Evaluating the economic benefits of complex interventions: accounting for non-health effects

UoM administered thesis: Phd

Abstract

Abstract Decisions about how to allocate finite NHS resources are based on the premise that overall health in the UK population should be maximised. Certain complex interventions, such as those seen in genomic medicine, can produce outcomes which go beyond health. This work aimed to characterise and quantify this non-health value to inform the future development of new decision-making frameworks. The opening chapters of the PhD thesis set out the context and rationale for the work and a critical appraisal is presented about how the term ‘personal utility’ has been used in relevant literatures. Qualitative evidence synthesis (metaethnography) was used to develop a taxonomy of non-health value deriving from genomic-based diagnostic information. Three sources of non-health value were identified relating to; 1) informed decision-making, 2) benefit to others and 3) the intrinsic value of knowing (reducing uncertainty). This taxonomy provided a starting point for the design of a stated preference study to quantify potential trade-offs people would make between elements of value from genomic tests. A proof-of-concept study, involving face-to-face interviews, confirmed that people could trade between non-health and health outcomes of testing. Health outcomes were presented as changes in health-related quality-of-life (HRQL) as measured by the EQ-5D. A discrete choice experiment (DCE) was deemed suitable to quantify trade-offs. The DCE design phase also involved the development and testing (using think-aloud technique) of training materials including a web-based animation. Pilot testing of the online survey determined the feasibility of running a definitive DCE in a general population sample of target size n=1000. Findings from the DCE demonstrated that people are willing to trade-off between health and non-health outcomes of genomic test information. On average, participants would give up 0.13 units of HRQL (95%CI: 0.12 to 0.14) if test information resulted in a unit improvement in their ability to make an informed decision. Improvement in the test’s ability to benefit others meant that 0.09 of HRQL (95%CI: 0.08 to 0.10) would be traded. A HRQL decrement of 0.07 (95%CI: 0.05 to 0.08) would be acceptable for reduced uncertainty. Preference heterogeneity in pooled data was described using latent class analysis. The quantification of elicited trade-offs could be used to refine CEA methods in this context. The work has implications for the economic evaluation of complex precision medicine interventions comprising a diagnostic component.

Details

Original languageEnglish
Awarding Institution
Supervisors/Advisors
Award date1 Aug 2020