Willingness to pay methods in health care: a sceptical view.

Recently, several experts in stated preference willingness to pay (WTP) methods have advocated greater use of these methods to facilitate cost–benefit analysis in health care [1–3]. The suggestion is that health economics lags behind other areas of economics that have embraced these methods, in particular environmental economics. A small but growing number of stated preference WTP studies have been conducted in the health field [4,5]. Two main methods have been employed: the ‘contingent valuation method’ (CVM) and ‘choice experiments’ (CE) – the method formerly known as conjoint analysis. These methods have generally been used to set a monetary value on a package of health and/or non-health benefits in the context of a specific intervention. Yet economic evaluation within the health care field remains dominated by cost-effectiveness and cost-per-QALY analysis. Health care payers have been reluctant to embrace cost–benefit analysis based on WTP methods [6,7]. And most health economists have preferred to refine the costeffectiveness approach rather than to develop new WTP methods [8,9]. Why is this? Advocates of WTP methods suggest it may be partly due to a common but erroneous perception that WTP studies are ‘somehow supportive of policies aimed at removing the provision of state-supplied health services’ [1]. It may also be due to the fact that stated preference WTP methods suffer from two serious (and possibly related) measurement biases that render them unattractive to health care decisionmakers. First, WTP responses tend to be undersensitive – although not necessarily totally insensitive – to the magnitude of benefit [10–12]. This includes both ‘scope effects’, involving different quantities of the same good, and ‘nesting effects’ (or ‘embedding effects’ or ‘part-whole bias’), involving one good incorporated within a larger bundle of goods. Scope effects are particularly strong in relation to health risks. Using high quality contingent valuation survey designs, and rigorous experimental methods, investigators have found that people tend to state a similar amount – roughly d50 – for any given magnitude of reduction in the risk of death or injury [13]. This has the effect of exaggerating implied monetary values for life and health for relatively small risk reductions. More generally, under-sensitivity to the magnitude of benefit tends to inflate valuations of interventions that yield relatively small benefits. Second, WTP methods tend to inflate valuations of the specific intervention that respondents are asked about, relative to interventions that respondents are not asked about [14]. Asking respondents to focus on one specific intervention in isolation acts as a kind of magnifying glass for stated WTP. When asked to consider an intervention in isolation, people are willing to pay sums of money far in excess of what they are willing to pay when asked to consider the same intervention in relation to a range of other interventions. This is sometimes known as ‘budget constraint bias’ [15]. Unlike the rational economic man of standard economic theory, survey respondents may be unable to budget simultaneously for the entire range of possible public and private goods and