Willingness-to-Pay Threshold

Willingness-to-Pay Threshold: Demand-Side Value of a QALY, λ and GDP-Based Benchmarks

In applied cost-effectiveness analysis, "willingness-to-pay threshold" is the everyday name for λ, the cost per QALY against which an ICER is judged. In its strict sense the term points to a demand-side quantity: how much other consumption individuals, or society acting on their behalf, would give up to gain one quality-adjusted life year. That consumption value of health is built up from people's preferences. It differs from a supply-side threshold, which measures the health displaced elsewhere when a fixed budget pays for something new. Both are forms of cost-effectiveness threshold, but they are estimated in different ways, usually give different numbers, and imply different things when a health technology assessment body compares an ICER with a benchmark.

This article explains the welfare economic idea behind the demand-side threshold and how it is estimated from stated-preference surveys and from the value of a statistical life, including the UK Social Value of a QALY study and the European EuroVaQ project. It sets out the evidence that preference-based values tend to exceed supply-side estimates, shows how the threshold enters the decision rule as the parameter λ in the ICER comparison and in net monetary benefit, and works through an illustrative example at two values of λ. It then covers the loose usage in which "willingness-to-pay threshold" names any decision threshold, including the NICE range, and the WHO convention of one to three times GDP per capita, with its main criticisms.

Consumption value versus opportunity cost

In welfare economics the value of a health gain is what the people who benefit, or the population that pays, would sacrifice to obtain it. A willingness-to-pay threshold in the strict sense applies this idea to a unit of health: it aggregates willingness to pay for health gains into a single amount per QALY, so that the threshold states the rate at which society is prepared to trade consumption for health. On this reading a technology is worth adopting when the monetary value of its health gain, at that rate, exceeds its cost.

A supply-side threshold asks a different question. Writing from the World Health Organization, Bertram and colleagues describe a willingness-to-pay threshold as an estimate of what a consumer of health care might be prepared to pay for a health benefit given the other demands on that consumer's resources, and contrast it with supply-side thresholds that estimate the health forgone when a payer spends part of a fixed budget on a new intervention and must reduce funding for existing care. Claxton and colleagues, in a 2015 report published in the NIHR Health Technology Assessment journal, define the cost-effectiveness threshold relevant to NICE in this second sense, as the additional cost that has to be imposed on the NHS to forgo one QALY through displacement, and put their central estimate at £12,936 per QALY (2008 expenditure, 2008 to 2010 mortality).

The two concepts can legitimately take different values. A systematic review by Vallejo-Torres and colleagues, published in 2016, grouped empirical studies into those that estimate the value society places on a QALY and those that estimate the opportunity cost to the health care system, and found that thresholds based on society's valuation of a QALY are generally larger. The authors drew the implication that some interventions with positive social net benefit, judged by individuals' preferences, might still not be an appropriate use of resources under a fixed budget. A gap of this kind says that people value a QALY more highly than the marginal cost at which current health spending produces one.

Estimating a value per QALY from preferences

There are two main empirical routes to a demand-side value. The first asks samples of the public, in contingent valuation or related stated-preference surveys, what they would pay for defined health gains and converts the answers into an amount per QALY. The second starts from values of a statistical life already used in public policy and converts them into a value per life year and then per QALY, using life expectancy, quality-of-life weights and discounting.

Stated-preference surveys

The Social Value of a QALY project, a UK study funded by the then National Coordinating Centre for Research Methodology and summarised by Donaldson and colleagues in 2011, tested the feasibility of combining respondents' answers to willingness-to-pay questions with their health state utility answers to arrive at values of a QALY. Most methods of aggregating the survey data produced values of £18,000 to £40,000 per QALY, while other aggregation methods produced implausibly high values. The authors concluded that neither their feasibility survey nor their modelling gave compelling evidence for moving the NICE threshold up or down.

The European Value of a Quality Adjusted Life Year project (EuroVaQ), funded under the EU Sixth Framework Programme and coordinated by Newcastle University from March 2007 to August 2010, set out to develop methods for a monetary value of a QALY across European member states, both by modelling from values of statistical lives and by survey research on societal willingness to pay. A 2015 analysis of its survey data by Pennington and colleagues covered 17,657 people in nine European countries, each valuing hypothetical one-QALY gains. Mean willingness to pay was PPP$11,000 for a QALY made up of quality-of-life improvements, PPP$19,000 for a life-extending scenario in which a coma is avoided, and PPP$29,000 when respondents faced imminent, premature death.

Reviews show how widely such estimates spread. Ryen and Svensson identified 24 studies containing 383 unique estimates of willingness to pay for a QALY, with a trimmed mean of €74,159 and a median of €24,226 at 2010 prices. The gap between the trimmed mean and the median shows how far a minority of high estimates can pull an average. Their regression analysis found higher values when the QALY gain came from life extension rather than quality-of-life improvement, and values that depended on the size of the gain being valued.

Values derived from the value of a statistical life

The value of a statistical life prices small changes in fatality risk, and several studies convert it into a value per QALY. Hirth and colleagues took 42 value-of-life estimates, converted them to 1997 US dollars and used life expectancy, age-specific QALY weights and a 3 per cent real discount rate to derive the implied value of a QALY. Median values differed sharply by method: $24,777 for human capital estimates, $93,402 for revealed preference studies of non-occupational safety, $161,305 for contingent valuation and $428,286 for revealed preference studies of job risk. The authors described the rules of thumb then used in US cost-effectiveness analysis, such as $50,000 per QALY, as arbitrary standards, and found that every median except the human capital one far exceeded them.

In the UK, Mason, Jones-Lee and Donaldson modelled a willingness-to-pay-based value of a QALY from the value of preventing a fatality used in public sector decision making, setting out one conventional and one new method. The Social Value of a QALY project combined that value with data on fatality age, life expectancy and age-related quality of life, and its models gave values of £10,000 to £70,000 per QALY. HM Treasury's Green Book (2026) takes the same demand-side approach for appraisal across central government: it values fatality risk with the value of a prevented fatality and points to supplementary guidance for the value of a QALY, and it notes that these values rest on average values from representative samples of the population.

Where λ enters the decision rule

In cost-effectiveness analysis the threshold appears as a single parameter, usually written λ, that converts health into money or money into health. The decision rules below are standard and hold whatever λ represents; what changes with the choice of λ is the meaning of a positive result.

$$\text{ICER} = \frac{\Delta C}{\Delta E}$$

where $\Delta C$ is the incremental cost and $\Delta E$ the incremental health effect in QALYs of the new option against its comparator. When $\Delta E$ is positive, the more effective, more costly option is preferred when its ICER is below $\lambda$, the threshold in money per QALY, and the two options are tied when the ICER equals $\lambda$.

$$\text{INMB} = \lambda \Delta E - \Delta C$$

where INMB is incremental net monetary benefit, the difference between the two options' net monetary benefits, each equal to $\lambda E - C$ with $E$ and $C$ that option's expected QALYs and costs. The new option is preferred when INMB is positive.

$$\text{NHB} = \Delta E - \frac{\Delta C}{\lambda}$$

where NHB is incremental net health benefit in QALYs, as set out by Stinnett and Mullahy, and $\Delta C / \lambda$ is the health that the extra cost is worth at the threshold. The new option is preferred when NHB is positive. When $\Delta E$ is positive the three rules always agree, since INMB and NHB are positive exactly when the ICER is below $\lambda$, and INMB equals NHB multiplied by $\lambda$.

With a demand-side λ, a positive INMB says that the population values the health gain, in consumption terms, more than the resources it uses, which is close to a cost-benefit judgement. With a supply-side λ, a positive NHB, or equally a positive INMB, says that the health gained exceeds the health displaced elsewhere in the budget. NICE's methods manual (PMG36, updated 31 March 2026) takes the supply-side reading. Section 6.3.1 states that, given the fixed NHS budget, the appropriate maximum acceptable ICER is the opportunity cost of programmes displaced by new, more costly technologies. Section 6.3.2 asks for net health benefits at values placed on a QALY gain of £25,000 and £35,000 when decision-making modifiers are applied, and states that negative net health benefits mean the technology's health benefits are not large enough to prevent overall health loss from healthcare not being funded elsewhere.

Worked example: one treatment judged at two values of λ

The figures in this example are illustrative. A new treatment costs £18,000 more per patient than standard care and yields 0.6 additional QALYs per patient. It is judged at two illustrative values of λ: £20,000 per QALY, standing for a supply-side estimate of the health displaced by extra spending, and £50,000 per QALY, standing for a preference-based consumption value of a QALY.

Step 1: the ICER.

$$\text{ICER} = \frac{18{,}000}{0.6} = 30{,}000$$

where 18,000 is $\Delta C$ in pounds and 0.6 is $\Delta E$ in QALYs, so the ICER is £30,000 per QALY gained. This is below £50,000 but above £20,000, so the treatment passes at the higher λ and fails at the lower one.

Step 2: incremental net monetary benefit.

$$\text{INMB} = 20{,}000 \times 0.6 - 18{,}000 = -6{,}000$$

where 20,000 is the lower λ in pounds per QALY, 0.6 is $\Delta E$ and 18,000 is $\Delta C$, giving an INMB of minus £6,000 per patient.

$$\text{INMB} = 50{,}000 \times 0.6 - 18{,}000 = 12{,}000$$

where 50,000 is the higher λ in pounds per QALY, giving an INMB of £12,000 per patient.

Step 3: net health benefit.

$$\text{NHB} = 0.6 - \frac{18{,}000}{20{,}000} = 0.6 - 0.9 = -0.3$$

where 0.9 is the QALYs that £18,000 is worth at the lower λ, giving an NHB of minus 0.3 QALYs per patient.

$$\text{NHB} = 0.6 - \frac{18{,}000}{50{,}000} = 0.6 - 0.36 = 0.24$$

where 0.36 is the QALYs that £18,000 is worth at the higher λ, giving an NHB of 0.24 QALYs per patient.

Each net benefit is the other multiplied or divided by λ: minus £6,000 divided by £20,000 is minus 0.3 QALYs, and £12,000 divided by £50,000 is 0.24 QALYs. The decision switches at λ = £30,000, the ICER itself. If £50,000 is read as the consumption value of a QALY and £20,000 as the health system's opportunity cost, the treatment is worth more to the population than it costs, yet funding it from a fixed budget would displace 0.3 QALYs more than it gains for every patient treated. That is the situation the Vallejo-Torres review describes, and it is why the label attached to λ has to match the question being asked.

"Willingness to pay" as a label for any threshold

In much applied work "willingness-to-pay threshold" names whatever value of λ an analysis applies, whatever its origin. Cost-effectiveness acceptability curves commonly label their horizontal axis "willingness to pay", and HTA thresholds set by policy are often described in the same words. The usage is convenient and widely understood, but it can suggest that a preference-based value, a budget-based estimate and an administrative benchmark are the same kind of number.

The NICE range shows why this matters. Under PMG36, below a most plausible ICER of £25,000 per QALY gained the decision is normally based on the cost-effectiveness estimate; as the ICER rises from £25,000 to £35,000 committees make explicit reference to factors such as uncertainty, uncaptured benefits and health inequalities; and above £35,000 an increasingly stronger case is needed. Section 6.3.1 presents the maximum acceptable ICER as the opportunity cost of displaced programmes, which NICE lacks the information to define precisely and so expresses as a range. The level of the range is also a policy decision: NICE reported on 1 December 2025 that the government would raise it from £20,000 to £30,000 per QALY to £25,000 to £35,000 from April 2026, a change designed to improve the operating environment for pharmaceutical companies. Describing the NICE range as society's willingness to pay for a QALY therefore misstates what the manual says it represents.

Blurring the two ideas also shifts arguments about the threshold's level. McCabe, Claxton and Culyer argued in 2008 that positive decisions above the NICE threshold on grounds of innovation reduce population health, and that reimbursement at the threshold transfers the full value of an innovation to the manufacturer. Survey evidence that the public values a QALY above the current benchmark bears on the value of health, and by itself gives no warrant for a fixed NHS budget to fund technologies at that price.

GDP per capita multiples and their critics

In low- and middle-income settings the best-known thresholds are multiples of gross domestic product per capita, and Bertram and colleagues describe three times GDP per capita per DALY averted as the most common. They trace these thresholds to the WHO Commission on Macroeconomics and Health in 2001 and its estimate of the economic value of a year of healthy life, which rested on assumptions about leisure time, non-health consumption, longevity and health-related quality of life. They note that these values can be compared with measures such as the value of a statistical life, which estimate an individual's willingness to pay to extend healthy life by one year, and Marseille and colleagues likewise treat a GDP-based threshold as an implied measure of a country's willingness to pay, which places the convention on the demand side.

As used by the WHO Choosing Interventions that are Cost-Effective project (WHO-CHOICE), an intervention costing less than GDP per capita per DALY averted was considered very cost-effective, and one costing less than three times GDP per capita per DALY averted was considered cost-effective. Bertram and colleagues, all WHO staff, state that none of these thresholds should be used alone as a decision rule for funding or as a measure of affordability, and that experience shows them to lack country specificity. They give the example of a WHO-CHOICE analysis in Peru that found trastuzumab for breast cancer cost-effective by the three-times rule, although adding it would have cost more than Peru's entire budget for breast cancer treatment.

Marseille and colleagues argued in the same journal in 2015 that the GDP-based thresholds have little theoretical justification and set so low a bar that very few interventions with evidence of efficacy can be ruled out. They pointed out that, although willingness to pay for health care is related to income, there is little evidence that the relationship is linear; if averted DALYs are valued more highly in high-income countries, the thresholds will tend to be too stringent there and too lax in low-income countries. They added that the rule does not adequately appraise affordability.

Misreadings of willingness-to-pay values

One error is to treat every threshold as a measure of society's willingness to pay. A threshold estimated from displaced NHS activity, a range written into a methods manual and a preference-based survey value answer different questions, and an analysis should state which one its λ represents.

A second error is to treat a preference-based value as a single, settled number. Published estimates vary with the elicitation method, as the Hirth review found; with whether the gain extends life or improves its quality and with the size of the gain, as Ryen and Svensson and the EuroVaQ survey found; and with how individual answers are aggregated, as the Social Value of a QALY survey showed. A third is to compare GDP-multiple thresholds, stated per DALY averted, directly with per-QALY thresholds from high-income HTA bodies, as if the outcome unit and the basis of the benchmark were the same.

A fourth is to conclude that a demand-side value above the current threshold means an HTA body should approve technologies at higher ICERs. Under a fixed budget the health consequences of an approval are governed by what it displaces, so a higher consumption value of health is an input to decisions about how much a society spends on health care, while each appraisal still has to account for the health its own costs displace.

Sources

• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). London: NICE; published 31 January 2022, last updated 31 March 2026. Sections 6.3.1 to 6.3.8.
• National Institute for Health and Care Excellence. Changes to NICE's cost-effectiveness thresholds confirmed. News article, 1 December 2025.
• Claxton K, Martin S, Soares M, Rice N, Spackman E, Hinde S, Devlin N, Smith PC, Sculpher M. Methods for the estimation of the National Institute for Health and Care Excellence cost-effectiveness threshold. Health Technology Assessment. 2015;19(14):1-503, v-vi. doi:10.3310/hta19140
• Vallejo-Torres L, García-Lorenzo B, Castilla I, Valcárcel-Nazco C, García-Pérez L, Linertová R, Polentinos-Castro E, Serrano-Aguilar P. On the estimation of the cost-effectiveness threshold: why, what, how? Value in Health. 2016;19(5):558-566. doi:10.1016/j.jval.2016.02.020
• Donaldson C, Baker R, Mason H, Jones-Lee M, Lancsar E, Wildman J, et al. The social value of a QALY: raising the bar or barring the raise? BMC Health Services Research. 2011;11:8. doi:10.1186/1472-6963-11-8
• European Commission, CORDIS. European value of a quality adjusted life year (EuroVaQ), project 44172, FP6 project fact sheet (1 March 2007 to 31 August 2010). Accessed 29 Sep 2026.
• Pennington M, Baker R, Brouwer W, Mason H, Hansen DG, Robinson A, Donaldson C, EuroVaQ Team. Comparing WTP values of different types of QALY gain elicited from the general public. Health Economics. 2015;24(3):280-293. doi:10.1002/hec.3018
• Ryen L, Svensson M. The willingness to pay for a quality adjusted life year: a review of the empirical literature. Health Economics. 2015;24(10):1289-1301. doi:10.1002/hec.3085
• Hirth RA, Chernew ME, Miller E, Fendrick AM, Weissert WG. Willingness to pay for a quality-adjusted life year: in search of a standard. Medical Decision Making. 2000;20(3):332-342. doi:10.1177/0272989X0002000310
• Mason H, Jones-Lee M, Donaldson C. Modelling the monetary value of a QALY: a new approach based on UK data. Health Economics. 2009;18(8):933-950. doi:10.1002/hec.1416
• HM Treasury. The Green Book (2026). Updated 5 February 2026.
• Stinnett AA, Mullahy J. Net health benefits: a new framework for the analysis of uncertainty in cost-effectiveness analysis. Medical Decision Making. 1998;18(2 Suppl):S68-S80. doi:10.1177/0272989X98018002S09
• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes, 4th edition. Oxford: Oxford University Press; 2015.
• McCabe C, Claxton K, Culyer AJ. The NICE cost-effectiveness threshold: what it is and what that means. PharmacoEconomics. 2008;26(9):733-744. doi:10.2165/00019053-200826090-00004
• Bertram MY, Lauer JA, De Joncheere K, Edejer T, Hutubessy R, Kieny MP, Hill SR. Cost-effectiveness thresholds: pros and cons. Bulletin of the World Health Organization. 2016;94(12):925-930. doi:10.2471/BLT.15.164418
• Marseille E, Larson B, Kazi DS, Kahn JG, Rosen S. Thresholds for the cost-effectiveness of interventions: alternative approaches. Bulletin of the World Health Organization. 2015;93(2):118-124. doi:10.2471/BLT.14.138206