A new method for valuing health: directly eliciting personal utility functions

A novel approach for health state valuation: Multiple bounded dichotomous choice compared to the traditional standard gamble

OBJECTIVE

to assess the feasibility of a new stated preference approach, the multiple bounded dichotomous choice (MBDC), designed to generate value sets for preference-based measurement of health-related quality of life.

METHODS

MBDC and standard gamble (SG) tasks were completed to derive SF-6Dv2 value sets from a sample of the general population in Quebec, Canada. Participants were randomized between the two approaches: 6 health states were evaluated in SG and 11 health states in MBDC. Several models were used to estimate data in each approach, and the preferred models were chosen by using mean absolute error (MAE), logical consistency of parameters, and significance levels. Results of MBDC were compared with SG in terms of acceptability (self-reported difficulty and quality levels in answering, and completion time), consistency (monotonicity of model coefficients), accuracy (standard errors), dimensions coefficient magnitude, correlation between the value sets estimated, and the range of estimated values. The intra-class correlation coefficient (ICC) was computed to assess value sets' consistency.

RESULTS

Out of 655 individuals who completed MBDC tasks and 828 who completed SG tasks, a total of 585 participants for MBDC and 714 for SG tasks were included for analysis. The preferred models for both approaches were GLS Tobit. No significant difference was observed in self-reported difficulties and qualities in answers among approaches, but MBDC had less excluded participants and was less prone to report difficulties in answering. Additionally, completion time in the MBDC group was significantly lower (99.80 vs 68.12 s). Most standard errors in the MBDC were lower than those in SG, and the number of non-significant parameters was also lower. The range of utility values generated by MBDC tended to be wider (-0.372 to 1) than those generated by the SG (-0.137 to 1) and the number of worse-than-dead states in MBDC (0.91%) was higher than for SG (0.08%). The Pain dimension was identified as the most significant, while the Vitality dimension showed the lowest significant decrement. Both approaches exhibited a tendency to overestimate severe health state values and underestimate better health state values. The correlation and ICC between the two value sets were 0.937 and 0.983, respectively.

CONCLUSION

Based on empirical evidence, it can be inferred that the MBDC method is not only feasible but also holds the potential to generate meaningful and well-informed preference data from respondents. This approach can be used to derive a value set for preference-based instrument.

Beyond 10-year lead-times in EQ-5D-5L: leveraging alternative lead-times in willingness-to-accept questions to capture preferences for worse-than-dead states and their implication

BACKGROUND

A fixed 10-year lead-time in composite time-trade-off (C-TTO) tasks might compromise the precision of utility values below - 1. This study explored how alternative lead-times (ALTs) influence EQ-5D-5L value sets and their implications in economic evaluations.

METHODS

Leveraging data from Taiwan's EQ-5D-5L valuation and capitalizing on its exploratory willingness-to-accept question, we explored participants' quantification of "worse-than-dead (WTD)" health states with ALTs up to 50 years. We then derived alternative value sets incorporating these ALTs through interval regression and compared them against those from conventional models. To evaluate their impact on health change valuation, we simulated utility differences for all possible EQ-5D-5L health-state-pairs using each value set.

RESULTS

With a salient floor effect observed in the C-TTO values, the model with ALT led to a wider range of predicted utilities ( - 2.3897 ~ 1), compared with those of conventional models (generalized least squares (GLS):  - 0.7773 ~ 1; Tobit-GLS:  - 0.9583 ~ 1). Compared to the Tobit-GLS model, the model with ALT increased the numerical distance in 80% of health-state-pairs, with 11% decreasing and 9% altering direction (e.g., positive to negative) in utility differences.

CONCLUSIONS

While ALTs offer insights into patient preferences, their integration into economic evaluations might require rescaling. Future research should prioritize advanced rescaling methods or enhanced elicitation strategies for populations with substantial censoring. This is pivotal for improving the elicitation of extreme WTD states and accurately discerning the relative distances between health states. Countries developing EQ-5D-5L value sets should consider pilot studies and incorporating region-specific questions on social determinants, especially where pronounced floor effects are suspected.

Using the Online Elicitation of Personal Utility Functions Approach to Derive a Patient-Based 5-Level Version of EQ-5D Value Set: A Study in 122 Patients With Rheumatic Diseases From Germany

OBJECTIVES

Traditional preference elicitation methods, such as discrete choice experiments or time trade-off, usually require large sample sizes. This can limit their applicability in patient populations, where recruiting enough participants can be challenging. The objective of this study was to test a new method, called the Online elicitation of Personal Utility Functions (OPUF) approach, to derive an EQ-5D-5L value set from a relatively small sample of patients with rheumatic diseases.

METHODS

OPUF is a new type of online survey that implements compositional preference elicitation techniques. Central to the method are 3 valuation steps: (1) dimension weighting, (2) level rating, and (3) anchoring. An English demo version of the OPUF survey can be accessed at https://valorem.health/eq5d5l. From the responses, a personal EQ-5D-5L utility function can be constructed for each participant, and a group-level value set can be derived by aggregating model coefficients across participants.

RESULTS

A total of 122 patients with rheumatic disease from Germany completed the OPUF survey. The survey was generally well received; most participants completed the survey in less than 20 minutes and were able to derive a full EQ-5D-5L value set. The precision of mean coefficients was high, despite the small sample size.

CONCLUSIONS

Our findings demonstrate that OPUF can be used to derive an EQ-5D-5L value set from a relatively small sample of patients. Although the method is still under development, we think that it has the potential to be a valuable preference elicitation tool and to complement traditional methods in several areas.

Appropriate design and reporting of superiority, equivalence and non-inferiority clinical trials incorporating a benefit-risk assessment: the BRAINS study including expert workshop

Background

Randomised controlled trials are designed to assess the superiority, equivalence or non-inferiority of a new health technology, but which trial design should be used is not always obvious in practice. In particular, when using equivalence or non-inferiority designs, multiple outcomes of interest may be important for the success of a trial, despite the fact that usually only a single primary outcome is used to design the trial. Benefit-risk methods are used in the regulatory clinical trial setting to assess multiple outcomes and consider the trade-off of the benefits against the risks, but are not regularly implemented in publicly funded trials.

Objectives

The aim of the project is to aid the design of clinical trials with multiple outcomes of interest by defining when each trial design is appropriate to use and identifying when to use benefit-risk methods to assess outcome trade-offs (qualitatively or quantitatively) in a publicly funded trial setting.

Methods

A range of methods was used to elicit expert opinion to answer the project objectives, including a web-based survey of relevant researchers, a rapid review of current literature and a 2-day consensus workshop of experts (in 2019).

Results

We created a list of 19 factors to aid researchers in selecting the most appropriate trial design, containing the following overarching sections: population, intervention, comparator, outcomes, feasibility and perspectives. Six key reasons that indicate a benefit-risk method should be considered within a trial were identified: (1) when the success of the trial depends on more than one outcome; (2) when important outcomes within the trial are in competing directions (i.e. a health technology is better for one outcome, but worse for another); (3) to allow patient preferences to be included and directly influence trial results; (4) to provide transparency on subjective recommendations from a trial; (5) to provide consistency in the approach to presenting results from a trial; and (6) to synthesise multiple outcomes into a single metric. Further information was provided to support the use of benefit-risk methods in appropriate circumstances, including the following: methods identified from the review were collated into different groupings and described to aid the selection of a method; potential implementation of methods throughout the trial process were provided and discussed (with examples); and general considerations were described for those using benefit-risk methods. Finally, a checklist of five pieces of information that should be present when reporting benefit-risk methods was defined, with two additional items specifically for reporting the results.

Conclusions

These recommendations will assist research teams in selecting which trial design to use and deciding whether or not a benefit-risk method could be included to ensure research questions are answered appropriately. Additional information is provided to support consistent use and clear reporting of benefit-risk methods in the future. The recommendations can also be used by funding committees to confirm that appropriate considerations of the trial design have been made.

Limitations

This research was limited in scope and should be considered in conjunction with other trial design methodologies to assess appropriateness. In addition, further research is needed to provide concrete information about which benefit-risk methods are best to use in publicly funded trials, along with recommendations that are specific to each method.

Study registration

The rapid review is registered as PROSPERO CRD42019144882.

Funding

Funded by the Medical Research Council UK and the National Institute for Health and Care Research as part of the Medical Research Council-National Institute for Health and Care Research Methodology Research programme.

Derivation of a UK preference-based value set for the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS) to allow estimation of Mental Well-being Adjusted Life Years (MWALYs)

BACKGROUND

The Mental Well-being Adjusted Life Year (MWALY) is an alternative outcome measure to the quality-adjusted life year (QALY) in economic evaluations of interventions aimed at improving mental well-being. However, there is a lack of preference-based mental well-being instruments for capturing population mental well-being preferences.

OBJECTIVES

To derive a UK preference-based value set for the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS).

METHODS

225 participants that were interviewed between December 2020 and August 2021 completed 10 composite time trade-off (C-TTO) and 10 discrete choice experiment (DCE) interviewer-administered exercises. Heteroskedastic Tobit and conditional logit models were used to model C-TTO and DCE responses respectively. The DCE utility values were rescaled to a C-TTO comparable scale through anchoring and mapping. An inverse variance weighting hybrid model (IVWHM) was used to derive weighted-average coefficients from the modelled C-TTO and DCE coefficients. Model performance was assessed using statistical diagnostics.

RESULTS

The valuation responses confirmed the feasibility and face validity of the C-TTO and DCE techniques. Apart from the main effects models, statistically significant associations were estimated between the predicted C-TTO value and participants' SWEMWBS scores, gender, ethnicities, education levels, and the interaction terms between age and useful feeling. The IVWHM was the most optimal model with the fewest logically inconsistent coefficients and the lowest pooled standard errors. The utility values generated by the rescaled DCE models and the IVWHM were generally higher than those of the C-TTO model. The predictive ability of the two DCE rescaling methods was similar according to the mean absolute deviation and root mean square deviation statistics.

CONCLUSIONS

This study has produced the first preference-based value set for a measure of mental well-being. The IVWHM provided a desirable blend of both C-TTO and DCE models. The value set derived by this hybrid approach can be used for cost-utility analyses of mental well-being interventions.

Efficient Designs for Valuation Studies That Use Time Tradeoff (TTO) Tasks to Map Latent Utilities from Discrete Choice Experiments to the Interval Scale: Selection of Health States for TTO Tasks

BACKGROUND

In eliciting utilities to value multiattribute utility instruments, discrete choice experiments (DCEs) administered online are less costly than interviewer-facilitated time tradeoff (TTO) tasks. DCEs capture utilities on a latent scale and are often coupled with a small number of TTO tasks to anchor utilities to the interval scale. Given the costly nature of TTO data, design strategies that maximize value set precision per TTO response are critical.

METHODS

Under simplifying assumptions, we expressed the mean square prediction error (MSE) of the final value set as a function of the number J of TTO-valued health states and the variance VJ of the states' latent utilities. We hypothesized that even when these assumptions do not hold, the MSE 1) decreases as VJ increases while holding J fixed and 2) decreases as J increases while holding VJ fixed. We used simulation to examine whether there was empirical support for our hypotheses a) assuming an underlying linear relationship between TTO and DCE utilities and b) using published results from the Dutch, US, and Indonesian EQ-5D-5L valuation studies.

RESULTS

Simulation set (a) supported the hypotheses, as did simulations parameterized using valuation data from Indonesia, which showed a linear relationship between TTO and DCE utilities. The US and Dutch valuation data showed nonlinear relationships between TTO and DCE utilities and did not support the hypotheses. Specifically, for fixed J, smaller values of VJ reduced rather than increased the MSE.

CONCLUSIONS

Given that, in practice, the underlying relationship between TTO and DCE utilities may be nonlinear, health states for TTO valuation should be placed evenly across the latent utility scale to avoid systematic bias in some regions of the scale.

HIGHLIGHTS

Valuation studies may feature a large number of respondents completing discrete choice tasks online, with a smaller number of respondents completing time tradeoff (TTO) tasks to anchor the discrete choice utilities to an interval scale.We show that having each TTO respondent complete multiple tasks rather than a single task improves value set precision.Keeping the total number of TTO respondents and the number of tasks per respondent fixed, having 20 health states directly valued through TTO leads to better predictive precision than valuing 10 health states directly.If DCE latent utilities and TTO utilities follow a perfect linear relationship, choosing the TTO states to be valued by weighting on the 2 ends of the latent utility scale leads to better predictive precision than choosing states evenly across the latent utility scale.Conversely, if DCE latent utilities and TTO utilities do not follow a linear relationship, choosing the states to be valued using TTO evenly across the latent utility scale leads to better predictive precision than weighted selection does.In the context of valuation of the EQ-5D-Y-3L, we recommend valuing 20 or more health states using TTO and placing them evenly across the latent utility scale.

Analytical Considerations When Anchoring Discrete Choice Experiment Values Using Composite Time Trade-Off Data: The Case of EQ-5D-Y-3L

Discrete choice experiments (DCEs) are becoming increasingly used to elicit preferences for children's health states. However, DCE data need to be anchored to produce value sets, and composite time trade-off (cTTO) data are typically used in the context of EQ-5D-Y-3L valuation. The objective of this paper is to compare different anchoring methods, summarise the characteristics of the value sets they produce, and outline key considerations for analysts. Three anchoring methods were compared using data from published studies: (1) rescaling using the mean value for the worst health state; (2) linear mapping; and (3) hybrid modelling. The worst state rescaling value set had the largest range. The worst state rescaling and linear mapping value sets preserved the relative importance of the dimensions from the DCE, whereas the hybrid model value set did not. Overall, the predicted values from the hybrid model value set were more closely aligned with the cTTO values. These findings are relatively generalisable. Deciding upon which anchoring approach to use is challenging, as there are numerous considerations. Where cTTO data are collected for more than one health state, anchoring on the worst health state will arguably be suboptimal. However, the final choice of approach may require value judgements to be made. Researchers should seek input from relevant stakeholders when commencing valuation studies to help guide decisions and should clearly set out their rationale for their preferred anchoring approach in study outputs.

Analyzing the Pain/Discomfort and Anxiety/Depression Composite Domains and the Meaning of Discomfort in the EQ-5D: A Mixed-Methods Study

OBJECTIVES

The EQ-5D has 2 composite domains: pain/discomfort (PD) and anxiety/depression (AD). This study aims to explore how respondents use the composites to self-report health and what the meaning of discomfort is in the EQ-5D for the general public.

METHODS

Both qualitative and quantitative data were collected in an online cross-sectional survey involving a nationally representative general population sample in Hungary (n = 1700). Respondents completed the 5-level version of EQ-5D, followed by the composites split into individual subdomains. Open-ended questions were asked to explore respondents' interpretations and experiences of discomfort.

RESULTS

Six different response behaviors were identified in the composites: "uniform" (21%-32%), "most severe" (30%-34%), "least severe" (16%-23%), "average" (2%-4%), "synergistic" (4%-5%), and "inconsistent" (13%-15%). Compared with the individual subdomains, many respondents under-reported their problems on both composites (PD 16%-22% and AD 6%-13%, P < .05). In respondents who scored differently in the 2 separate domains, mainly problems with the first subdomain determined responses in the composites (PD 66% and AD 61%). The discomfort subdomain in the EQ-5D captured more than 100 different problems, including pain, nonpain physical discomfort (eg, tiredness, dizziness, and nausea), and psychological discomfort (eg, anxiety, nervousness, and sadness). Women, older adults, and those in worse general health status more often considered discomfort as pain (P < .05).

CONCLUSIONS

We found empirical evidence of measurement error in the composite responses on the EQ-5D, including under- and inconsistent reporting, ordering effects, potential differential item functioning, and interdomain dependency. Our findings contribute new knowledge to the development of new and refinement of existing self-reported health status instruments, also beyond the EQ-5D.

Valuing EQ-5D-Y: the current state of play

BACKGROUND

For nearly a decade, value sets for the EQ-5D-Y were not available, reflecting challenges in valuing child HRQoL. A methodological research programme led to publication of a valuation protocol in 2020, which was rapidly taken up by local study teams. By the end of 2022, between 11 and 17 EQ-5D-Y value sets will be available, more than for any other child HRQoL measure. It is timely to review the experience of those using the protocol to identify early learnings and remaining issues where more research is needed.

METHODS

In June 2021, the EuroQol Group organised a three-day workshop, bringing together all those involved in EQ-5D-Y value set studies and related methodological research concerning EQ-5D-Y and valuation. Workshop discussions were captured by note taking and recording all sessions and online chat. A narrative summary of all sessions was produced and synthesised to identify points of agreement and aspects of methods where uncertainty remains.

RESULTS

There was broad agreement that DCE is working well as the principal valuation method. However, the most appropriate means of anchoring the latent scale values produced by DCE remains unclear. Some studies have deviated from the protocol by extending the number of states included in TTO tasks, to better support modelling of DCE and TTO. There is ongoing discussion about the relative merits of alternative variants of TTO and other methods for anchoring. Very few studies have consulted with local end-users to gauge the acceptability of methods used to value EQ-5D-Y.

CONCLUSIONS

Priority areas for research include testing alternative methods for anchoring DCE data; exploring the preferences of adolescents; and scale differences in values for EQ-5D-Y and adult EQ-5D states, and implications of such differences for the use of EQ-5D-Y values in HTA. Given the normative elements of the protocol, engaging with HTA bodies and other local users should be the first step for all future value set studies. Value sets undertaken to date are for the three-level EQ-5D-Y. However, the issues discussed in this paper are equally relevant to valuation of the five-level version of EQ-5D-Y; indeed, similar challenges are encountered valuing any measure of child HRQoL.

The Online Elicitation of Personal Utility Functions (OPUF) tool: a new method for valuing health states

Introduction Standard valuation methods, such as TTO and DCE are inefficient. They require data from hundreds if not thousands of participants to generate value sets. Here, we present the Online elicitation of Personal Utility Functions (OPUF) tool; a new type of online survey for valuing EQ-5D-5L health states using more efficient, compositional elicitation methods, which even allow estimating value sets on the individual level. The aims of this study are to report on the development of the tool, and to test the feasibility of using it to obtain individual-level value sets for the EQ-5D-5L. Methods We applied an iterative design approach to adapt the PUF method, previously developed by Devlin et al., for use as a standalone online tool. Five rounds of qualitative interviews, and one quantitative pre-pilot were conducted to get feedback on the different tasks. After each round, the tool was refined and re-evaluated. The final version was piloted in a sample of 50 participants from the UK. A demo of the EQ-5D-5L OPUF survey is available at: https://eq5d5l.me Results On average, it took participants about seven minutes to complete the OPUF Tool. Based on the responses, we were able to construct a personal EQ-5D-5L value set for each of the 50 participants. These value sets predicted a participants' choices in a discrete choice experiment with an accuracy of 80%. Overall, the results revealed that health state preferences vary considerably on the individual-level. Nevertheless, we were able to estimate a group-level value set for all 50 participants with reasonable precision. Discussion We successfully piloted the OPUF Tool and showed that it can be used to derive a group-level as well as personal value sets for the EQ-5D-5L. Although the development of the online tool is still in an early stage, there are multiple potential avenues for further research.

Self vs. other, child vs. adult. An experimental comparison of valuation perspectives for valuation of EQ-5D-Y-3L health states

OBJECTIVES

EQ-5D-Y-3L health states are valued by adults taking the perspective of a 10-year-old child. Compared to valuation of adult EQ-5D instruments, this entails two changes to the perspective: (i) child health states are valued instead of adult health states and: (ii) health states are valued for someone else instead of for oneself. Although earlier work has shown that these combined changes yield different values for child and adult health states that are otherwise equal, it currently remains unclear why. Hence, we aimed to disentangle the effects of both changes.

METHODS

A sample of 205 students (mean age: 19.48) was surveyed. Each respondent completed visual analogue scale (VAS) and time trade-off (TTO) tasks for five EQ-5D-Y-3L states, using four randomly ordered perspectives: (i) self-adult (themselves), (ii) other-adult (someone their age), (iii) self-child (themselves as a 10-year-old), (iv) other-child (a child of 10 years old). We compared how each perspective impacted outcomes, precision and quality of EQ-5D-Y-3L valuation.

RESULTS

Overall, differences between perspectives were consistent, with their direction being dependent on the health states and respondents. For VAS, the effect on outcomes of valuation depended on severity, but variance was higher in valuation with child perspectives. For TTO, we observed that EQ-5D-Y-3L states valued on behalf of others (i.e., children or adults) received higher valuations, but lower variances.

CONCLUSION

The use of a different perspective appears to yield systematic differences in EQ-5D-Y-3L valuation, with considerable heterogeneity between health states and respondents. This may explain mixed findings in earlier work.

Impact of Including Carer Information in Time Trade-Off Tasks: Results from a Pilot Study

INTRODUCTION

Carer quality of life (QoL) can be included in economic evaluations and captured using EQ-5D. Traditional valuation tasks require participants to imagine living in a health state for a number of years, without being told what to consider. This pilot study sought to investigate whether participants implicitly consider the impact of the health state on others, and the extent to which this may impact health state valuations.

METHODS

Composite time trade-off (TTO) interviews were conducted with a convenience sample. Each interview included a 'traditional' TTO exercise to value three health states, and a 'combined' TTO exercise, where participants valued the same health states again, having been informed that they would require a carer living in a particular health state. Qualitative feedback was collected after each exercise. Paired t-test comparisons of the utilities elicited in each exercise were made.

RESULTS

Thirty-three participants enrolled in the pilot. Mean differences between exercises were not statistically significant and differed in direction, although considerable heterogeneity was observed in individual response trajectories. Overall, 36% (n = 12) of participants expressed an unprompted concern about being a burden on others in the traditional exercise, and 67% (n = 22) of participants would have responded differently had the carer been in full health in the combined exercise.

CONCLUSION

Providing contextual information about carers may impact valuations. Further research is required to better understand the reasons behind the variation in individual response trajectories observed in this pilot study. The insights from this study may be useful for informing the design of related future studies.

Transforming challenges into opportunities: conducting health preference research during the COVID-19 pandemic and beyond

The disruptions to health research during the COVID-19 pandemic are being recognized globally, and there is a growing need for understanding the pandemic’s impact on the health and health preferences of patients, caregivers, and the general public. Ongoing and planned health preference research (HPR) has been affected due to problems associated with recruitment, data collection, and data interpretation. While there are no “one size fits all” solutions, this commentary summarizes the key challenges in HPR within the context of the pandemic and offers pragmatic solutions and directions for future research. We recommend recruitment of a diverse, typically under-represented population in HPR using online, quota-based crowdsourcing platforms, and community partnerships. We foresee emerging evidence on remote, and telephone-based HPR modes of administration, with further studies on the shifts in preferences related to health and healthcare services as a result of the pandemic. We believe that the recalibration of HPR, due to what one would hope is an impermanent change, will permanently change how we conduct HPR in the future.

Social tariffs and democratic choice-Do population-based health state values reflect the will of the people?

In economic evaluations of health technologies, health outcomes are commonly measured in terms of quality-adjusted life years (QALYs). QALYs are the product of time and health-related quality of life. Health-related quality of life, in turn, is determined by a social tariff, which is supposed to reflect the public's preference over health states. This study argues that, because of the tariff's role in the societal decision-making process, it should not be understood as merely an operational (statistical) definition of health, but as a major instrument of democratic participation. I outline what implications this might have for both the method used to aggregate individual preferences, and the set of individuals whose preferences should count. Alternative tariff specifications and decision rules are explored, and future research directions are proposed.

Comparison of online and face-to-face valuation of the EQ-5D-5L using composite time trade-off

OBJECTIVE

The aim of this study was to compare online, unsupervised and face-to-face (F2F), supervised valuation of EQ-5D-5L health states using composite time trade-off (cTTO) tasks.

METHODS

The official EuroQol experimental design and valuation protocol for the EQ-5D-5L of 86 health states were implemented in interviewer-assisted, F2F and unsupervised, online studies. Validity of preferences was assessed using prevalence of inconsistent valuations and expected patterns of TTO values. Respondent task engagement was measured using number of trade-offs and time per task. Trading patterns such as better-than-dead only was compared between modes. Value sets were generated using linear regression with a random intercept (RILR). Value set characteristics such as range of scale and dimension ranking were evaluated between modes.

RESULTS

Five hundred one online and 1,134 F2F respondents completed the surveys. Mean elicited TTO values were higher online than F2F when compared by health state severity. Compared to F2F, a larger proportion of online respondents did not assign the poorest EQ-5D-5L health state (i.e., 55555) the lowest TTO value ([Online] 41.3% [F2F] 12.2%) (p < 0.001). A higher percentage of online cTTO tasks were completed in 3 trade-offs or fewer ([Online] 15.8% [F2F] 3.7%), (p < 0.001). When modeled using the RILR, the F2F range of scale was larger than online ([Online] 0.600 [F2F] 1.307) and the respective dimension rankings differed.

CONCLUSIONS

Compared to F2F data, TTO tasks conducted online had more inconsistencies and decreased engagement, which contributed to compromised data quality. This study illustrates the challenges of conducting online valuation studies using the TTO approach.

An exploration of methods for obtaining 0 = dead anchors for latent scale EQ-5D-Y values

OBJECTIVES

Discrete choice experiments (DCEs) can be used to obtain latent scale values for the EQ-5D-Y, but these require anchoring at 0 = dead to meet the conventions of quality-adjusted life year (QALY) estimation. The primary aim of this study is to compare four preference elicitation methods for obtaining anchors for latent scale EQ-5D-Y values.

METHODS

Four methods were tested: visual analogue scale (VAS), DCE (with a duration attribute), lag-time time trade-off (TTO) and the location-of-dead (LOD) approach. In computer-assisted personal interviews, UK general public respondents valued EQ-5D-3L health states from an adult perspective and EQ-5D-Y health states from a 10-year-old child perspective. Respondents completed valuation tasks using all four methods, under both perspectives.

RESULTS

349 interviews were conducted. Overall, respondents gave lower values under the adult perspective compared to the child perspective, with some variation across methods. The mean TTO value for the worst health state (33333) was about equal to dead in the child perspective and worse than dead in the adult perspective. The mean VAS rescaled value for 33333 was also higher in the child perspective. The DCE produced positive child perspective values and negative adult perspective values, though the models were not consistent. The LOD median rescaled value for 33333 was negative under both perspectives and higher in the child perspective.

DISCUSSION

There was broad agreement across methods. Potential criteria for selecting a preferred anchoring method are presented. We conclude by discussing the decision-making circumstances under which utilities and QALY estimates for children and adults need to be commensurate to achieve allocative efficiency.

International Valuation Protocol for the EQ-5D-Y-3L

The EQ-5D-Y-3L is a generic, health-related, quality-of-life instrument for use in younger populations. Some methodological studies have explored the valuation of children's EQ-5D-Y-3L health states. There are currently no published value sets available for the EQ-5D-Y-3L that are appropriate for use in a cost-utility analysis. The aim of this article was to describe the development of the valuation protocol for the EQ-5D-Y-3L instrument. There were several research questions that needed to be answered to develop a valuation protocol for EQ-5D-Y-3L health states. Most important of these were: (1) Do we need to obtain separate values for the EQ-5D-Y-3L, or can we use the ones from the EQ-5D-3L? (2) Whose values should we elicit: children or adults? (3) Which valuation methods should be used to obtain values for child's health states that are anchored in Full health = 1 and Dead = 0? The EuroQol Research Foundation has pursued a research programme to provide insight into these questions. In this article, we summarized the results of the research programme concluding with the description of the features of the EQ-5D-Y-3L valuation protocol. The tasks included in the protocol for valuing EQ-5D-Y-3L health states are discrete choice experiments for obtaining the relative importance of dimensions/levels and composite time-trade-off for anchoring the discrete choice experiment values on 1 = Full Health and 0 = Dead. This protocol is now available for use by research teams to generate EQ-5D-Y-3L value sets for their countries allowing the implementation of a cost-utility analysis for younger populations.

Using EQ-5D Data to Measure Hospital Performance: Are General Population Values Distorting Patients' Choices?

Background. The English National Health Service publishes hospital performance indicators based on average postoperative EQ-5D index scores after hip replacement surgery to inform prospective patients' choices of hospital. Unidimensional index scores are derived from multidimensional health-related quality-of-life data using preference weights estimated from a sample of the UK general population. This raises normative concerns if general population preferences differ from those of the patients who are to be informed. This study explores how the source of valuation affects hospital performance estimates. Methods. Four different value sets reflecting source of valuation (general population v. patients), valuation technique (visual analog scale [VAS] v. time tradeoff [TTO]), and experience with health states (currently experienced vs. experimentally estimated) were used to derive and compare performance estimates for 243 hospitals. Two value sets were newly estimated from EQ-5D-3L data on 122,921 hip replacement patients and 3381 members of the UK general public. Changes in hospital ranking (nationally) and performance outlier status (nationally; among patients' 5 closest hospitals) were compared across valuations. Results. National rankings were stable under different valuations (rank correlations >0.92). Twenty-three (9.5%) hospitals changed outlier status when using patient VAS valuations instead of general population TTO valuations, the current approach. Outlier status also changed substantially at the local level. This was explained mostly by the valuation technique, not the source of valuations or experience with the health states. Limitations. No patient TTO valuations were available. The effect of value set characteristics could be established only through indirect comparisons. Conclusion. Different value sets may lead to prospective patients choosing different hospitals. Normative concerns about the use of general population valuations are not supported by empirical evidence based on VAS valuations.

Quality of life in chronic conditions using patient-reported measures and biomarkers: a DEA analysis in type 1 diabetes

BACKGROUND

A chronic disease impacts a patient's daily life, with the burden of symptoms and managing the condition, and concerns of progression and disease complications. Such aspects are captured by Patient-Reported Outcomes Measures (PROM), assessments of e.g. wellbeing. Patient-Reported Experience Measures (PREM) assess patients' experiences of healthcare and address patient preferences. Biomarkers are useful for monitoring disease activity and treatment effect and determining risks of progression and complications, and they provide information on current and future health. Individuals may differ in which among these aspects they consider important. We aimed to develop a measure of quality of life using biomarkers, PROM and PREM, that would provide an unambiguous ranking of individuals, without presuming any specific set of importance weights. We anticipated it would be useful for studying needs and room for improvement, estimating the effects of interventions and comparing alternatives, and for developing healthcare with a broad focus on the individual. We wished to examine if efficiency analysis could be used for this purpose, in an application to individuals with type 1 diabetes.

RESULTS

We used PROM and PREM data linked to registry data on risk factors, in a large sample selected from the National Diabetes Registry in Sweden. Efficiency analysis appears useful for evaluating the situation of individuals with type 1 diabetes. Quality of life was estimated as efficiency, which differed by age. The contribution of different components to quality of life was heterogeneous, and differed by gender, age and duration of diabetes. Observed quality of life shortfall was mainly due to inefficiency, and to some extent due to the level of available inputs.

CONCLUSIONS

The efficiency analysis approach can use patient-reported outcomes measures, patient-reported experience measures and comorbidity risk factors to estimate quality of life with a broad focus on the individual, in individuals with type 1 diabetes. The approach enables ranking and comparisons using all these aspects in parallel, and allows each individual to express their own view of which aspects are important to them. The approach can be used for policy regarding interventions on inefficiency as well as healthcare resource allocation, although currently limited to type 1 diabetes.

Describing the Development of a Health State Valuation Protocol to Obtain Community-Derived Disability Weights

For the prioritization of the allocation of national resources, estimating the burden of disease studies play a critical role. Hence the first Global Burden of Disease study conducted in the 1990s was done for this particular estimation. By the means of introducing disability-adjusted life year (DALY) metric, the burden of various diseases was calculated using disability weights (DWs)—a component of DALY. DWs are values that capture individuals' perception regarding the severity of diseases that involve valuation tools and health state descriptions. Various studies have been conducted over the past few decades to evaluate health states and derive disease-specific disability weights using Person-Trade off, Time-trade off, etc. However, use of these complex and cognitively demanding methods has been carried out in developed countries where the bulk of the populace is more educated. Few attempts have been made in low- and middle-income countries such as India, where not only the majority is less educated but also the social construction of diseases and health conditions are diverse. Therefore, due to the absence of methodological protocols of health state valuations for application at the community-level in the developing world, we attempted to systematically describe the procedure that can be used universally and cross-culturally for various health states. We began with the tentative selection of health states and health states valuation methods by conducting a meticulous literature review, followed by community exploration and medical consultations. This led to developing vignettes (clinical description) and 6D5L pictorial narrations (functional status description). Two field tests for checking the usability and refinement of the tools was done. Final consultation by an expert panel comprising of medical and non-medical professionals was held/conducted to finalize the health state labels and functional status profiles of each health state. The methodical approach provides a robust and thorough procedure for guiding researchers to implement health state valuation studies at community level.

MCDA-based deliberation to value health states: lessons learned from a pilot study

BACKGROUND

Health economists have shown a growing interest in deliberation and multi-criteria decision analysis (MCDA) as possible pathways to transparently integrate value judgments in cost-utility analyses. In line with these developments, this study piloted a consensus process to derive a German value set for the Short-Form Six-Dimension (SF-6D). In a conference setting, a group was tasked to deliberate on scores and weights for the SF-6D from the perspective of a self-determined and independent life.

METHODS

The one-day consensus conference was based on a deliberative process in combination with the MCDA method MACBETH (Measuring Attractiveness by a Categorical Based Evaluation Technique). According to MACBETH, participants were asked to qualitatively rate pairwise comparisons of SF-6D health states. The scoring within each dimension was conducted in parallel group sessions. Final agreement on the scores as well as weights for the SF-6D dimensions were derived in a subsequent plenary assembly. Results were analyzed using the software M-MACBETH and qualitative content analysis.

RESULTS

A total of 34 participants were recruited. While each of the 6 small groups presented a consented score, the plenary assembly reached consensus on all dimensions apart from pain. Concerning dimension weights, some participants favored prioritizing pain and mental health. Others disputed that trade-offs between dimensions and thus assigning weights were not acceptable in a context where this may involve withholding care from someone. As a consequence, no consensus on a value set was reached. Participants identified the group size of the plenary session and the applied weighting procedure as main obstacles to the process.

CONCLUSIONS

This pilot study presents a consensus-based approach for valuing health-related quality of life. However, further research is needed on deliberative processes that yield quantifiable results. Future conferences should explore smaller group sizes, longer durations of the deliberative process and alternatives to the additive value function applied in MACBETH.

A Norwegian 15D value algorithm: proposing a new procedure to estimate 15D value algorithms

Purpose

So far there is no Norwegian value algorithm to inform healthcare decision making. The 15D health state values estimated with the original 15D valuation procedure tend to be higher than the values of other generic preference-based health-related quality of life (HRQoL) instruments. The main purpose of this study was to use a new 15D valuation procedure to estimate Norwegian 15D health state values and to explore their empirical performance.

Methods

The visual analogue scale was used to collect 15D valuation data in a representative sample of the Norwegian general population. The new procedure used fewer valuation tasks and anchored the 15D health state values in an empirically assessed range. The Norwegian 15D health state values were compared to the values of five HRQoL instruments which were provided by Norwegian residents belonging to seven disease groups and a healthy population.

Results

The Norwegian 15D health state values ranged from 1 to − 0.52. Compared to 15D health state values estimated with the original procedure, the Norwegian 15D health state values were lower and more in line with values of other HRQoL instruments.

Conclusions

The new 15D valuation procedure is simpler, links the 15D health state values better to the requirements of the QALY model, and provides an empirically-based range. We recommend using the new valuation procedure in future 15D valuation studies, and the Norwegian health state values for use in 15D-based health economic analyses in Norway.