A linear index for predicting joint health-states utilities from single health-states utilities

Estimating Joint Health State Utility Algorithms Under Partial Information

OBJECTIVES

We explored the performance of existing joint health state utility estimators when data are not available on utilities that isolate single-condition health states excluding any co-occurring condition.

METHODS

Using data from the National Epidemiologic Survey on Alcohol and Related Conditions-III, we defined 2 information sets: (1) a full-information set that includes the narrowly defined health state utilities used in most studies that test the performance of joint health state utility estimators, and (2) a limited information set that includes only the more broadly defined health state utilities more commonly available to researchers. We used an example of alcohol use disorder co-occurring with cirrhosis of the liver, depressive disorder, or nicotine use disorder to illustrate our analysis.

RESULTS

We found that the performance of joint health state utility estimators is appreciably different under limited information than under full information. Full-information estimators typically overestimate the joint state utility, whereas limited-information estimators underestimate the joint state utility, except for the minimum estimator, which is overestimated in all cases.

CONCLUSIONS

Researchers using joint health state utility estimators should understand the information set available to them and use methodological guidance appropriate for that information set. We recommend the minimum estimator under limited information based on its ease of use, consistency (and therefore a predictable direction of bias), and lower root mean squared error.

A National Catalog of Mapped Short-Form Six-Dimension Utility Scores for Chronic Conditions in the United States From 2010 to 2015

OBJECTIVES

This study examined health preference utility weights and utility decrements associated with different types of chronic conditions in the United States.

METHODS

We used the 2010-2015 Medical Expenditure Panel Survey data for persons aged ≥ 18 years with 12-Item Short-Form Survey Physical and Mental Component Summary scores. 12-Item Short-Form Survey scores were converted to Short-Form Six-Dimension (SF-6D) preference scores to measure utilities of different chronic diseases. We used the Clinical Classification Code to identify 30 chronic diseases from 12 categories, such as cardiovascular diseases, cerebrovascular diseases, hypertension, hyperlipidemia, obesity, cancers, musculoskeletal diseases, endocrine or metabolic diseases, oral diseases, respiratory diseases, and mental disorders. A generalized linear model was used to quantify the utility decrements for 30 chronic diseases, controlling for demographic characteristics.

RESULTS

We identified 132 737 adults (mean age 47.2 years, 52.2% female, 80% white); 73% had at least one identified chronic disease, and the mean SF-6D was 0.786. Among 30 chronic diseases, the unadjusted mean SF-6D scores of patients with cognitive disorder (0.607) were the lowest, followed by congestive heart failure (0.629), rheumatoid arthritis (0.654), and lung cancer (0.662). After controlling for demographic variables (ie, age, sex) and comorbidities, cognitive disorders (-0.116), mood disorders (-0.099), rheumatoid arthritis (-0.090), liver cancer (-0.078), and stroke (-0.063) showed the highest decrements in the SF-6D scores (P < .05).

CONCLUSIONS

This study provides a nationally representative catalog of utility weights for major chronic diseases in the US general population. The utility decrements will enable researchers to calculate the health utilities of patients with multiple comorbid diseases.

Cost burden and net monetary benefit loss of neonatal hypoglycaemia

BACKGROUND

Neonatal hypoglycaemia is a common but treatable metabolic disorder that affects newborn infants and which, if not identified and treated adequately, may result in neurological sequelae that persist for the lifetime of the patient. The long-term financial and quality-of-life burden of neonatal hypoglycaemia has not been previously examined.

METHODS

We assessed the postnatal hospital and long-term costs associated with neonatal hypoglycaemia over 80 year and 18 year time horizons, using a health-system perspective and assessing impact on quality of life using quality-adjusted life year (QALYs). A decision analytic model was used to represent key outcomes in the presence and absence of neonatal hypoglycaemia.

RESULTS

The chance of developing one of the outcomes of neonatal hypoglycaemia in our model (cerebral palsy, learning disabilities, seizures, vision disorders) was 24.03% in subjects who experienced neonatal hypoglycaemia and 3.56% in those who do did not. Over an 80 year time horizon a subject who experienced neonatal hypoglycaemia had a combined hospital and post-discharge cost of NZ$72,000 due to the outcomes modelled, which is NZ$66,000 greater than a subject without neonatal hypoglycaemia. The net monetary benefit lost due to neonatal hypoglycaemia, using a value per QALY of NZ$43,000, is NZ$180,000 over an 80 year time horizon.

CONCLUSIONS

Even under the most conservative of estimates, neonatal hypoglycaemia contributes a significant financial burden to the health system both during childhood and over a lifetime. The combination of direct costs and loss of quality of life due to neonatal hypoglycaemia means that this condition warrants further research to focus on prevention and effective treatment.

Cost-effectiveness of surveillance with CT colonography after resection of colorectal cancer

OBJECTIVE

Surveillance following colorectal cancer (CRC) resection uses optical colonoscopy (OC) to detect intraluminal disease and CT to detect extracolonic recurrence. CT colonography (CTC) might be an efficient use of resources in this situation because it allows for intraluminal and extraluminal evaluations with one test.

DESIGN

We developed a simulation model to compare lifetime costs and benefits for a cohort of patients with resected CRC. Standard of care involved annual CT for 3 years and OC for years 1, 4 and every 5 years thereafter. For the CTC-based strategy, we replace CT+OC at year 1 with CTC. Patients with lesions greater than 6 mm detected by CTC underwent OC. Detection of an adenoma 10 mm or larger was followed by OC at 1 year, then every 3 years thereafter. Test characteristics and costs for CTC were derived from a clinical study. Medicare costs were used for cancer care costs as well as alternative test costs. We discounted costs and effects at 3% per year.

RESULTS

For persons with resected stage III CRC, the standard-of-care strategy was more costly (US$293) and effective (2.6 averted CRC cases and 1.1 averted cancer deaths per 1000) than the CTC-based strategy, with an incremental cost-effectiveness ratio of US$55 500 per quality-adjusted life-year gained. Our analysis was most sensitive to the sensitivity of CTC for detecting polyps 10 mm or larger and assumptions about disease progression.

CONCLUSION

In a simulation model, we found that replacing the standard-of-care approach to postdiagnostic surveillance with a CTC-based strategy is not an efficient use of resources in most situations.

A Systematic Review and Meta-Analysis of Prostate Cancer Utility Values of Patients and Partners Between 2007 and 2016

Background. There is widespread agreement that both the length and quality of life matter when assessing new technologies and/or models of care in the treatment for cancer patients. Quality of life for partners/carers also matters, particularly for prostate cancer. Purpose. This systematic review aims to provide up-to-date utility values along the prostate cancer care continuum (i.e., from prescreening through to palliative care) for use where future trial-based or modelled economic evaluations cannot collect primary data from men and/or partners. Data Sources. A protocol was developed and registered on the international register of systematic reviews—PROSPERO. Databases searched included EBSCO Information Services (CINAHL, EconLit, Global Health, HEED, MEDLINE Complete, PsycINFO), Cochrane Database of Systematic Reviews, Web of Science, and Embase. Study Selection. Study selection terms included health-related quality of life, prostate cancer, and partners or carers. Data Extraction. The authors identified articles published between 2007 and 2016 that provided health state utility values, with statistical uncertainty, for men with or at risk of prostate cancer and/or their partner/carers. Data Synthesis and Results. Study quality and generalizability of utilities was evaluated and meta-analysis conducted against prespecified criteria. From 906 original articles, 29 recent primary studies met the inclusion/exclusion criteria. We tabulate all the utility values with uncertainty, along with considerable methodological detail and patient population characteristics. Limitations. Utility values pertaining to carers/partners were limited to one study. Conclusions. Studies varied in design, measurement instruments utilized, quality, and generalizability. There is sufficient qualitative and quantitative detail for the reported utility values to be readily incorporated into economic evaluations. More research is needed with carers/partners and with newly developing prostate cancer-specific quality of life tools.

Estimating Joint Health Condition Utility Values

OBJECTIVES

To predict health state utility values (HSUVs) for individuals with up to 4 conditions simultaneously.

METHODS

Person-level data were taken from the General Practice Patient Survey, a national survey of adult patients registered with general practices in England. Individuals reported whether they had any 1 of 16 chronic conditions and completed the 3-level EuroQol 5-dimensional questionnaire. Four nonparametric methods (additive, multiplicative, minimum, and the adjusted decrement estimator) and 1 parametric estimator (the linear index) were used to predict HSUVs for individuals with a joint health condition (JHC). Predicted and actual utility scores were compared for precision using root mean square error and mean absolute error. Bias was assessed using mean error.

RESULTS

The analysis included 929,565 individuals, of which 30.5% had at least 2 conditions. Of the nonparametric estimators, the multiplicative approach produced estimates with the lowest bias and most precision for 2 JHCs. For populations with a long-term mental health condition within the JHC, the multiplicative approach overestimated utility scores. All nonparametric methods produced biased results when estimating HSUVs for 3 or 4 JHCs. The linear index generally produced unbiased results with the highest precision.

CONCLUSIONS

The multiplicative approach was the best nonparametric estimator when estimating HSUVs for 2 JHCs. None of the nonparametric approaches for estimating HSUVs can be recommended with more than 2 JHCs. The linear index was found to have good predictive properties but needs external validation before being recommended for routine use.

A SYSTEMATIC APPROACH FOR ASSESSING, IN THE ABSENCE OF FULL EVIDENCE, WHETHER MULTICOMPONENT INTERVENTIONS CAN BE MORE COST-EFFECTIVE THAN SINGLE COMPONENT INTERVENTIONS

OBJECTIVES

Multicomponent interventions (MCIs), consisting of at least two interventions, are common in rehabilitation and other healthcare fields. When the effectiveness of the MCI versus that of its single interventions is comparable or unknown, evidence of their expected incremental cost-effectiveness can be helpful in deciding which intervention to recommend. As such evidence often is unavailable this study proposes an approach to estimate what is more cost-effective; the MCI or the single intervention(s).

METHODS

We reviewed the literature for potential methods. Of those identified, headroom analysis was selected as the most suitable basis for developing the approach, based on the criteria of being able to estimate the cost-effectiveness of the single interventions versus that of the MCI (a) within a limited time frame, (b) in the absence of full data, and (c) taking into account carry-over and interaction effects. We illustrated the approach with an MCI for cancer survivors.

RESULTS

The approach starts with analyzing the costs of the MCI. Given a specific willingness-to-pay-value, it is analyzed how much effectiveness the MCI would need to generate to be considered cost-effective, and if this is likely to be attained. Finally, the cost-effectiveness of the single interventions relative to the potential of the MCI for being cost-effective can be compared.

CONCLUSIONS

A systematic approach using headroom analysis was developed for estimating whether an MCI is likely to be more cost effective than one (or more) of its single interventions.

Joint Utility Estimators in Substance Use Disorders

BACKGROUND

Although co-occurring conditions are common with substance use disorders (SUDs), estimation methods for joint health state utilities have not yet been tested in this context.

OBJECTIVES

To compare joint health state utility estimators in SUD to inform economic evaluation.

METHODS

We conducted two Internet-based surveys of US adults to collect community perspective standard gamble utilities for SUD and common co-occurring conditions. We evaluated six conditions as they occur individually and four combinations of these as they occur in tandem. We applied joint utility estimators using the six individual conditions' utilities to compare their performance relative to the observed combination states' utilities. We assessed performance with bias (estimated utility minus observed utility) and root mean square error (RMSE).

RESULTS

Using 3892 utilities from 1502 respondents, the minimum estimator was statistically unbiased (i.e., the 95% confidence interval included 0) for all combination states that we measured. The maximum estimator was unbiased for two states and the linear index and adjusted decrement estimators were unbiased for one state. The maximum estimator had the smallest RMSE for two combination states (back pain and prescription opioid misuse [0.0004] and injection crack and injection opioid use [0.0007]); the linear index and minimum estimators had the smallest RMSE for one combination state each. The additive and multiplicative estimators had the largest RMSE for all states.

CONCLUSIONS

Our results demonstrate the usefulness of the minimum estimator in this context, and confirm the inadequacy of the additive and multiplicative estimators. Further research is needed to extend these results to other SUD states.

Clinical impact of treatment timing for chronic hepatitis C infection: a decision model

Recent advances in the treatment of hepatitis C virus (HCV) infection have led to the availability of both highly efficacious interferon-containing and interferon-sparing regimens. However, the use of such therapies faces restrictions due to high costs. For patients who are medically eligible to receive interferon, the choice between the two will likely be impacted by preferences surrounding interferon, severity of disease, coverage policies and out-of-pocket costs. We developed a decision model to quantify the trade-offs between immediate, interferon-containing therapy and delayed, interferon-free therapy for patients with chronic, genotype 1 HCV infection. We projected the quality-adjusted life expectancy stratified by the presence or absence of cirrhosis for four strategies: (i) no treatment; (ii) immediate, one-time treatment with an interferon-containing regimen; (iii) immediate treatment as above with the opportunity for retreatment in patients who fail to achieve sustained virologic response with interferon-free therapy in 1 year; and (iv) delayed therapy with interferon-free therapy in 1 year. When compared to one-time immediate treatment with the interferon-containing regimen, delayed treatment with the interferon-free regimen in 1 year resulted in longer life expectancy, with a 0.2 quality-adjusted life year (QALY) increase in noncirrhotic patients, and a 1.1 QALY increase in patients with cirrhosis. This superiority in health benefits was lost when wait time for interferon-free therapy was greater than 3-3.2 years. In this modelling analysis, interferon-free therapy resulted in superior health benefits compared to immediate therapy with interferon until wait time exceeded 3-3.2 years. Such data can inform decision-making regarding treatment initiation for HCV as healthcare financing evolves.

Examining the relationship between health-related quality of life and increasing numbers of diagnoses

Purpose

Little is known about estimating utilities for comorbid (or ‘joint’) health states. Several joint health state prediction models have been suggested (for example, additive, multiplicative, best-of-pair, worst-of-pair, etc.), but no general consensus has been reached. The purpose of the study is to explore the relationship between health-related quality of life (HRQoL) and increasing numbers of diagnoses.

Methods

We analyzed a large dataset containing respondents’ ICD-9 diagnoses and preference-based HRQoL (EQ-5D and SF-6D). Data were stratified by the number of diagnoses, and mean HRQoL values were estimated. Several adjustments, accounting for the respondents’ age, sex, and the severity of the diagnoses, were carried out. Our analysis fitted additive and multiplicative models to the data and assessed model fit using multiple standard model selection methods.

Results

A total of 39,817 respondents were included in the analyses. Average HRQoL values were represented well by both linear and multiplicative models. Although results across all analyses were similar, adjusting for severity of diagnoses, age, and sex strengthened the linear model’s performance measures relative to the multiplicative model. Adjusted R² values were above 0.99 for all analyses (i.e., all adjusted analyses, for both HRQoL instruments), indicating a robust result.

Conclusions

Additive and multiplicative models perform equally well within our analyses. A practical implication of our findings, based on the presumption that a linear model is simpler than an additive model, is that an additive model should be preferred unless there is compelling evidence to the contrary.

Sourcing quality-of-life weights obtained from previous studies: theory and reality in Korea

BACKGROUND

The quality-of-life weights obtained in previous studies are frequently used in cost-utility analyses. The purpose of this study is to describe how the values obtained in previous studies are incorporated into the industry submissions requesting listing at the Korean National Health Insurance (NHI), focusing on the issues discussed in theoretical studies and national guidelines.

METHODS

The industry submissions requesting listing at the Korean NHI from January 2007 until December 2009 were evaluated by two independent researchers at the Health Insurance Review and Assessment Service (HIRA). Specifically, we observed the methods that were used to pool, predict joint health state utilities, and retain consistency within submissions in terms of the issues discussed in methodological research papers and recommendations from national guidelines.

RESULTS

More than half of the submissions used QALY as an outcome measure, and most of these submissions were sourced from prior studies. Heterogeneous methodologies were frequently used within a submission, with the inconsistent use of upper and lower anchors being prevalent. Assumptions behind measuring joint health state utilities or pooling multiple values for single health states were omitted in all submissions. Most national guidelines were rather vague regarding how to predict joint health states, how to select the best available value, how to maintain consistency within a submission, and how to generalize values obtained from prior studies.

CONCLUSIONS

Previously-generated values were commonly sourced, but this practice was frequently related to inconsistencies within and among submissions. Attention should be paid to the consistency and transparency of the value, especially if the value is sourced from prior studies.

Using health state utility values in models exploring the cost-effectiveness of health technologies

BACKGROUND

To improve comparability of economic data used in decision making, some agencies recommend that a particular instrument should be used to measure health state utility values (HSUVs) used in decision-analytic models. The methods used to incorporate HSUVs in models, however, are often methodologically poor and lack consistency. Inconsistencies in the methodologies used will produce discrepancies in results, undermining policy decisions informed by cost per quality-adjusted life-years.

OBJECTIVE

To provide an overview of the current evidence base relating to populating decision-analytic models with HSUVs.

FINDINGS

Research exploring suitable methods to accurately reflect the baseline or counterfactual HSUVs in decision-analytic models is limited, and while one study suggested that general population data may be appropriate, guidance in this area is poor. Literature describing the appropriateness of different methods used to estimate HSUVs for combined conditions is growing, but there is currently no consensus on the most appropriate methodology. While exploratory analyses suggest that a statistical regression model might improve accuracy in predicted values, the models require validation and testing in external data sets. Until additional research has been conducted in this area, the current evidence suggests that the multiplicative method is the most appropriate technique. Uncertainty in the HSUVs used in decision-analytic models is rarely fully characterized in decision-analytic models and is generally poorly reported.

CONCLUSIONS

A substantial volume of research is required before definitive detailed evidence-based practical advice can be provided. As the methodologies used can make a substantial difference to the results generated from decision-analytic models, the differences and lack of clarity and guidance will continue to lead to inconsistencies in policy decision making.

Estimating Health State Utility Values for Joint Health Conditions

Background. Analysts frequently estimate the health state utility values (HSUVs) for joint health conditions (JHCs) using data from cohorts with single health conditions. The methods can produce very different results, and there is currently no consensus on the most appropriate technique. Objective. To conduct a detailed critical review of existing empirical literature to gain an understanding of the reasons for differences in results and identify where uncertainty remains that may be addressed by further research. Results. Of the 11 studies identified, 10 assessed the additive method, 10 the multiplicative method, 7 the minimum method, and 3 the combination model. Two studies evaluated just 1 of the techniques, whereas the others compared results generated using 2 or more. The range of actual HSUVs can influence general findings, and methods are sometimes compared using descriptive statistics that may not be appropriate for assessing predictive ability. None of the methods gave consistently accurate results across the full range of possible HSUVs, and the values assigned to normal health influence the accuracy of the methods. Conclusions. Within the limitations of the current evidence base, we would advocate the multiplicative method, conditional on adjustment for baseline utility, as the preferred technique to estimate HSUVs for JHCs when using mean values obtained from cohorts with single conditions. We would recommend that a range of sensitivity analyses be performed to explore the effect on results when using the estimated HSUVs in economic models. Although the linear models appeared to give more accurate results in the studies we reviewed, these models require validating in external data before they can be recommended.

Comparing EQ-5D scores for comorbid health conditions estimated using 5 different methods

BACKGROUND

There is currently no consensus on the most appropriate method to estimate health state utility values (HSUVs) for comorbid health conditions.

OBJECTIVE

The objective of the study was to assess the accuracy by applying 5 different methods to an EQ-5D dataset.

METHODS

EQ-5D data (n=41,174) from the Health Survey for England were used to compare HSUVs generated using the additive, multiplicative and minimum methods, the adjusted decrement estimator, and a linear regression.

RESULTS

The additive and multiplicative methods underestimated the majority of HSUVs and the magnitude of the errors increased as the actual HSUV increased. Conversely, the minimum and adjusted decrement estimator methods overestimated the majority of HSUVs and the magnitude of errors increased as the actual HSUV decreased. Although the simple linear model produced the most accurate results, there was a tendency to underpredict higher HSUVs and overpredict lower HSUVs. The magnitude and direction of mean errors could be driven by the actual scores being estimated in addition to the technique used and the HSUVs estimated using an adjusted baseline were generally more accurate.

CONCLUSIONS

The additive and minimum methods performed very poorly in our data. Although the simple linear model gave the most accurate results, the model requires validating in external data obtained from the EQ-5D and other preference-based measures. Based on the current evidence base, we would recommend the multiplicative method is used together with a range of univariate sensitivity analyses.

The impact of co-morbidity burden on preference-based health-related quality of life in the United States

BACKGROUND

Many statistical analyses, clinical trials and cost-utility analyses designed to measure the impact of a particular disease on utility scores often overlook the important influence of co-morbidity burden.

OBJECTIVES

This study aims to examine the impact of co-morbidity burden on EQ-5D index scores in a nationally representative sample of the US.

METHODS

The pooled 2001 and 2003 Medical Expenditure Panel Survey was used. The total number of chronic conditions for each individual was calculated based on Clinical Classification Categories codes. Spline regression was used to identify nonlinear age effects: individuals were separated into four quartiles based on age. Censored least absolute deviation was used to regress EQ-5D index scores on age and chronic co-morbidity, controlling for income, gender, race, ethnicity, education, physical activity and smoking status. Interactions between age and chronic conditions were also explored.

RESULTS

The coefficients for chronic co-morbidities were highly statistically significant with large magnitudes for those with two or more chronic conditions (coefficient two chronic conditions=-0.16; coefficient nine chronic conditions=-0.28). After controlling for chronic co-morbidities and other confounders, age was not statistically significant except for those aged>58 years and the magnitude of this coefficient was very small (coefficient aged>58 years=-0.0006). The interactions between age and chronic co-morbidity were significant, but the deleterious impact of their interaction was largely dominated by the existence and number of chronic conditions.

CONCLUSIONS

Chronic conditions have a significant deleterious impact on EQ-5D index scores that is much more pronounced than age and other sociodemographic and behavioural characteristics. Future analyses and cost-utility models should incorporate the impact of multiple morbidity.

Estimating health state utility values for comorbid health conditions using SF-6D data

INTRODUCTION

When health state utility values for comorbid health conditions are not available, data from cohorts with single conditions are used to estimate scores. The methods used can produce very different results and there is currently no consensus on which is the most appropriate approach.

OBJECTIVE

The objective of the current study was to compare the accuracy of five different methods within the same dataset.

METHOD

Data collected during five Welsh Health Surveys were subgrouped by health status. Mean short-form 6 dimension (SF-6D) scores for cohorts with a specific health condition were used to estimate mean SF-6D scores for cohorts with comorbid conditions using the additive, multiplicative, and minimum methods, the adjusted decrement estimator (ADE), and a linear regression model.

RESULTS

The mean SF-6D for subgroups with comorbid health conditions ranged from 0.4648 to 0.6068. The linear model produced the most accurate scores for the comorbid health conditions with 88% of values accurate to within the minimum important difference for the SF-6D. The additive and minimum methods underestimated or overestimated the actual SF-6D scores respectively. The multiplicative and ADE methods both underestimated the majority of scores. However, both methods performed better when estimating scores smaller than 0.50. Although the range in actual health state utility values (HSUVs) was relatively small, our data covered the lower end of the index and the majority of previous research has involved actual HSUVs at the upper end of possible ranges.

CONCLUSIONS

Although the linear model gave the most accurate results in our data, additional research is required to validate our findings.

A value set for the EQ-5D based on experienced health states: development and testing for the German population

BACKGROUND

Decision makers responsible for allocation of healthcare resources may require that health states are valued by the population for whom they are making decisions. To achieve this, health-state descriptions can be combined with a value set that reflects the valuations of the target population. In the decision-utility approach, such a value set is at least partly based on wants and expectations regarding given health states. This may reflect aspects different from the health state experienced and valued by a respondent.

OBJECTIVES

To derive a value set that is completely based on experienced health states, emphasising the patient's perspective, and test its predictive performance in comparison with established approaches.

METHODS

Problem descriptions and rating scale valuations of the EQ-5D were drawn from two representative German population surveys in 2006 and 2007. Two models based on given health states but differing in valuation method (1a, b) were analysed, along with three models based on experienced health states: (2) ordinary least squares regression; (3) scale-transformed regression; and (4) a generalized linear model with binomial error distribution and constraint parameter estimation. The models were compared with respect to issues in specification, and accuracy in predicting the actual valuations of experienced health states in a new data set, using correlation, mean error and ranking measures for the latter. In addition, the impact of standardizing experience-based index models for age and sex of the subjects was investigated.

RESULTS

Models 1 (a, b), 2 and 3 partly led to plausible and comparable parameter estimates, but also led to problems of insignificance and inconsistencies in some of the estimates. Model 4 achieved consistency and featured partly equivalent and partly better predictive accuracy. Using this model, mean valuations of health states were much better predicted by the experience-based approach than by the decision-utility approach, especially for health states that frequently (>10) occurred in the population sample. Standardizing the experience-based index models for age and sex further improved predictive accuracy and strengthened the position of model 4.

CONCLUSIONS

A value set for the EQ-5D can be plausibly estimated from experience-based valuations. The approach offers an alternative to decision makers who prefer experience-based valuation over decision utilities in the measurement of health outcome. Although usefulness in population samples was shown, use in a clinical context will first require indication-specific tests. Current limitations include use in a general population only, and a restricted range of health states covered.

Populating an economic model with health state utility values: moving toward better practice

BACKGROUND

The methods used to estimate health-state utility values (HSUV) for multiple health conditions can produce very different values. Economic results generated using baselines of perfect health are not comparable with those generated using baselines adjusted to reflect the HSUVs associated with the health condition. Despite this, there is no guidance on the preferred techniques and little research describing the effect on cost per quality adjusted life-year (QALY) results when using the different methods.

METHODS

Using a cardiovascular disease (CVD) model and cost per QALY thresholds, we assess the consequence of using different baseline health-state utility profiles (perfect health, no history of CVD, general population) in conjunction with models (minimum, additive, multiplicative) frequently used to approximate scores for health states with multiple health conditions. HSUVs are calculated using the EQ-5D UK preference-based algorithm.

RESULTS

Assuming a baseline of perfect health ignores the natural decline in quality of life associated with age, overestimating the benefits of treatment. The results generated using baselines from the general population are comparable to those obtained using baselines from individuals with no history of CVD. The minimum model biases results in favor of younger-aged cohorts. The additive and multiplicative models give similar results.

CONCLUSION

Although further research in additional health conditions is required to support our findings, our results highlight the need for analysts to conform to an agreed reference case. We demonstrate that in CVD, if data are not available from individuals without the health condition, HSUVs from the general population provide a reasonable approximation.

Predicting Utility for Joint Health States: A General Framework and a New Nonparametric Estimator

Measuring utility is important in clinical decision making and cost-effectiveness analysis because utilities are often used to compute quality-adjusted life expectancy, a metric used in measuring the effectiveness of health care programs and medical interventions. Predicting utility for joint health states has become an increasingly valuable research topic because of the aging of the population and the increasing prevalence of comorbidities. Although multiplicative, minimum, and additive estimators are commonly used in practice, research has shown that they are all biased. In this study, the authors propose a general framework for predicting utility for joint health states. This framework includes these 3 nonparametric estimators as special cases. A new simple nonparametric estimator, the adjusted decrement estimator, [Uij = Umin - Umin (1 - Ui )(1 - Uj )], is introduced under the proposed framework. When applied to 2 independent data sources, the new nonparametric estimator not only generated unbiased prediction of utilities for joint health states but also had the least root mean squared error and highest concordance when compared with other nonparametric and parametric estimators. Further research and validation of this new estimator are needed.