In search of a common currency: A comparison of seven EQ-5D-5L value sets

Olsen, Jan Abel; Lamu, Admassu Nadew; Cairns, John

doi:10.1002/hec.3606

Cited by 60 publications

(60 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…New Zealand. It should also be noted that a comparison of value sets across 7 countries, including the UK [33], found them to have similar characteristics, suggesting that the use of the EQ-5D-5L UK value set in this study is unlikely to have influenced the findings.…”

Section: Discussionmentioning

confidence: 68%

Testing the validity and responsiveness of a new cancer-specific health utility measure (FACT-8D) in relapsed/refractory mantle cell lymphoma, and comparison to EQ-5D-5L

Herdman

Kerr

Pavesi

et al. 2020

J Patient Rep Outcomes

View full text Add to dashboard Cite

Background: The FACT-8D is a new cancer-specific, preference-based measure (PBM) of health, derived from the Functional Assessment of Cancer Therapy-General (FACT-G) questionnaire. The FACT-8D's measurement properties have not been tested to date. We assessed it's validity and responsiveness in relapsed/refractory mantle cell lymphoma (RR MCL) and compared the results to the EQ-5D-5L. Methods: Blinded analysis of pooled data from a phase 3 clinical trial. FACT-8D baseline and follow-up data (weeks 4, 7, 16, 31) were scored using Australian preference weights, the first available value-set. Convergent validity was assessed by estimating baseline correlations with the FACT-Lym total score, Trial Outcome Index (TOI), FACT-Lym lymphoma-specific sub-scale (LymS), EQ-5D Visual Analog Scale (VAS), and haemoglobin (HgB). Relevant clinical variables were used to categorise patients to test known groups' validity and responsiveness was investigated using data from baseline (n = 250) and week 31 (n = 130). Results were compared with EQ-5D-5L, scored using the UK 3L crosswalk and the 5L England value-sets. Results: The FACT-8D showed good convergent validity and responsiveness; baseline Pearson correlation coefficients between FACT-8D Index scores and other PRO measures were moderate to very strong (range: 0.49 for the EQ-VAS to 0.79 for FACT TOI) and the size of the change in FACT-8D Index scores at week 31 differed significantly (p < 0.005) between patients categorised as improved, worsened or stable using the FACT-Lym total score, LymS, and HgB. However, when assessing known groups' validity, FACT-8D failed to discriminate between patients categorised by health status on four of the seven variables analysed. Overall, FACT-8D and EQ-5D-5L performed similarly, although EQ-5D-5L showed better known groups' validity.

show abstract

Section: Discussionmentioning

confidence: 68%

Testing the validity and responsiveness of a new cancer-specific health utility measure (FACT-8D) in relapsed/refractory mantle cell lymphoma, and comparison to EQ-5D-5L

Herdman

Kerr

Pavesi

et al. 2020

J Patient Rep Outcomes

View full text Add to dashboard Cite

show abstract

“…This general problem of mapping studies is due to regression to the mean [36]. Furthermore, this might also be explained by the preference pattern of EQ-5D-5L, where substantial decrements in preference weights occur at severe health states [30]. The conceptual difference between the source and target instruments may be another explanation.…”

Section: Discussionmentioning

confidence: 99%

“…For ease of exposition, the English value set would be emphasised, though results for other countries' value sets were discussed. The scale length differs across countries with the worst health state or the 'pits' (55555) ranges from -0.446 in the Netherlands to -0.025 in Japan [30].…”

Section: Measures Of Variablesmentioning

confidence: 99%

Do country-specific preference weights matter in the choice of mapping algorithms? The case of mapping the Diabetes-39 onto eight country-specific EQ-5D-5L value sets

et al. 2018

Self Cite

View full text Add to dashboard Cite

Purpose: To develop mapping algorithms that transform Diabetes-39 (D-39) scores onto EQ-5D-5L utility values for each of eight recently published country-specific EQ-5D-5L value sets, and to compare mapping functions across the EQ-5D-5L value sets.Methods: Data include 924 individuals with self-reported diabetes from six countries. The D-39 dimensions, age and gender were used as potential predictors for EQ-5D-5L utilities, which were scored using value sets from eight countries (England, Netherland, Spain, Canada, Uruguay, China, Japan and Korea). Ordinary least squares, generalized linear model, beta binomial regression, fractional regression, MM-estimation, and censored least absolute deviation were used to estimate the mapping algorithms. The optimal algorithm for each country-specific value set was primarily selected based normalized root mean square error (NRMSE), normalized mean absolute error (NMAE) and adjusted-r 2 . Cross-validation with 5-fold approach was conducted to test the generalizability of each model. Results:The fractional regression model with loglog as a link function consistently performed best in all country-specific value sets. For instance, the NRMSE (0.1282) and NMAE (0.0914) were the lowest, while adjusted-r 2 was the highest (52.5%) when the English value set was considered. Among D-39 dimensions, the energy and mobility was the only one that was consistently significant for all models. Conclusions:The D-39 can be mapped onto the EQ-5D-5L utilities with good predictive accuracy. The fractional regression model, which is appropriate for handling bounded outcomes, outperformed other candidate methods in all country-specific value sets. However, the regression coefficients differed reflecting preference heterogeneity across countries.

show abstract

“…Despite these concerns, another paper in this issue (Olsen, Lamu, & Cairns, ) provides some reassurance. In a comparison of value sets for the EQ‐5D‐5L across seven countries, the authors found that the four “Western” countries of Canada, England, the Netherlands, and Spain showed remarkable similarities considering the use of different modelling methods that produced the algorithms.…”

mentioning

confidence: 99%

EQ-5D-5L: Smaller steps but a major step change?

2018

View full text Add to dashboard Cite

), potentially has important implications for technology coverage decision making in many jurisdictions around the world.The 5-level version of EQ-5D (EQ-5D-5L) was developed in response to criticism centring on questions of sensitivity of the original 3-level version; many patients would indicate no problems on all five dimensions (i.e., State 11111), despite many of these respondents reporting they do have problems on similar dimensions of other measures (Brazier, Ratcliffe, Tsuchiya, & Solomon, 2017). At the same time, for some dimensions, particularly mobility and self-care, very few patients reported Level 3. It was increasingly recognised that the move from "none" to "some" to "unable"/"severe" does not have the sensitivity to reflect many of the important differences or changes found in patient populations. Despite this sensitivity limitation, the EQ-5D has nevertheless compared reasonably favourably with other generic measures, in terms of validity, for many conditions (Longworth et al., 2014). Perhaps less well recognised and understood have been the concerns with the preference elicitation method used in the UK valuation survey conducted in the early 1990s. This was a variant of the time trade-off (TTO) technique and adopted an unusual approach to valuing and transforming states worse than dead. In addition, simplistic modelling methods were used and, of course, general population preferences may well have changed over subsequent 25 years.Reassuringly, the increase to the five levels of "none," "slightly," "moderate," "severe," and "extreme"/"unable" has resulted in a measure that is better descriptively, with reduced "ceiling effects" (Herdman et al., 2011), more even distribution of responses across the levels (Mulhern et al., 2017), and this should result in greater sensitivity to differences between groups of patients and to changes over time. The replacement of the label "confined to bed" (in the EQ-5D-3L) by "unable to walk about" on the mobility dimension has been another descriptive improvement, although the continued reliance on the term "walking" has been challenged (Hannan et al., 2017). Further, concerns from a linguistic perspective have been expressed in relation to the level descriptors; unsurprisingly some respondents have been unclear on the differences between "slight" and "moderate" or "severe" and "extreme."As developers, the Euroqol Group have used the new descriptive system as an opportunity to improve, test new approaches and standardise the valuation methods, in part reflecting the development of research methods over the past 20 years since the publication of the EQ-5D-3L value sets. They chose to keep with TTO but made a number of important modifications to the original protocol. States worse than dead are now valued using a lead time TTO task, for example, where respondents are given an extra 10 years of good health before the tenth, they are asked to assume they will live in the health state being valued (Devlin, Tsuchiya, Buckingham, & Tilling, 2011). This task, although mo...

show abstract

In search of a common currency: A comparison of seven EQ-5D-5L value sets

Cited by 60 publications

References 12 publications

Testing the validity and responsiveness of a new cancer-specific health utility measure (FACT-8D) in relapsed/refractory mantle cell lymphoma, and comparison to EQ-5D-5L

Testing the validity and responsiveness of a new cancer-specific health utility measure (FACT-8D) in relapsed/refractory mantle cell lymphoma, and comparison to EQ-5D-5L

Do country-specific preference weights matter in the choice of mapping algorithms? The case of mapping the Diabetes-39 onto eight country-specific EQ-5D-5L value sets

EQ-5D-5L: Smaller steps but a major step change?

Contact Info

Product

Resources

About