Harrison MJ, Davies LM, Bansback NJ, McCoy MJ, Verstappen SM, Watson K, Symmons DP; The British Society for Rheumatology Biologics Register Control Centre Consortium, on behalf of the BSRBR

PURPOSE: Comparative evidence regarding the responsiveness of the EQ-5D and SF-6D in arthritis patients is conflicting and insufficient across the range of disease severity. We examined the comparative responsiveness of the EQ-5D and SF-6D in cohorts of patients with early inflammatory disease through to severe rheumatoid arthritis (RA). METHODS: Responsiveness was tested using the effect size (ES) and standardised response mean (SRM). Correlation of change in EQ-5D and SF-6D with disease specific measures was tested using Pearson correlations and the Steiger’s Z test. Treatment response and self-reported change were used as anchors of important change. RESULTS: The EQ-5D was more responsive to deterioration (ES ratio (EQ-5D/SF-6D): 1.6-3.0) and the SF-6D more responsive to improvement (ES ratio (SF-6D/EQ-5D): 1.1-1.8) in health. The SF-6D did not respond well to deterioration in patients with established severe RA (ES and SRM 0.08). The EQ-5D provided larger absolute mean change estimates but with greater variance compared to the SF-6D. CONCLUSIONS: The comparative responsiveness of the EQ-5D and SF-6D differs according to the direction of change. The level of mean change of the EQ-5D relative to the SF-6D has implications for cost-effectiveness analysis. Use of the SF-6D in patients with severe progressive disease may be inappropriate.

Qual Life Res. 2009 Sep 24

Nosyk B, Sun H, Bansback N, Guh DP, Li X, Barnett P, Bayoumi A, Griffin S, Joyce V, Holodniy M, Owens DK, Anis AH

OBJECTIVES: To assess the concurrent validity and responsiveness of the Health Utility Index 3 (HUI3) in patients with advanced HIV/AIDS, and to determine the responsiveness of this measure, the MOS-HIV and EQ-5D to HIV-related clinical events. METHODS: Data from the OPTIMA (OPTions In Management with Antiretrovirals) trial was analyzed. Two aspects of the validity of the HUI3 were considered: concurrent validity was evaluated using Spearman correlations with MOS-HIV component and summary scores. Responsiveness to AIDS-defining events (ADE) and all adverse events (our external change criterion) was assessed using area under the receiver operating characteristic (AUROC) curves. RESULTS: The study enrolled 368 patients (mean follow-up: 3.66 years); 82% had at least one severe adverse event and 27% had at least one ADE. The HUI3 scale and items showed good concurrent validity, with 85% of the expected relationships with the MOS-HIV subscales verified. The HUI3 was responsive to both adverse events (AUROC [95%CI]: 0.68 [0.57, 0.80]) and ADEs (0.62 [0.51, 0.74]). The EQ-5D was responsive to ADEs (0.66 [0.56, 0.76]), but not responsive to adverse events (0.56 [0.46, 0.68]). CONCLUSION: The HUI3 is a valid and responsive measure of the change in HRQoL associated with clinical events in an advanced HIV/AIDS population.

Qual Life Res. 2009 Sep;18(7):815-24

Harrison MJ, Davies LM, Bansback NJ, McCoy MJ, Farragher TM, Verstappen SM, Hassell A, Symmons DP

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 84 health 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 >or=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.

Value Health. 2009 Sep;12(6):1026-34

Anis AH, Nosyk B, Sun H, Guh DP, Bansback N, Li X, Barnett PG, Joyce V, Swanson KM, Kyriakides TC, Holodniy M, Cameron DW, Brown ST; OPTIMA Team1

OBJECTIVE: To investigate the relative magnitude and duration of impact of AIDS-defining events (ADEs) and non-AIDS serious adverse events (SAEs) on health-related quality of life (HRQoL) among patients with advanced HIV/AIDS. METHODS: We use data from OPTIMA (OPTions In Management with Antiretrovirals), a multinational, randomized, open, control, clinical management trial of treatment strategies for patients with multidrug-resistant HIV and advanced immune disease. Longitudinal models were used to determine the effects of ADEs and SAEs on HRQoL across periods before, during, and after event onset. The Medical Outcomes Study HIV Health Survey (MOS-HIV) physical and mental health summary scores (MOS-PHS and MOS-MHS), EQ-5D, and the Health Utilities Index Mark 3 HRQoL measures were all assessed at regular follow-up intervals during the trial. RESULTS: ADEs occurred much less frequently than SAEs (n = 147 vs. n = 821) in the study sample population of 368 patients, during median follow-up of 3.96 years. Although both ADEs and SAEs had significant negative impacts on HRQoL, SAEs had at least as large an impact upon HRQoL as ADEs when both were included in a multivariate linear regression model, controlling for other covariates. However, the effect of ADEs on HRQoL was more persistent, with larger magnitude of effect across all instruments in time intervals further from the onset of the event. CONCLUSIONS: Non-AIDS SAEs occurring in patients with late-stage HIV/AIDS seem to have at least as important an immediate impact on patient HRQoL as ADEs; however, the impact of ADEs seems to be more persistent. Our findings call for a greater emphasis on the detection and active prevention of non-AIDS SAEs in patients with late-stage HIV/AIDS.

J Acquir Immune Defic Syndr. 2009 Aug 15;51(5):631-9

Anis AH, Zhang W, Bansback N, Guh DP, Amarsi Z, Birmingham CL

Summary This study is to update the estimates of the economic burden of illness because of overweight and obesity in Canada by incorporating the increase in prevalence of overweight and obesity, findings of new related comorbidities and rise in the national healthcare expenditure. The burden was estimated from a societal perspective using the prevalence-based cost-of-illness methodology. Results from a literature review of the risks of 18 related comorbidities were combined with prevalence of overweight and obesity in Canada to estimate the extent to which each comorbidity is attributable to overweight and obesity. The direct costs were extracted from the National Health Expenditure Database and allocated to each comorbidity using weights principally from the Economic Burden of Illness in Canada. The study showed that the total direct costs attributable to overweight and obesity in Canada were $6.0 billion in 2006, with 66% attributable to obesity. This corresponds to 4.1% of the total health expenditures in Canada in 2006. The inclusion of newly identified comorbidities increased the direct cost estimates of obesity by 25%, while the rise in national healthcare expenditure accounted for a 19% increase. Policies to reduce being overweight and obese could potentially save the Canadian healthcare system millions of dollars.

Obes Rev. 2009 Apr 1.

Bansback N, Sun H, Guh DP, Li X, Nosyk B, Griffin S, Barnett PG, Anis AH; OPTIMA TEAM.

The impact of healthcare interventions on health utility values is most frequently measured using a preference-based instrument. Each of the available instruments instructs the respondent to report their health status over different recall periods ranging from the current day to the past month. In an ongoing randomised controlled trial in patients with advanced HIV disease, the impact of using a preference-based instrument with a 1-week recall period vs a 1-day recall period (e.g. today) for capturing recently resolved serious adverse events was measured. The results suggest that the instrument with a 1-week recall period gave lower utility values for recently resolved events in comparison with the instrument with a 1-day recall period. A plausible interpretation of these results is that the recall period was adhered to; for example, patients ignored the impact of recently resolved events in their response if the questionnaire asked them only about their health today. While there are limitations to our study, we believe further consideration should be given to the recall period used for preference-based instruments, and future research should examine other patient groups using a single instrument with multiple recall periods.

Health Econ. 2008 Dec;17(12):1413-9

Bansback N, Sun H, Guh DP, Li X, Nosyk B, Griffin S, Barnett PG, Anis AH; OPTIMA TEAM

The impact of healthcare interventions on health utility values is most frequently measured using a preference-based instrument. Each of the available instruments instructs the respondent to report their health status over different recall periods ranging from the current day to the past month. In an ongoing randomised controlled trial in patients with advanced HIV disease, the impact of using a preference-based instrument with a 1-week recall period vs a 1-day recall period (e.g. today) for capturing recently resolved serious adverse events was measured. The results suggest that the instrument with a 1-week recall period gave lower utility values for recently resolved events in comparison with the instrument with a 1-day recall period. A plausible interpretation of these results is that the recall period was adhered to; for example, patients ignored the impact of recently resolved events in their response if the questionnaire asked them only about their health today. While there are limitations to our study, we believe further consideration should be given to the recall period used for preference-based instruments, and future research should examine other patient groups using a single instrument with multiple recall periods.

Health Econ. 2008 Dec;17(12):1413-9

Bansback N, Ara R, Karnon J, Anis A

We reviewed the clinical measures used in rheumatoid arthritis (RA) economic evaluations with respect to their relevance and sensitivity to changes in survival, health-related quality of life (HR-QOL) and costs. We compared the measures from the economic perspective and discussed the validity of methods used to extrapolate beyond the trial data. Cost-effectiveness evaluations of disease-modifying antirheumatic drugs in RA were identified by searching MEDLINE, EMBASE, Econlit and NHS EED databases. Studies were retained if they extrapolated beyond randomized controlled trial evidence using relationships between clinical measures, costs and utilities. In the 22 studies identified, clinical severity was measured using the Health Assessment Questionnaire (HAQ) Disability Index, the American College of Rheumatology (ACR) response criteria, the Disease Activity Score (DAS) or a combination of the HAQ and DAS. The HAQ is correlated with mortality, costs and HR-QOL instruments, and several studies used linear relationships to model these associations. However, a polynomial relationship or discrete states may be more appropriate for patients at the extremes of the disease spectrum, and numerous HAQ health states may be required to capture differences in mortality risk. While the ACR response criteria is a more comprehensive measure than the HAQ, it is a relative measure, which creates difficulties when estimating absolute changes in HR-QOL, costs and mortality risk. The evidence base linking DAS scores with HR-QOL instruments, costs and mortality is less robust, possibly due to the comparatively recent development of the measure and the limited number of possible scores (mild/moderate/severe). While there is some evidence of a relationship between DAS scores and costs, the DAS does not capture all aspects of HR-QOL, and no significant relationship has been established with mortality risk. Evidence suggests the HAQ to be the primary clinical measure for use in economic evaluations as it is measured in almost all clinical studies, and is closely correlated to health utilities, mortality and costs. While new developments suggest the sensitivity of health states may be improved by combining the HAQ with measures such as the DAS, further research is required in this area. Further research is also required to explore the advantages in using either continuous or discrete health states.

Pharmacoeconomics. 2008;26(5):395-408

Bansback N, Harrison M, Brazier J, Davies L, Kopec J, Marra C, Symmons D, Anis A.

Arthritis Rheum. 2008 Jul 15;59(7):1018-26

Harrison MJ, Davies LM, Bansback NJ, Ingram M, Anis AH, Symmons DP

OBJECTIVE: Cost-utility analysis is increasingly important as healthcare providers aim to invest scarce resources in interventions offering the greatest health benefit. The ability to attach utility values to health states is essential, and is increasingly performed using generic scales. However, the evidence regarding the validity of generic utility scales in rheumatoid arthritis (RA) is unclear. We summarize and review evidence on the validity and comparative performance of generic utility scales in RA. METHODS: We searched the English-language medical literature for studies using utilities in RA between 1980 and mid-2006. Reports describing primary evidence of the validity or performance of a generic utility scale in RA were selected, summarized, and reviewed using the OMERACT filter. RESULTS: In total 923 articles were identified, of which 228 reported the use of utility scales in RA; 26 studies related to the validation or evidence of generic utility scales in RA, the EQ-5D, Health Utility Index-2 (HUI2) and HUI3, SF-6D, and Quality of Well-Being Scale. The EQ-5D, HUI2 and HUI3, and SF-6D all have consistent evidence of construct validity and responsiveness in RA, but each has limitations. CONCLUSION:The EQ-5D and HUI3 have been the most extensively studied instruments and show validity and responsiveness for use in RA, but both instruments have limitations. The SF-6D is relatively new and appears to have potential for use in milder RA, but needs further evaluation. More longitudinal head-to-head evaluation of measures is needed across the spectrum of RA disease severity to further investigate their comparative properties, and to seek consensus on the best utility measure for use in economic evaluation.

J Rheumatol. 2008 Apr;35(4):592-602

Bansback N, Marra C, Tsuchiya A, Anis A, Guh D, Hammond T, Brazier J

OBJECTIVE: To estimate the relationship between preference-based measures, EuroQol (EQ-5D) and SF-6D, and the Health Assessment Questionnaire (HAQ) disability index (DI) in patients with rheumatoid arthritis (RA), and to characterize components that are predictors of health utility. METHODS: Patients with RA participating in 2 studies in the UK (n = 151) and Canada (n = 319) completed the HAQ, EQ-5D, and Short Form 36 (SF-36). The SF-36, a generic measure of quality of life, was converted into the preference-based SF-6D. From these results we developed models of the relationship between the HAQ and SF-6D and EQ-5D using various regression analyses. RESULTS: The optimal model developed for the EQ-5D entered levels for each item as independent variables (model 5). A root mean square error (RMSE) of 0.18 suggested relatively good predictive ability. For the SF-6D, RMSEs were lower (0.09), suggesting better predictions than for the EQ-5D, but models with more explanatory variables did not improve results (model 2 or 4 optimal). The models were able to predict actual SF-6D and EQ-5D across the range of the HAQ DI. CONCLUSION: Our approach enabled calculations of quality-adjusted life years from existing trials where only the HAQ was measured. All aspects of the HAQ may not be reflected in the preference-based measures, and this method is suboptimal to direct measurement of health state utility in clinical trials. Given this limitation, our approach provides an alternative for researchers who need health-state utility values, but had not included a preference-based measure in their clinical study because of resource constraints or a desire to limit patient burden.

Arthritis Rheum. 2007 Aug 15;57(6):963-71

Bansback N, Czoski-Murray C, Carlton J, Lewis G, Hughes L, Espallargues M, Brand C, Brazier J

BACKGROUND: There has been increasing interest in the use of measures of health related quality of life (HRQoL) and health state utility values in Age Related Macular Degeneration (ARMD). Visual acuity has been found to be an important determinant of such measures in previous studies. More recently, another measure of visual impairment, contrast sensitivity has received considerable attention. We designed a study to examine whether the contribution of contrast sensitivity in explaining HRQoL and health utilities over and above that of visual acuity. METHODS: 209 patients with unilateral or bilateral ARMD were recruited into a cross-sectional study of patients from a large teaching hospital. Patients underwent visual tests (near and distant visual acuity, contrast sensitivity) and completed a vision function questionnaire, the VF-14, HUI3, and time trade-off. RESULTS: Using multivariate regression analysis, the study revealed that contrast sensitivity remained a statistically significant predictor of all outcome measures even when visual acuity was included. This result was supported by the correlation coefficients between measures. CONCLUSIONS: The measurement of contrast sensitivity appears to be better related to a person’s HRQoL and health utility. Future studies should consider incorporating contrast sensitivity in addition to visual acuity. Studies, in particular economic evaluations, may underestimate the effect of treatment unless contrast sensitivity is considered.

Qual Life Res. 2007 Apr;16(3):533-43

Bansback NJ, Ara R, Barkham N, Brennan A, Fraser AD, Conway P, Reynolds A, Emery P

OBJECTIVES: Tumour necrosis factor (TNF) has been shown to improve the outcomes in patients with psoriatic arthritis (PsA). We estimate the long-term impact on health status of prescribing the TNF antagonist etanercept, and evaluate the cost-effectiveness in a health economic model. METHODS: The relationship between disability (Health Assessment Questionnaire) and health state utility was explored to estimate the quality-adjusted life years (QALYs) gained from the TNF antagonist etanercept. A model was then used to compare sequences of treatments for PsA after failure of two conventional disease modifying anti-rheumatic drugs (DMARDs). One arm commences on etanercept therapy and this is compared with a strategy commencing with combination therapy of methotrexate and ciclosporin and another commencing with leflunomide. Individual patient data from Phase III etanercept trials is used to populate the model supported by published evidence from extensive literature searches. By incorporating a life table specific for a PsA population, and using a number of evidence- and expert opinion-based assumptions for disease progression, the model was extended beyond the trial duration to a 10-yr time horizon. Cost offsets were produced by avoiding surgery through delayed progression; drug and monitoring costs were also modelled. RESULTS: Over the 10 yrs, modelled etanercept treatment gave 0.82 more QALYs when compared with combination therapy with methotrexate and ciclosporin, and 0.65 more QALYs in comparison with leflunomide. This equates to a central estimate for the cost per QALY of pound28 189 and pound28 189 for ciclosporin and leflunomide, respectively. Sensitivity analyses demonstrated this could vary by as much as +/-28%. CONCLUSIONS: With limited data currently available, the potential cost-effectiveness of etanercept in DMARD failures for adults with PsA appears encouraging. The result for other TNF antagonists will depend on how their relative efficacy and drug price compares with etanercept. A number of limitations are described and priorities for further research suggested.

Rheumatology (Oxford). 2006 Aug;45(8):1029-38

Espallargues M, Czoski-Murray CJ, Bansback NJ, Carlton J, Lewis GM, Hughes LA, Brand CS, Brazier JE

PURPOSE: To estimate health status utility values in patients with age-related macular degeneration (ARMD) associated with visual impairments, by using preference-based measures of health. METHOD: This was a cross-sectional study involving patients with unilateral or bilateral ARMD who attended a large teaching hospital. Patients underwent visual tests (near and distant visual acuity [VA] and contrast sensitivity [CS]) and completed health status questionnaires including the Index of Visual Function (VF)-14 and three preference-based measures (the Health Utilities Index Mark III [HUI-3], the EuroQoL Health Questionnaire [EQ-5D], and the Short Form 6D Health Status Questionnaire [SF-6D]) and the time tradeoff (TTO). The mean health status is presented for five groups, defined according to the VA in the better-seeing eye and for four CS groups. RESULTS: Two hundred nine patients were recruited with substantial loss of visual function as obtained by visual tests (mean decimal VA in the better-seeing eye: 0.2) and self-report (mean VF-14 score: 41.5). The mean (+/-SD) utilities were 0.34 +/- 0.28 for HUI-3, 0.66 +/- 0.14 for SF-6D, 0.72 +/- 0.22 for EQ-5D, and 0.64 +/- 0.31 for TTO. The HUI-3 had the highest correlation with VA and CS (0.40 and -0.34), followed by TTO (0.25 and -0.21). Across the VA and CS groups, only HUI3 and TTO had a significant linear trend (P < 0.05). In a regression model with CS and VA as explanatory variables, only the coefficient on CS was statistically significant. CONCLUSIONS: ARMD is associated with a substantial impact on patients’ health status, but this was not reflected in two of the generic preference-based measures used. The HUI-3 seems to be the instrument of choice for use in economic evaluations in which community data are needed. It may be more appropriate to base economic models on CS or some combination of CS and VA rather than on VA alone.

Invest Ophthalmol Vis Sci. 2005 Nov;46(11):4016-23

This paper is currently in submission and available as a discussion paper at the Health Economics and Decision Sciences website at the University of Sheffield

Download (PDF, 570.87KB)

A preliminary set of values for EQ-5D health states is available here. A paper is currently in submission describing the methods. A recent presentation I gave at the CADTH conference is available below. Please contact me before using any of the values, as this study is still ongoing and there may be some changes to final models.