https://orcid.org/0000-0001-8860-3823
Figures
Abstract
Objectives
Multinational clinical trials are increasingly common in the European Economic Area, yet no guideline exists on which EQ-5D value set(s) should be used for health economic evaluations alongside multinational trials. Either a single value set or country-specific value sets can be applied, and previous studies have shown that the choice of value set can impact the estimated utility scores. For the EQ-5D-5L, the impact on European pooled cost-utility analyses has not been established. This study evaluates the impact of EQ-5D-5L value sets on cost-utility outcomes in two multinational trials.
Methods
Data from two multinational randomized controlled trials were used: (i) supervised exercise compared to usual care for patients with metastatic breast cancer (PREFERABLE-EFFECT), (ii) hemodiafiltration compared to hemodialysis for patients with kidney failure (CONVINCE). EQ-5D-5L was assessed at baseline and during follow-up. Utility scores and quality-adjusted life years (QALYs) were calculated with the country-specific value set for each participant, and with the value set of each included country. The probability of cost-effectiveness was estimated using bootstrapping.
Results
Mean assigned utility scores per year alive ranged between 0.762 and 0.889 for PREFERABLE-EFFECT, and 0.673 and 0.806 for CONVINCE. The difference in utility scores is largest when participants report low quality of life. Estimated QALY gains ranged between 0.013 and 0.020 for PREFERABLE-EFFECT and 0.045 and 0.058 for CONVINCE. The maximum difference in probability of cost-effectiveness between the value sets was Δ8.3% at €80,000/QALY in PREFERABLE-EFFECT, and Δ11.1% at €40,000/QALY for CONVINCE.
Conclusions
Choice of value set led to substantial variation in absolute utility scores and QALYs, which may influence cost-utility outcomes. This impact could be greater when an intervention prevents or aids recovery of health conditions associated with low quality of life, or results in large mortality differences. Scenario analyses using multiple value sets should be conducted for multinational trials.
Citation: Schouten AEM, Frederix GWJ, Stuiver MM, Fischer F, Hiensch AE, Cromm K, et al. (2026) Choice and impact of EQ-5D-5L value set in cost-utility analyses alongside multinational trials: Insights from PREFERABLE-EFFECT and CONVINCE. PLoS One 21(6): e0350019. https://doi.org/10.1371/journal.pone.0350019
Editor: Henry Hugh Bailey, The University of the West Indies at St Augustine, TRINIDAD AND TOBAGO
Received: January 14, 2026; Accepted: May 7, 2026; Published: June 1, 2026
Copyright: © 2026 Schouten et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data requests can be sent to the following institutional email address: HTAsecretariaat@umcutrecht.nl. Requests for the R scripts of this methodological article can be directed to the corresponding author. The data that support the findings of this study are not openly available owing to reasons of confidentiality. Researchers can direct requests for the PREFERABLE-EFFECT data to Professor Anne May. Requests should also be sent to a.m.may@umcutrecht.nl. Pseudonymized data (including data dictionaries) will be made available through the Digital Research Environment, which is a trusted digital research environment that can be accessed at https://mydre.org. This will be carried out after the review and approval of a methodologically sound proposal by the General Assembly of PREFERABLE, with a signed data access agreement, which is in line with Ethics Committee requirements (The Ethics Committee of University Medical Center Utrecht, The Netherlands). Requests will be processed within 6 weeks. These files will be available from the date of publication until the date stated in the approved request. Once the PREFERABLE project has been fully completed, the database will be anonymized and shared using DataverseNL. The study protocol is available as an open access publication (https://doi.org/10.1186/s13063-022-06556-7). Individual investigators can request CONVINCE data by sending a proposal directed to the Convince Scientific Committee. Proposals should also be sent to r.w.m.vernooij-2@umcutrecht.nl. To gain access to data, requestors will need to sign a data access agreement. Proposed use of data can be made available for the complete deidentified patient data set, along with study protocol, statistical analyses plan, informed consent, analytical code (R scripts), and Excel document with cohort model.
Funding: The author(s) received no specific funding for this work. PREFERABLE-EFFECT was supported by the European Union’s Horizon 2020 research and innovation program (grant agreement 825677) and the National Health and Medical Research Council of Australia (2018/GNT1170698). CONVINCE was exclusively supported by a grant from the European Commission (Horizon 2020 program, grant agreement 754803).
Competing interests: GF reports a consulting role for Illumina (Inst). KC reports holding Fresenius Medical Care company stocks (ordinary shares) and being an employee of Fresenius Medical Care. BC reports being CEO of MTX Consulting Int, Montpellier; Emeritus Medical Officer at FMC Deutschland Global Medical Office; and Emeritus Professor of Medicine, Montpellier, France. GS reports receiving honoraria for educational events from Fresenius Medical Care Deutschland GmbH. All other authors declared no competing interests. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
The EQ-5D is an instrument for estimating health-related quality of life (HRQoL) [1]. This instrument is frequently used to calculate the effect outcome ‘quality-adjusted life years’ (QALYs) in health economic evaluations [2,3], and is preferred in the majority of national guidelines [4,5]. People’s perceived problems with mobility, self-care, usual activities, pain/discomfort, and anxiety/depression are estimated using a questionnaire with either a 3-level or a 5-level scale (EQ-5D-3L and EQ-5D-5L, respectively). These descriptive answers can be translated into a utility score using a value set [6]. This utility score reflects how good or bad someone perceives their health status to be, with “1” being perfect health and “0” death [1], with negative values indicating a health state considered worse than death. By combining utility scores with length of life, QALYs can be estimated, whereby one QALY equals one year of perfect health. Use of EQ-5D-5L is preferred over the EQ-5D-3L, due to improved psychometric performance [7], and more recently derived value sets using a standardized protocol [8,9].
Country-specific value sets have been developed for both the EQ-5D-3L and the EQ-5D-5L to capture local preferences of different health states [10]. Health economic evaluation guidelines recommend using the value set of the jurisdiction of interest for single-country studies [11]. However, there is currently no guideline on which value set(s) should be used for (pooled) multinational health economic evaluations. Over recent years, there has been an increase in multinational clinical trials relative to single-country trials, particularly in the European Economic Area [12]. Statistical power in multinational trials is often not sufficient to perform country-specific cost-utility subgroup analyses [11,13,14]. For pooled analyses, there is no standard practice for estimating utilities [15]. In practice, a single value set is most commonly applied, often that of the UK as it was the first developed value set [6,16]. An alternative is applying the country-specific value set of each participant. A single value set, reflecting the preferences of a larger region, has been proposed as solution [17–20], but this is currently not commonly adopted [10].
Previous studies have shown that choice of value set can impact the absolute value of estimated utility scores [16,21–31]. Moreover, van Dongen et al. (2021) showed that both incremental effects and probability of being cost-effective could vary significantly between value sets [23]. These differences were particularly large between countries of different continents (e.g., between Zimbabwe and Taiwan, or Canada and Singapore). Oppong et al. (2010) found that the choice between European, country-specific, or UK value sets in country-specific sub-group analyses had little impact on cost-utility outcomes when using the EQ-5D-3L [16]. For the EQ-5D-5L, the impact on European pooled multinational cost-utility analyses has not yet been established.
This study aims to evaluate the impact of the choice of EQ-5D-5L value set on pooled cost-utility outcomes alongside European multinational trials, using data from two studies involving patients with chronic diseases. The two studies include distinct patient populations (patients with metastatic breast cancer and kidney failure), to assess whether the observed impact is consistent across different demographic groups. Utility scores per year alive and QALYs, estimated using both single and country-specific value sets, were assessed for the entire study population, as well as stratified by country and randomization group. The findings provide evidence to support informed decision-making for health economic evaluations conducted alongside multinational clinical trials.
Methods
The impact of EQ-5D-5L value sets on cost-utility outcomes was estimated using collected EQ-5D-5L data from two multinational, randomized controlled trials: PREFERABLE-EFFECT and CONVINCE. Cost-effectiveness analyses of these trials have been performed [32,33]. Utility scores of the full sample were calculated with the value sets of each included country, and with the country-specific value set of each participant. Differences in utility scores per year alive and quality-adjusted life years (QALYs) were evaluated at study population level and randomization level. The probability of cost-effectiveness was estimated using bootstrapping.
Study population
The PREFERABLE-EFFECT trial (ClinicalTrials.gov, NCT04120298) included patients with metastatic breast cancer (diagnosis of breast cancer stage IV) from 8 sites in 6 different countries. Participating countries were the Netherlands (NL), Germany (DE), Spain (ES), Sweden (SE), Poland (PL) and Australia (AU) (S1 Table in S1 File). The intervention group received a structured and individualized exercise program in addition to usual care for 9 months, while the control group received usual care, as well as general physical activity advice and an activity tracker. Recruitment took place from January 8th, 2020 to August 3rd, 2022. Participants (N = 357) were randomized to either intervention (N = 178) or control (N = 179) and data was collected for 9 months. Patients valued their health status with the EQ-5D-5L at baseline and in 3-month increments. Majority of patients were female (99%) and average age was 55.4 (SD = 11.1) years old. The trial protocol, primary study results and cost-effectiveness analysis have been published elsewhere [33–35].
The CONVINCE trial (Netherlands National Trial Register, NTR 7138) recruited patients with kidney failure who had received hemodialysis (HD) for at least 3 months. Patients were included in 61 centers in 8 different countries. Participating countries were the Netherlands (NL), Germany (DE), Spain (ES), France (FR), the United Kingdom (UK), Portugal (PT), Hungary (HU) and Romania (RO) (S1 Table in S1 File). The intervention group received high-dose hemodiafiltration (HDF), while the control group received high-flux hemodialysis. Recruitment took place from November 7th, 2018, until March 24th, 2021. Participants (N = 1360) were randomized to either intervention (N = 677) or control (N = 683) and data of participants who completed the trial were collected for at least 24 months. Patients valued their health status with the EQ-5D-5L at baseline and in 6-month increments. Majority of patients were male (62.9%) and average age was 62.4 (SD = 13.5) years old. The trial protocol, primary study results and cost-effectiveness analysis have been published elsewhere [32,36,37].
The study protocols of both PREFERABLE-EFFECT (19–524/M) and CONVINCE (18–260/D) were approved by the institutional review board of the University Medical Center Utrecht, the Netherlands, and by the local ethical review boards of all participating institutions. The studies were conducted in accordance with Good Clinical Practice and the Declaration of Helsinki. All patients of both PREFERABLE-EFFECT and CONVINCE provided written informed consent.
Data analysis
Data analysis was performed with R (R Foundation for Statistical Computing), version 4.4.1. Across all timepoints, 10.7% and 23.9% of the EQ-5D-5L data of PREFERABLE-EFFECT and CONVINCE were missing in total, respectively. Missing values were predicted using multiple imputation (m = 20) [38], applying the R mice package (see S2 Table in S1 File for predictors) [39]. The imputed datasets were analyzed using Rubin’s rules, a method that combines (pools) results from multiple imputed datasets and adjusts the standard error to reflect the uncertainty introduced by the missing data [40].
Unlike the published cost-effectiveness analyses of both trials, the EQ-5D domains were imputed rather than the calculated utility scores to ensure we investigated the difference in value sets, not the difference in imputation. Based on the imputed domains, utilities were estimated multiple times for the entire study population by applying the EQ-5D-5L value set of each included country of the study separately [41–51], as well as the appropriate country-specific value set per patient (CS). For the UK, the crosswalk version of the EQ-5D-3L value set was applied using the index values at the EuroQol website [52,53], following the UK national guideline [54]. A new value set for the EQ-5D-5L has been published in March 2026, which we have included as additional analysis [55]. While it has been proposed that this new EQ-5D-5L value set will replace the EQ-5D-3L cross-walk version as the preferred value set in the UK, this is not yet the case at the time of writing [56]. Therefore, results for both value sets are presented. The utility score was set to 0 for patients who died from date of death onwards. QALYs were calculated using an area under the curve approach with linear interpolation [57].
Descriptive statistics of baseline utility scores, utility per year alive and QALYs were calculated, reporting the mean values of the entire study population, per country, and per randomization group. Uncertainty was estimated by bootstrapping (5000 iterations for each of the 20 imputed data sets) with two seemingly unrelated regressions (SUR) to correct for confounders [58]. For each value set, the probability of cost-effectiveness was explored using the lowest and highest willingness-to-pay thresholds of the Netherlands (i.e., €20,000/QALY to €80,000/QALY) [59], as well as the thresholds in between. For CONVINCE, results were also extrapolated over lifetime using a Markov cohort model with a cycle length of 1 year, simulating 1000 patients. The model contained three health states: dialysis (HD or HDF), renal transplant, and dead. Half-cycle correction was applied. This model is the same as in the cost-effectiveness article where more details can be found [32], apart from the utility scores.
Scenario analyses
A scenario analysis was conducted to facilitate interpretation of the magnitude of impact related to choice of value set. Additional multiple imputation approaches were evaluated, all of which could be used in practice. First, only the seed was changed, with three different seed variations [60]. Second, the EQ-5D-5L answers were translated into utility scores first and then imputed, also with three different seed variations.
For all scenarios, the difference in utility alive between intervention and control per value set was calculated (i.e., utility alive gain). We then assessed: (1) the difference between the maximum and minimum utility alive gain across the different value sets for each imputation scenario; (2) the difference between the maximum and minimum utility alive gain across the different imputation scenario for each value set.
Results
EQ-5D-5L domain outcomes over time
For both studies, quality of life decreased across all EQ-5D-5L domains between baseline and last follow-up (S3 Table in S1 File). The domain most impacted at baseline was pain/discomfort, with 66.6% and 62.6% of participants of PREFERABLE-EFFECT and CONVINCE, respectively, indicating experiencing moderate, severe or extreme problems. Imputing the missing values at 9 months resulted in a smaller proportion of ‘no problem’ scores across both studies and all domains, suggesting that health-related quality of life is likely lower for participants that do not fill in the questionnaires.
Utilities across value sets, entire study population
First, utilities were estimated with the different value sets for the entire study population of both trials, without stratification. Table 1 shows the differences between the highest and lowest baseline utility scores (Δmax0.121 and Δmax0.133, respectively) and utility score per year alive (Δmax0.127 and Δmax0.133, respectively) when applying the different value sets for PREFERABLE-EFFECT and CONVINCE. In PREFERABLE-EFFECT, the mean QALYs ranged between 0.555 QALY (NL value set) and 0.647 QALY (PL value set). In CONVINCE, mean QALYs ranged between 1.287 QALY (UK 3L value set) and 1.537 QALY (RO value set). Using the country-specific value sets of each participant resulted in utility scores close to the average.
For the UK, the crosswalk version of the EQ-5D-3L value set was applied using the index values at the EuroQol website, following the national guideline. As an additional analysis, we also applied the 2026 EQ-5D-5L UK value set, although this is not officially adopted for national use at time of writing. For the country-specific value set, the EQ-5D-3L crosswalk value is used. DE, Germany; NL, the Netherlands; ES, Spain; SE, Sweden; PL, Poland; AU, Australia; CS, country-specific; PT, Portugal; FR, France; UK1, United Kingdom; HU, Hungary; RO, Romania.
Fig 1 shows the distribution of the baseline utility scores of both study populations using the different value sets in violin plots (see S1 & S2 Figs in S1 File for the distributions of QALYs and utility scores per year alive). The plots are skewed to higher utility scores for all value sets, but most for the Swedish and Polish value sets in PREFERABLE-EFFECT, and the French value set in CONVINCE. Fig 2 shows the difference between the German value set (close to the average utility score across value sets) and other value sets from low to high utility scores. Most variation between value sets occurs at the lower utility scores, while variation is smaller for the higher utility scores, indicating that the impact of choice of value set is larger at lower quality of life scores.
Fig 2 represents the differences in baseline EQ-5D-5L utility scores between the various value sets and the German value set (Y-axis) across the range of baseline scores estimated using the German value set (X-axis). On the Y-axis, positive values indicate higher utility scores compared to the DE value set, negative values indicate lower scores. Estimates are pooled per participant aross the imputation sets. Lines are estimated using LOESS (Locally Estimated Scatterplot Smoothing), and represent a smoothed trend line with a 95% confidence interval (CI). DE, Germany; NL, the Netherlands; ES, Spain; SE, Sweden; PL, Poland; AU, Australia; CS, country-specific; PT, Portugal; FR, France; UK, United Kingdom; HU, Hungary; RO, Romania.
Utilities across value sets, participants stratified by country
Mean utility alive scores of patients split up by country are shown in Table 2, and S4 Table in S1 File. For PREFERABLE-EFFECT, the Swedish and Polish patients have the lowest mean utility alive scores when using the German value set, while the Dutch and the Spanish patients have the highest mean utility alive scores. For CONVINCE, patients from the UK and Germany had the lowest utility alive scores when using the German value set, and patients from Hungary and Portugal the highest. With the country-specific value set the order of the countries in terms of utility score (highest>lowest) changes. For PREFERABLE-EFFECT, the ranking changed from (ES > NL > DE > AU > PL > SE) with the German value set to (PL > AU > DE > SE > ES > NL) with the country-specific value set. For CONVINCE, the ranking changed from (HU > PT > RO > ES > FR > NL > DE > UK) with the German value set to (RO > FR > HU > PT > DE > ES > NL > UK) with the country-specific value set.
Utilities across value sets, participants stratified by randomization group
After bootstrapping and correcting for control variables, the mean QALYs are higher for the intervention group than the control group in both studies, regardless of value set (Table 3). This is both due to more life years; 0.725 for intervention and 0.710 for control in PREFERABLE (max 0.75 year), and 1.870 for intervention and 1.845 for control in CONVINCE (max 2 years), as well as the higher utility scores alive for the intervention group (S5 Table in S1 File).
For PREFERABLE-EFFECT, the difference between the highest and lowest incremental QALYs of the different value sets is 0.007 QALY (0.020–0.013 QALY: ES and SE value set). For CONVINCE, the difference between the highest and lowest incremental QALY in the trial-based analysis is 0.013 QALY (0.058–0.045: NL and RO value set). The CS value set resulted in a QALY gain of 0.017 (95%CI: −0.010–0.044) and 0.047 (95%CI: 0.004–0.090) for PREFERABLE and CONVINCE respectively. The difference in incremental QALYs for the model-based analysis was 0.105 (0.980–0.875 QALY: FR and UK 3L value set).
Fig 3 shows the impact of value set on the probability of cost-effectiveness. For the trial-based analyses, the maximum difference in probability of cost-effectiveness is between the value sets with the maximum difference in incremental QALY (Table 3). For the CONVINCE model-based analysis, the maximum difference is between the value sets with the highest and lowest average utility score alive (Table 1). In PREFERABLE-EFFECT, the intervention was dominant, and the difference in probability between value sets is largest at higher willingness-to-pay (WTP) thresholds (e.g., 8.3% difference at €80.000/QALY). For CONVINCE, the incremental cost-effectiveness ratio (ICER) ranged between €41,121/QALY and €53,000/QALY for the trial-based analysis, and €37,471/QALY and €41,957 QALY for the model-based analysis. The difference in probability of cost-effectiveness is largest when the WTP threshold approximates the ICER (Δmax20.1% at WTP of €40.000/QALY for the model-based analysis).
For the United Kingdom, the crosswalk version of the EQ-5D-3L value set was applied using the index values at the EuroQol website, following the national guideline. As an additional analysis, we also applied the 2026 EQ-5D-5L United Kingdom value set, although this is not officially adopted for national use at time of writing. For the country-specific value set, the EQ-5D-3L crosswalk value is used. QALY, quality-adjusted life year; DE, Germany; NL, the Netherlands; ES, Spain; SE, Sweden; PL, Poland; AU, Australia; CS, country-specific; PT, Portugal; FR, France; UK, United Kingdom; HU, Hungary; RO, Romania.
Scenario analyses
Complete results of all imputation scenarios are given in S6 & S7 Table in S1 File. The average of the difference between the highest and lowest utility alive gain using different value sets across the imputation scenarios was 0.013 (range: 0.011–0.016) for PREFERABLE-EFFECT. For CONVINCE, this was 0.008 (range: 0.007–0.010). The mean variation in utility alive gain across imputation scenarios within one value set was 0.004 for PREFERABLE-EFFECT (range: 0.002–0.006) and 0.003 for CONVINCE (range: 0.002–0.005).
Discussion
Choice of EQ-5D-5L value set impacted the estimated absolute utility scores per year alive (Δmax0.127 and Δmax0.133 for PREFERABLE-EFFECT and CONVINCE, respectively). In the trial-based analyses, the included value sets resulted in a maximum difference in the probability of cost-effectiveness of 8.3% at a WTP of €80,000/QALY in PREFERABLE-EFFECT and 11.1% at €40,000/QALY for CONVINCE. In the model-based analysis of CONVINCE, the maximum difference in incremental QALYs was 0.105, and the maximum difference in probability of cost-effectiveness was 20.1% at a WTP of €40,000/ QALY. These results suggest that cost-utility outcomes may be impacted by the choice of value set in different patient populations.
Consistent with previous research, our results showed substantial variation in average utility scores between the value sets (maximum difference in utility score per year alive of 0.127 and 0.133 for PREFERABLE and CONVINCE, respectively) [16,21–31], with largest variation at the lower utility scores [16,22,25,27,28,61,62]. Similar to the findings of Oppong et al. (2011) and van Dongen et al. (2021), we found that for the trial-based analyses, value sets that resulted in higher absolute utility scores did not directly lead to a higher or lower incremental difference between intervention and control [16,23]. The variation in QALY gains is primarily driven by the EQ-5D-5L domains impacted and related country-specific preferences [26]. In the lifetime Markov cohort model of the CONVINCE trial, however, higher utility scores did result in larger QALY losses for the same reduction in life-years. This correlation indicates that choice of value set is more impactful when mortality differences are large. These findings reinforce the ISPOR Task Force recommendation that when several utility scores are used in an economic model, they should as much as possible be elicited using the same methods, population, and value set [63].
The UK crosswalk value set has been suggested for multinational analyses and when a national value set is missing [6,16]. This value set resulted in much lower average utility scores compared to those of other countries [23,25,27–30], as well as a lower probability of cost-effectiveness in the model-based analysis. The maximum difference in probability was 20.1% at a WTP threshold of €40,000, compared to 9.1% when this outlier was excluded. This difference in probability implies that applying the UK crosswalk value set might underestimate the probability of cost-effectiveness [22]. Although the 2026 UK EQ-5D-5L value set is not officially adopted for national use at time of writing [56], the additional analysis showed that estimates based on this set did fall between those of the other value sets. The EQ-5D-3L value set is likely obsolete, as it was developed using methods currently considered outdated, and the underlying population preferences may have changed since 1990 [64,65]. Therefore, we recommend against using the UK EQ-5D-3L crosswalk value set to estimate pooled cost-utility outcomes alongside a multinational trial.
General cost-effectiveness principles apply. When interventions are expected to be both cost-saving and effective, the impact of choice of value set will be considerably less than when the intervention yields higher QALYs & higher costs, especially when the ICER is close to the willingness-to-pay threshold [16,29]. For example, if the willingness-to-pay threshold was €50,000/QALY gained, then different value sets could have resulted in different cost-effectiveness conclusions for the CONVINCE trial-based analyses. The probability of cost-effectiveness will be more strongly impacted by choice of value set when the uncertainty surrounding QALYs is smaller, than when there is a lot of uncertainty. Furthermore, the sensitivity analysis showed that different imputation strategies resulted in a change of utility gain per year alive. This finding highlights that although the impact of choice of value set is relevant, other factors also influence the exact value of the estimated treatment effect, such as imputation strategy when there is a substantial number of missing values. Acknowledging these caveats improves interpretation and must be considered when comparing QALY outcomes between value sets.
Strengths and limitations
This study has several strengths and limitations to consider when interpreting the findings. One strength is the use of two empirical multinational studies, thereby evaluating the impact of different interventions, patient populations, and time horizons. One limitation of this study is that the value sets of the countries included in these trials follow similar preference patterns [7]. The impact of choice of value set might be more substantial in multinational trials with countries whose preferences deviate more, as shown by van Dongen et al (2021) [23]. As a result, these findings might not hold outside the European context. Another consideration is that this study implicitly assumes that there is measurement equivalence, i.e., the answers to the EQ-5D-5L questionnaire are interpreted the same across countries, and are not influenced by, e.g., cultural differences in self-reporting. Whether this assumption is valid is up for debate [24,66–68], but falls outside of the scope of this article. Finally, the number of included participants per country differs for both studies. The responses from countries with higher recruitment numbers have a larger influence on the estimated pooled values than countries with lower participation.
Implications
While our findings are similar to Oppong et al. (2011) [16], we do not fully agree with their conclusion that choice of value set does not matter for cost-utility outcomes. Most variation between value sets occurs at the lower utility scores. Consequently, the impact of choice of value set might be stronger for interventions preventing or aiding recovery of health conditions associated with low quality of life, e.g., prevention of moderate to severe strokes [69], than for supportive interventions for patients with chronic diseases, such as PREFERABLE-EFFECT and CONVINCE. This hypothesis is supported by the study conducted by Karlsson et al. on rheumatoid arthritis patients [25], where treatment improved the low baseline utility scores substantially. Karlsson et al. found that the 1-year adjusted QALY gain ranged between 0.06 and 0.09 when using the US and UK value set respectively. Future research should further investigate which underlying factors and conditions are determinants for changes in outcomes, e.g., by means of a simulation study.
The maximum difference in probability of cost-effectiveness was around 10% for both trial-based analyses, as well as for the model-based analysis when excluding the outlier value. For these empirical studies, using country-specific value sets, or selecting a value set that results in QALY gains closest to the average across all sets, appears to be a pragmatic and appropriate approach. However, if the aim is to inform national reimbursement decisions [15], applying country-specific preferences is essential to allow for comparability across interventions [70]. Health economic evaluations will likely continue to require localization, as these results are generally not considered transferable between countries and are therefore excluded from the European Union’s joint clinical assessment framework of health technologies [71]. For future multinational health economic analyses, we recommend performing sensitivity analyses using multiple value sets to see the impact on outcomes. As discussed above, previous research suggests that when (1) preferences patterns between included countries are dissimilar, (2) the intervention substantially improves or prevents severe health states, and/or (3) mortality is strongly impacted by the intervention, the choice of value set may have a stronger influence on outcomes. Being aware of the potential impact is crucial, as over- or underestimating quality-of-life improvements due to the choice of value set may lead to suboptimal policy or clinical decisions.
Conclusion
The choice of EQ-5D-5L value set led to substantial variation in absolute utility scores and QALYs. In these two multinational studies of supportive interventions for patients with chronic diseases, the maximum difference in probability of cost-effectiveness was around 10%. For short-term analyses, applying country-specific value sets or selecting a single value set that yields QALY estimates close to the average across all options appears to be a pragmatic and appropriate choice. These findings suggest that the commonly used UK value set may be less appropriate, as it consistently resulted in lower utility scores compared to the other value sets. The impact on cost-utility outcomes may be larger in interventions resulting in large mortality differences, if the intervention prevents or aids recovery of health conditions associated with low quality of life, and/or when the preferences patterns between included countries are very dissimilar. Scenario analyses using multiple value sets should be incorporated into a checklist for multinational trials, as over- or underestimating quality-of-life improvements may lead to suboptimal policy or clinical decisions.
Supporting information
S1 File.
Supplementary Material EQ5D PLOS One. Contains Supplementary Table S1-S7 and Supplementary Figure S1-S2.
https://doi.org/10.1371/journal.pone.0350019.s001
(DOCX)
Acknowledgments
The authors thank all participants and involved professional staff at all participating centers for their contribution to the PREFERABLE-EFFECT and CONVINCE randomized controlled trials.
PREFERABLE-EFFECT Scientific Committee:
• Anne May (lead author), Department of Epidemiology and Health Economics, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands, a.m.may@umcutrecht.nl
• Ander Urroticoechearibate, Gipuzkoa Cancer Unit, OSID-Onkologikoa, BioGipuzkoa, Osakidetza, San Sebastian, Spain
• Eva Zopf, Mary MacKillop Institute for Health Research, Australian Catholic University, Melbourne, Australia AND Cabrini Cancer Institute, Cabrini Health, Melbourne, Victoria, Australia
• Joachim Wiskemann, Working Group Exercise Oncology, Division of Medical Oncology, Heidelberg University Hospital and NCT Heidelberg, a partnership between DKFZ and University Medical Center Heidelberg, Heidelberg, Germany
• Karen Steindorf, Division of Physical Activity, Prevention and Cancer, German Cancer Research Center (DKFZ) and National Center for Tumor Diseases (NCT) Heidelberg, a partnership between DKFZ and University Medical Center Heidelberg, Heidelberg, Germany
• Martijn Stuiver, Division of Psychosocial Research and Epidemiology & Center for Quality of Life, Netherlands Cancer Institute, Amsterdam, The Netherlands
• Wilhelm Bloch, Department of Molecular and Cellular Sports Medicine, Institute of Cardiovascular Research and Sports Medicine, German Sport University Cologne, Cologne, Germany
• Yvonne Wengstrom, Karolinska Institutet and Karolinska Comprehensive Cancer Center, Karolinska University Hospital, Stockholm, Sweden
• Elzbieta Senkus, Department of Oncology and Radiotherapy, Medical University of Gdańsk, Gdańsk, Poland
CONVINCE Scientific Committee:
• Michiel Bots (lead author), Department of Epidemiology and Health Economics, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands, M.L.Bots@umcutrecht.nl
• Claudia Barth, Medical Scientific Affairs, B. Braun Avitum, Melsungen, Germany
• Peter Blankestijn, Department of Nephrology and Hypertension, University Medical Center Utrecht, Utrecht University, Utrecht, the Netherlands
• Bernard Canaud, Fresenius Medical Care Deutschland GmbH, Global Medical Office, Bad Homburg, Germany; Montpellier University School of Medicine, Montpellier, France
• Krister Cromm, Fresenius Medical Care Deutschland GmbH, Global Medical Office, Bad Homburg, Germany
• Andrew Davenport, Department of Renal Medicine, Division of Medicine, Royal Free Hospital London, University College London, London, England
• Kathrin Fischer, Centre of Internal Medicine and Dermatology, Department of Psychosomatic Medicine, Charité Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Berlin, Germany
• Jörgen Hegbrant, Division of Nephrology, Department of Clinical Sciences, Lund University, Lund, Sweden
• Matthias Rose, Center for Patient-Centered Outcomes Research and Centre of Internal Medicine and Dermatology, Department for Psychosomatic Medicine, Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Berlin, Germany
• Giovanni Strippoli, The Department of Precision and Regenerative Medicine and Ionian Area, University of Bari, Bari, Italy; The School of Public Health, University of Sydney, Sydney, Australia
• Mariëtta Török, Corporate Medical Office Diaverum, Malmö, Sweden
• Mark Woodward, George Institute for Global Health, School of Public Health, Imperial College London, London, England; George Institute for Global Health, University of New South Wales, Sydney, Australia
References
- 1. Devlin N, Pickard S, Busschbach J. The Development of the EQ-5D-5L and its Value Sets. In: Devlin N, Roudijk B, Ludwig K, editors. Value Sets for EQ-5D-5L: A Compendium, Comparative Review & User Guide. Cham (CH): Springer; 2022. p. 1–12.
- 2. Oppong R, Jowett S, Roberts TE. Economic Evaluation alongside Multinational Studies: A Systematic Review of Empirical Studies. PLoS One. 2015;10(6):e0131949. pmid:26121465
- 3. Rivero-Arias O, Gray A. The multinational nature of cost-effectiveness analyses alongside multinational clinical trials. Value Health. 2010;13(1):34–41. pmid:20667068
- 4. Mauskopf J, Walter J, Birt J, Bowman L, Copley-Merriman C, Drummond M. Differences among formulary submission guidelines: implications for health technology assessment. Int J Technol Assess Health Care. 2011;27(3):261–70. pmid:21756414
- 5. Zisis K, Naoum P, Athanasakis K. Qualitative comparative analysis of health economic evaluation guidelines for health technology assessment in European countries. Int J Technol Assess Health Care. 2020;37:e2. pmid:33298238
- 6. Devlin N, Finch AP, Parkin D. Guidance to Users of EQ-5D-5L Value Sets. In: Devlin N, Roudijk B, Ludwig K, editors. Value Sets for EQ-5D-5L: A Compendium, Comparative Review & User Guide. Cham (CH): Springer; 2022. p. 213–33.
- 7. Buchholz I, Janssen MF, Kohlmann T, Feng Y-S. A Systematic Review of Studies Comparing the Measurement Properties of the Three-Level and Five-Level Versions of the EQ-5D. Pharmacoeconomics. 2018;36(6):645–61. pmid:29572719
- 8. Stolk E, Ludwig K, Rand K, van Hout B, Ramos-Goñi JM. Overview, Update, and Lessons Learned From the International EQ-5D-5L Valuation Work: Version 2 of the EQ-5D-5L Valuation Protocol. Value Health. 2019;22(1):23–30. pmid:30661630
- 9. Stolk E, Ramos-Goñi JM, Ludwig K, Oppe M, Norman R. The Development and Strengthening of Methods for Valuing EQ-5D-5L – An Overview. In: Devlin N, Roudijk B, Ludwig K, editors. Value Sets for EQ-5D-5L: A Compendium, Comparative Review & User Guide. Cham (CH): Springer; 2022. p. 13–27.
- 10. Roudijk B, Janssen B, Olsen JA. How Do EQ-5D-5L Value Sets Differ?. In: Devlin N, Roudijk B, Ludwig K, editors. Value Sets for EQ-5D-5L: A Compendium, Comparative Review & User Guide. Cham (CH): Springer; 2022. p. 235–58.
- 11. Drummond M, Barbieri M, Cook J, Glick HA, Lis J, Malik F, et al. Transferability of economic evaluations across jurisdictions: ISPOR Good Research Practices Task Force report. Value Health. 2009;12(4):409–18. pmid:19900249
- 12.
IQVIA, European Federation of Pharmaceutical Industries and Associations, Vaccines Europe. Assessing the clinical trial ecosystem in Europe. 2024. Accessed January 12, 2025. https://efpia.eu/media/3edpooqp/assessing-the-clinical-trial-ecosystem-in-europe.pdf
- 13. Reed SD, Anstrom KJ, Bakhai A, Briggs AH, Califf RM, Cohen DJ, et al. Conducting economic evaluations alongside multinational clinical trials: toward a research consensus. Am Heart J. 2005;149(3):434–43. pmid:15864231
- 14. Manca A, Rice N, Sculpher MJ, Briggs AH. Assessing generalisability by location in trial-based cost-effectiveness analysis: the use of multilevel models. Health Econ. 2005;14(5):471–85. pmid:15386662
- 15. Manca A, Sculpher MJ, Goeree R. The Analysis of Multinational Cost-Effectiveness Data for Reimbursement Decisions. PharmacoEconomics. 2010;28(12):1079–96.
- 16. Oppong R, Kaambwa B, Nuttall J, Hood K, Smith RD, Coast J. The impact of using different tariffs to value EQ-5D health state descriptions: an example from a study of acute cough/lower respiratory tract infections in seven countries. Eur J Health Econ. 2013;14(2):197–209. pmid:22057665
- 17. Sajjad A, Versteegh MM, Santi I, Busschbach J, Simon J, Roijen LH. In search of a “pan-European value set”; application for EQ-5D-3L. BMC Med Res Methodol. 2023;23(1):13. pmid:36635625
- 18. Olsen JA, Lamu AN, Cairns J. In search of a common currency: A comparison of seven EQ-5D-5L value sets. Health Econ. 2018;27(1):39–49. pmid:29063633
- 19. Łaszewska A, Sajjad A, Busschbach J, Simon J, Hakkaart-van Roijen L. Conceptual Framework for Optimised Proxy Value Set Selection Through Supra-National Value Set Development for the EQ-5D Instruments. Pharmacoeconomics. 2022;40(12):1221–34. pmid:36201130
- 20. Greiner W, Weijnen T, Nieuwenhuizen M, Oppe S, Badia X, Busschbach J, et al. A single European currency for EQ-5D health states. Results from a six-country study. Eur J Health Econ. 2003;4(3):222–31. pmid:15609189
- 21. Zhao Y, Li S-P, Liu L, Zhang J-L, Chen G. Does the choice of tariff matter?: A comparison of EQ-5D-5L utility scores using Chinese, UK, and Japanese tariffs on patients with psoriasis vulgaris in Central South China. Medicine (Baltimore). 2017;96(34):e7840. pmid:28834893
- 22. Gerlinger C, Bamber L, Leverkus F, Schwenke C, Haberland C, Schmidt G, et al. Comparing the EQ-5D-5L utility index based on value sets of different countries: impact on the interpretation of clinical study results. BMC Res Notes. 2019;12(1):18. pmid:30642397
- 23. van Dongen JM, Jornada Ben Â, Finch AP, Rossenaar MMM, Biesheuvel-Leliefeld KEM, Apeldoorn AT, et al. Assessing the Impact of EQ-5D Country-specific Value Sets on Cost-utility Outcomes. Med Care. 2021;59(1):82–90. pmid:33201085
- 24. Bernert S, Fernández A, Haro JM, König H-H, Alonso J, Vilagut G, et al. Comparison of different valuation methods for population health status measured by the EQ-5D in three European countries. Value Health. 2009;12(5):750–8. pmid:19490564
- 25. Karlsson JA, Nilsson J-Å, Neovius M, Kristensen L-E, Gülfe A, Saxne T, et al. National EQ-5D tariffs and quality-adjusted life-year estimation: comparison of UK, US and Danish utilities in south Swedish rheumatoid arthritis patients. Ann Rheum Dis. 2011;70(12):2163–6. pmid:21859684
- 26. Joelson A, Wildeman P, Sigmundsson FG, Rolfson O, Karlsson J. Properties of the EQ-5D-5L when prospective longitudinal data from 28,902 total hip arthroplasty procedures are applied to different European EQ-5D-5L value sets. Lancet Reg Health Eur. 2021;8:100165. pmid:34557854
- 27. Kiadaliri AA, Eliasson B, Gerdtham U-G. Does the choice of EQ-5D tariff matter? A comparison of the Swedish EQ-5D-3L index score with UK, US, Germany and Denmark among type 2 diabetes patients. Health Qual Life Outcomes. 2015;13:145. pmid:26374709
- 28. Johnson JA, Luo N, Shaw JW, Kind P, Coons SJ. Valuations of EQ-5D health states: are the United States and United Kingdom different? Med Care. 2005;43(3):221–8. pmid:15725978
- 29. Lien K, Tam VC, Ko YJ, Mittmann N, Cheung MC, Chan KKW. Impact of country-specific EQ-5D-3L tariffs on the economic value of systemic therapies used in the treatment of metastatic pancreatic cancer. Curr Oncol. 2015;22(6):e443-52. pmid:26715881
- 30. Huang I-C, Willke RJ, Atkinson MJ, Lenderking WR, Frangakis C, Wu AW. US and UK versions of the EQ-5D preference weights: does choice of preference weights make a difference? Qual Life Res. 2007;16(6):1065–72. pmid:17415683
- 31. Zrubka Z, Beretzky Z, Hermann Z, Brodszky V, Gulácsi L, Rencz F, et al. A comparison of European, Polish, Slovenian and British EQ-5D-3L value sets using a Hungarian sample of 18 chronic diseases. Eur J Health Econ. 2019;20(Suppl 1):119–32. pmid:31104218
- 32. Schouten AEM, Fischer F, Blankestijn PJ, Vernooij RWM, Hockham C, Strippoli GFM, et al. A health economic evaluation of the multinational, randomized controlled CONVINCE trial: cost-utility of high-dose online hemodiafiltration compared to high-flux hemodialysis. Kidney Int. 2025;107(4):728–39. pmid:39848405
- 33. Schouten AEM, Hiensch AE, Frederix GWJ, et al. Supervised exercise for patients with metastatic breast cancer: a cost-utility analysis alongside the PREFERABLE-EFFECT randomized controlled trial. J Clin Oncol. 2025.
- 34. Hiensch AE, Monninkhof EM, Schmidt ME, Zopf EM, Bolam KA, Aaronson NK, et al. Design of a multinational randomized controlled trial to assess the effects of structured and individualized exercise in patients with metastatic breast cancer on fatigue and quality of life: the EFFECT study. Trials. 2022;23(1):610. pmid:35906659
- 35. Hiensch AE, Depenbusch J, Schmidt ME, Monninkhof EM, Pelaez M, Clauss D, et al. Supervised, structured and individualized exercise in metastatic breast cancer: a randomized controlled trial. Nat Med. 2024;30(10):2957–66. pmid:39054374
- 36. Blankestijn PJ, Fischer KI, Barth C, Cromm K, Canaud B, Davenport A, et al. Benefits and harms of high-dose haemodiafiltration versus high-flux haemodialysis: the comparison of high-dose haemodiafiltration with high-flux haemodialysis (CONVINCE) trial protocol. BMJ Open. 2020;10(2):e033228. pmid:32029487
- 37. Blankestijn PJ, Vernooij RWM, Hockham C, Strippoli GFM, Canaud B, Hegbrant J, et al. Effect of Hemodiafiltration or Hemodialysis on Mortality in Kidney Failure. N Engl J Med. 2023;389(8):700–9. pmid:37326323
- 38. Mutubuki EN, El Alili M, Bosmans JE, Oosterhuis T, J Snoek F, Ostelo RWJG, et al. The statistical approach in trial-based economic evaluations matters: get your statistics together!. BMC Health Serv Res. 2021;21(1):475. pmid:34011337
- 39. van Buuren S, Groothuis-Oudshoorn K. Mice: Multivariate imputation by chained equations in R. J Stat Softw. 2011;45:1–67.
- 40. Rubin DB. Multiple imputation for nonresponse in surveys. 1987.
- 41. Ludwig K, Graf von der Schulenburg J-M, Greiner W. German Value Set for the EQ-5D-5L. Pharmacoeconomics. 2018;36(6):663–74. pmid:29460066
- 42. M Versteegh M, M Vermeulen K, M A A Evers S, de Wit GA, Prenger R, A Stolk E. Dutch Tariff for the Five-Level Version of EQ-5D. Value Health. 2016;19(4):343–52. pmid:27325326
- 43. Ramos-Goñi JM, Craig BM, Oppe M, Ramallo-Fariña Y, Pinto-Prades JL, Luo N, et al. Handling Data Quality Issues to Estimate the Spanish EQ-5D-5L Value Set Using a Hybrid Interval Regression Approach. Value Health. 2018;21(5):596–604. pmid:29753358
- 44. Sun S, Chuang L-H, Sahlén K-G, Lindholm L, Norström F. Estimating a social value set for EQ-5D-5L in Sweden. Health Qual Life Outcomes. 2022;20(1):167. pmid:36564844
- 45. Golicki D, Jakubczyk M, Graczyk K, Niewada M. Valuation of EQ-5D-5L Health States in Poland: the First EQ-VT-Based Study in Central and Eastern Europe. Pharmacoeconomics. 2019;37(9):1165–76. pmid:31161586
- 46. Norman R, Mulhern B, Lancsar E, Lorgelly P, Ratcliffe J, Street D, et al. The Use of a Discrete Choice Experiment Including Both Duration and Dead for the Development of an EQ-5D-5L Value Set for Australia. Pharmacoeconomics. 2023;41(4):427–38. pmid:36720793
- 47. Ferreira PL, Antunes P, Ferreira LN, Pereira LN, Ramos-Goñi JM. A hybrid modelling approach for eliciting health state preferences: the Portuguese EQ-5D-5L value set. Qual Life Res. 2019;28(12):3163–75. pmid:31201730
- 48. Rencz F, Brodszky V, Gulácsi L. Parallel valuation of the EQ-5D-3L and EQ-5D-5L by time trade-off in Hungary. Value in Health. 2020;23:1235–45.
- 49. Andrade LF, Ludwig K, Goni JMR, Oppe M, de Pouvourville G. A French Value Set for the EQ-5D-5L. Pharmacoeconomics. 2020;38(4):413–25. pmid:31912325
- 50. Olariu E, Mohammed W, Oluboyede Y, Caplescu R, Niculescu-Aron IG, Paveliu MS, et al. EQ-5D-5L: a value set for Romania. Eur J Health Econ. 2023;24(3):399–412. pmid:35688994
- 51. Dolan P. Modeling valuations for EuroQol health states. Med Care. 1997;35(11):1095–108. pmid:9366889
- 52. van Hout B, Janssen MF, Feng Y-S, Kohlmann T, Busschbach J, Golicki D, et al. Interim scoring for the EQ-5D-5L: mapping the EQ-5D-5L to EQ-5D-3L value sets. Value Health. 2012;15(5):708–15. pmid:22867780
- 53. van Reenen M, Janssen B, Stolk E, et al. EQ-5D-5L user guide — basic information on how to use the EQ-5D-5L instrument. Rotterdam; 2019. Accessed May 16, 2025. https://euroqol.org/wp-content/uploads/2023/11/EQ-5D-5LUserguide-23-07.pdf
- 54.
National Institute for Health and Care Excellence (NICE). Position statement on use of the EQ-5D-5L value set for England (updated October 2019). Accessed April 29, 2025. https://www.nice.org.uk/about/what-we-do/our-programmes/nice-guidance/technology-appraisal-guidance/eq-5d-5l
- 55. Rowen D, Mukuria C, Bray N, Carlton J, Longworth L, Meads D, et al. A UK value set for the EQ-5D-5L. Value in Health. 2026.
- 56.
National Institute for Health and Care Excellence (NICE). Have your say on NICE’s adoption of the new EQ-5D-5L value set (updated 2026 Apr 15). Accessed April 17, 2026. Available from: https://www.nice.org.uk/news/articles/have-your-say-on-nices-adoption-of-the-new-eq-5d-5l-value-set
- 57.
Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford University Press; 2015.
- 58. Ben ÂJ, van Dongen JM, El Alili M, Esser JL, Broulíková HM, Bosmans JE. Conducting Trial-Based Economic Evaluations Using R: A Tutorial. Pharmacoeconomics. 2023;41(11):1403–13. pmid:37458913
- 59. Zorginstituut Nederland, Zwaap J, Knies S, Staal P, van der Heiden L. Kosteneffectiviteit in de praktijk. Diemen; 2015. Accessed January 30, 2025.https://zoek.officielebekendmakingen.nl/blg-543402.pdf
- 60. Dikta G, Scheer M. Bootstrap Methods: With Applications in R. 1st ed. Cham: Springer Nature Switzerland AG; 2021.
- 61. Roudijk B, Donders ART, Stalmeier PFM, Cultural Values Group. Cultural Values: Can They Explain Differences in Health Utilities between Countries? Med Decis Making. 2019;39(5):605–16. pmid:31257997
- 62. Sakthong P, Charoenvisuthiwongs R, Shabunthom R. A comparison of EQ-5D index scores using the UK, US, and Japan preference weights in a Thai sample with type 2 diabetes. Health Qual Life Outcomes. 2008;6:71. pmid:18811935
- 63. Brazier J, Ara R, Azzabi I, Busschbach J, Chevrou-Séverac H, Crawford B, et al. Identification, Review, and Use of Health State Utilities in Cost-Effectiveness Models: An ISPOR Good Practices for Outcomes Research Task Force Report. Value Health. 2019;22(3):267–75. pmid:30832964
- 64. Rowen D, Mukuria C, Bray N, et al. UK valuation of EQ-5D-5L, a generic measure of health-related quality of life: a study protocol. Value in Health. 2023;26:1625–35.
- 65. Norman R, Roudijk B, Jonker M, Stolk E, Knies S, Pwu R-F, et al. A Taxonomy for Assessing Whether HRQoL Value Sets Are Obsolete. Pharmacoeconomics. 2025;43(5):473–81. pmid:39960567
- 66. Luo N, Wang Y, How CH, Wong KY, Shen L, Tay EG, et al. Cross-cultural measurement equivalence of the EQ-5D-5L items for English-speaking Asians in Singapore. Qual Life Res. 2015;24(6):1565–74. pmid:25394895
- 67. Hayes A, Arima H, Woodward M, Chalmers J, Poulter N, Hamet P, et al. Changes in Quality of Life Associated with Complications of Diabetes: Results from the ADVANCE Study. Value Health. 2016;19(1):36–41. pmid:26797234
- 68. Salomon JA, Patel A, Neal B, Glasziou P, Grobbee DE, Chalmers J, et al. Comparability of patient-reported health status: multicountry analysis of EQ-5D responses in patients with type 2 diabetes. Med Care. 2011;49(10):962–70. pmid:21918400
- 69. Luengo-Fernandez R, Gray AM, Bull L, Welch S, Cuthbertson F, Rothwell PM. Quality of life after TIA and stroke. Neurology. 2013;81(18):1588–95.
- 70. Knies S, Evers SMAA, Candel MJJM, Severens JL, Ament AJHA. Utilities of the EQ-5D: transferable or not?. Pharmacoeconomics. 2009;27(9):767–79. pmid:19757870
- 71. Johns A, Andrade A, Johal S, Ryan J. EU Joint Clinical Assessment: A Framework for Optimising Use with Cost-Effectiveness Decision-Making. J Mark Access Health Policy. 2025;13(4):52. pmid:41142925