Incorporating Mortality in Health Utility Measures

The creation of multiattribute health utility systems requires design choices that have profound effects on the utility model, many of which have been documented and studied in the literature. Here we describe one design choice that has, to the best of our knowledge, been unrecognized and therefore ignored. It can emerge in any multiattribute decision analysis in which one or more essential outcomes cannot be described in terms of the multiattribute space. In health applications, the state of being dead is such an outcome. When the remaining health is conceptualized as a multidimensional space, determining the utility of the state of being dead requires using the interval-scale properties of cardinal utility, combined with elicited utilities for the state of being dead and the all-worst state, to produce a utility function in which the state of being dead has a utility of 0 and full health has a utility of 1 (i.e., the quality-adjusted life-year scale). Although previously unrecognized, there are two approaches to accomplish that step, and they produce different results in almost all cases. As a corollary, the choice of approach determines the proportion of states rated as worse than dead by the system. For example, in the Health Utility Index 3 (HUI3), the method used classifies 78% of the 972,000 unique health states in the classification system as worse than dead, and that proportion increases to 85% when the HUI3 is recalculated using the alternative approach. Studies of populations with significant morbidity are the most likely to be sensitive to the design choice. Those who design utility measures should be aware that they are using a researcher degree of freedom when they decide how to scale the state of being dead.

Quality of Life in End-Stage Renal Disease

Survival in end-stage renal disease seems to depend more on age and associated diseases than on the form of treatment. Previously published comments on the quality of life experienced by patients treated by hemodialysis, peritoneal dialysis and transplantation were based on subjective estimates or on the use of unvalidated instruments. We have used a modified form of a “time trade-off” technique to estimate the patient's perception of the utility or worth of their ESRD health state. It was used successfully in 42 of 50 stable chronic hemodialysis patients. The correlation between paired observations six weeks apart (a test of temporal stability) was 0.80. Evidence for the validity of this instrument was obtained by comparing it with the physicians’ assessment of the patient's health state.

Health was scored on a scale from 0 for death to 1 for perfect health. The mean values for hospital-based hemodialysis, CAPD and transplantation were 0.57, 0.57 and 0.80 respectively. For each form of treatment the distribution of scores was wide suggesting that factors other than form of treatment are important determinants of the per ceived quality of life for these patients.

What were they thinking when providing preference measurements for generic health states? The evidence for HUI3

Background

Multi-attribute generic preference-based measures of health-related quality of life are used as comprehensive outcome measures. Typically preferences for health states defined by these systems are elicited from a representative sample of the general population. An important element in that elicitation process is the information that respondents were instructed to consider in providing their responses.

Methods

A random sample of community-dwelling respondents in Canada was surveyed in face-to-face interviews. Respondents provided preference scores for selected Health Utilities Index Mark 3 (HUI3) health states. Respondents also answered questions about the most important attributes and the importance of various impacts of the health states in providing their preference scores.

Results

Fifty per cent of respondents reported that they focussed on two, and 21% on three, attributes of the eight HUI3 attributes. Each of the eight attributes was identified as important; pain (49%), vision (37%), cognition (34%), emotion (28%), and ambulation (28%) were the most important. The null hypothesis that all of the attributes were equally important was rejected (p < 0.001). With respect to the impacts, 89% of respondents indicated that the ability to take care of oneself was quite or very important; similarly 76% reported the same for impact on family life, 69% for impact on the happiness of others, 61% for the impact on their ability to work, and 42% for the impact on their leisure activities. The null hypothesis that all of the impacts were equally important was rejected (p < 0.001).

Conclusions

In providing preference scores for HUI3 health states, respondents thoughtfully examined the implications of the health states for their ability to live, work, socialize, and function.

Guidelines for conducting pharmaceutical budget impact analyses for submission to public drug plans in Canada

Until now, there has been no standardized method of performing and presenting budget impact analyses (BIAs) in Canada. Nevertheless, most drug plan managers have been requiring this economic data to inform drug reimbursement decisions. This paper describes the process used to develop the Canadian BIA Guidelines; describes the Guidelines themselves, including the model template; and compares this guidance with other guidance on BIAs. The intended audience includes those who develop, submit or use BIA models, and drug plan managers who evaluate BIA submissions. The Patented Medicine Prices Review Board (PMPRB) initiated the development of the Canadian BIA Guidelines on behalf of the National Prescription Drug Utilisation Information System (NPDUIS). The findings and recommendations from a needs assessment with respect to BIA submissions were reviewed to inform guideline development. In addition, a literature review was performed to identify existing BIA guidance. The detailed guidance was developed on this basis, and with the input of the NPDUIS Advisory Committee, including drug plan managers from multiple provinces in Canada and a representative from the Canadian Agency for Drugs and Technologies in Health. A Microsoft Excel-based interactive model template was designed to support BIA model development. Input regarding the guidelines and model template was sought from each NPDUIS Advisory Committee member to ensure compatibility with existing drug plan needs. Decisions were made by consensus through multiple rounds of review and discussion. Finally, BIA guidance in Canadian provinces and other countries were compared on the basis of multiple criteria. The BIA guidelines consist of three major sections: Analytic Framework, Inputs and Data Sources, and Reporting Format. The Analytic Framework section contains a discussion of nine general issues surrounding BIAs (model design, analytic perspective, time horizon, target population, costing, scenarios to be compared, the characterisation of uncertainty, discounting, and validation methods). The Inputs and Data Sources section addresses methods for market size estimation, comparator selection, scenario forecasting and drug price estimation. The Reporting Format section describes methods for BIA reporting. The new Canadian BIA Guidelines represent a significant departure from the limited guidance that was previously available from some of the provinces, because they include specific details of the methods of performing BIAs. The Canadian BIA Guidelines differ from the Principles of Good Research Practice for BIAs developed by the International Society for Pharmacoeconomic and Outcomes Research (ISPOR), which provide more general guidance. The Canadian BIA Guidelines and template build upon existing guidance to address the specific requirements of each of the participating drug plans in Canada. Both have been endorsed by the NPDUIS Steering Committee and the PMPRB for the standardization of BIA submissions.

A model to estimate health utilities index mark 3 utility scores from WOMAC index scores in patients with osteoarthritis of the knee

OBJECTIVE

To develop a formula to translate Western Ontario and McMaster University Osteoarthritis Index (WOMAC) scores collected in clinical trials of patients with osteoarthritis (OA) into Health Utilities Index Mark 3 (HUI3) utility scores for application in economic evaluation.

METHODS

Data from a previously published open-label randomized controlled trial of appropriate care with hylan G-F 20 versus appropriate care without hylan G-F 20 in 255 outpatients with knee OA. We estimated linear regression models of HUI3 scores using various functions of WOMAC, demographics, and clinical variables. Out-of-sample predictive performance of the models was assessed using the mean absolute error and several other criteria.

RESULTS

The preferred formula included WOMAC pain, stiffness, function subscales, demographic variables; it accounted for almost 40% of the variation in the HUI3 utility scores. At the group level, absolute differences between predicted and actual overall HUI3 utility scores were < 0.001 and not statistically significantly different from zero.

CONCLUSION

A formula was derived from the WOMAC index to estimate overall utility scores based on the HUI3 for studies of patients with OA for whom utility has not been recorded. Researchers can estimate overall utility scores, compute quality-adjusted life-years, and perform cost-utility analyses within a defined range of certainty.

BLISS index for analyzing knee osteoarthritis trials data

OBJECTIVES

Different pain thresholds were investigated, using the WOMAC Pain Scale (WOMAC-P) to determine if they could differentiate between treatment groups (hylan G-F 20 vs. appropriate care) at low and very low levels of state attainment in patients with knee osteoarthritis (OA). A method, termed the BLISS (Bellamy et al. Low Intensity Symptom State-attainment) Index, for analyzing OA knee clinical trials data, was proposed.

STUDY DESIGN AND SETTING

Five analyses were performed: time to first BLISS day, BLISS days over 12 months, patients with a BLISS response at month 12, patients with a BLISS response at any time, and number of BLISS periods over 12 months. For each analysis, five levels of WOMAC-P were examined: <or=5 normalized units (NU), <or=10, <or=15, <or=20, and <or=25 (higher=more pain).

RESULTS

More patients in the hylan G-F 20 group achieved BLISS states in all five analyses. These differences were statistically significant for all pain threshold levels except <or=5 NU.

CONCLUSION

Five methods of measuring BLISS attainment using four prespecified threshold levels of pain were able to statistically discriminate between treatment groups. This method may potentially provide an approach, to defining which patients not only improve but also achieve a good state of health, at low and very low levels of pain intensity. BLISS-10 is a therapeutically attainable very low symptom state at which clinically important, statistically significant between-group differences are detectable, and therefore may provide a benchmark against which therapeutic interventions can be assessed. However, the value to patients, of this and other low and very low intensity pain states, requires further elaboration.

Measuring quality of life: The development and initial validation of the Patient-Reported Erectile Function Assessment instrument

Objectives: Erectile dysfunction (ED) is a complex condition, which is variously influenced by physical, emotional, societal, and relationship factors. ED has serious implications for the quality of life (QoL) enjoyed by an affected male and his partner. It is very important, therefore, to understand the impact of ED on the QoL of those affected by it. Our objective was to determine if the eight-question Patient Reported Erectile Function Assessment (PREFA) could act as an independent, comprehensive disease-specific instrument in the assessment of QoL as it is impacted by ED.

Methods: During the development and validation of the Erectile Function–Visual Analog Scale (EF-VAS) (14), a new ED-specific preference-based instrument, a series of questions were included at the beginning of the assessment that would act as a way to encourage respondents to focus on their own experience with ED. Upon analysis of the EF-VAS data, it became apparent that the eight-question “warm up” section might act as a stand-alone assessment. Accordingly, the eight questions were named PREFA, and a validation analysis was undertaken to determine their consistency, feasibility, reliability, validity, and responsiveness.

Results: The PREFA questionnaire was found to be feasible and simple to complete, reliable, and valid, with excellent responsiveness. Overall, the PREFA has demonstrated that it can perform as a stand alone, validated assessment of the impact of ED on QoL, assessing areas of QoL not previously captured in existing instruments.

Conclusions: The PREFA is suitable for use in clinical and research settings as a disease-specific QoL assessment tool.

Economic evaluation of controlled-release oxycodone vs oxycodone-acetaminophen for osteoarthritis pain of the hip or knee

OBJECTIVE

To examine, in routine practice, the effectiveness and cost-effectiveness of oxycodone (OxyContin) compared with standard therapy for osteoarthritis pain.

STUDY DESIGN

Open-label active-controlled randomized naturalistic 4-month study of oxycodone vs a combination of oxycodone-acetaminophen (Percocet).

METHODS

Outcomes and health resource utilization data were collected by telephone interview. Effectiveness was measured among 485 patients as the proportion having at least 20% improvement from baseline in the Western Ontario and McMaster Universities Osteoarthritis Index pain score. Quality-adjusted life-years (QALYs) were calculated from the Health Utilities Index 3 score. Cost-effectiveness was measured as cost per patient improved and the QALYs gained, using generic oxycodone-acetaminophen in the base case for the healthcare and societal perspectives. Uncertainty was evaluated using multiple 1-way sensitivity analyses and cost-effectiveness acceptability curves.

RESULTS

Improvement occurred in 62.2% of patients with oxycodone and in 45.9% of patients with oxycodone-acetaminophen (P < .001). After adjustment for baseline differences, 0.0105 QALYs were gained with oxycodone compared with oxycodone-acetaminophen (P = .17). The mean societal costs per patient during 4 months were 7379 US dollars and 7528 US dollars for oxycodone and oxycodone-acetaminophen, respectively (P = .33). Oxycodone was more effective and less costly than oxycodone-acetaminophen based on the societal perspective (including costs associated with time lost). Based on the healthcare perspective (excluding costs associated with time lost), the cost-effectiveness of oxycodone was 4883 US dollars per patient improved and 75,810 US dollars per QALY gained. The base-case results were robust.

CONCLUSIONS

From the societal perspective, oxycodone was more effective and less costly than oxycodone-acetaminophen. From the healthcare perspective, oxycodone (compared with generic oxycodone-acetaminophen) fell within the acceptable range of cost-effectiveness between 50,000 US dollars and 100,000 US dollars per QALY gained.

The Erectile Function Visual Analog Scale (EF-VAS): a disease-specific utility instrument for the assessment of erectile function

This article presents the responsiveness results of the Erectile Function Visual Analog Scale (EF-VAS) and reports, for the first time, utilities associated with erectile dysfunction (ED), as calculated by a disease-specific utility assessment. The EF-VAS is a new quality of life (QoL) instrument specific to ED that combines the strengths of the disease-specific approach to measuring QoL (greater disease relevance and responsiveness, with relevance to clinicians and patients) with those of preference-based assessments (generalizability and relevance to decision makers). The EF-VAS has demonstrated feasibility, reliability, and validity as reported in a recent publication.

METHODS

Standard instrument development methodology was utilized and the finalized content was integrated into a preference based scoring instrument comprised of two visual analogue scales (VAS). The EF-VAS was implemented in a clinical trial and data from the trial was subjected to validation analysis. Three methods were used to evaluate the responsiveness of the EF-VAS: Spearman correlations, effect size and standardized response means. VAS scores were converted to von Neumann-Morgenstern (vNM) utilities through a conversion curve.

RESULTS

The EF-VAS was established to be responsive to changes in disease state within and between patients with ED. The EF-VAS allowed the calculation of vNM utility values and a significant increase in utility was observed in the sildenafil group compared to placebo at study end.

CONCLUSION

The EF-VAS represents an important advance in the understanding of the impact of ED on patients' QoL and in providing a mechanism to allow the quantification of the health status that patients associate with ED. Based on its responsiveness, the EF-VAS will provide an important clinical tool to assess and contribute to the understanding of the impact of treatment for ED. The EF-VAS represents a major advance in the science of health-related quality of life (HRQol) assessment, as it is the first validated ED-specific utility assessment reported in the literature.

Utility measurement in healthcare: the things I never got to

The present article provides a brief historical background on the development of utility measurement and cost-utility analysis in healthcare. It then outlines a number of research ideas in this field that the author never got to. The first idea is extremely fundamental. Why is health economics the only application of economics that does not use the discipline of economics? And, more importantly, what discipline should it use? Research ideas are discussed to investigate precisely the underlying theory and axiom systems of both Paretian welfare economics and the decision-theoretical utility approach. Can the two approaches be integrated or modified in some appropriate way so that they better reflect the needs of the health field? The investigation is described both for the individual and societal levels. Constructing a 'Robinson Crusoe' society of only a few individuals with different health needs, preferences and willingness to pay is suggested as a method for gaining insight into the problem. The second idea concerns the interval property of utilities and, therefore, QALYs. It specifically concerns the important requirement that changes of equal magnitude anywhere on the utility scale, or alternatively on the QALY scale, should be equally desirable. Unfortunately, one of the original restrictions on utility theory states that such comparisons are not permitted by the theory. It is shown, in an important new finding, that while this restriction applies in a world of certainty, it does not in a world of uncertainty, such as healthcare. Further research is suggested to investigate this property under both certainty and uncertainty. Other research ideas that are described include: the development of a precise axiomatic basis for the time trade-off method; the investigation of chaining as a method of preference measurement with the standard gamble or time trade-off; the development and training of a representative panel of the general public to improve the completeness, coherence and consistency of measured preferences; and the investigation, using a model of a very small society, of the conflict between the patient perspective and the societal perspective regarding preferences. Finally, it is suggested that an important area of research, which the author never got to, would be to work closely with specific decision makers on specific decision problems, to help them formulate the problem, provide useful analyses, and to publish these as case studies to give the field a better understanding of the problems and the needs of decision makers.

The effectiveness of hylan G-F 20 in patients with knee osteoarthritis: an application of two sets of response criteria developed by the OARSI and one set developed by OMERACT-OARSI

OBJECTIVE

Secondary analyses of a previously conducted 1-year randomized controlled trial were performed to assess the application of responder criteria in patients with knee osteoarthritis (OA) using different sets of responder criteria developed by the Osteoarthritis Research Society International (OARSI) (Propositions A and B) for intra-articular drugs and Outcome Measures in Arthritis Clinical Trials (OMERACT)-OARSI (Proposition D).

METHODS

Two hundred fifty-five patients with knee OA were randomized to "appropriate care with hylan G-F 20" (AC+H) or "appropriate care without hylan G-F 20" (AC). A patient was defined as a responder at month 12 based on change in Western Ontario and McMaster Universities Osteoarthritis Index pain and function (0-100 normalized scale) and patient global assessment of OA in the study knee (at least one-category improvement in very poor, poor, fair, good and very good). All propositions incorporate both minimum relative and absolute changes.

RESULTS

Results demonstrated that statistically significant differences in responders between treatment groups, in favor of hylan G-F 20, were detected for Proposition A (AC+H=53.5%, AC=25.2%), Proposition B (AC+H=56.7%, AC=32.3%) and Proposition D (AC+H=66.9%, AC=42.5%). The highest effectiveness in both treatment groups was observed with Proposition D, whereas Proposition A resulted in the lowest effectiveness in both treatment groups. The treatment group differences always exceeded the required 20% minimum clinically important difference between groups established a priori, and were 28.3%, 24.4% and 24.4% for Propositions A, B and D, respectively.

CONCLUSION

This analysis provides evidence for the capacity of OARSI and OMERACT-OARSI responder criteria to detect clinically important statistically detectable differences between treatment groups.

Effectiveness and safety of repeat courses of hylan G-F 20 in patients with knee osteoarthritis

OBJECTIVE

To compare the effectiveness and safety of repeat treatment with hylan G-F 20 based on data from a randomized, controlled trial [Raynauld JP, Torrance GW, Band PA, Goldsmith CH, Tugwell P, Walker V, et al. A prospective, randomized, pragmatic, health outcomes trial evaluating the incorporation of hylan G-F 20 into the treatment paradigm for patients with knee osteoarthritis (Part 1 of 2): clinical results. Osteoarthritis Cartilage 2002;10:506-17]. The hypotheses tested were whether the single-course and repeat-course subgroups would be superior to appropriate care and not different from each other.

METHOD

A total of 255 patients with knee osteoarthritis were randomized to "appropriate care with hylan G-F 20" or "appropriate care without hylan G-F 20". The hylan G-F 20 group was partitioned into two subgroups: (1) patients who received a single course of hylan G-F 20; and (2) patients who received two or more courses of hylan G-F 20.

RESULTS

For the primary effectiveness measure, change in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain score as a percent of baseline, the single-course subgroup improved by 41%, the repeat-course subgroup by 35%, and the appropriate care group by 14%. Both subgroups improved significantly more than the appropriate care group (P<0.05), and were not statistically significantly different from each other (70% power to detect a 20% difference). Secondary effectiveness measures showed similar results. In the repeat-course subgroup, no statistically significant differences were found in the number of local adverse events, the number of patients with local adverse events, or arthrocentesis rates between the first and repeat courses of treatment.

CONCLUSIONS

Although the study was neither designed nor powered to examine repeat treatment, this a posteriori analysis provides support for a favorable effectiveness and safety profile of hylan G-F 20 in repeat course patients.

Evaluation of WOMAC 20, 50, 70 response criteria in patients treated with hylan G-F 20 for knee osteoarthritis

OBJECTIVE

A secondary analysis of a previously conducted one year randomised controlled trial to evaluate the capacity of responder criteria based on the WOMAC index to detect between treatment group differences.

METHODS

255 patients with knee osteoarthritis were randomised to "appropriate care with hylan G-F 20" (AC+H) or "appropriate care without hylan G-F 20" (AC). In the original analysis, two definitions of patient response from baseline to month 12 were used: (1) at least a 20% reduction in WOMAC pain score (WOMAC 20P); (2) at least a 20% reduction in WOMAC pain score and at least a 20% reduction in either WOMAC function or stiffness score (WOMAC 20PFS). For this analysis, a responder was identified using 50% and 70% minimum clinically important response levels to investigate how increasing response affects the ability to detect treatment group differences.

RESULTS

The hylan G-F 20 group had numerically more responders using all patient responder criteria. Increasing the response level from 20% to 50% detected similar differences between treatment groups (25% to 29%). Increasing the response level to 70% reduced the differences between treatment groups (11% to 12%) to a point where the differences were not significant after Bonferroni adjustment.

CONCLUSIONS

These results provide evidence for incorporating response levels (WOMAC 50) in clinical trials. While differences at the highest threshold (WOMAC 70) were not statistically detectable, an appropriately powered study may be capable of detecting differences even at this very high level of improvement.

Development and initial validation of a new preference-based disease-specific health-related quality of life instrument for erectile function

Health-related quality of life instruments may be generic or specific. In general, only generic instruments use preference-based scoring. We report on a novel approach to combine in one instrument the strengths of the specific approach, greater disease relevance and responsiveness, with those of preference-based scoring, generalizability through utilities.

OBJECTIVES

The primary objective was to develop a self-administered, preference-based instrument capable of measuring utilities in the disease-specific context of erectile dysfunction (ED).

METHODS

Content derivation/validation began with a literature review. Eight attributes (domains) were selected to provide clinical experts structure for focus group discussion. Four levels describing a continuum of dysfunction-function were defined for each domain. Each domain, including functional levels, was reviewed and modified until consensus was achieved regarding content. This content was then integrated into a preference based scoring instrument using two visual analogue scales (VAS) with which patients rated three 'marker' health states (representing mild, moderate and severe ED), their self-state and a previously validated external marker state. The instrument was pilot tested, and implemented in a clinical trial. Initial validation analyses have been performed.

RESULTS

A self-administered, preference-based, VAS instrument was developed for use in the ED population, and the instrument was feasible to complete, was reliable beyond the threshold of acceptability established a priori and demonstrated good validity. Evidence of these properties accumulates over time and this study begins that process with this instrument. Responsiveness is being assessed in the context of a clinical trial.

Improvement in health utility among patients with rheumatoid arthritis treated with adalimumab (a human anti-TNF monoclonal antibody) plus methotrexate

OBJECTIVES

To compare health-related quality of life (HRQoL), as measured by health utility, in patients with rheumatoid arthritis (RA) treated with adalimumab (a human anti-tumour necrosis factor (anti-TNF) monoclonal antibody) plus methotrexate or placebo plus methotrexate.

METHODS

HRQoL data were obtained in two randomized, double-blind, placebo-controlled, multidose clinical trials conducted in the United States and Canada. The Health Utilities Index Mark 3 (HUI3) was administered in both studies at baseline, at the end of the study and at two time points in between. Patients' HUI3 scores were compared with population norm scores. Change in HUI3 was defined as the end-of-study score minus the baseline score. Utility gained throughout the study was measured by area under the utility curve and expressed as quality-adjusted life years (QALYs). Statistical testing adjusted for confounders and used the Dunnett test to account for multiple comparisons.

RESULTS

Patients' utility scores at baseline were low (range of treatment group means 0.38-0.44) compared with population norms (0.88). HUI3 mean changes from baseline scores for adalimumab-treated patients were 0.22 and 0.21 in the two trials, whereas placebo patients' changes were 0.04 and 0.07. The rate of QALYs gained per year in the treatment group compared with the placebo group were 0.145 in the ARMADA trial and 0.104 in the DE019 trial. All gains were clinically important and statistically significant.

CONCLUSIONS

Treatment with adalimumab plus methotrexate provides clinically important and statistically significant improvements in HRQoL as measured by health utility in patients with RA. This translates into measurable and important gains in QALYs.

Switching therapy in health economics trials: confronting the confusion

Should patients in a randomized, pragmatic health economics trial be allowed to switch therapy in mid-trial to that provided in the other arm? Specifically, should patients in the treatment arm (T) be allowed to switch to the therapy of the comparator arm (C) if they need a change of therapy--that is, should TC switches be allowed? Also, should patients in the comparator arm be allowed to switch to the therapy of the treatment arm if they need changes of therapy--should CT switches be allowed? This is a nontrivial issue in study design that has been debated in the clinical trials literature and is currently being handled inconsistently in the health economics literature. In this article, the authors argue that TC switches should always be allowed and that CT switches should be allowed or not depending on the economic question. They further argue that the most common economic question is one that would lead to CT switches not being allowed.

Health-related quality-of-life assessment of prenatal diagnosis: chorionic villi sampling and amniocentesis

This study assesses the health-related quality-of-life (HRQL) effects of chorionic villi sampling (CVS) and genetic amniocentesis (GA), including both process and outcomes of prenatal diagnosis. The HRQL of 126 women participating in a randomized controlled clinical trial of CVS versus GA in Toronto and Hamilton, Ontario, was assessed in four interviews at weeks 8, 13, 18, and 22 of pregnancy. Statistical analyses included analysis of variance, repeated measures analysis of covariance, chi-square, Fisher's exact test, Student's t-tests, and paired t-tests. Utility scores for patients undergoing CVS exceeded those for GA patients at week 18 (p = 0.04). Utility scores for hypothetical health states did not differ significantly by trial arm. CVS results in slightly improved HRQL during prenatal diagnosis. This advantage needs to be weighed against the high disutility patients attach to infrequent outcomes associated with pregnancy losses, equivocal diagnoses, and diagnostic inaccuracy.

Willingness to pay for what? A note on alternative definitions of health care program benefits for contingent valuation studies

The authors examine a number of ways in which willingness to pay (WTP) can be defined for measurement and use in a cost-benefit analysis (CBA) of a collectively funded health care program. They show how ambiguous specification of the program consequences that respondents should consider in their WTP responses can lead to problems of double counting or zero countingin a subsequent CBA. An example is whether the value of lost time from work because of poor health should be included by a CBA analyst (e.g., valued at the wage rate) as a separate cost item or whether this has already been monetized and included in respondents' WTP data. The authors highlight how differences in assumed or actual institutional structures are often ignored in measures of WTP and the consequences of this for the interpretation of WTP data.

A prospective, randomized, pragmatic, health outcomes trial evaluating the incorporation of hylan G-F 20 into the treatment paradigm for patients with knee osteoarthritis (Part 2 of 2): economic results

OBJECTIVE

Viscosupplementation with hylan G-F 20 has recently become registered for treatment of patients with osteoarthritis (OA) of the knee in most parts of the world. The cost effectiveness and cost utility of this new therapeutic modality were determined as part of a Canadian prospective, randomized, 1-year, open-label, multicentered trial.

DESIGN

A total of 255 patients were randomized to 'Appropriate care with hylan G-F 20' (AC+H) or 'Appropriate care without hylan G-F 20' (AC). Costs (1999 Canadian dollars) were collected from the societal viewpoint and included all costs related to OA of the knee and OA in all joints. Patients completed a number of outcomes questionnaires including the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the Health Utilities Index Mark 3 (HUI3). Data were collected at clinic visits (baseline, 12 months) and by telephone (1, 2, 4, 6, 8, 10, and 12 months).

RESULTS

The AC+H group over the year had higher costs ($2125-$1415=$710, P< 0.05), more patients improved (69%-40%=29%,P =0.0001), greater increases in HUI3 (0.13-0.03=0.10, P< 0.0001) and increased quality-adjusted life years (QALYs) (0.071, P< 0.05). The incremental cost-effectiveness ratio was $2505/patient improved. The incremental cost-utility ratio was $10000/QALY gained. Sensitivity analyses and a second cost perspective gave similar results.

CONCLUSION

The cost-utility ratio is below the suggested Canadian adoption threshold. The results provide strong evidence for adoption of treatment with hylan G-F 20 in the patients and settings studied in the trial.

A prospective, randomized, pragmatic, health outcomes trial evaluating the incorporation of hylan G-F 20 into the treatment paradigm for patients with knee osteoarthritis (Part 1 of 2): clinical results

OBJECTIVE

First, to assess the clinical effectiveness of hylan G-F 20 in an appropriate care treatment regimen (as defined by the American College of Rheumatology (ACR) 1995 guidelines) as measured by validated disease-specific outcomes and health-related quality of life endpoints for patients with osteoarthritis (OA) of the knee. Second, to utilize the measures of effectiveness and costs in an economic evaluation (see accompanying manuscript).

DESIGN

A total of 255 patients with OA of the knee were enrolled by rheumatologists or orthopedic surgeons into a prospective, randomized, open-label, 1-year, multi-centred trial, conducted in Canada. Patients were randomized to 'Appropriate care with hylan G-F 20' (AC+H) or 'Appropriate care without hylan G-F 20' (AC). Data were collected at clinic visits (baseline, 12 months) and by telephone (1, 2, 4, 6, 8, 10, and 12 months).

RESULTS

The AC+H group was superior to the AC group for all primary (% reduction in mean Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain scale: 38% vs 13%,P =0.0001) and secondary effectiveness outcome measures. These differences were all statistically significant and exceeded the 20% difference between groups set a priori by the investigators as the minimum clinically important difference. Health-related quality of life improvements in the AC+H group were statistically superior for the WOMAC pain, stiffness and physical function (all P< 0.0001), the SF-36 aggregate physical component (P< 0.0001) and the Health Utilities Index Mark 3 (HUI3) overall health utility score (P< 0.0001). Safety (adverse events and patient global assessments of side effects) differences favoured the AC+H group.

CONCLUSION

The data presented here indicate that the provision to patients with knee OA of viscosupplementation with hylan G-F 20 within an appropriate care treatment regimen provides benefits in the knee, overall health and health related quality of life at reduced levels of co-therapy and systemic adverse reactions.

The burden of illness of severe nausea and vomiting of pregnancy in the United States

OBJECTIVE

Our purpose was to determine the extent to which nausea and vomiting of pregnancy affects a woman's quality of life (QOL), ability to function, and health care resource use.

STUDY DESIGN

We conducted an observational, multicenter, prospective cohort study by gathering data on the symptoms, QOL, and health care resource use from women who have nausea and vomiting of pregnancy.

RESULTS

All 8 domains of health measured by the Short Form-36 QOL survey were limited by patient symptoms. This limitation manifested itself as patient-time loss from work and other normal activities, unpaid caregiver-time loss from work, and use of health care resources (eg, hospitalization). All types of time loss were correlated to severity of symptoms.

CONCLUSIONS

Nausea and vomiting of pregnancy can severely reduce a woman's QOL and ability to function. The degree of limitation is associated with the severity of symptoms.

Multiattribute and single-attribute utility functions for the health utilities index mark 3 system

BACKGROUND

The Health Utilities Index Mark 3 (HUI3) is a generic multiattribute preference-based measure of health status and health-related quality of life that is widely used as an outcome measure in clinical studies, in population health surveys, in the estimation of quality-adjusted life years, and in economic evaluations. HUI3 consists of eight attributes (or dimensions) of health status: vision, hearing, speech, ambulation, dexterity, emotion, cognition, and pain with 5 or 6 levels per attribute, varying from highly impaired to normal.

OBJECTIVES

The objectives are to present a multiattribute utility function and eight single-attribute utility functions for the HUI3 system based on community preferences.

STUDY DESIGN

Two preference surveys were conducted. One, the modeling survey, collected preference scores for the estimation of the utility functions. The other, the direct survey, provided independent scores to assess the predictive validity of the utility functions.

MEASURES

Preference measures included value scores obtained on the Feeling Thermometer and standard gamble utility scores obtained using the Chance Board.

RESPONDENTS

A random sample of the general population (> or =16 years of age) in Hamilton, Ontario, Canada.

RESULTS

Estimates were obtained for eight single-attribute utility functions and an overall multiattribute utility function. The intraclass correlation coefficient between directly measured utility scores and scores generated by the multiattribute function for 73 health states was 0.88.

CONCLUSIONS

The HUI3 scoring function has strong theoretical and empirical foundations. It performs well in predicting directly measured scores. The HUI3 system provides a practical way to obtain utility scores based on community preferences.

The development and use of quality-of-life measures to evaluate health outcomes in Japan

Health outcomes measurement forms a key aspect of data collection for economic evaluations. Quality-of-life instruments provide a measure of patient-focused health outcomes. Such instruments, including the EuroQoL questionnaire (EQ-5D) and the Short-form 36 (SF-36) are already being used in Japan. Utility instruments provide an alternative way of measuring quality of life for use in economic analyses. At present, utility measurement in Japan is limited to scores derived from the EQ-5D. Although the SF-36 was designed as a health profile measure, it has since been tested for use as a utility measure, with valuation of items now being undertaken in Japan. Utility measurement in Japan is likely to be advanced further by the validation and introduction of the Health Utilities Index. The experiences of other countries provide the opportunity for Japan to learn how utility scores may be used in cost-utility analyses of healthcare technologies.

Why has the use of health economic evaluation in Japan lagged behind that in other developed countries?

The aging population and the increasing availability of new medical technologies, particularly pharmaceuticals, have led to growing pressure on governments worldwide to contain healthcare costs. Increasingly, economic evaluation is used to aid decisions on the reimbursement and formulary access of drugs, and pharmaceutical companies are often required to demonstrate the cost effectiveness of their products. Canada and the UK are examples of countries that have successfully incorporated mandatory requirements for economic evaluations into the decision-making process in healthcare. Japan faces cost-containment issues for its health and welfare system similar to those seen elsewhere in the world. Despite this, economic assessments are not currently used in the allocation of drug budgets. Reasons why economic evaluations for healthcare have not yet been used routinely in Japan include governmental approaches to healthcare cost containment, the pricing of pharmaceuticals, the organisation of the healthcare system, attitudes of the medical profession, and limited knowledge and expertise. However, small but encouraging steps are now being taken towards the introduction of economic evaluations in Japanese medicine.

Information barriers to the implementation of economic evaluations in Japan

With increasing cost-containment pressures within healthcare systems worldwide, economic evaluations of medical technologies, particularly pharmaceuticals, are used to aid the allocation of expenditure and resources. Facing similar pressures, Japan will probably also introduce economic evaluation of health technology. However, the structure of the healthcare system in Japan does not lend itself naturally to the collection of the epidemiological and cost data required for economic evaluations in medicine. In addition, there are no formal methodological guidelines in place for these analyses. To overcome these information barriers in Japan, progress may be aided by the adoption of approaches used in other countries for data collection and guideline development.

Cost-effectiveness of Apligraf in the treatment of venous leg ulcers

Venous ulcers are the most common chronic wounds of the lower leg. Skin substitutes recently have been introduced to stimulate nonhealing wounds. To conduct an incremental cost-effectiveness analysis, a model was developed to compare the four-layer bandage system, with and without one application of skin substitute, for the outpatient treatment of venous leg ulcers. The model estimated the costs and consequences of treatment with and without the skin substitute application. Two analytic horizons were explored: 3 months and 6 months. Determined by seven physicians, data and assumptions for the 3-month model were based on information from a clinical trial, published studies, and clinical experience. Data for the 6-month model were extrapolated from the shorter model. The model results indicate that over 3 months, the use of the skin substitute provided a benefit of 22 ulcer days averted per patient at an incremental cost of $304 (societal). The incremental cost-effectiveness ratio was $14 per ulcer day averted. Over 6 months, the incremental cost-effectiveness ratio was less than $5 per ulcer-day averted. The skin substitute plus a four-layer bandage was more costly and more effective than the four-layer bandage alone. The skin substitute is increasingly cost-effective over a longer analytic horizon and in a subgroup of patients with ulcers of long duration (greater than 1-year duration at baseline). The results come from a model that is based on a series of estimates and assumptions, and accordingly, confirmation of this finding in a prospective study is encouraged.

Visual analog scales: do they have a role in the measurement of preferences for health states?

Visual analog scales (VASs) have long been used as a method of measuring preferences for health outcomes. They are easy and inexpensive to implement, can be administered quickly, and lend themselves to self-completion. Over time, however, disturbing questions have emerged concerning the validity of the VAS approach. This article reviews briefly the history, theory, practice, problems, and advantages of VASs; presents some suggestions to improve the validity of VASs; and recommends a limited but useful role for VASs in the process of measuring preferences for health states.

The Health Utilities Index (HUI) system for assessing health-related quality of life in clinical studies

This paper reviews the Health Utilities Index (HUI) systems as means to describe health status and obtain utility scores reflecting health-related quality of life (HRQoL). The HUI Mark 2 (HUI2) and Mark 3 (HUI3) classification and scoring systems are described. The methods used to estimate multiattribute utility functions for HUI2 and HUI3 are reviewed. The use of HUI in clinical studies for a wide variety of conditions in a large number of countries is illustrated. HUI provides a comprehensive description of the health status of subjects in clinical studies. HUI has been shown to be a reliable, responsive and valid measure in a wide variety of clinical studies. Utility scores provide an overall assessment of the HRQoL of patients. Utility scores are also useful in cost-utility analyses and related studies. General population norm data are available. The widespread use of HUI facilitates the interpretation of results and permits comparisons. HUI is a useful tool for assessing health status and HRQoL in clinical studies.

What should be reported in a methods section on utility assessment?

BACKGROUND

The measurement of utilities, or preferences, for health states may be affected by the technique used. Unfortunately, in papers reporting utilities, it is often difficult to infer how the utility measurement was carried out.

PURPOSE

To present a list of components that, when described, provide sufficient detail of the utility assessment.

METHODS

An initial list was prepared by one of the authors. A panel of 8 experts was formed to add additional components. The components were drawn from 6 clusters that focus on the design of the study, the administration procedure, the health state descriptions, the description of the utility assessment method, the description of the indifference procedure, and the use of visual aids or software programs. The list was updated and redistributed among a total of 14 experts, and the components were judged for their importance of being mentioned in a Methods section.

RESULTS

More than 40 components were generated. Ten components were identified as necessary to include even in an article not focusing on utility measurement: how utility questions were administered, how health states were described, which utility assessment method(s) was used, the response and completion rates, specification of the duration of the health states, which software program (if any) was used, the description of the worst health state (lower anchor of the scale), whether a matching or choice indifference search procedure was used, when the assessment was conducted relative to treatment, and which (if any) visual aids were used. The interjudge reliability was satisfactory (Cronbach's alpha = 0.85).

DISCUSSION

The list of components important for utility papers may be used in various ways, for instance, as a checklist while writing, reviewing, or reading a Methods section or while designing experiments. Guidelines are provided for a few components.

Cost-utility analysis in depression: the McSad utility measure for depression health states

Cost-utility analysis, used increasingly over the past decade to analyze costs and effects in treating physical diseases, has received little attention in psychiatry. This article briefly introduces the concepts and methods of utility measurement and illustrates it using depression as an example. The authors describe the McSad health state classification system for depression, a direct utility measure for depression, and report results of an application of McSad among 105 patients who had a recent history of depression. Utility measures express patient preferences for specific health states on a scale ranging from 0, representing death, to 1, representing perfect health. These scores provide the weights used to calculate the number of quality-adjusted life-years gained by an intervention or service. McSad allows a patient's depression health state to be classified according to level of functioning in six dimensions of depression and to be compared with other hypothetical depression health states in order to produce utility scores indicating the patient's relative preferences.

Cost and health related quality of life consequences of multiple sclerosis

OBJECTIVES

To (i) quantify the cost of multiple sclerosis (MS) to the Canadian health care system and society; (ii) measure health utility in MS patients, and (iii) examine the influence of disability on patient utility and health care costs.

MATERIALS AND METHODS

A comprehensive patient survey and chart review of relapsing MS patients in remission, relapse and recalling a relapse.

RESULTS

Annual remission costs increased with EDSS level ($7596 at EDSS 1, $33 206 at EDSS 6). At all EDSS levels the largest costs were due to inability to work, which increased with EDSS. The average relapse cost for all EDSS levels was $1367. An inverse correlation was found between EDSS level and patient utility for patients in remission and relapse. The decrease in remission health utility from EDSS 1 to 6 was 0.24, which is 25% greater than the difference in health status between an average 25 and 85 year-old.

CONCLUSIONS

This study demonstrates that MS produces substantial health care costs and reductions in patient quality of life and ability to work, losses that can be avoided or delayed if disease progression is slowed. These data provide health-care decision-makers with the opportunity to consider the full impact of MS when faced with budget allocation decisions.

Development and testing of a utility measure for major, unipolar depression (McSad)

OBJECTIVE

To develop and test a direct utility measure (McSad) for major, unipolar depression.

METHODS

A depression specific, multi-attribute health state classification system was created; clinical validity was evaluated by experts using specially designed structured exercises; a cross-sectional survey was conducted to obtain directly measured utilities for depression health states.

SETTING

Tertiary care, university medical centre.

PARTICIPANTS

Three psychiatrists, 3 psychiatric nurses and 3 social workers assessed depression health state clinical validity. Survey participants were referred by psychiatrists and consisted of 105 outpatients, currently in remission with at least one episode of major, unipolar depression in the past two years.

SURVEY RESULTS

Respondent self-health state utility (mean and 95% confidence interval (CI)) was 0.79 (0.74-0.83). Utilities for hypothetical, untreated depression health states were: mild depression, 0.59 (0.55-0.62); moderate depression, 0.32 (0.29-0.34); severe depression, 0.04 (0.01-0.07). Fifty-six percent of respondents rated severe depression worse than being dead. Utilities for the hypothetical health states were not correlated with self-health utility. The intra-class correlation coefficient (ICC) was satisfactory for 13 of the 14 health states assessed.

CONCLUSIONS

McSad was feasible and acceptable in patients with a history of major unipolar depression. The utilities for mild, moderate and severe untreated depression show the low health-related quality of life associated with depression. Initial assessments of test-retest reliability and validity yielded satisfactory results but further studies are needed to extend our knowledge of the measurement properties of McSad.

The revised Canadian Guidelines for the Economic Evaluation of Pharmaceuticals

The first edition of the Guidelines for Economic Evaluation of Pharmaceuticals: Canada was published in November 1994. At that time, the Canadian Coordinating Office for Health Technology Assessment (CCOHTA) was assigned the task of maintaining and regularly updating the Canadian Guidelines. Since their introduction, a great deal of experience has been gained with the practical application of the guidelines. Their role has also evolved over time, from being a framework for pharmacoeconomic research to the point where a wide variety of decision-makers use economic evaluations based on the principles set out in the guidelines as a means of facilitating their formulary decisions. In addition, methodologies in certain areas (and the body of related research literature in general) have developed considerably over time. Given these changes in the science and the experience gained, CCOHTA convened a multi-disciplinary committee to address the need for revisions to the guidelines. The underlying principles of the review process were to keep the guidance nature of the document, to focus on the needs of 'doers' (so as to meet the information needs of 'users') and to provide information and advice in areas of controversy, with sound direction in areas of general agreement. The purpose of this review is three-fold: (i) to outline the process which lead to the revision of the Canadian Guidelines; (ii) to describe the major changes made to the second edition of this document; and (iii) to consider the 'next steps' as they relate to the impact of such guidelines and the measurement of outcomes related to economic assessments of pharmaceuticals in general.

Cost effectiveness of multi-therapy treatment strategies in the prevention of vertebral fractures in postmenopausal women with osteoporosis

OBJECTIVE

The aim of this study was to evaluate the cost effectiveness of multi-therapy treatment strategies in the prevention of vertebral fractures in postmenopausal women with osteoporosis.

DESIGN

A retrospective, incremental cost-effectiveness analysis was conducted from a societal perspective. It compared 9 treatment strategies over 3 years and incorporated the willingness of patients to initiate and continue each therapy.

MAIN OUTCOME MEASURES AND RESULTS

Four nondominated strategies formed the efficient frontier in the following order: (i) calcium-->no therapy; (ii) ovarian hormone therapy (OHT)-->calcium-->no therapy [166 Canadian dollars ($Can)]; (iii) OHT-->etidronate-->calcium-->no therapy ($Can2331); and (iv) OHT-->alendronate-->calcium-->no therapy ($Can40,965). The figures in parentheses are the incremental costs per vertebral fracture averted to move to that strategy from the previous strategy for patients who had undergone a hysterectomy.

CONCLUSIONS

We identified 4 efficient multi-therapy strategies for the treatment of vertebral osteoporosis in postmenopausal women, 2 of which were consistent with the practice guidelines of the Osteoporosis Society of Canada. Decision-makers may select from among these efficient strategies on the basis of incremental cost effectiveness.

Assessing the value of a new pharmaceutical. A feasibility study of contingent valuation in managed care

OBJECTIVES

The authors assessed the feasibility and construct validity of the contingent valuation method for measuring the monetary value to healthy enrollees in a health maintenance organization of a new drug, filgrastim, as prophylaxis against febrile neutropenia after chemotherapy treatment for cancer.

METHODS

A random sample of 220 enrollees from a closed-panel staff-model health maintenance organization who did not have cancer were interviewed. Chemotherapy, febrile neutropenia and filgrastim were described by video and decision board. Questions were asked in two different scenarios: (1) User-based: Assuming they were at the point of consumption and about to receive chemotherapy, what is the maximum they would be willing to pay to receive filgrastim? and (2) Insurance-based: Given they were at risk of cancer in the future, what is the maximum they would be willing to pay in additional monthly insurance premiums to add filgrastim to the plan? In a second insurance scenario where respondents were told that filgrastim was covered, what is the minimum reduction in premium that persons were willing to accept to relinquish coverage of the drug? A 2 x 2 factorial design was used to contrast two bidding algorithms to test for starting point bias and two 5-yearly prior risks of cancer, 1/200 versus 1/100. Main effects were tested by ANCOVA controlling for age, sex, health, and income.

RESULTS

Demographics of experimental cells were similar. No evidence was found of significant starting point bias. For user-based questions, as expected, willingness-to-pay increases with febrile neutropenia risk reduction, but at a declining marginal rate. Despite careful presentation of information to respondents, willingness-to-pay for insurance was higher in the lower prior risk group. Consistent with previous contingent valuation studies, the authors of the present study found evidence that willingness-to-accept exceeds willingness-to-pay for coverage of the same benefit.

CONCLUSIONS

An insurance-based contingent valuation study is feasible in a health maintenance organization. Construct validation evidence was encouraging, with the exception of the test for prior risk of cancer; however, this was a between-person contrast and may have been confounded by other factors.

Health state utilities in knee replacement surgery: the development and evaluation of McKnee

OBJECTIVE

1. To develop McKnee, a classification system and direct utility measure for health states associated with knee replacement (KR) surgery. 2. To apply McKnee in a before-after study of KR surgery to: (i) gain experience with McKnee in an elderly population; (ii) confirm the practicality and usefulness of the McKnee system; (iii) assess self-health utility one week before and 3 mo after surgery; (iv) evaluate the stability of 3 clinical marker health states describing mild, moderate, and severe knee disability; (v) compare self-health utility scores with Short Form 36 (SF-36).

METHODS

1. Instrument development: The McKnee modified Health Utilities Index was developed and used to describe self-health and clinical marker health states: the clinical validity of the clinical marker states was evaluated by 5 clinicians involved in the care off KR patients. 2. Instrument evaluation: McKnee and the SF-36 were administered to 48 patients with osteoarthritis one week before and 3 mo after KR surgery.

RESULTS

Before-after study: McKnee was feasible and acceptable in the older patient group studied (mean age in years, SD: 69.9, 8.6). No change in self-health utility (mean, SD) was observed at 3 mo postsurgery: before -0.78, 0.17; after -0.78, 0.21. On the SF-36, only the change scores for pain and health transition were statistically significant. Utilities (mean, SD) for the clinical marker health states were: mild -0.80, 0.20; moderate -0.55, 0.28; and severe -0.48, 0.31. The clinical marker mean utility scores were stable between the baseline and 3 mo assessment, but the intraclass correlation coefficients for individual scores were low.

CONCLUSION

McKnee provides a preference based measure of health related quality of life that can be used to obtain and interpret clinically the knee disability utility scorers needed for cost-utility studies and medical decision-making models about KR surgery. The McKnee system provides a practical and useful method for classifying knee disability health states and obtaining direct measurements of utility scores for selected health states.

Preferences for health outcomes and cost-utility analysis

Economic evaluation of health programs consists of the comparative analysis of alternative courses of action in terms of both costs and consequences. The five analytic techniques are cost-consequence analysis, cost-minimization analysis, cost-effectiveness analysis, cost-utility analysis, and cost-benefit analysis. Although all techniques have the same objective of informing decision making in the health programs, they come from different theoretic backgrounds and relate differently to the discipline of economics. Cost-utility analysis formally incorporates the measured preferences of individuals for the health outcome consequences of the alternative programs. The individuals may be actual patients who are experiencing or have experienced the outcomes, or they may be a representative sample of the community, many of whom may someday face the outcomes. The health outcomes, at the most general level, consist of changes in the quantity and quality of life; that is, changes in mortality and morbidity. Changes in quantity of life are measured with mortality; changes in quality of life are measured with health-related quality-of-life instruments. Utilities represent a particular approach to the measurement of health-related quality of life that is founded on a well specified theory and provides an interval scale metric. Changes in quantity of life, as measured in years, can be combined with changes in quality of life, as measured in utilities, to determine the number of quality-adjusted life years gained by a particular health program. This can be compared with the incremental cost of the program to determine the cost per quality-adjusted life-year gained. Utilities may be measured directly on patients or other respondents by means of techniques such as visual analog scaling, standard gamble, or time trade-off. Utilities may be determined indirectly by means of a preference-weighted multi-attribute health status classification system such as the health utilities index. The health utilities index is actually a complete system for use in studies. It consists of questionnaires in various formats and languages, scoring manuals, and descriptive health status classification systems. The health utilities index is useful in clinical studies and in population health surveys, as well as in cost-utility analyses.

Guidelines for pharmacoeconomic studies. Recommendations from the panel on cost effectiveness in health and medicine. Panel on cost Effectiveness in Health and Medicine

This article reports the recommendations of the Panel on Cost Effectiveness in Health and Medicine, sponsored by the US Public Health Service, on standardised methods for conducting cost-effectiveness analyses. Although not expressly directed at analyses of pharmaceutical agents, the Panel's recommendations are relevant to pharmacoeconomic studies. The Panel outlines a 'Reference Case' set of methodological practices to improve quality and comparability of analyses. Designed for studies that inform resource-allocation decisions, the Reference Case includes recommendations for study framing and scope, components of the numerator and denominator of cost-effectiveness ratios, discounting, handling uncertainty and reporting. The Reference Case analysis is conducted from the societal perspective, and includes all effects of interventions on resource use and health. Resource use includes 'time' resources, such as for caregiving or undergoing an intervention. The quality-adjusted life-year (QALY) is the common measure of health effect across Reference Case studies. Although the Panel does not endorse a measure for obtaining quality-of-life weights, several recommendations address the QALY. The Panel recommends a 3% discount rate for costs and health effects. Pharmacoeconomic studies have burgeoned in recent years. The Reference Case analysis will improve study quality and usability, and permit comparison of pharmaceuticals with other health interventions.

Multiattribute utility function for a comprehensive health status classification system. Health Utilities Index Mark 2

The Health Utilities Index Mark 2 (HUI:2) is a generic multiattribute, preference-based system for assessing health-related quality of life. Health Utilities Index Mark 2 consists of two components: a seven-attribute health status classification system and a scoring formula. The seven attributes are sensation, mobility, emotion, cognition, self-care, pain, and fertility. A random sample of general population parents were interviewed to determine cardinal preferences for the health states in the system. The health states were defined as lasting for a 60-year lifetime, starting at age 10. Values were measured using visual analogue scaling. Utilities were measured using a standard gamble technique. A scoring formula is provided, based on a multiplicative multiattribute utility function from the responses of 194 subjects. The utility scores are death-anchored (death = 0.0) and form an interval scale. Health Utilities Index Mark 2 and its utility scores can be useful to other researchers in a wide variety of settings who wish to document health status and assign preference scores.

Canadian guidelines for economic evaluation of pharmaceuticals. Canadian Collaborative Workshop for Pharmacoeconomics

In 1994, Canada became the second country to release national guidelines for the economic evaluation of pharmaceuticals. The guidelines were developed over a period of 18 months through an elaborate process of broad consultation with a wide variety of relevant stakeholders. The intent of the guidelines is to provide guidance to doers and users of studies, by laying out the general 'state of the art' regarding methods, and by providing specific methodological advice on many matters. The aim is to improve the scientific quality and integrity of studies, and to enhance consistency and comparability across studies. This article presents the Canadian guidelines, both in summary and in detail. Because the techniques of economic evaluation are widely applicable beyond pharmaceuticals, the guidelines will be of interest to researchers and decision makers in all fields of healthcare. Because the methods are not country specific, the guidelines will be of interest to those in other countries as well as in Canada.

Multi-attribute health status classification systems. Health Utilities Index

In this article, multi-attribute approaches to the assessment of health status are reviewed with a special focus on 2 recently developed systems, the Health Utilities Index (HUI) Mark II and Mark III systems. The Mark II system consists of 7 attributes: sensation, mobility, emotion, cognition, self-care, pain and fertility. The Mark III system contains 8 attributes: vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain. Each attribute consists of multiple levels of functioning. A combination of levels across the attributes constitutes a health state. The HUI systems are deliberately focused on the fundamental core attributes of health status, and on the capacity of individuals to function with respect to these attributes. Thus, the measure obtained constitutes a pure description of health status, uncontaminated by differential opportunity or preference. Multi-attribute systems provide a compact but comprehensive framework for describing health status for use in population health and programme evaluation studies. An important advantage of such systems is their ability to simultaneously provide detail on an attribute-by-attribute basis and to capture combinations of deficits among attributes. An additional advantage is their compatibility with multi-attribute preference functions, which provide a method for computing a summary health-related quality-of-life score for each health state.

Reliability of the Health Utilities Index--Mark III used in the 1991 cycle 6 Canadian General Social Survey Health Questionnaire

This study presents information on the test-retest reliability of the Health Utility Index--Mark III (HUI) system used in cycle 6 of the Canadian General Social Survey (GSS). The HUI system used in this reliability study consists of an eight-attribute health status classification system (HSCS) and a function for generating a summary score of health-related quality of life. To estimate test-retest reliability, a stratified random sample of individuals (n = 506) completing GSS telephone interviews during August and September, 1991 were interviewed again 1 month later. Weighting adjustments based on the probability of selection were invoked during the analyses to provide unbiased estimates of test-retest reliability for all GSS respondents in the August-September period. The results indicate that the individual questions, attributes and provisional index scores generally provided reliable information on health status in the GSS. The exceptions to this were limitations in speech and dexterity which were reported very infrequently. Kappa estimates of test-retest reliability for individual questions varied from 0.184 to 0.766. For the eight attributes, kappa estimates varied from 0.137 to 0.728. Using the provisional index scores to quantify health overall, a test-retest reliability of 0.767 was obtained (intra-class correlation coefficient).

Multi-attribute preference functions. Health Utilities Index

Multi-attribute utility theory, an extension of conventional utility theory, can be applied to model preference scores for health states defined by multi-attribute health status classification systems. The type of preference independence among the attributes determines the type of preference function required: additive, multiplicative or multilinear. In addition, the type of measurement instrument used determines the type of preference score obtained: value or utility. Multi-attribute utility theory has been applied to 2 recently developed multi-attribute health status classification systems, the Health Utilities Index (HUI) Mark II and Mark III systems. Results are presented for the Mark II system, and ongoing research is described for the Mark III system. The theory is also discussed in the context of other well known multi-attribute systems. The HUI system is an efficient method of determining a general public-based utility score for a specified health outcome or for the health status of an individual. In clinical populations, the scores can be used to provide a single summary measure of health-related quality of life. In cost-utility analyses, the scores can be used as quality weights for calculating quality-adjusted life years. In general populations, the measure can be used as quality weights for determining population health expectancy.

A study to determine the efficacy and safety of tenoxicam versus piroxicam, diclofenac and indomethacin in patients with osteoarthritis: a meta-analysis

OBJECTIVE

To obtain a better quantitative and qualitative estimate of the effect of tenoxicam (Tx) compared to piroxicam (Px), diclofenac (Dcl) and indomethacin (Ind) in the treatment of osteoarthritis (OA).

METHODS

Relevant studies were identified using computerized Medline search, manual search of cited references and correspondence with investigators, colleagues and the manufacturer of Tx. Once the studies were selected and chosen on the basis of predetermined methodologic criteria, the required data were extracted by 2 authors, independently. Eighteen studies met the required eligibility criteria. Meta-analyses were undertaken on 12 studies of Tx vs Px, 3 studies of Tx vs Dcl, and 2 studies of Tx vs Ind. Efficacy was measured in 2 ways: (1) physician global rating scale and (2) pain scale. Safety was measured in 3 ways: (1) physician global rating scale, (2) number of patients with adverse events, and (3) dropouts due to adverse events.

RESULTS

The following findings of the meta-analysis were statistically significant: In Tx vs Px comparisons, efficacy-(1), safety-(1) and safety-(3) were all better with Tx; in Tx vs Ind comparisons, safety-(1) and safety-(2) were better with Tx. All other findings showed no statistically significant differences between Tx and the comparison drug.

CONCLUSIONS

Compared to Px, Tx performs better on physician assessment of efficacy and tolerability, but the other comparisons remain inconclusive. Compared to Dcl, there appears not to be a difference. Compared to Ind, Tx is safer.