ISSUE: By convention, values for generic preference-based measures (such as the EQ-5D) are anchored at 1 = 'full health' and 0 = 'dead'. Consequently, stated preference methods used to value health states often refer to 'death' and result in negative values for states 'worse than dead'. Evidence shows that including 'dead' in health state valuation could bias responses. The need to identify negative values creates methodological difficulties that have not been fully resolved. An alternative lower anchor (e.g. 'worst health state imaginable') could be used in place of 'dead'. However, this would constitute a dramatic shift in methodology and could have profound implications for the valuation of health states.

OVERVIEW: The panel will debate the motion that 'dead' should be removed from health state valuation. Jacoline Bouvy will introduce the issue, describing the role of 'dead' and identifying related challenges for health technology assessment and appraisal decisions. Audience members will vote on whether they support the motion. Liv Ariane Augested will provide an overview of relevant empirical work, outlining evidence demonstrating the problems caused by the use of 'dead' as an anchor. Chris Sampson will argue in favour of the 'drop dead' motion, describing the origins of the use of 'dead', arguing that it lacks foundations in either utility theory or applied cost-effectiveness analysis. He will suggest some practical alternatives. Bram Roudijk will argue in favour of maintaining 'dead', as a practical basis for estimating quality-adjusted life years. Bram will describe the 'zero condition' and its implications for health state valuation, highlighting the value of identifying health states as being 'worse than dead'. Jacoline will summarise the arguments and outstanding questions. At the end of the session, 15 minutes will be allocated for audience discussion, including a series of polls and a vote on the overall motion to 'drop dead'.