EORTC QLU-C10D

EORTC QLU-C10D Cancer-Specific Utilities for QALYs

The QLU-C10D turns answers to the EORTC QLQ-C30, the core cancer quality of life questionnaire, into a single utility on the scale that weights time in a QALY. Its classification uses ten dimensions with four levels each, and every country value set gives the utility of each combination as full health minus a set of level decrements. For a health economist the attraction is practical, since an oncology trial that collected the QLQ-C30 without the EQ-5D can still be scored with the QLU-C10D directly, instead of predicting EQ-5D values through a mapping function. The first value set, for Australia, put the worst state the classification describes at minus 0.096, slightly worse than dead, and the UK value set put it at minus 0.083.

How the ten-dimension classification was built from the QLQ-C30

King and colleagues set out to derive a health state classification system from the QLQ-C30 that could carry utility weights. Their conceptual model followed the questionnaire's existing domain structure, and expert opinion and patient input informed the choice of dimensions, with 79 cancer patients rating the relative importance of items.

Confirmatory factor analysis supported the model across primary cancer sites. The ten dimensions are physical, role, social and emotional functioning, pain, fatigue, sleep, appetite, nausea and bowel problems. The EORTC user manual describes the classification as 13 QLQ-C30 items forming the 10 dimensions, so some dimensions draw on more than one item, and each dimension has four levels. The development paper by King and colleagues reported 12 items for the ten dimensions, and the published erratum to that paper corrects only a reference.

Ten dimensions with four levels give a large descriptive system.

$$N = 4^{10} = 1048576$$

where $N$ is the number of distinct health states the QLU-C10D can describe, 4 is the number of levels per dimension and 10 is the number of dimensions. No valuation study can value each state directly, so value sets are estimated from a sample of states and a statistical model.

The QLU-C10D is used with the QLQ-C30 and needs no extra questions. A trial that administered the core questionnaire can be scored after the event, which is why the measure suits secondary analysis of oncology trials for cost-utility analysis.

Valuation by discrete choice experiment with duration

QLU-C10D value sets come from a discrete choice experiment in which every option combines a health state with a length of life. Norman and colleagues first tested the format in an online panel, in which 449 people were recruited, 430 completed at least one choice set and 422 completed all 16, alongside eight telephone interviews. Only four of the ten dimensions differed between the two options in each choice set, and the authors judged the task feasible for deriving utility weights.

In the Australian valuation, reported by King and colleagues, an online panel was quota-sampled by sex and age to match the adult population, and respondents completed 16 choice pairs, each made of two QLU-C10D health states plus life expectancy. Of 1,979 panel members who opted in, 1,904 completed at least one choice task. A conditional logit model interacted the dimension levels with survival time, which fixes the value of being dead at zero as the QALY model requires.

$$V_{ij} = \alpha T_{ij} + \sum_{k} \beta_{k} X_{ijk} T_{ij} + \varepsilon_{ij}$$

where $V_{ij}$ is the value respondent $i$ places on option $j$, $T_{ij}$ is the survival time in that option, $X_{ijk}$ is a dummy variable equal to 1 when the option's health state includes level $k$ (a level below the best level of a dimension), $\alpha$ is the coefficient on time, $\beta_{k}$ is the coefficient on the interaction of level $k$ with time and $\varepsilon_{ij}$ is the random error.

The utility weight of each level is the ratio $\beta_{k} / \alpha$, a negative number for levels worse than the best. Scoring a patient's state starts from 1, full health, and subtracts the decrement for each dimension level the patient reports, as the EORTC user manual describes. Because respondents answered 16 choice sets each, King and colleagues adjusted standard errors with a clustered sandwich estimator.

Results of the first value sets

The Australian and UK value sets show what the measure gives in practice. Both used the same discrete choice design, and both produced a worst state slightly below zero.

King and colleagues found that physical functioning, pain, role functioning and emotional functioning, in that order, had the most impact on choices in Australia. Among the cancer-specific dimensions, nausea and bowel problems had moderate impact while fatigue, trouble sleeping and appetite had relatively small impact, and the worst health state was valued at minus 0.096.

For the UK, Norman and colleagues analysed 2,187 respondents who completed at least one choice set. Nine of the ten dimensions showed the expected ordering across levels, and adjacent levels of appetite were merged to achieve it in the final model. The largest utility decrements were for physical functioning and pain, and the worst possible state was valued at minus 0.083.

National value sets and choosing between them

The EORTC Quality of Life Group lists value sets for many countries. Its user manual says that for an economic evaluation the best choice is the value set of the country where the decision will be made, and otherwise the set of the closest country in culture, care setting and language. The EORTC's guidance also allows an existing set to be applied where no national set exists, provided the reasoning and limitations are clearly outlined.

The choice of value set changes the numbers, because the same patient responses can give different utilities and different QALY gains in two countries whose weights differ.

Illustrative worked example of QLU-C10D utilities in a QALY calculation

The decrements below are illustrative and are not taken from any national value set. A cancer patient on standard care reports physical functioning at level 3, pain at level 2, fatigue at level 2 and level 1 (no problems) on every other dimension. On a new treatment the same patient reports physical functioning at level 2, no pain and fatigue at level 2.

1. Decrements. Physical functioning level 3 carries 0.12, physical functioning level 2 carries 0.06, pain level 2 carries 0.05 and fatigue level 2 carries 0.02. Level 1 on any dimension carries no decrement.

2. Utility on standard care. 1 − 0.12 − 0.05 − 0.02 = 0.81.

3. Utility on the new treatment. 1 − 0.06 − 0.02 = 0.92.

4. QALYs over six months. On standard care 0.81 × 0.5 = 0.405 QALYs, and on the new treatment 0.92 × 0.5 = 0.46 QALYs, a gain of 0.46 − 0.405 = 0.055 QALYs.

5. Cost per QALY. If the new treatment adds GBP 2,750 in costs, 2,750 / 0.055 = 50,000, an incremental cost-utility ratio of GBP 50,000 per QALY.

The gain comes from physical functioning and pain, the dimensions with the largest decrements in both the Australian and UK sets. Had the patient's improvement been in appetite or sleep, dimensions with small weights, the same clinical change would have produced a much smaller QALY gain, so the dimensions a treatment affects matter as much as the size of the change.

Place in HTA and comparisons with the EQ-5D

In NICE appraisals the QLU-C10D sits outside the reference case for adults. NICE's interim methods statement (PMG51), published on 27 August 2026, says the UK EQ-5D-5L value set should be used to derive all utility values for adults in reference-case analyses, and it takes precedence over the NICE technology appraisal and highly specialised technologies manual (PMG36) until the manual is updated. PMG36 keeps a separate route in sections 4.3.10 to 4.3.12 and figure 4.1, unchanged by PMG51, under which another measure can be used when evidence shows that the EQ-5D is inappropriate, starting with evidence on its content validity and then on construct validity and responsiveness. A case for QLU-C10D utilities in a NICE appraisal rests on that evidence.

Head-to-head studies give mixed signals. Kim and colleagues fed both sets of utilities into a published cost-utility model of immunotherapy for metastatic melanoma. Mean baseline utilities were 0.744 with the QLU-C10D and 0.735 with the EQ-5D-3L, not statistically different, and the instruments correlated at 0.74 although concordance was low. The model gave 1.87 QALYs gained with the EQ-5D-3L and 1.74 with the QLU-C10D, ICERs of USD 30,500 and USD 32,700, and cost-effectiveness acceptability curves that were almost indistinguishable. The authors found no additional benefit from the QLU-C10D with Australian weights in that setting.

Pilz and colleagues reached a more favourable view in the Netherlands, analysing 602 patients from four randomised trials with Dutch value sets for both instruments. The QLU-C10D detected 63% of expected clinical group differences and 50% of changes over time in patients on treatment, with a higher detection rate and measurement efficiency than the EQ-5D-3L, while both instruments performed poorly in survivors.

How the QLU-C10D differs from neighbouring measures

The QLU-C10D is a preference-based measure for one disease area, which sets it apart from both its parent questionnaire and generic utility instruments. The QLQ-C30 gives 15 descriptive scores on 0 to 100 scales, and the QLU-C10D gives one preference-weighted utility from 13 of its items.

The EQ-5D is a generic instrument whose utilities can be compared across disease areas, while the QLU-C10D names cancer symptoms such as nausea and bowel problems that the EQ-5D does not. The EORTC-8D, derived by Rowen and colleagues, is an earlier eight-dimension measure built from QLQ-C30 items and valued by time trade-off with 350 members of the UK general population. A mapping function predicts EQ-5D utilities from QLQ-C30 answers with a regression model, whereas the QLU-C10D scores the answers directly with its own valuation.

Limitations and common misreadings

The QLU-C10D answers a narrower question than a generic measure, and its values depend on choices made in valuation. Most of the common errors concern how utilities are reported and compared.

• Assuming one QLU-C10D utility. Each country value set gives different weights, so a utility needs its value set named, and a set from another country is a substitute with stated limitations.
• Comparing across disease areas. Cancer-specific dimensions improve sensitivity to cancer symptoms, but a QALY built on them is not on the same descriptive basis as QALYs from generic measures used elsewhere in a health budget.
• Ignoring non-monotonic levels. The UK set needed adjacent appetite levels merged to keep worse levels from scoring higher, so the appetite weights in that set distinguish fewer levels than the classification describes.
• Mistaking the questionnaire for the utility. A trial that reports the QLQ-C30 has not reported QLU-C10D utilities until the items are scored with a named value set.
• Expecting the EQ-5D answer. Utilities and QALY gains can differ from EQ-5D-based results even when means look similar, as in the melanoma comparison where QALYs gained were 1.74 against 1.87.

Sources

• EORTC Quality of Life Group. EORTC Quality of Life Utility-Core 10 Dimensions (QLU-C10D) user manual. Accessed 17 October 2026. https://www.eortc.org/app/uploads/sites/2/2023/09/Manual-QLUC10D-Final_V5.pdf
• EORTC Quality of Life Group. EORTC Quality of Life Utility Core 10 Dimensions (QLU-C10D), value sets. Accessed 17 October 2026. https://qol.eortc.org/eortc-qlu-c10d/
• Kim H, Cook G, Goodall S, Liew D. Comparison of EQ-5D-3L with QLU-C10D in metastatic melanoma using cost-utility analysis. PharmacoEconomics Open. 2021;5(3):459-467. doi:10.1007/s41669-021-00265-8
• King MT, Costa DSJ, Aaronson NK, et al. QLU-C10D: a health state classification system for a multi-attribute utility measure based on the EORTC QLQ-C30. Quality of Life Research. 2016;25(3):625-636. doi:10.1007/s11136-015-1217-y
• King MT, et al. Australian utility weights for the EORTC QLU-C10D, a multi-attribute utility instrument derived from the cancer-specific quality of life questionnaire, EORTC QLQ-C30. PharmacoEconomics. 2018;36(2):225-238. doi:10.1007/s40273-017-0582-5
• National Institute for Health and Care Excellence. Interim methods statement: implementing the EQ-5D-5L value set (PMG51). Published 27 August 2026. https://www.nice.org.uk/process/pmg51/chapter/statement-on-using-the-eq-5d-5l-value-set
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 31 January 2022, last updated 13 October 2026. https://www.nice.org.uk/process/pmg36
• Norman R, Viney R, Aaronson NK, et al. Using a discrete choice experiment to value the QLU-C10D: feasibility and sensitivity to presentation format. Quality of Life Research. 2016;25(3):637-649. doi:10.1007/s11136-015-1115-3
• Norman R, Mercieca-Bebber R, Rowen D, et al. U.K. utility weights for the EORTC QLU-C10D. Health Economics. 2019;28(12):1385-1401. doi:10.1002/hec.3950
• Pilz MJ, et al. The EORTC QLU-C10D is a valid cancer-specific preference-based measure for cost-utility and health technology assessment in the Netherlands. European Journal of Health Economics. 2024. doi:10.1007/s10198-024-01670-6
• Rowen D, Brazier J, Young T, et al. Deriving a preference-based measure for cancer using the EORTC QLQ-C30. Value in Health. 2011;14(5):721-731. doi:10.1016/j.jval.2011.01.004