Capability Index Values

Capability Index Values in Economic Evaluation Beyond Health

A completed capability questionnaire describes what a person is able to do and be, attribute by attribute, and an economic evaluation needs one number per person per time point. Capability index values supply that number by attaching a population weight to every attribute level, so that a profile on a measure such as ICECAP-A or ICECAP-O becomes a single score. The ICECAP values were estimated with best-worst scaling and sit on a scale where 1 is full capability and 0 is no capability, which is why a capability score multiplied by time gives years of full capability and not QALYs. The term here means the scoring of capability wellbeing measures built on the capability approach, not the process capability index used in statistical process control.

What a capability index value scores

The capability approach judges a person's position by the freedom to achieve valued doings and beings, not by health alone or by satisfaction with life. Capability measures turn that idea into a short descriptive system. ICECAP-A has five attributes for adults, ICECAP-O has five for older people, and the ICECAP Supportive Care Measure (ICECAP-SCM) has seven attributes for use at the end of life. The developers' guidance codes full capability as the top level and no capability as level 1.

The University of Bristol, which hosts the ICECAP measures, states that the value of a whole profile is found by adding the values of its attribute levels. The index is therefore additive.

$$I = \sum_{k=1}^{K} v_k(l_k)$$

where $I$ is the index value of one completed questionnaire, $K$ is the number of attributes, $l_k$ is the level reported on attribute $k$ and $v_k(l_k)$ is the published value for that level of that attribute. In a main effects set the value of a level on one attribute does not depend on the levels of the others.

Interactions are where the instruments part company. Best-worst scaling of single profiles cannot estimate interaction effects, so Huynh and colleagues went beyond best-worst scaling alone when they valued the ICECAP-SCM in the UK, and produced two value sets, one of which includes interaction terms.

Eliciting the values with best-worst scaling

ICECAP-A and ICECAP-O were valued with best-worst scaling alone. In the profile case a respondent sees one complete capability state and picks the best and the worst attribute level within it, and the pattern of choices across many states gives the relative value of every level on a common scale. The UK ICECAP-A values come from a survey of 413 people analysed by Flynn and colleagues and then rescaled into a tariff, and Coast and colleagues valued ICECAP-O with people aged 65 and over.

The ICECAP-SCM design was carried into Germany. Dams, Huynh and colleagues gave an online sample of 2,996 participants both a best-worst task, in which participants chose the best and worst statement within the same capability state, and a discrete choice experiment, in which they traded off between two capability states. They recommended tariffs based on the combined data for German evaluations. The tasks follow the general principles of choice experiment design, and the estimates come from logit-family models with latent classes for preference and scale.

Himmler, van Exel and Brouwer took a different route. They estimated ICECAP-O and ICECAP-A weights from experienced utility, using regression on a composite subjective wellbeing measure in two cross-sectional samples, and compared them with the existing decision utility weights. The experienced utility values were similar for some dimensions but deviated for others, most for enjoyment on ICECAP-O and for attachment and autonomy on ICECAP-A.

Anchoring at full capability and no capability

Best-worst scaling yields values that are relative to one another, with no built-in link to length of life, and rescaling with time trade-off tasks is one proposed route to a scale where death is zero. The developers' frequently asked questions describe the ICECAP index values as anchored on a full capability and no capability scale, with 1 for full capability and 0 for no capability. ASCOT took the time trade-off route, and the University of Kent ASCOT SCT4 Easy Read guidance v2.1 anchors its score at 0.00 for being dead and 1.00 for the ideal state.

The zero point decides what the score can be combined with. A health utility of zero means a state as bad as being dead, so the utility can weight survival time in a QALY. An ICECAP score of zero means the lowest level on every attribute. The ICECAP valuation tasks did not ask respondents to compare that state with death, and the developers instead treat death as a state of no capability, so their guidance gives a person who dies zero years of full capability.

Combining index values with time

An evaluation needs capability accumulated over follow-up, in the way that utilities accumulate into QALYs. The product of an ICECAP index and time is called Years of Full Capability (YFC), which the developers state is not equivalent to a QALY.

$$YFC = \sum_{j=1}^{J} I_j \, t_j$$

where $YFC$ is the years of full capability accrued over $J$ periods, $I_j$ is the capability index value in period $j$ and $t_j$ is the length of period $j$ in years.

Mitchell and colleagues developed Years of Sufficient Capability (YSC), which counts capability only up to a sufficient level. Kinghorn and Afentou describe the difference from YFC as attaching no additional societal value to improvements in wellbeing beyond the sufficiency level. In deliberative citizens' workshops in the UK, reported in 2019, Kinghorn defined the sufficient state on ICECAP-A as level 3 on all five attributes, the state 33333. In 2021 Kinghorn and Afentou reported a deliberative threshold of GBP 33,500 per additional year of sufficient capability.

NICE has not adopted a threshold for a cost per YFC or per YSC. Its guideline manual, Developing NICE guidelines (PMG20), names capability measures such as ICECAP in table 7.1, and NICE has said that section 7.4, which holds the table, will be updated after the EQ-5D-5L change that took effect on 27 August 2026.

National tariffs and other care-related value sets

Capability index values are country-specific in the same sense as a country-specific value set for a health measure, because they rest on the preferences of one population. Rohrbach and colleagues produced a Dutch ICECAP-A tariff from a profile case best-worst task completed by 1,002 members of the general population, following the methods of the UK tariff. Their selected model had three preference classes and two scale classes. Stability, attachment and enjoyment weighed more than autonomy and achievement, and moving a capability off a low level mattered more than moving it up from an already high level.

Wang, Sicsic and Chauvin reported a French ICECAP-A value set in 2026 from profile case best-worst tasks completed by a representative sample of 3,546 people, noting that ICECAP-A value sets had already been developed for the UK, the Netherlands and Hungary. Their best model had four preference classes, one of which strongly favoured autonomy. Sado and colleagues reported a Japanese scoring system for ICECAP-A in 2025, from best-worst scaling interviews with 400 participants across 16 hypothetical scenarios, in which enjoyment carried the largest share of the value space.

Care-related measures that are not capability measures have their own tariffs. Hoefman and colleagues derived a Dutch tariff for the CarerQol-7D, which describes the situation of an informal carer, from a discrete choice experiment with 992 adults from the general population. The tariff sums the relative utility weights of the seven dimensions plus an interaction term between physical and mental health problems. Baji and colleagues later estimated CarerQol tariffs for Hungary, Poland and Slovenia from about 1,000 respondents per country and concluded that the differences show the limited transferability of CarerQol tariffs across European regions. Nguyen and colleagues estimated Finnish population-based weights for the service-user ASCOT.

Illustrative worked example scoring one ICECAP-A profile

All values in this example are illustrative and are not the published UK ICECAP-A values, which should be taken from the developers' scoring files. Suppose a best-worst study yields, for each attribute, raw estimates measured relative to level 1, so level 1 is 0 on every attribute. Setting level 1 to 0 everywhere is itself a simplification, since the published UK tariff includes, for example, a level 1 value of minus 0.003.

1. Rescale to full and no capability. The level 4 estimates sum to 2.2 + 2.2 + 1.9 + 1.8 + 1.9 = 10.0, so one linear rescaling that puts full capability at 1 and no capability at 0 divides every estimate by 10.0.

$$v_k(l) = \frac{r_k(l)}{R}$$

where $v_k(l)$ is the index value of level $l$ on attribute $k$, $r_k(l)$ is the raw estimate for that level relative to level 1 and $R$ is the sum of the five level 4 estimates, here 10.0.

2. Score one profile. A person reports stability at level 3, attachment at level 4, autonomy at level 2, achievement at level 2 and enjoyment at level 3. The index is 0.18 + 0.22 + 0.09 + 0.09 + 0.16 = 0.74.

3. Score the profile after an intervention. A community programme lifts autonomy to level 3 and enjoyment to level 4. The new index is 0.18 + 0.22 + 0.16 + 0.09 + 0.19 = 0.84, a gain of 0.84 − 0.74 = 0.10.

4. Years of full capability. If the gain lasts 3 years, undiscounted, the person gains 0.10 × 3 = 0.30 YFC. At an extra cost of GBP 2,700, the ratio is 2,700 ÷ 0.30 = 9,000, or GBP 9,000 per YFC.

5. A simplified sufficiency count. On these illustrative weights the state 33333 scores 0.18 + 0.18 + 0.16 + 0.15 + 0.16 = 0.83. Counting gains only up to that index value, the credited gain is 0.83 − 0.74 = 0.09 a year. Over 3 years that is 0.09 × 3 = 0.27, and 2,700 ÷ 0.27 = 10,000, or GBP 10,000 per unit. This index-level cap shows the idea of sufficiency and is not the published YSC method.

The same change in one person's answers gives two different ratios, because under the sufficiency rule the part of the gain above the 33333 level earns no credit. Neither figure is a cost per QALY. The 0.74 means the person holds 74 per cent of the value space between no capability and full capability on these weights, and a health utility of 0.74 would mean something else entirely, so the result belongs beside QALYs in a cost-consequence analysis or in a separate capability analysis, not inside a cost-utility analysis.

How capability index values differ from neighbouring ideas

A value set for a health measure such as the EQ-5D converts health states into utilities anchored at full health and dead, and the resulting health state utility value can weight survival time directly. Capability index values play the same scoring role for capability measures but on different anchors. Best-worst scaling is the elicitation method, while the index values are its output after estimation and rescaling, so the same method can produce values for health measures and for capability measures.

Wellbeing scores are a further neighbour. The WEMWBS and SWEMWBS ask about positive aspects of mental wellbeing, whereas a capability measure asks what people are able to do, and the capability values come from preferences over profiles. The experienced utility weights of Himmler and colleagues sit between the two, scoring capability states by their association with subjective wellbeing. ASCOT weights score social care-related quality of life and, unlike the ICECAP values, are anchored at dead, while CarerQol tariffs score the carer's care situation.

Common misreadings of capability index values

The errors common to single ICECAP instruments, such as calling YFC a QALY, reading level codes in the EQ-5D direction or adding raw codes, apply to every capability tariff. Two errors arise only when instruments or countries are compared.

The first is to treat all care-related scores as sharing one scale. An ASCOT score of 0 is anchored at being dead and can fall below zero, while an ICECAP score of 0 is the lowest capability state, so the two cannot be pooled or compared point for point. The second is to treat a tariff as portable. The Dutch, French and Japanese ICECAP-A studies report weightings that differ from the UK set, and the CarerQol studies point the same way, so the tariff used should match the population whose preferences the analysis is meant to reflect.

Sources

• Baji P, Farkas M, Golicki D, et al. Development of population tariffs for the CarerQol instrument for Hungary, Poland and Slovenia: a discrete choice experiment study to measure the burden of informal caregiving. PharmacoEconomics. 2020. doi:10.1007/s40273-020-00899-2
• Coast J, Flynn TN, Natarajan L, Sproston K, Lewis J, Louviere JJ, Peters TJ. Valuing the ICECAP capability index for older people. Social Science and Medicine. 2008;67(5):874-882. doi:10.1016/j.socscimed.2008.05.015
• Dams J, Huynh E, et al. German tariffs for the ICECAP-Supportive Care Measure (ICECAP-SCM) for use in economic evaluations at the end of life. European Journal of Health Economics. 2021. doi:10.1007/s10198-020-01260-2
• Flynn TN, Huynh E, Peters TJ, Al-Janabi H, Clemens S, Moody A, Coast J. Scoring the ICECAP-A capability instrument. Estimation of a UK general population tariff. Health Economics. 2015;24(3):258-269. doi:10.1002/hec.3014
• Himmler S, van Exel J, Brouwer W. Happy with your capabilities? Valuing ICECAP-O and ICECAP-A states based on experienced utility using subjective well-being data. Medical Decision Making. 2020;40(4):498-510. doi:10.1177/0272989X20923015
• Hoefman RJ, van Exel J, Rose JM, van de Wetering EJ, Brouwer WBF. A discrete choice experiment to obtain a tariff for valuing informal care situations measured with the CarerQol instrument. Medical Decision Making. 2014;34(1):84-96. doi:10.1177/0272989X13492013
• Huynh E, Coast J, Rose J, Kinghorn P, Flynn T. Values for the ICECAP-Supportive Care Measure (ICECAP-SCM) for use in economic evaluation at end of life. Social Science and Medicine. 2017;189:114-128. doi:10.1016/j.socscimed.2017.07.012
• Kinghorn P. Using deliberative methods to establish a sufficient state of capability well-being for use in decision-making in the contexts of public health and social care. Social Science and Medicine. 2019;240. https://www.sciencedirect.com/science/article/abs/pii/S0277953619305404
• Kinghorn P, Afentou N. Eliciting a monetary threshold for a year of sufficient capability to inform resource allocation decisions in public health and social care. Social Science and Medicine. 2021;279. https://pubmed.ncbi.nlm.nih.gov/33964588/
• Mitchell PM, Roberts TE, Barton PM, Coast J. Assessing sufficient capability: a new approach to economic evaluation. Social Science and Medicine. 2015;139:71-79. doi:10.1016/j.socscimed.2015.06.037
• National Institute for Health and Care Excellence. Developing NICE guidelines: the manual (PMG20), chapter 7 Incorporating economic evaluation. Last updated 23 October 2025. https://www.nice.org.uk/process/pmg20/chapter/incorporating-economic-evaluation
• Nguyen L, Jokimäki H, Linnosmaa I, et al. Do you prefer safety to social participation? Finnish population-based preference weights for the Adult Social Care Outcomes Toolkit (ASCOT) for service users. MDM Policy and Practice. 2021. doi:10.1177/23814683211027902
• Rohrbach PJ, Dingemans AE, Groothuis-Oudshoorn CGM, et al. The ICEpop Capability Measure for Adults instrument for capabilities: development of a tariff for the Dutch general population. Value in Health. 2022;25(1). https://pubmed.ncbi.nlm.nih.gov/35031091/
• Sado M, Nagashima, Koreki, et al. Scoring system for the Japanese version of the ICECAP-A. Psychology Research and Behavior Management. 2025. https://pubmed.ncbi.nlm.nih.gov/40129962/
• University of Bristol. ICECAP frequently asked questions. https://www.bristol.ac.uk/population-health-sciences/projects/icecap/frequently-asked-questions/
• University of Bristol. Using ICECAP in decision making. https://www.bristol.ac.uk/population-health-sciences/projects/icecap/using-icecap-in-decision-making/
• University of Kent. ASCOT SCT4 Easy Read guidance, version 2.1. https://research.kent.ac.uk/ascot/wp-content/uploads/sites/2976/2026/07/ASCOT-ER_Guidance-v2.1.pdf
• Wang Z, Sicsic J, Chauvin P. Developing French population tariffs for the ICECAP-A using best-worst scaling. PharmacoEconomics. 2026;44(8):963-974. doi:10.1007/s40273-026-01614-3