Mapping the Cancer-Specific EORTC QLQ-C30 to the Preference-Based EQ-5D, SF-6D, and 15D Instruments

Objectives:  To estimate models, via ordinary least squares regression, for predicting Euro Qol 5D (EQ-5D), Short Form 6D (SF-6D), and 15D utilities from scale scores of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30).
Methods:  Forty-eight gastric cancer patients, split up into equal subgroups by age, sex, and chemotherapy scheme, were interviewed, and the survey included the QLQ-C30, SF-36, EQ-5D, and 15D instruments, along with sociodemographic and clinical data. Model predictive ability and explanatory power were assessed by root mean square error (RMSE) and adjusted R ² values, respectively. Pearson's r between predicted and reported utility indices was compared. Three random subsamples, half in size the initial sample, were created and used for "external" validation of the modeling equations.
Results:  Explanatory power was high, with adjusted R ² reaching 0.909, 0.833, and 0.611 for 15D, SF-6D, and EQ-5D, respectively. After normalization of RMSE to the range of possible values, the prediction errors were 12.0, 5.4, and 5.6% for EQ-5D, SF-6D, and 15D, respectively. The estimation equations produced a range of utility scores similar to those achievable by the standard scoring algorithms. Predicted and reported indices from the validation samples were comparable thus confirming the previous results.
Conclusions:  Evidence on the ability of QLQ-C30 scale scores to validly predict 15D and SF-6D utilities, and to a lesser extent, EQ-5D, has been provided. The modeling equations must be tried in future studies with larger and more diverse samples to confirm their appropriateness for estimating quality-adjusted life-year in cancer-patient trials including only the QLQ-C30.     

Mapping Utility Scores from a Disease-Specific Quality-of-Life Measure in Bariatric Surgery Patients

Objectives:  To develop algorithms for a conversion of disease-specific quality-of-life into health state values for morbidly obese patients before or after bariatric surgery.
Methods:  A total of 893 patients were enrolled in a prospective cross-sectional multicenter study. In addition to demographic and clinical data, health-related quality-of-life (HRQoL) data were collected using the disease-specific Moorehead-Ardelt II questionnaire (MA-II) and two generic questionnaires, the EuroQoL-5D (EQ-5D) and the Short Form-6D (SF-6D). Multiple regression models were constructed to predict EQ-5D- and SF-6D-based utility values from MA-II scores and additional demographic variables.
Results:  The mean body mass index was 39.4, and 591 patients (66%) had already undergone surgery. The average EQ-5D and SF-6D scores were 0.830 and 0.699. The MA-IIwas correlated to both utility measures (Spearman's r  = 0.677 and 0.741). Goodness-of-fit was highest ( R ² = 0.55 in the validation sample) for the following item-based transformation algorithm: utility (MA-II-based) = 0.4293 + (0.0336 × MA1) + (0.0071 × MA2) + (0.0053 × MA3) + (0.0107 × MA4) + (0.0001 × MA5). This EQ-5D-based mapping algorithm outperformed a similar SF-6D-based algorithm in terms of mean absolute percentage error ( P  = 0.045).
Conclusions:  Because the mapping algorithm estimated utilities with only minor errors, it appears to be a valid method for calculating health state values in cost-utility analyses. The algorithm will help to define the role of bariatric surgery in morbid obesity.     

Comparison of EQ-5D and SF-6D utilities in mental health patients

Various preference-based measures of health are available for use as an outcome measure in cost-utility analysis. The aim of this study is to compare two such measures EQ-5D and SF-6D in mental health patients. Baseline data from a Dutch multi-centre randomised trial of 616 patients with mood and/or anxiety disorders were used. Mean and median EQ-5D and SF-6D utilities were compared, both in the total sample and between severity subgroups based on quartiles of SCL-90 scores. Utilities were expected to decline with increased severity.Both EQ-5D and SF-6D utilities differed significantly between patients of adjacent severity groups. Mean utilities increased from 0.51 at baseline to 0.68 at 1.5 years follow-up for EQ-5D and from 0.58 to 0.70 for SF-6D. For all severity subgroups, the mean change in EQ-5D utilities as well as in SF-6D utilities was statistically significant. Standardised response means were higher for SF-6D utilities. We concluded that both EQ-5D and SF-6D discriminated between severity subgroups and captured improvements in health over time. However, the use of EQ-5D resulted in larger health gains and consequent lower cost-utility ratios, especially for the subgroup with the highest severity of mental health problems.     

A comparison of the performance of the EQ-5D and SF-6D for individuals aged >or= 45 years

We sought to compare the performance of the EQ-5D and SF-6D with regard to the criteria of practicality, convergent validity, and construct validity, the level of agreement between the two measures was also assessed. Responses from 1865 individuals aged >or= 45 years in one general practice were analysed. Of these, 93.1% completed the EQ-5D, compared with 86.4% for the SF-6D, where individuals who were older, female, of a lower occupational skill level, from an area of lower deprivation, or used prescribed medication were significantly less likely to complete the SF-6D. The performance of both measures was comparable with regard to both convergent and construct validities, as both the EQ-5D and SF-6D scores were closely related to scores on the EuroQol visual analogue scale (VAS) (p<0.001) and able to discriminate between people who did and did not take: (i) analgesics and (ii) other prescribed medication. Despite EQ-5D and SF-6D scores being highly correlated (p<0.001), individuals who were healthier (according to the VAS) had higher mean scores on the EQ-5D (p<0.001), whereas less healthy individuals had higher mean scores on the SF-6D (individuals with knee pain, osteoarthritis, back pain, rheumatoid arthritis, and hip pain had significantly lower mean scores on the EQ-5D, p<0.001).     

Why Do Patients with Inflammatory Arthritis Often Score States “Worse than Death” on the EQ-5D? An Investigation of the EQ-5D Classification System

Objective:  Using inflammatory arthritis patients as an example, we investigate EuroQol-5D (EQ-5D) profiles resulting in states worse than death (WTD), and the heath status of patients occupying these states.
Methods:  Baseline data from two UK trials were used that reflected the range of arthritis states/severity found in routine practice. EQ-5D profiles resulting in negative valuations (i.e., states WTD) based on UK weights were identified. EQ-5D scores for these profiles from alternative valuation sets, including a reanalysis of the UK weights, were compared. The health status and characteristics of patients, and factors associated with patients in the low distribution of the EQ-5D and those with WTD EQ-5D scores were identified.
Results:  Seven hundred patients were included in the analysis. Sixty-two (9%) patients occupied states WTD. Patients occupied 9 of the possible 84health profiles with negative scores (53% occupied one profile); this profile was not rated WTD by any of the alternative EQ-5D scoring algorithms. All WTD profiles included severe pain/discomfort plus moderate problems in ≥3 other domains. Patients with WTD valuations reported higher levels of pain, and feeling downhearted and low on alternative health status measures.
Conclusions:  Pain was the predominant factor in the WTD EQ-5D profiles occupied by arthritis patients. Patients occupying states WTD have poorer health-related quality of life than patients in low "better than death" states. Valuations of profiles vary according to how sets of preference weights for health profiles were developed. Further research should explore whether WTD valuations are supported by qualitative evidence and reflect the patient's health and experience of disease.     

Mapping between the Roland Morris Questionnaire and Generic Preference-Based Measures

ObjectivesThe Roland Morris Questionnaire (RMQ) is a widely used health status measure for low back pain (LBP). It is not preference-based, and there are currently no established algorithms for mapping between the RMQ and preference-based health-related quality-of-life measures. Using data from randomized controlled trials of treatment for LBP, we sought to develop algorithms for mapping between RMQ scores and health utilities derived using either the EuroQol five-dimensional questionnaire (EQ-5D) or the six-dimensional health state short form (derived from Medical Outcomes Study 36-Item Short-Form Health Survey) (SF-6D).MethodsThis study is based on data from the Back Skills Training Trial in which data were collected from 701 patients at baseline and subsequently at 3, 6, and 12 months postrandomization using a range of outcome measures, including the RMQ, the EQ-5D, and the Short Form 12 item Health Survey (SF-12) (from which SF-6D utilities can be derived). We used baseline trial data to estimate models using both direct and response mapping approaches to predict EQ-5D and SF-6D health utilities and dimension responses. A multistage model selection process was used to assess the predictive accuracy of the models. We then explored different techniques and mapping models that made use of repeated follow-up observations in the data. The estimated mapping algorithms were validated using external data from the UK Back Pain Exercise and Manipulation trial.ResultsA number of models were developed that accurately predict health utilities in this context. The best performing model for RMQ to EQ-5D mapping was a beta regression with Bayesian quasi-likelihood estimation that included 24 dummy variables for RMQ responses, age, and sex as covariates (mean squared error 0.0380) based on repeated data. The model selected for RMQ to SF-6D mapping was a finite mixture model that included the overall RMQ score, age, sex, RMQ score squared, age squared, and an interaction term for age and RMQ score as covariates (mean squared error 0.0114) based on repeated data.ConclusionsIt is possible to reasonably predict EQ-5D and SF-6D health utilities from RMQ scores and responses using regression methods. Our regression equations provide an empirical basis for estimating health utilities when EQ-5D or SF-6D data are not available. They can be used to inform future economic evaluations of interventions targeting LBP.     

Mapping the English and Chinese Versions of the Functional Assessment of Cancer Therapy–General to the EQ-5D Utility Index

Objective:  This study aims to develop a function for mapping the English and Chinese versions of the Functional Assessment of Cancer Therapy–General (FACT–G) scores to the EuroQoL Group's EQ-5D utility index and to test whether a single function is sufficient for the two language versions.
Methods:  A baseline survey of 558 cancer patients in Singapore using the FACT–G and EQ-5D was conducted (308 English and 250 Chinese questionnaires). Regression models were used to predict the EQ-5D utility index values based on the FACT–G scores and thus derive a mapping equation. Data from a follow-up survey of the patients were used to validate the results.
Results:  The FACT–G Social/Family scale was not associated with the EQ-5D utility index ( P  = 0.701). There was no interaction between language version and the predictors (each P  > 0.1). An equation that maps the FACT–G Physical, Emotional, and Functional well-being scales to the EQ-5D utility index was derived. In the validation sample, the mean observed utility values was larger than the mapped by only 0.005 (95% confidence interval [CI]−0.006 to 0.016), but the mean absolute difference was 0.083 (95% CI 0.076 to 0.090).
Conclusions:  At the group level, but not individual level, the equation developed can accurately map the English and Chinese versions of the FACT–G scores to the EQ-5D utility index.     

Mapping from Disease-Specific Measures to Utility: An Analysis of the Relationships between the Inflammatory Bowel Disease Questionnaire and Crohn''s Disease Activity Index in Crohn''s Disease and Measures of Utility

OBJECTIVES: To examine the relationship between the Inflammatory Bowel Disease Questionnaire (IBDQ), Crohn's Disease Activity Index (CDAI) and measures of utility (EQ-5D and the SF-6D indexes), and to estimate algorithms to map the two utility values from IBDQ and CDAI scores. METHODS: A large data set from clinical trials in Crohn's disease provided contemporaneous patient responses to all four questionnaires. Paired observations from multiple time-points were analyzed. We calculated mean utility scores by IBDQ and CDAI score deciles; Spearman correlation coefficients for paired observations between IBDQ and EQ-5D (n = 3320) and IBDQ and SF-6D (n = 3230), and explored regression models using maximum likelihood estimation. The IBDQ/SF-6D model was validated against paired observations from an independent data set. RESULTS: The IBDQ decile analysis demonstrated a consistent positive relationship with both utility indexes. Correlations between the IBDQ and both the EQ-5D and SF-6D were statistically significant (P < 0.0001), with correlation coefficients of 0.76 and 0.85, respectively. A simple linear model between EQ-5D and IBDQ explained 45% of the variance. The residuals plot for the IBDQ/SF-6D model suggested some nonlinearity and a nonlinear model explained 69% of the variance. In the validation analysis, no statistically significant difference was observed between the mean observed SF-6D and the SF-6D scores estimated using the IBDQ/SF-6D regression model. CONCLUSIONS: Given the strength, consistency, and predictable characteristics of the relationships, the algorithms appear to provide valuable and valid methods to estimate utilities from IBDQ scores (but not CDAI) in trials of Crohn's disease patients that have collected IBDQ scores but not utilities.     

Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D

Background: The SF-6D and EQ-5D are both preference-based measures of health. Empirical work is required to determine what the smallest change is in utility scores that can be regarded as important and whether this change in utility value is constant across measures and conditions. Objectives: To use distribution and anchor-based methods to determine and compare the minimally important difference (MID) for the SF-6D and EQ-5D for various datasets. Methods: The SF-6D is scored on a 0.29–1.00 scale and the EQ-5D on a −0.59–1.00 scale, with a score of 1.00 on both, indicating ‘full health’. Patients were followed for a period of time, then asked, using question 2 of the SF-36 as our anchor, if their general health is much better (5), somewhat better (4), stayed the same (3), somewhat worse (2) or much worse (1) compared to the last time they were assessed. We considered patients whose global rating score was 4 or 2 as having experienced some change equivalent to the MID. This paper describes and compares the MID and standardised response mean (SRM) for the SF-6D and EQ-5D from eight longitudinal studies in 11 patient groups that used both instruments. Results: From the 11 reviewed studies, the MID for the SF-6D ranged from 0.011 to 0.097, mean 0.041. The corresponding SRMs ranged from 0.12 to 0.87, mean 0.39 and were mainly in the ‘small to moderate’ range using Cohen’s criteria, supporting the MID results. The mean MID for the EQ-5D was 0.074 (range −0.011–0.140) and the SRMs ranged from −0.05 to 0.43, mean 0.24. The mean MID for the EQ-5D was almost double that of the mean MID for the SF-6D. Conclusions: There is evidence that the MID for these two utility measures are not equal and differ in absolute values. The EQ-5D scale has approximately twice the range of the SF-6D scale. Therefore, the estimates of the MID for each scale appear to be proportionally equivalent in the context of the range of utility scores for each scale. Further empirical work is required to see whether or not this holds true for other utility measures, patient groups and populations.     

Predicting EuroQoL EQ-5D preference scores from the SF-12 Health Survey in a nationally representative sample.

PURPOSE: To predict the EuroQoL EQ-5D utility index from the SF-12 Health Survey for a US national sample of adults. METHODS: The authors used the 2000 Medical Expenditure Panel Survey to examine the relationship between instruments. Linear regression was used to predict EQ-5D scores from Physical Component Summary (PCS) and Mental Component Summary (MCS) scores of the SF-12. A prediction model was derived in one half of the sample and validated in the other half. RESULTS: Complete responses to both measures were available for 14,580 adults; 7313 (50.2%) surveys were used for the derivation set. The 2-variable model predicted 61% of the variance in EQ-5D scores and provided reasonable ability to predict mean EQ-5D scores from mean PCS and MCS scores. Confidence intervals are dependent on sample size and variance of PCS and MCS scores. CONCLUSIONS: EQ-5D scores can be reasonably predicted from the SF-12. This model allows researchers to estimate utility data for use in decision and cost-utility analyses.     

Estimating health utilities in patients with asthma and COPD: evidence on the performance of EQ-5D and SF-6D

Objective  The objective of this study was to understand systematic differences in utility values derived from the EQ-5D and the SF-6D in two respiratory populations with heterogeneous disease severity. Methods  This study involved secondary analysis of data from two cross-sectional surveys of patients with asthma (N = 228; Hungary) and COPD (N = 176; Sweden). Disease severity was defined according to GINA and GOLD guidelines for asthma and COPD, respectively. EQ-5D and SF-6D scores and their distributional characteristics were compared across the two samples by disease severity level. Results  Within each patient population, mean EQ-5D and SF-6D scores were similar for the overall group and for those with moderate disease. Mean scores varied for patients with mild and severe disease. EQ-5D versus SF-6D scores in the asthma group by severity levels were 0.89 versus 0.80, 0.70 versus 0.73, 0.63 versus 0.64, and 0.51 versus 0.63, respectively. EQ-5D versus SF-6D scores in the COPD group by severity levels were 0.85 versus 0.80, 0.73 versus 0.73, 0.74 versus 0.73, and 0.53 versus 0.62, respectively. Conclusions  Results suggest the EQ-5D and SF-6D do not yield consistent utility values in patients with asthma and COPD due to differences in underlying valuation techniques and the EQ-5D’s limited response options relative to mild disease.     

An investigation into the empirical validity of the EQ-5D and SF-6D based on hypothetical preferences in a general population

BACKGROUND: An important consideration for studies that derive utility scores using multi-attribute utility measures is the psychometric integrity of the measurement instrument. Of particular importance is the requirement to establish the empirical validity of multi-attribute utility measures; that is, whether they generate utility scores that, in practice, reflect people's preferences. We compared the empirical validity of EQ-5D versus SF-6D utility scores based on hypothetical preferences in a large, representative sample of the English population. METHODS: Adult participants in the 1996 Health Survey for England (n=16 443) formed the basis of the investigation. The subjects were asked to complete the EQ-5D and SF-36 measures. Their responses were converted into utility scores using the York A1 tariff set and the SF-6D utility algorithm, respectively. One-way analysis of variance was used to test the hypothetically constructed preference rule that each set of utility scores differs significantly by self-reported health status (categorised as very good, good, fair, bad or very bad). The degree to which EQ-5D and SF-6D utility scores reflect alternative configurations of self-reported health status; illness, disability or infirmity, and medication use was tested using the relative efficiency statistic and receiver operating characteristic (ROC) curves. RESULTS: The mean utility score for the EQ-5D was 0.845 (95% CI: 0.842, 0.849), whilst the mean utility score for the SF-6D was 0.799 (95% CI: 0.797, 0.802), representing a mean difference in utility score of 0.046 (95% CI: 0.044, 0.049; p<0.001). Bland-Altman plots displayed considerable lack of agreement between the two measures, particularly at the lower end of the utility scale. Both measures demonstrated statistically significant differences between subjects who described their health status as very good, good, fair, bad or very bad (p<0.001), as well as monotonically decreasing utility scores (test for linear trend: p<0.001). The SF-6D was between 30.9 and 100.4% more efficient than the EQ-5D at detecting differences in self-reported health status, and between 10.4 and 45.6% more efficient at detecting differences in illness, disability or infirmity and medication use. The area under the curve scores generated by the ROC curves were significantly higher for the SF-6D at the 0.1% significance level when self-reported health status was dichotomised as very good versus good, fair, bad or very bad. However, the AUC scores did not reveal any significant differences in the discriminatory powers of the measures when alternative configurations of illness, disability or infirmity and medication use were examined. CONCLUSIONS: This study provides evidence that the SF-6D is an empirically valid and efficient alternative multi-attribute utility measure to the EQ-5D, and is capable of discriminating between external indicators of health status. However, health economists should also consider other psychometric properties, such as practicality and reliability, when selecting either measure for evaluative purposes.     

An empirical comparison of EQ-5D and SF-6D in liver transplant patients

There remains disagreement about the preferred utility-based measure of health-related quality of life for use in constructing quality-adjusted life years (QALYs). The recent development of a new measure, the SF-6D, has highlighted this issue. The SF-6D and EuroQol EQ-5D measure health-related utilities on a scale where 0 represents death and 1 represents full health, and both have utility scores generated from random samples of the general UK population. This study explored whether, in a large sample of liver transplant patients, the two instruments provide similar results. The empirical data highlight important variation in the results generated from the use of the two instruments. The data are consistent with a view that the SF-6D does not describe health states at the lower end of the utility scale but is more sensitive than EQ-5D in detecting small changes towards the top of the scale.     

Mapping the SF-12 to the EuroQol EQ-5D Index in a national US sample.

BACKGROUND: Preference scores for the Medical Outcomes Study (MOS) SF-12 would enable its use in cost-effectiveness analyses. Previous mapping studies of MOS instruments top reference-based instruments have not examined performance in national samples. PARTICIPANTS: 15,000 adults in the 2000 Medical Expenditure Panel Survey annual survey including the SF-12 and EQ-5D Index. METHODS: Regression of the EQ-5D Index scores onto the physical and mental component summary scores of the SF-12, testing 2nd-4th degree polynomial and spline models, including and excluding sociodemographics. RESULTS: A 2nd degree polynomial model explained 63% of the variance in EQ-5D scores, with robust internal and external validation. More complex mod-els explained minimally additional variance. Compared with EQ-5D valuations, prediction models overestimated the lowest health states (6% of the population). CONCLUSIONS: The mapped SF-12 yields usable preference-scaled scores, with some caution for the lowest health states.     

Comparing the Health Utilities Index Mark 3 (HUI3) with the Short Form-36 Preference-Based SF-6D in Chronic Kidney Disease

Objective:  Assess within-subject agreement and compare discriminative abilities between the SF-6D and the Health Utilities Index Mark 3 (HUI3) in patients with chronic kidney disease (CKD).
Methods:  The HUI3 and Short Form-36 were self-completed by 185 CKD patients enrolled in a prospective study of incident patients with stage 4 and 5 CKD.
Results:  The mean preference-based score for the SF-6D was 0.67 ± 0.13 compared to 0.58 ± 0.26 for the HUI3 ( P  < 0.01). There was a strong association between SF-6D and HUI3 scores (Pearson correlation coefficient 0.55, 95% CI 0.43–0.65) and moderate agreement with an intraclass correlation coefficient of 0.44. The HUI3 was better able to capture more severe burden of illness with fewer floor effects. The SF-6D was better at capturing differences among patients at the top range of the scale with fewer ceiling effects. Both the HUI3 and SF-6D were able to discriminate between patient groups differing in disease severity defined as predialysis versus dialysis dependent and depressive symptoms using a Beck Depression Inventory II score of ≥14 as the cutoff. The HUI3 was better able to discriminate greater depressive symptoms.
Conclusion:  The SF-6D and the HUI3 generate different preference-based scores for patients with CKD and any comparison between their scores should be made with caution. The HUI3 appears more suitable for measuring the health of populations with greater disability such as patients with CKD. It remains to be determined whether these differences will remain when one compares within-instrument differences in preference scores over time.