EQ-HWB-S

EQ-HWB-S Short Form and Its UK Pilot Value Set

The EQ-HWB-S, which EuroQol renamed the EQ-HWB-9 in 2025, is the part of the EQ Health and Wellbeing programme that can produce QALYs. The full EQ-HWB describes health and wellbeing in 25 items with no value set for the 25 items as a whole, so the developers chose nine items, embedded in the long form, that could be valued by members of the public and turned into utilities on the scale where full health is 1 and dead is 0. Mukuria and colleagues published a UK pilot value set in 2023, estimated from composite time trade-off and discrete choice data, with values running as low as minus 0.384 for the worst states. For a health economist the short form is the route by which loneliness, exhaustion, anxiety and a sense of control could enter a cost-utility analysis for patients, social care users and carers, although the instrument and its value set are still described as experimental and interim.

Nine items chosen for valuation

The website of the Extending the QALY project, hosted by the University of Sheffield, reports that the new measure has 25 items with a shorter version, the EQ-HWB-S, designed for valuation in the project's next stage. Brazier and colleagues describe the final selection of the 25-item and 9-item versions as following two rounds of stakeholder consultation. Papers published from 2025 onwards cite a EuroQol notice changing the name to EQ-HWB-9 and describe the 2022 version as experimental.

McDool, Mukuria and Peasgood list the nine items as mobility, activities, exhaustion, loneliness, cognition, anxiety, sadness or depression, control and physical pain. Several of these, such as loneliness, exhaustion and control, have no direct counterpart in a health-only descriptive system. None of the nine items comes from the self-identity domain of the long form, so the short form covers every domain of the EQ-HWB except self-identity. Because the items are embedded in the long form, Kinchin and colleagues could derive EQ-HWB-S responses from EQ-HWB answers in an Irish general population survey of 1,220 adults.

Like the long form, the short form asks about the last seven days and uses five-level response scales describing frequency, severity or difficulty. With nine items of five levels each, the number of response patterns is five raised to the ninth power.

$$5^{9} = 1953125$$

where 5 is the number of levels per item, 9 is the number of items and 1953125 is the number of distinct EQ-HWB-S states a value set has to score.

How the UK pilot value set was estimated

Stage 6 of the Extending the QALY project set out the valuation of the short form with the EQ-VT protocol, including a feasibility study and a planned deliberative workshop with members of the NICE Citizens Council. Mukuria and colleagues reported the feasibility study in 2023. Its aim was to test whether composite time trade-off and a discrete choice experiment, based on an international protocol, could derive utilities for the EQ-HWB-S and generate a pilot value set.

The study recruited a representative UK general population sample and ran the tasks in online videoconference interviews using EuroQol Portable Valuation Technology, with quality control used to assess interviewer performance. Composite time trade-off and discrete choice data were available for 520 participants. Participants understood the tasks and the EQ-HWB-S states, and interviewers reported high levels of understanding and engagement.

The data were modelled with Tobit, probit and hybrid models. Raw composite time trade-off values ranged between minus 1 and 1, with greater disutility for more severe states. The authors selected a hybrid Tobit heteroscedastic model for the pilot value set, which gives values of minus 0.384 to 1. Pain, mobility, daily activities and feeling sad or depressed carried the largest disutilities. The authors concluded that the EQ-HWB-S can be valued with composite time trade-off and a discrete choice experiment, recommended the pilot value set for testing the validity of the EQ-HWB-S, and called for further methodological work on a valuation protocol specific to the measure.

A hybrid model uses both kinds of evidence at once. Time trade-off answers place states on the scale anchored at full health and dead, while discrete choice answers give information on the relative value of many more states, and a joint model estimates one set of values from both.

From EQ-HWB-S responses to QALYs

A value set turns a nine-digit EQ-HWB-S profile into a single utility, and the utility is then weighted by time to give QALYs, as with any preference-based measure. The equation below is a simplified additive form, used only to show how item levels lower utility. It is not the specification of the pilot hybrid model.

$$U = 1-\sum_{k=1}^{9} d_k(l_k)$$

where $U$ is the utility of a state, $k$ indexes the nine items, $l_k$ is the level reported on item $k$ and $d_k(l_k)$ is the decrement the value set attaches to that level of that item, with $d_k(1) = 0$ for the best level.

The QALY gain for a person or group then follows the standard area calculation used for any utility instrument.

$$\Delta Q = \sum_{t} \Delta U_t \times T_t$$

where $\Delta Q$ is the QALY gain, $\Delta U_t$ is the difference in mean utility between the arms in period $t$ and $T_t$ is the length of that period in years.

Illustrative worked example with a carer support service

All decrements, utilities, costs and the threshold in this example are illustrative and do not come from the pilot value set. A carer support service costs GBP 2,400 per carer more than usual support and is evaluated over 2 years with no discounting.

1. Utility before the service. A typical carer reports moderate pain, some sadness and marked exhaustion, and every other item at the best level. With illustrative decrements of 0.12 for the pain level, 0.05 for the sadness level and 0.06 for the exhaustion level, the total decrement is 0.12 + 0.05 + 0.06 = 0.23 and the utility is 1 − 0.23 = 0.77.

2. Utility with the service. Pain is unchanged, sadness falls to the best level and exhaustion falls one level, with an illustrative decrement of 0.03. The total decrement is 0.12 + 0.03 = 0.15 and the utility is 1 − 0.15 = 0.85.

3. QALY gain. The utility gain is 0.85 − 0.77 = 0.08, worth 0.08 QALYs a year, and over 2 years 0.08 × 2 = 0.16 QALYs.

4. Cost per QALY. The ratio is 2,400 ÷ 0.16 = 15,000, or GBP 15,000 per QALY.

5. The same service measured with a health-only profile. In an illustrative alternative, an EQ-5D-5L profile completed by the same carers records a gain of only 0.02 QALYs a year. The QALY gain is 0.02 × 2 = 0.04 and the ratio is 2,400 ÷ 0.04 = 60,000, or GBP 60,000 per QALY.

Against an illustrative threshold of GBP 30,000 per QALY, the service looks cost-effective when measured with the EQ-HWB-S and not when measured with the health-only profile. The choice of instrument is in effect a decision about which effects count. Results from the EQ-HWB-S are most informative when shown beside results from the reference-case instrument, because a committee can then see how much of the gain comes from the wider items.

Evidence on validity and comparison with the EQ-5D-5L

Several studies have compared the short form with the EQ-5D-5L. Monteiro, Kuharic and Pickard surveyed 903 people on a US online panel, 400 cancer survivors and 503 members of the general population, with a preliminary version of the EQ-HWB-S. They found moderate to strong associations between conceptually overlapping dimensions of the two instruments. Only EQ-HWB-S items distinguished caregivers from non-caregivers and carers with low and high caregiver burden, although the effect sizes were small, and the authors concluded that the EQ-HWB-S performed similarly to or better than the EQ-5D-5L among patient groups.

McDool, Mukuria and Peasgood used two UK general population datasets, of 429 and 248 respondents. Conceptually overlapping dimensions correlated strongly with the EQ-5D-5L, utility scores were strongly associated (Spearman correlations of at least 0.5) and the two instruments separated known groups similarly. In 2026 McDool and colleagues tested the experimental 2022 version of the EQ-HWB-9 in a UK general population sample of 11,383, describing it as a preference-weighted instrument with an interim value set. They compared its items with the Short Warwick-Edinburgh Mental Well-being Scale, the ICECAP-A, the HUI3, the ONS-4 and the SIPHER-7, and reported substantive evidence of convergent and known-group validity.

How the EQ-HWB-S differs from related measures and value sets

The EQ-HWB-S and the EQ-HWB share items, recall period and response scales, but only the nine short-form items carry a value set. Utilities can come from a short-form administration or from the nine items embedded in a 25-item administration, while only the long form gives the full 25-item profile.

The EQ-5D and its value sets, such as the EQ-5D-5L value set, describe health. The EQ-HWB-S adds wellbeing items and was intended to give comparable utilities for patients, social care users and carers. ICECAP-A and ASCOT also reach beyond health, but ICECAP-A scores capability and ASCOT scores social care-related quality of life, so neither describes the same mix of health and wellbeing items that the EQ-HWB-S values. Instruments of caregiver quality of life are designed for carers alone, whereas the EQ-HWB-S is designed to be completed in the same form by carers and the people they care for.

NICE and the status of the value set

NICE describes the EQ-HWB as a new standardised measure created after the Extending the QALY project, says the EQ-HWB has not been developed as a replacement for the EQ-5D, and reports that research is under way to validate it and compare it with existing instruments, research that NICE says will help it decide whether and how to use the EQ-HWB in its evaluations in the future. NICE has not adopted the EQ-HWB-S or its pilot value set in the reference case for technology appraisal.

The link between NICE and the measure runs through its development. NICE says it contributed to the Extending the QALY project, and the project's valuation stage planned a deliberative workshop with members of the NICE Citizens Council. A QALY gain measured with the EQ-HWB-S would therefore sit outside the reference case.

Limitations and common misreadings

The pilot value set and the instrument version both carry caveats. These matter whenever an EQ-HWB-S utility or QALY result is interpreted.

• A pilot, not a final value set. Mukuria and colleagues called the 2023 set a pilot value set from a feasibility study, and McDool and colleagues in 2026 described the EQ-HWB-9 value set as interim. Utilities from it should be read as provisional.
• Version drift. The 2022 version is described as experimental, and a modified EQ-HWB EMV2.1 has since been tested. Modified items may produce versions whose responses do not match the valued states exactly.
• Different scale floors. The pilot set reaches minus 0.384, so QALY gains for severe states are not directly comparable with gains measured on instruments whose value sets have a different lowest value.
• Utility differences are not only preference differences. When the EQ-HWB-S and the EQ-5D-5L give different QALY gains for the same people, part of the gap comes from different descriptive systems and part from different valuation studies, and the two sources cannot be separated by comparing the final numbers alone.

Sources

• Brazier J, Peasgood T, Mukuria C, et al. The EQ-HWB: overview of the development of a measure of health and wellbeing and key results. Value in Health. 2022;25(4):482-491. doi:10.1016/j.jval.2022.01.009
• Brazier J, Ratcliffe J, Salomon JA, Tsuchiya A. Measuring and valuing health benefits for economic evaluation. 2nd ed. Oxford University Press; 2017.
• Kinchin et al. A comparative study of health and well-being measures in Ireland using EQ Health and Wellbeing (EQ-HWB) and its short version, EQ-5D-5L, and ICEpop Capability Measure for Adults (ICECAP-A). Value in Health. 2025. https://pubmed.ncbi.nlm.nih.gov/40320237/
• McDool E, Mukuria C, Peasgood T. Psychometric performance of the EQ Health and Wellbeing Short in a United Kingdom population sample. Value in Health. 2024;27(9):1215-1224. https://pubmed.ncbi.nlm.nih.gov/38795959/
• McDool E, Mukuria C, Wickramasekera N, Tsuchiya A. Evaluating the validity of the EQ-HWB-9 in a large United Kingdom general population sample. Value in Health. 2026;29(8):1452-1462. doi:10.1016/j.jval.2026.02.006
• Monteiro AL, Kuharic M, Pickard AS. A comparison of a preliminary version of the EQ-HWB Short and the 5-level version EQ-5D. Value in Health. 2022;25(4):534-543. https://pubmed.ncbi.nlm.nih.gov/35279371
• Mukuria C, Peasgood T, McDool E, Norman R, Rowen D, Brazier J. Valuing the EQ Health and Wellbeing Short using time trade-off and a discrete choice experiment: a feasibility study. Value in Health. 2023;26(7):1073-1084. doi:10.1016/j.jval.2023.02.008
• National Institute for Health and Care Excellence. Science policy and methods projects (web page, EQ-HWB entry). https://www.nice.org.uk/what-nice-does/our-research-work/our-projects-and-partners/science-policy-projects (accessed 17 October 2026)
• University of Sheffield. Extending the QALY (E-QALY) project website, Findings, stage 5 and stage 6. https://e-qaly.sites.sheffield.ac.uk/findings/stage-6 (accessed 17 October 2026)