclusions to a range of plausible missingness\nassumptions. This paper highlights the importance of adopting a comprehensive\nmodelling approach to economic evaluations and the strategic advantages of\nbuilding these complex models within a Bayesian framework.\n