How do preferences for public health interventions differ? a case study using a weight loss maintenance intervention

Abstract

Preference information is routinely used in healthcare decision-making, such as in health technology assessment, where preferences for health states are used to estimate quality-adjusted life years. However, there is debate over whose preferences should be considered. This debate extends to other areas in health economics where preferences are elicited to inform decision-making. Common stated preference methodologies include discrete choice experiments (DCEs) and contingent valuation (CV), which produce information that can be used in economic evaluations and benefit-risk assessments. This thesis expands the ‘preference debate’ beyond health utilities by considering the use of DCE and CV studies in decision-making. It also extends the debate by considering more than ‘patients’ and the general population; introducing four user groups that differ in their relative experience. The thesis’ aim is to explore whether, and how, differences in individuals’ experience of a healthcare intervention/health issue might influence preferences. It uses a weight loss maintenance intervention from the NULevel trial (ISRCTN14657176) as a case study. Chapter One introduces the thesis premise and Chapter Two outlines the economic theory underpinning the methodologies used. Chapter Three discusses the use of economic evaluation, providing a review of the ‘preference debate’ and recent developments. CV and DCE methodologies are next outlined and the use of information from DCEs in healthcare decision-making reviewed. Chapter Four systematically reviews the DCE literature in health economics to determine whether respondents can be classified as either a patient or general population sample. Building upon previous chapters, Chapter Five introduces an alternative framework for classifying respondent samples in DCE studies in terms of their relative experience; this defines the user groups for the empirical work (conducted via NULevel and an online panel). The research questions are next outlined. Chapter Six summarises the NULevel trial, the recruitment plan and the online survey design (containing both a DCE and CV task) used to address the research questions. Chapters Seven to Nine present results and Chapter Ten provides a discussion of implications, strengths and limitations of the work. The empirical work suggests that preferences differ according to the relative experience of respondents; raising concerns surrounding the generalisability of DCE studies. This is important given the increasing application of preference studies in decision-making. If this trend continues the implications will be even greater in future.