Mapping the SF-12 to Preference-Based Instruments

Current Status of Research on the Mapping Function of Health Utility Values in the Asia Pacific Region: A Systematic Review

OBJECTIVES

This systematic review aimed to analyze the published studies on the use of the mapping method between generic scales and disease-specific scales as well as between 2 universal scales.

METHODS

A systematic literature search was conducted using PubMed, ScienceDirect, Web of Science, CNKI, Weipa Database, Wanfang Database, and HERC Database to collect articles about the application of the mapping method to the measurement of health utility value from January 2000 to December 2019.

RESULTS

Overall, 59 articles met the inclusion requirements, and most of them were a mapping study between a disease-specific scale and a generic scale. Then all these articles were classified by the following study types: a clear functional relationship; unclear functional relationship; disease-specific scale and universality; mapping between generic scales and disease-specific scales, and mapping between universal scales. Most studies derived the best mapping model from the ordinary least squares regression, and fewer studies chose to use new regression methods. Sample sizes in the retrieved studies generally affected the reliability of the study results.

CONCLUSIONS

In recent years, as more attention has been paid to the research of the mapping method, a large number of problems have followed, such as the selection of scale types, the coverage of the study sample, and the selection of evaluation index of model performance and sample size. It is hoped that these problems can be properly solved in the future research.

Behavioural modification interventions for medically unexplained symptoms in primary care: systematic reviews and economic evaluation

BACKGROUND

The term 'medically unexplained symptoms' is used to cover a wide range of persistent bodily complaints for which adequate examination and appropriate investigations do not reveal sufficiently explanatory structural or other specified pathologies. A wide range of interventions may be delivered to patients presenting with medically unexplained symptoms in primary care. Many of these therapies aim to change the behaviours of the individual who may have worsening symptoms.

OBJECTIVES

An evidence synthesis to determine the clinical effectiveness and cost-effectiveness of behavioural modification interventions for medically unexplained symptoms delivered in primary care settings was undertaken. Barriers to and facilitators of the effectiveness and acceptability of these interventions from the perspective of patients and service providers were evaluated through qualitative review and realist synthesis.

DATA SOURCES

Full search strategies were developed to identify relevant literature. Eleven electronic sources were searched. Eligibility criteria - for the review of clinical effectiveness, randomised controlled trials were sought. For the qualitative review, UK studies of any design were included. For the cost-effectiveness review, papers were restricted to UK studies reporting outcomes as quality-adjusted life-year gains. Clinical searches were conducted in November 2015 and December 2015, qualitative searches were conducted in July 2016 and economic searches were conducted in August 2016. The databases searched included MEDLINE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), PsycINFO and EMBASE. Updated searches were conducted in February 2019 and March 2019.

PARTICIPANTS

Adult participants meeting the criteria for medically unexplained symptoms, including somatoform disorders, chronic unexplained pain and functional somatic syndromes.

INTERVENTIONS

Behavioural interventions were categorised into types. These included psychotherapies, exercise-based interventions, multimodal therapies (consisting of more than one intervention type), relaxation/stretching/social support/emotional support, guided self-help and general practitioner interventions, such as reattribution. Evidence synthesis: a network meta-analysis was conducted to allow a simultaneous comparison of all evaluated interventions in a single coherent analysis. Separate network meta-analyses were performed at three time points: end of treatment, short-term follow-up (< 6 months since the end of treatment) and long-term follow-up (≥ 6 months after the end of treatment). Outcomes included physical and psychological symptoms, physical functioning and impact of the illness on daily activities. Economic evaluation: within-trial estimates of cost-effectiveness were generated for the subset of studies where utility values (or quality-adjusted life-years) were reported or where these could be estimated by mapping from Short Form questionnaire-36 items or Short Form questionnaire-12 items outcomes.

RESULTS

Fifty-nine studies involving 9077 patients were included in the clinical effectiveness review. There was a large degree of heterogeneity both between and within intervention types, and the networks were sparse across all outcomes. At the end of treatment, behavioural interventions showed some beneficial effects when compared with usual care, in particular for improvement of specific physical symptoms [(1) pain: high-intensity cognitive-behavioural therapy (CBTHI) standardised mean difference (SMD) 0.54 [95% credible interval (CrI) 0.28 to 0.84], multimodal SMD 0.52 (95% CrI 0.19 to 0.89); and (2) fatigue: low-intensity cognitive-behavioural therapy (CBTLI) SMD 0.72 (95% CrI 0.27 to 1.21), relaxation/stretching/social support/emotional support SMD 0.87 (95% CrI 0.20 to 1.55), graded activity SMD 0.51 (95% CrI 0.14 to 0.93), multimodal SMD 0.52 (95% CrI 0.14 to 0.92)] and psychological outcomes [(1) anxiety CBTHI SMD 0.52 (95% CrI 0.06 to 0.96); (2) depression CBTHI SMD 0.80 (95% CrI 0.26 to 1.38); and (3) emotional distress other psychotherapy SMD 0.58 (95% CrI 0.05 to 1.13), relaxation/stretching/social support/emotional support SMD 0.66 (95% CrI 0.18 to 1.28) and sport/exercise SMD 0.49 (95% CrI 0.03 to 1.01)]. At short-term follow-up, behavioural interventions showed some beneficial effects for specific physical symptoms [(1) pain: CBTHI SMD 0.73 (95% CrI 0.10 to 1.39); (2) fatigue: CBTLI SMD 0.62 (95% CrI 0.11 to 1.14), relaxation/stretching/social support/emotional support SMD 0.51 (95% CrI 0.06 to 1.00)] and psychological outcomes [(1) anxiety: CBTHI SMD 0.74 (95% CrI 0.14 to 1.34); (2) depression: CBTHI SMD 0.93 (95% CrI 0.37 to 1.52); and (3) emotional distress: relaxation/stretching/social support/emotional support SMD 0.82 (95% CrI 0.02 to 1.65), multimodal SMD 0.43 (95% CrI 0.04 to 0.91)]. For physical functioning, only multimodal therapy showed beneficial effects: end-of-treatment SMD 0.33 (95% CrI 0.09 to 0.59); and short-term follow-up SMD 0.78 (95% CrI 0.23 to 1.40). For impact on daily activities, CBTHI was the only behavioural intervention to show beneficial effects [end-of-treatment SMD 1.30 (95% CrI 0.59 to 2.00); and short-term follow-up SMD 2.25 (95% CrI 1.34 to 3.16)]. Few effects remained at long-term follow-up. General practitioner interventions showed no significant beneficial effects for any outcome. No intervention group showed conclusive beneficial effects for measures of symptom load (somatisation). A large degree of heterogeneity was found across individual studies in the assessment of cost-effectiveness. Several studies suggested that the interventions produce fewer quality-adjusted life-years than usual care. For those interventions that generated quality-adjusted life-year gains, the mid-point incremental cost-effectiveness ratios (ICERs) ranged from £1397 to £129,267, but, where the mid-point ICER fell below £30,000, the exploratory assessment of uncertainty suggested that it may be above £30,000.

LIMITATIONS

Sparse networks meant that it was not possible to conduct a metaregression to explain between-study differences in effects. Results were not consistent within intervention type, and there were considerable differences in characteristics between studies of the same type. There were moderate to high levels of statistical heterogeneity. Separate analyses were conducted for three time points and, therefore, analyses are not repeated-measures analyses and do not account for correlations between time points.

CONCLUSIONS

Behavioural interventions showed some beneficial effects for specific medically unexplained symptoms, but no one behavioural intervention was effective across all medically unexplained symptoms. There was little evidence that these interventions are effective for measures of symptom load (somatisation). General practitioner-led interventions were not shown to be effective. Considerable heterogeneity in interventions, populations and sparse networks mean that results should be interpreted with caution. The relationship between patient and service provider is perceived to play a key role in facilitating a successful intervention. Future research should focus on testing the therapeutic effects of the general practitioner-patient relationship within trials of behavioural interventions, and explaining the observed between-study differences in effects within the same intervention type (e.g. with more detailed reporting of defined mechanisms of the interventions under study).

STUDY REGISTRATION

This study is registered as PROSPERO CRD42015025520.

FUNDING

This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 24, No. 46. See the NIHR Journals Library website for further project information.

Developing a crosswalk between the RAND-12 and the health utilities index for multiple sclerosis

BACKGROUND

Researchers studying health-related quality of life (HRQOL) in multiple sclerosis (MS) can choose from many instruments, but findings from studies which use different instruments cannot be easily combined. We aimed to develop a crosswalk that associates scores from the RAND-12 to scores on the Health Utilities Index-Mark III (HUI3) in persons with MS.

METHODS

In 2018, participants in the North American Research Committee on Multiple Sclerosis (NARCOMS) registry completed the RAND-12 and the HUI3 to assess HRQOL. We used item-response theory (IRT) and equipercentile linking approaches to develop a crosswalk between instruments. We compared predicted scores for the HUI3 from each crosswalk to observed scores using Pearson correlations, intraclass correlation coefficients (ICCs), and Bland-Altman plots.

RESULTS

Of 11,389 invited participants, 7129 (62.6%) responded. Predicted and observed values of the HUI3 from the IRT-linking method were moderately correlated (Pearson r = 0.76) with good concordance (ICC = 0.72). However, the Bland-Altman plots suggested biased prediction. Predicted and observed values from the equipercentile linking method were also moderately correlated (Pearson r = 0.78, ICC = 0.78). The Bland-Altman plots suggested no bias.

CONCLUSION

We developed a crosswalk between the RAND-12 and the HUI3 in the MS population which will facilitate data harmonization efforts.

Valuing health-related quality of life in heart failure: a systematic review of methods to derive quality-adjusted life years (QALYs) in trial-based cost-utility analyses

The accurate measurement of health-related quality of life (HRQoL) and the value of improving it for patients are essential for deriving quality-adjusted life years (QALYs) to inform treatment choice and resource allocation. The objective of this review was to identify and describe the approaches used to measure and value change in HRQoL in trial-based economic evaluations of heart failure interventions which derive QALYs as an outcome. Three databases (PubMed, CINAHL, Cochrane) were systematically searched. Twenty studies reporting economic evaluations based on 18 individual trials were identified. Most studies (n = 17) utilised generic preference-based measures to describe HRQoL and derive QALYs, commonly the EQ-5D-3L. Of these, three studies (from the same trial) also used mapping from a condition-specific to a generic measure. The remaining three studies used patients’ direct valuation of their own health or physician-reported outcomes to derive QALYs. Only 7 of the 20 studies reported significant incremental QALY gains. Most interventions were reported as being likely to be cost-effective at specified willingness to pay thresholds. The substantial variation in the approach applied to derive QALYs in the measurement of and value attributed to HRQoL in heart failure requires further investigation.

Mapping the kidney disease quality of life 36-item short form survey (KDQOL-36) to the EQ-5D-3L and the EQ-5D-5L in patients undergoing dialysis

OBJECTIVES

To develop algorithms mapping the Kidney Disease Quality of Life 36-Item Short Form Survey (KDQOL-36) onto the 3-level EQ-5D questionnaire (EQ-5D-3L) and the 5-level EQ-5D questionnaire (EQ-5D-5L) for patients with end-stage renal disease requiring dialysis.

METHODS

We used data from a cross-sectional study in Europe (France, n = 299; Germany, n = 413; Italy, n = 278; Spain, n = 225) to map onto EQ-5D-3L and data from a cross-sectional study in Singapore (n = 163) to map onto EQ-5D-5L. Direct mapping using linear regression, mixture beta regression and adjusted limited dependent variable mixture models (ALDVMMs) and response mapping using seemingly unrelated ordered probit models were performed. The KDQOL-36 subscale scores, i.e., physical component summary (PCS), mental component summary (MCS), three disease-specific subscales or their average, i.e., kidney disease component summary (KDCS), and age and sex were included as the explanatory variables. Predictive performance was assessed by mean absolute error (MAE) and root mean square error (RMSE) using 10-fold cross-validation.

RESULTS

Mixture models outperformed linear regression and response mapping. When mapping to EQ-5D-3L, the ALDVMM model was the best-performing one for France, Germany and Spain while beta regression was best for Italy. When mapping to EQ-5D-5L, the ALDVMM model also demonstrated the best predictive performance. Generally, models using KDQOL-36 subscale scores showed better fit than using the KDCS.

CONCLUSIONS

This study adds to the growing literature suggesting the better performance of the mixture models in modelling EQ-5D and produces algorithms to map the KDQOL-36 onto EQ-5D-3L (for France, Germany, Italy, and Spain) and EQ-5D-5L (for Singapore).

Impact of mapped EQ-5D utilities on cost-effectiveness analysis: in the case of dialysis treatments

Objectives

This study aimed to evaluate the performance of EQ-5D data mapped from SF-12 in terms of estimating cost effectiveness in cost-utility analysis (CUA). The comparability of SF-6D (derived from SF-12) was also assessed.

Methods

Incremental quality-adjusted life years (QALYs) and incremental cost-effectiveness ratios (ICERs) were calculated based on two Markov models assessing the cost effectiveness of haemodialysis (HD) and peritoneal dialysis (PD) using utility values based on EQ-5D-5L, EQ-5D using three direct-mapping algorithms and two response-mapping algorithms (mEQ-5D), and SF-6D. Bootstrap method was used to estimate the 95% confidence interval (percentile method) of incremental QALYs and ICERs with 1000 replications for the utilities.

Results

In both models, compared to the observed EQ-5D values, mEQ-5D values expressed much lower incremental QALYs (range − 14.9 to − 33.2%) and much higher ICERs (range 17.5 to 49.7%). SF-6D also estimated lower incremental QALYs (− 29.0 and − 14.9%) and higher ICERs (40.9 and 17.5%) than did the observed EQ-5D. The 95% confidence interval of incremental QALYs and ICERs confirmed the lower incremental QALYs and higher ICERs estimated using mEQ-5D and SF-6D.

Conclusion

Compared to observed EQ-5D, EQ-5D mapped from SF-12 and SF-6D would under-estimate the QALYs gained in cost-utility analysis and thus lead to higher ICERs. It would be more sensible to conduct CUA studies using directly collected EQ-5D data and to designate one single preference-based measure as reference case in a jurisdiction to achieve consistency in healthcare decision-making.

Electronic supplementary material

The online version of this article (10.1007/s10198-018-0987-x) contains supplementary material, which is available to authorized users.

Do racial and ethnic disparities in health care use vary with health?

OBJECTIVE

To determine the relationship between health status and the magnitude of black-white and Hispanic-white disparities in the likelihood of having any office-based or hospital outpatient department visits, as well as number of visits.

DATA SOURCE

2010-2014 Medical Expenditure Panel Survey.

STUDY DESIGN

The probability of having a visit is modeled using a Probit model, and the number of visits using a negative binomial model. We use a nonlinear rank-and-replace method to adjust minority health status to be comparable to that of whites, and predict utilization at different levels of health by fixing an indicator of health status. We compare estimated differences in predicted utilization across racial/ethnic groups for each level of health status to map out the relationship between the racial/ethnic disparity and health status, also stratifying by health insurance coverage.

EXTRACTION METHODS

We subset to nonelderly adults.

PRINCIPAL FINDINGS

We find that Hispanic-white differences in the probability of having an office-based or hospital outpatient department were widest among adults in excellent health (27 percentage points, 95% CI: [23, 31]) and narrowest when reporting poor or fair health (15 p.p. [13, 17]). Black-white and Hispanic-white differences in the number of visits were wider for adults who report poor or fair health (5.3 visits [4.0, 6.6] and 5.7 [4.3, 7.0], respectively) compared to excellent health (1.7 [1.2, 2.1] and 1.5 [1.1, 2.0], respectively) among adults who are full-year privately insured.

CONCLUSIONS

The magnitudes of racial/ethnic disparities vary with level of health.

Cost-Effectiveness of Primary and Revision Surgery for Adult Spinal Deformity

STUDY DESIGN

Retrospective comparative study.

OBJECTIVE

The purpose of this study is to compare functional outcomes, hospital resource utilization, and spine-related costs during 2 years in patients who had undergone primary or revision surgery for adult spinal deformity (ASD).

SUMMARY OF BACKGROUND DATA

After surgery for ASD, patients may require revision for pseudarthrosis, implant complications, or deformity progression. Data evaluating cost-effectiveness of primary and, in particular, revision surgery, for ASD are sparse.

METHODS

We retrospectively reviewed records for 119 consecutive patients who had undergone primary or revision surgery for ASD. Two-year total spine-related medical costs were derived from hospital charge data. Functional outcome scores were extracted from prospectively collected patient data. Cost utility ratios (cost/quality-adjusted life-year [QALY]) at 2 years were calculated and assessed against a threshold of $154,458/QALY gained (three times the 2015 US per-capita gross domestic product).

RESULTS

The primary surgery cohort (n = 56) and revision cohort (n = 63) showed significant improvements in health-related quality-of-life scores at 2 years. Median surgical and spine-related 2-year follow-up costs were $137,990 (interquartile range [IQR], $84,186) for primary surgery and $115,509 (IQR, $63,753) for revision surgery and were not significantly different between the two groups (P = 0.12). We report 2-year QALY gains of 0.36 in the primary surgery cohort and 0.40 in the revision group (P = 0.71). Primary instrumented fusion was associated with a median 2-year cost per QALY of $197,809 (IQR, $187,350) versus $129,950 (IQR, $209,928) for revision surgery (P = 0.31).

CONCLUSION

Revision surgery had lower total 2-year costs and higher QALY gains than primary surgery for ASD, although the differences were not significant. Although revision surgery for ASD is known to be technically challenging and to have a higher rate of major complications than primary surgery, revision surgery was cost-effective at 2 years. The cost/QALY ratio for primary surgery for ASD exceeded the threshold for cost effectiveness at 2 years.

LEVEL OF EVIDENCE

3.

Estimating the Association between SF-12 Responses and EQ-5D Utility Values by Response Mapping

Background. Reliably mapping from generic or diseasespecific health status measures into health state utilities would assist health economists. Existing studies mainly use ordinary least squares (OLS) regression equations to predict utility values for particular health states. The authors examine an alternative approach tomap between 2 generic health status instruments, the SF-12 and the EQ-5D. Methods. Multinomial logit regression is used to estimate the probability that a respondent will select a particular level of response to questions in the EQ-5D, using individual question responses and summary scores from the SF-12 as predictors. Monte Carlo simulation methods are used to generate predicted EQ-5D responses, and utility scores (tariffs) are then attached. Results are comparedwithanalternativeapproach based on direct mapping to utility scores using OLS. Data. The authors estimate equations using 12,967 adult survey responses-from the 2000 US Medical Expenditure Panel Survey. They report mean squared error (MSE) andmean absolute error (MAE) of their predicted utilitieswithin this sample, and out-of-sample using 13,304 adults from the 1996 Health Survey for England. Results. The authors obtain an in-sample and out-of-sample MSE of 0.03, compared with 0.02 for the OLS approach. Their MAE of 0.11 is similar to OLS results. The authors’ method predicts groupmean utility scores and differentiates between groups with or without known existing illness. Conclusions. The authors’ approach has higher MSE than the direct OLS approach but givesmore descriptive data on domains of health effects. Further outofsample prediction work will help test the validity of these methods.

Mapping the SF-12 to the EuroQol EQ-5D Index in a National US Sample

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

Using QALYs in telehealth evaluations: a systematic review of methodology and transparency

Background

The quality-adjusted life-year (QALY) is a recognised outcome measure in health economic evaluations. QALY incorporates individual preferences and identifies health gains by combining mortality and morbidity into one single index number. A literature review was conducted to examine and discuss the use of QALYs to measure outcomes in telehealth evaluations.

Methods

Evaluations were identified via a literature search in all relevant databases. Only economic evaluations measuring both costs and QALYs using primary patient level data of two or more alternatives were included.

Results

A total of 17 economic evaluations estimating QALYs were identified. All evaluations used validated generic health related-quality of life (HRQoL) instruments to describe health states. They used accepted methods for transforming the quality scores into utility values. The methodology used varied between the evaluations. The evaluations used four different preference measures (EQ-5D, SF-6D, QWB and HUI3), and utility scores were elicited from the general population. Most studies reported the methodology used in calculating QALYs. The evaluations were less transparent in reporting utility weights at different time points and variability around utilities and QALYs. Few made adjustments for differences in baseline utilities. The QALYs gained in the reviewed evaluations varied from 0.001 to 0.118 in implying a small but positive effect of telehealth intervention on patient’s health. The evaluations reported mixed cost-effectiveness results.

Conclusion

The use of QALYs in telehealth evaluations has increased over the last few years. Different methodologies and utility measures have been used to calculate QALYs. A more harmonised methodology and utility measure is needed to ensure comparability across telehealth evaluations.

Peer Support for Achieving Independence in Diabetes (Peer-AID): design, methods and baseline characteristics of a randomized controlled trial of community health worker assisted diabetes self-management support

BACKGROUND & OBJECTIVES

Community health workers (CHWs) may be an important mechanism to provide diabetes self-management to disadvantaged populations. We describe the design and baseline results of a trial evaluating a home-based CHW intervention.

METHODS & RESEARCH DESIGN

Peer Support for Achieving Independence in Diabetes (Peer-AID) is a randomized, controlled trial evaluating a home-based CHW-delivered diabetes self-management intervention versus usual care. The study recruited participants from 3 health systems. Change in A1c measured at 12 months is the primary outcome. Changes in blood pressure, lipids, health care utilization, health-related quality of life, self-efficacy and diabetes self-management behaviors at 12 months are secondary outcomes.

RESULTS

A total of 1438 patients were identified by a medical record review as potentially eligible, 445 patients were screened by telephone for eligibility and 287 were randomized. Groups were comparable at baseline on socio-demographic and clinical characteristics. All participants were low-income and were from diverse racial and ethnic backgrounds. The mean A1c was 8.9%, mean BMI was above the obese range, and non-adherence to diabetes medications was high. The cohort had high rates of co-morbid disease and low self-reported health status. Although one-third reported no health insurance, the mean number of visits to a physician in the past year was 5.7. Trial results are pending.

CONCLUSIONS

Peer-AID recruited and enrolled a diverse group of low income participants with poorly controlled type 2 diabetes and delivered a home-based diabetes self-management program. If effective, replication of the Peer-AID intervention in community based settings could contribute to improved control of diabetes in vulnerable populations.

Modelling disparities in health services utilisation for older Blacks: a quantile regression framework

With the on-going ageing of the United States population, resolving health disparities continues to be a prominent and worthwhile goal, particularly in the areas of promoting minority health and reducing racial/ethnic disparities. This analysis employs the 2004 and 2005 Household Component records from the Medical Expenditures Panel Survey, which correspond to data files H89 and H97, to examine utilisation by race across the entire distribution function; more specifically, applying the behavioural model of health services utilisation and employing a Quantile Regression (QR) framework. This is a noteworthy contribution because the conditional mean may not be the best approximation for a skewed-location distribution. In contrast, QR is robust to outliers and scale effects since the estimation minimises least absolute deviation. The sample consists of 2,525 older adults at least 65 years of age with 303 corresponding to Black and 2,222 corresponding to White. Results suggest older Blacks continue to utilise health services (i.e. office or clinic visits with a physician or medical provider) at lower levels and this is more pronounced at and below the median quantile (i.e. below the 50th cut-off). Usual source of care (USC) continues to play an important role. Beliefs surrounding the need for insurance and medical intervention are also significant and explain some of the racial disparities. Although utilisation disparities persist for older Blacks, collaborative and flexible models of care can reach this group.

Cost-effectiveness of bariatric surgical procedures for the treatment of severe obesity

OBJECTIVE

One-third of Americans are obese and an increasing number opt for bariatric surgery. This study estimates the cost-effectiveness of common bariatric surgical procedures from a healthcare system perspective.

METHODS

We evaluated the three most common bariatric surgical procedures in the US: laparoscopic gastric bypass (LRYGB), conventional (open) Roux-en-Y gastric bypass (ORYGB), and laparoscopic adjustable gastric banding (LAGB) compared to no surgery. The reference case was defined as a 53-year old female with body mass index (BMI) of 44 kg/m(2). We developed a two-part model using a deterministic approach for the first 5-year period post-surgery and separate empirical forecasts for the natural history of BMI, costs and outcomes in the remaining years. We used a combination of datasets including Medicare and MarketScan(®) together with estimates from the literature to populate the model.

RESULTS

Bariatric surgery produced additional life expectancy (80-81 years) compared to no surgery (78 years). The incremental cost-effectiveness ratios (ICERs) of the surgical procedures were US $6,600 per quality-adjusted life expectancy (QALY) gained for LRYGB, US $6,200 for LAGB, and US $17,300 for ORYGB (3 % discount rate for cost and QALYs). ICERs varied according to choice of BMI forecasting method and clinically plausible variation in parameter estimates. In most scenarios, the ICER did not exceed a threshold of US $50,000 per QALY gained.

Mapping the FACT-G cancer-specific quality of life instrument to the EQ-5D and SF-6D

Objective

To help facilitate economic evaluations of oncology treatments, we mapped responses on cancer-specific instrument to generic preference-based measures.

Methods

Cancer patients (n = 367) completed one cancer-specific instrument, the FACT-G, and two preference-based measures, the EQ-5D and SF-6D. Responses were randomly divided to form development (n = 184) and cross-validation (n = 183) samples. Relationships between the instruments were estimated using ordinary least squares (OLS), generalized linear models (GLM), and censored least absolute deviations (CLAD) regression approaches. The performance of each model was assessed in terms of how well the responses to the cancer-specific instrument predicted EQ-5D and SF-6D utilities using mean absolute error (MAE) and root mean squared error (RMSE).

Results

Physical, functional, and emotional well-being domain scores of the FACT-G best explained the EQ-5D and SF-6D. In terms of accuracy of prediction as measured in RMSE, the CLAD model performed best for the EQ-5D (RMSE = 0.095) whereas the GLM model performed best for the SF-6D (RMSE = 0.061). The GLM predicted SF-6D scores matched the observed values more closely than the CLAD and OLS.

Conclusion

Our results demonstrate that the estimation of both EQ-5D and SF-6D utility indices using the FACT-G responses can be achieved. The CLAD model for the EQ-5D and the GLM model for the SF-6D are recommended. Thus, it is possible to estimate quality-adjusted life years for economic evaluation from studies where only cancer-specific instrument have been administered.

Probabilistic mapping of the health status measure SF-12 onto the health utility measure EQ-5D using the US-population-based scoring models

PURPOSE

Probabilistic mapping of the health status instrument SF-12 onto the health utility instrument EuroQol-5 dimensions (EQ-5D)-3L using the UK-population-based scoring model showed encouraging results as compared to other mapping methods, although its predictive performance using the US-population-based EQ-5D scoring models has not been investigated. In addition, a new and improved US-population-based EQ-5D scoring method has recently been developed and suggested for use in applications that required US societal health state values. In this study, we assessed predictive performance of the probabilistic mapping approach using the US-population-based scoring models on EQ-5D utility scores based on SF-12 responses and compared the results with those of other mapping methods.

METHODS

Using a sample of 19,678 adults from the 2003 Medical Expenditure Panel Survey, we evaluated the predictive performance of probabilistic mapping using Bayesian networks, response mapping using multinomial logistic regression, ordinary least squares, and censored least absolute deviations models by implementing a fivefold cross-validation method. The EQ-5D utility scores were generated using two US-population-based models: D1 and MM-OC.

RESULTS

Overall, the probabilistic mapping approach using Bayesian networks consistently outperformed other mapping methods with mean squared errors (MSE) of 0.007 and 0.007, mean absolute errors (MAE) of 0.057 and 0.039, and overall R (2) of 0.773 and 0.770 for the US-population-based EQ-5D scoring D1 and MM-OC models, respectively.

CONCLUSION

The probabilistic mapping approach can be useful to estimate EQ-5D utility scores from SF-12 responses with better predictive measures in terms of MSE, MAE, and R (2) than other common mapping methods.

Quality-of-life loss of people admitted to burn centers, United States

PURPOSE

To estimate quality-of-life loss per serious burn survivor in a large U.S. cohort.

METHODS

Longitudinal functional assessments of all 1,587 people receiving primary treatment in 5 burn centers between 2000 and 2009 included pre-burn (retrospective), at time of discharge, and 6, 12, and 24 months post-injury. We assessed adults with RAND Short Form (SF) 12 and children with SF-10 or Child Health Questionnaire, the child surveys scored using standard norms-based scoring. A literature review identified 20 quality-adjusted life year utility scorings for SF-12 and 27 scorings for EQ-5d response distributions predicted from SF-12 scores. We computed composite scores for each patient and time period by applying 32 scorings that met quality/non-duplication criteria.

RESULTS

Mean quality-of-life scores were 0.805 4 weeks pre-burn, 0.562 at discharge, rebounded through 1 year, and stabilized at 0.735 (0.750 for TBSA burned below 25 %, 0.722 for TBSA burned of 25-50 %, and 0.695 for larger burns). As a percentage of initial levels, burns reduced short-term quality of life by 30 %. Long-term loss averaged 11 %, ranging from 9 % for TBSA burned below 25-13 % for TBSA burned above 50 %. Children recovered faster and more fully.

CONCLUSION

Burns cause substantial losses in quality of life, with long-term losses comparable to traumatic brain injury.

The independent effect of body mass index on health-related quality of life among racial and ethnic subgroups

PURPOSE

To evaluate the impact of race/ethnicity on the relationship between body mass index (BMI) and health-related quality of life (HRQOL) among blacks, Hispanics, and whites.

METHODS

We used the Sinai Urban Health Institute's Improving Community Health Survey dataset to measure physical and mental HRQOL using the Physical Component Score (PCS-12) and the Mental Component Score (MCS-12) of the Short Form-12. Multivariate linear regression models were applied to the overall sample and in models stratified by race/ethnicity to evaluate the effects of BMI on physical and mental HRQOL outcome variables while controlling for confounders.

RESULTS

Considering physical HRQOL, increasing BMI was independently associated with worse PCS-12 (β = -0.22, p value <0.001) in the overall sample; the magnitude was not significantly different across racial/ethnic subgroups (blacks: β = -0.18, p value = 0.02; Hispanics: β = -0.28, p value = 0.01; whites: β = -0.20, p value = 0.02). Overall, Hispanic participants reported a worse PCS-12 compared to whites (β = -3.06, p value = 0.002). Considering mental HRQOL, BMI was not significantly associated with MCS-12 in the overall sample (β = -0.06, p value = 0.21) nor was BMI significantly associated with MCS-12 in any racial/ethnic subgroups. Overall, black participants reported better MCS-12 compared to whites (β = 2.51, p value = 0.001).

CONCLUSIONS

BMI was associated with worse physical HRQOL to a similar degree among blacks, Hispanics, and whites. This finding suggests that interventions leading to obesity reduction should be associated with substantial and equal improvements in the physical HRQOL of all race/ethnicity groups.

Estimating utilities for chronic kidney disease, using SF-36 and SF-12-based measures: challenges in a population of veterans with diabetes

PURPOSE

Using transformations of existing quality-of-life data to estimate utilities has the potential to efficiently provide investigators with utility information. We used within-method and across-method comparisons and estimated disutilities associated with increasing chronic kidney disease (CKD) severity.

METHODS

In an observational cohort of veterans with diabetes (DM) and pre-existing SF-36/SF-12 responses, we used six transformation methods (SF-12 to EQ-5D, SF-36 to HUI2, SF-12 to SF-6D, SF-36 to SF-6D, SF-36 to SF-6D (Bayesian method), and SF-12 to VR-6D) to estimate unadjusted utilities. CKD severity was staged using glomerular filtration rate estimated from serum creatinines, with the modification of diet in renal disease formula. We then used multivariate regression to estimate disutilities specifically associated with CKD severity stage.

RESULTS

Of 67,963 patients, 22,273 patients had recent-onset DM and 45,690 patients had prevalent DM. For the recent-onset group, the adjusted disutility associated with CKD derived from the six transformation methods ranged from 0.0029 to 0.0045 for stage 2; -0.004 to -0.0009 for early stage 3; -0.017 to -0.010 for late stage 3; -0.023 to -0.012 for stage 4; -0.078 to -0.033 for stage 5; and -0.012 to -0.001 for ESRD/dialysis.

CONCLUSION

Disutility did not increase monotonically as CKD severity increased. Differences in disutilities estimated using the six different methods were found. Both findings have implications for using such estimates in economic analyses.

The development of a model for translation of the Neck Disability Index to utility scores for cost-utility analysis in cervical disorders

BACKGROUND CONTEXT

The Neck Disability Index (NDI) is a commonly used disease-specific instrument for cervical spine disorders with good responsiveness and psychometric properties compared with general health status measures. However, NDI scores are unitless and do not have an intrinsic value that is comparable to other health status measures, and these scores have limited value in cost-utility analysis. The translation of disease-specific measures to Short Form-6 Dimensions (SF-6D) utility scores may be useful in cost-utility analysis.

PURPOSE

The purpose of this study is to present a model for translating the NDI to SF-6D utility scores, permitting the use of NDI scores in the cost-utility analysis of cervical disorders.

STUDY DESIGN/SETTING

A secondary analysis of a multicenter prospective clinical trial of the Synthes ProDisc-C (Synthes, West Chester, PA, USA) was performed.

PATIENT SAMPLE

Patients included were randomized to receive either a total disc arthroplasty or anterior cervical discectomy and fusion for treatment of symptomatic cervical disc disease involving one vertebral level between C3 and C7. All subjects completed NDI and 36-Item Short Form Health Survey (SF-36) self-assessments at preoperative and postoperative follow-ups of 6 weeks, 3, 6, 12, 18, and 24 months.

OUTCOME MEASURES

The NDI is a validated and widely used self-reported questionnaire designed to assess patient-determined disability resulting from neck pain, including pain level and effects on activities of daily living. The SF-6D is a preference-based health state classification system derived from six health dimensions of the SF-36 self-reported questionnaire, including the domains of physical functioning, role limitation, social functioning, bodily pain, mental health, and vitality.

METHODS

The collected data points were divided into two cohorts: one for model formation and one for the assessment of model validity. SF-36 scores were converted to SF-6D utilities via three previously published methods. Correlation analyses and linear regression modeling between SF-6D and NDI created the models for translating scores. For validation, Spearman and Pearson correlations were calculated between the observed and predicted SF-6D utilities, and prediction errors were calculated.

RESULTS

Four hundred thirty patients with 2,137 time points were used for creation and validation of the model. Pearson and Spearman correlation coefficients between the NDI and the SF-6D derived from each conversion method were found to be between -0.8255 and -0.8504 (p<.01). R(2) values ranged from 0.68 to 0.71 and root mean squared error (RMSE) from 0.092 to 0.084. Correlations between estimated and observed SF-6D scores ranged from 0.8325 to 0.8372 (p<.01). The mean prediction error was less than 0.006, with standard deviation (SD) between 0.082 and 0.093.

DISCUSSION

Correlations between NDI and SF-6D utility scores are strong and statistically significant. The model has a large R(2) and small RMSE. The prediction models produce a small mean prediction error, but the SD of the prediction errors is large. High correlations between NDI and SF-6D permit these models to be used to calculate overall utilities, changes in utilities, and quality-adjusted life-years for large data samples. However, the relatively large observed prediction error SDs may limit the accuracy of translation of individual data points or small sample sizes.

Predicting SF-6D utility scores from the neck disability index and numeric rating scales for neck and arm pain

STUDY DESIGN

Cross-sectional cohort.

OBJECTIVE

This study aims to provide an algorithm to estimate Short Form-6D (SF-6D) utilities using data from the Neck Disability Index (NDI), neck pain, and arm pain scores.

SUMMARY OF BACKGROUND DATA

Although cost-utility analysis is increasingly used to provide information about the relative value of alternative interventions, health state values or utilities are rarely available from clinical trial data. The Neck Disability Index (NDI) and numerical rating scales for neck and arm pain are widely used disease-specific measures in patients with cervical degenerative disorders. The purpose of this study is to provide an algorithm to allow estimation of SF-6D utilities using data from the NDI, and numerical rating scales for neck and arm pain.

METHODS

SF-36, NDI, neck and arm pain rating scale scores were prospectively collected before surgery, at 12 and 24 months after surgery in 2080 patients undergoing cervical fusion for degenerative disorders. SF-6D utilities were computed, and Spearman correlation coefficients were calculated for paired observations from multiple time points between NDI, neck and arm pain scores, and SF-6D utility scores. SF-6D scores were estimated from the NDI, neck and arm pain scores were estimated using a linear regression model. Using a separate, independent dataset of 396 patients in which NDI scores were available, SF-6D was estimated for each subject and compared to their actual SF-6D.

RESULTS

The mean age for those in the development sample was 50.4 ± 11.0 years and 33% were male. In the validation sample, the mean age was 53.1 ± 9.9 years and 35% were male. Correlations between the SF-6D and the NDI, neck and arm pain scores were statistically significant (P < 0.0001) with correlation coefficients of 0.82, 0.62, and 0.50, respectively. The regression equation using NDI aloneto predict SF-6D had an R of 0.66 and a root mean square error of 0.056. In the validation analysis, there was no statistically significant difference (P 5 0.961) between actual mean SF-6D (0.49 6 0.08) and the estimated mean SF-6D score (0.49 6 0.08), using the NDI regression model.

CONCLUSION

This regression-based algorithm may be a useful tool to predict SF-6D scores in studies of cervical degenerative disease that have collected NDI but not utility scores.

Predicting the EuroQol Group's EQ-5D index from CDC's "Healthy Days" in a US sample

BACKGROUND

Obtaining reliable preference-based scores from the widely used Healthy Days measures would enable calculation of quality-adjusted life years (QALYs) and cost-utility analyses in many US community populations and over time. Previous studies translating the Healthy Days to the EQ-5D, a preference-based measure, relied on an indirect method because of a lack of population-based survey data that asked both sets of questions of the same respondents.

METHOD

Data from the 2005-2006 National Health Measurement Study (NHMS; n = 3844 adults 35 years old or older) were used to develop regression-based models to estimate EQ-5D index scores from self-reported age, self-rated general health, and numbers of unhealthy days.

RESULTS

The models explained up to 52% of the variance in the EQ-5D. Estimated EQ-5D scores matched well to the observed EQ-5D scores in mean scores overall and by age, gender, race/ethnicity, income, education, body mass index, smoking, and disease categories. The average absolute differences were 0.005 to 0.006 on a health utility scale. After estimating mean EQ-5D index scores overall and for various subgroups in a large representative US sample of Healthy Days respondents, the authors found that these mean scores also closely matched the corresponding mean scores of EQ-5D respondents obtained from another large US representative sample with an average absolute difference of 0.013 points.

CONCLUSIONS

This study yielded a mapping algorithm to estimate EQ-5D index scores from the Healthy Days measures for populations of adults 35 years old and older. Such analysis confirms it is feasible to estimate mean EQ-5D index scores with acceptable validity for use in calculating QALYs and cost-utility analyses based on the overall model fit and relatively small differences between the observed and the estimated mean scores.

A review of studies mapping (or cross walking) non-preference based measures of health to generic preference-based measures

Clinical studies use a wide variety of health status measures to measure health related quality of life, many of which cannot be used in cost-effectiveness analysis using cost per quality adjusted life year (QALY). Mapping is one solution that is gaining popularity as it enables health state utility values to be predicted for use in cost per QALY analysis when no preference-based measure has been included in the study. This paper presents a systematic review of current practice in mapping between non-preference based measures and generic preference-based measures, addressing feasibility and validity, circumstances under which it should be considered and lessons for future mapping studies. This review found 30 studies reporting 119 different models. Performance of the mappings functions in terms of goodness-of-fit and prediction was variable and unable to be generalised across instruments. Where generic measures are not regarded as appropriate for a condition, mapping does not solve this problem. Most testing in the literature occurs at the individual level yet the main purpose of these functions is to predict mean values for subgroups of patients, hence more testing is required.

Mapping the modified Rankin scale (mRS) measurement into the generic EuroQol (EQ-5D) health outcome

BACKGROUND

Mapping disease-specific instruments into generic health outcomes or utility values is an expanding field of interest in health economics. This article constructs an algorithm to translate the modified Rankin scale (mRS) into EQ-5D utility values.

METHODS

mRS and EQ-5D information was derived from stroke or transient ischemic attack (TIA) patients identified as part of the Oxford Vascular study (OXVASC). Ordinary least squares (OLS) regression was used to predict UK EQ-5D tariffs from mRS scores. An alternative method, using multinomial logistic regression with a Monte Carlo simulation approach (MLogit) to predict responses to each EQ-5D question, was also explored. The performance of the models was compared according to the magnitude of their predicted-to-actual mean EQ-5D tariff difference, their mean absolute and mean squared errors (MAE and MSE), and associated 95% confidence intervals (CIs). Out-of-sample validation was carried out in a subset of coronary disease and peripheral vascular disease (PVD) patients also identified as part of OXVASC but not used in the original estimation.

RESULTS

The OLS and MLogit yielded similar MAE and MSE in the internal and external validation data sets. Both approaches also underestimated the uncertainty around the actual mean EQ-5D tariff producing tighter 95% CIs in both data sets.

CONCLUSIONS

The choice of algorithm will be dependent on the study aim. Individuals outside the United Kingdom may find it more useful to use the multinomial results, which can be used with different country-specific tariff valuations. However, these algorithms should not replace prospective collection of utility data.

Predicting SF-6D utility scores from the Oswestry disability index and numeric rating scales for back and leg pain

STUDY DESIGN

Cross-sectional cohort.

OBJECTIVE

The purpose of this study is to provide a model to allow estimation of utility from the Short Form (SF)-6D using data from the Oswestry Disability Index (ODI), Back Pain Numeric Rating Scale (BPNRS), and the Leg Pain Numeric Rating Scale (LPNRS).

SUMMARY OF BACKGROUND DATA

Cost-utility analysis provides important information about the relative value of interventions and requires a measure of utility not often available from clinical trial data. The ODI and numeric rating scales for back (BPNRS) and leg pain (LPNRS), are widely used disease-specific measures for health-related quality of life in patients with lumbar degenerative disorders. The purpose of this study is to provide a model to allow estimation of utility from the SF-6D using data from the ODI, BPNRS, and the LPNRS.

METHODS

SF-36, ODI, BPNRS, and LPNRS were prospectively collected before surgery, at 12 and 24 months after surgery in 2640 patients undergoing lumbar fusion for degenerative disorders. Spearman correlation coefficients for paired observations from multiple time points between ODI, BPNRS, and LPNRS, and SF-6D utility scores were determined. Regression modeling was done to compute the SF-6D score from the ODI, BPNRS, and LPNRS. Using a separate, independent dataset of 2174 patients in which actual SF-6D and ODI scores were available, the SF-6D was estimated for each subject and compared to their actual SF-6D.

RESULTS

In the development sample, the mean age was 52.5 +/- 15 years and 34% were male. In the validation sample, the mean age was 52.9 +/- 14.2 years and 44% were male. Correlations between the SF-6D and the ODI, BPNRS, and LPNRS were statistically significant (P < 0.0001) with correlation coefficients of 0.82, 0.78, and 0.72, respectively. The regression equation using ODI, BPNRS,and LPNRS to predict SF-6D had an R of 0.69 and a root mean square error of 0.076. The model using ODI alone had an R of 0.67 and a root mean square error of 0.078. The correlation coefficient between the observed and estimated SF-6D score was 0.80. In the validation analysis, there was no statistically significant difference (P = 0.11) between actual mean SF-6D (0.55 +/- 0.12) and the estimated mean SF-6D score (0.55 +/- 0.10) using the ODI regression model.

CONCLUSION

This regression-based algorithm may be used to predict SF-6D scores in studies of lumbar degenerative disease that have collected ODI but not utility scores.

Predicting an SF-6D preference-based score using MCS and PCS scores from the SF-12 or SF-36

BACKGROUND

The SF-6D preference-based scoring system was developed several years after the SF-12 and SF-36 instruments. A method to predict SF-6D scores from information in previous reports would facilitate backwards comparisons and the use of these reports in cost-effectiveness analyses.

METHODS

This report uses data from the 2001-2003 Medical Expenditures Panel Survey (MEPS), the Beaver Dam Health Outcomes Survey, and the National Health Measurement Study. SF-6D scores were modeled using age, sex, mental component summary (MCS) score, and physical component summary (PCS) score from the 2002 MEPS. The resulting SF-6D prediction equation was tested with the other datasets for groups of different sizes and groups stratified by age, MCS score, PCS score, sum of MCS and PCS scores, and SF-6D score.

RESULTS

The equation can be used to predict an average SF-6D score using average age, proportion female, average MCS score, and average PCS score. Mean differences between actual and predicted average SF-6D scores in out-of-sample tests was -0.001 (SF-12 version 1), -0.013 (SF-12 version 2), -0.007 (SF-36 version 1), and -0.010 (SF-36 version 2). Ninety-five percent credible intervals around these point estimates range from +/-0.045 for groups with 10 subjects to +/-0.008 for groups with more than 300 subjects. These results were consistent for a wide range of ages, MCS scores, PCS scores, sum of MCS and PCS scores, and SF-6D scores. SF-6D scores from the SF-36 and SF-12 from the same data set were found to be substantially different.

CONCLUSIONS

Simple equation predicts an average SF-6D preference-based score from widely published information.

Cost-effectiveness of eszopiclone for the treatment of adults with primary chronic insomnia

STUDY OBJECTIVE

To assess the cost-effectiveness of treatment with eszopiclone for chronic primary insomnia in adults.

METHODS

A model using patient-level data from a 6-month, double-blind, placebo-controlled, clinical trial (n = 824), combined with data from a claims database and published literature, was used to assess the quality-adjusted life years (QALYs) gained and costs associated with eszopiclone versus placebo in adults with primary insomnia. Quality of life data were collected during the trial via the SF-36, from which preference-based utility scores were derived using published algorithms. Medical and absenteeism costs, estimated via a retrospective analysis of a claims and absenteeism database, were assigned to patients based on the degree of severity of their insomnia, assessed via the Insomnia Severity Index collected in the clinical trial. Presenteeism costs (lost productivity while at work) were estimated from responses to the Work Limitation Questionnaire collected during the trial. Six-month gains in QALYs and costs for each treatment group were calculated to derive cost-effectiveness ratios. Uncertainty was addressed via univariate and multivariate sensitivity analyses.

RESULTS

Over the 6-month period, eszopiclone use resulted in a net gain of 0.0137 QALYs over placebo at an additional cost of $67, resulting in an incremental cost per QALY gained of slightly less than $5,000. When absenteeism and presenteeism costs were excluded, the cost-effectiveness ratio increased to approximately $33,000 per QALY gained, which is below the commonly used threshold of $50,000 used to define cost-effectiveness. Extensive sensitivity analyses indicate the results are generally robust.

CONCLUSION

Our model, based on efficacy data from a clinical trial, demonstrated eszopiclone was cost-effective for the treatment of primary insomnia in adults, especially when lost productivity costs were included.

Can we derive an 'exchange rate' between descriptive and preference-based outcome measures for stroke? Results from the transfer to utility (TTU) technique

Background

Stroke-specific outcome measures and descriptive measures of health-related quality of life (HRQoL) are unsuitable for informing decision-makers of the broader consequences of increasing or decreasing funding for stroke interventions. The quality-adjusted life year (QALY) provides a common metric for comparing interventions over multiple dimensions of HRQoL and mortality differentials. There are, however, many circumstances when – because of timing, lack of foresight or cost considerations – only stroke-specific or descriptive measures of health status are available and some indirect means of obtaining QALY-weights becomes necessary. In such circumstances, the use of regression-based transformations or mappings can circumvent the failure to elicit QALY-weights by allowing predicted weights to proxy for observed weights. This regression-based approach has been dubbed 'Transfer to Utility' (TTU) regression. The purpose of the present study is to demonstrate the feasibility and value of TTU regression in stroke by deriving transformations or mappings from stroke-specific and generic but descriptive measures of health status to a generic preference-based measure of HRQoL in a sample of Australians with a diagnosis of acute stroke. Findings will quantify the additional error associated with the use of condition-specific to generic transformations in stroke.

Methods

We used TTU regression to derive empirical transformations from three commonly used descriptive measures of health status for stroke (NIHSS, Barthel and SF-36) to a preference-based measure (AQoL) suitable for attaching QALY-weights to stroke disease states; based on 2570 observations drawn from a sample of 859 patients with stroke.

Results

Transformations from the SF-36 to the AQoL explained up to 71.5% of variation in observed AQoL scores. Differences between mean predicted and mean observed AQoL scores from the 'severity-specific' item- and subscale-based SF-36 algorithms and from the 'moderate to severe' index- and item-based Barthel algorithm were neither clinically nor statistically significant when 'low severity' SF-36 transformations were used to predict AQoL scores for patients in the NIHSS = 0 and NIHSS = 1–5 subgroups and when 'moderate to severe severity' transformations were used to predict AQoL scores for patients in the NIHSS ≥ 6 subgroup. In contrast, the difference between mean predicted and mean observed AQoL scores from the NIHSS algorithms and from the 'low severity' Barthel algorithms reached levels that could mask minimally important differences on the AQoL scale.

Conclusion

While our NIHSS to AQoL transformations proved unsuitable for most applications, our findings demonstrate that stroke-relevant outcome measures such as the SF-36 and Barthel Index can be adequately transformed to preference-based measures for the purposes of economic evaluation.

Converting the SF-12 into the EQ-5D: an empirical comparison of methodologies

BACKGROUND

For cost-utility analysis, analysts need a measure that summarizes health-status utilities in a single index of health-related quality of life (HR-QOL). It is common to find in clinical studies that only an HR-QOL profile measure such as the SF-36 is included, but not the summary HR-QOL index. Therefore, the economist's usual practice is to reprocess the profile data into a single index format. Several 'after-market' tools are available to convert the SF-36 or SF-12 into a single form with or without utility-weighting metric property. However, there has been no consensus with regard to a regression method that should be recommended for such a mapping task.

OBJECTIVE

To report on the performance of different regression methods that have previously been applied to the conversion of SF-12 data in the analysis of a single common dataset. The mapping between the SF-12 and EQ-5D is the focus.

METHODS

The data were adopted from the Medical Expenditure Panel Survey 2003 where 19 678 adults completed both EQ-5D and SF-12 questionnaires. Four econometric techniques, namely ordinary least squares (OLS), censored least absolute deviation, multinomial logit model and two-part model regressions were investigated together with two main types of model specifications: item-based and summary score-based. The performance of each examined model was judged by various criteria, including its estimated mean, the size of mean absolute error and the number of errors.

RESULTS

Among four compared econometric techniques, OLS regression was the most accurate model in estimating the group mean. Models with item-based model specification performed better than those with summary score-based regardless of the chosen econometric technique. Nevertheless, the accuracy of OLS deteriorates in older and less healthy subgroups. The results also suggested that the two-part model, which addresses the heterogeneity issue, performs better in these vulnerable subgroups.

CONCLUSIONS

None of the mapping methods included in the current study are suitable for estimating at the individual level. The methodology exemplified here has wider applicability and might just as readily be applied to other members of the SF family or indeed to other profile measures of HR-QOL. However, it is recommended that a preference-based, single index measure of HR-QOL should be included in the clinical studies for the purpose of economic evaluation.

Deriving an algorithm to convert the eight mean SF-36 dimension scores into a mean EQ-5D preference-based score from published studies (where patient level data are not available)

OBJECTIVE

The objective of the study was to derive a method to predict a mean cohort EQ-5D preference-based index score using published mean statistics of the eight dimension scores describing the SF-36 health profile.

METHODS

Ordinary least square regressions models are derived using patient level data (n = 6350) collected during 12 clinical studies. The models were compared for goodness of fit using standard techniques such as variance explained, the magnitude of errors in predicted values, and the proportion of values within the minimal important difference of the EQ-5D. Predictive abilities were also compared using summary statistics from both within-sample subgroups and published studies.

RESULTS

The models obtained explained more than 56% of the variance in the EQ-5D scores. The mean predicted EQ-5D score was correct to within two decimal places for all models and the absolute error for the individual predicted values was approximately 0.13. Using summary statistics to predict within-sample subgroup mean EQ-5D scores, the mean errors (mean absolute errors) ranged from 0.021 to 0.077 (0.045-0.083). These statistics for the out-of-sample published data sets ranged from 0.048 to 0.099 (0.064-0.010).

CONCLUSIONS

The models provided researchers with a mechanism to estimate EQ-5D utility data from published mean dimension scores. This research is unique in that it uses mean statistics from published studies to validate the results. While further research is required to validate the results in additional health conditions, the algorithms can be used to derive additional preference-based measures for use in economic analyses.

Sociodemographic and clinical correlates of utility scores in Alzheimer's disease

OBJECTIVES

To examine the relationship between psychiatric symptoms, cognitive performance, functional capacity and quality of life in Alzheimer's disease (AD), and change in the Health Utilities Index (HUI)-Mark III, a widely used generic, multiattribute preference-based health-status classification system.

METHODS

Follow-up data were obtained from caregiver proxy raters at 3, to 6, and 9-months postrandom assignment concerning 421 patients with AD, living with at least one caregiver in a noninstitutional setting, who participated in the Clinical Antipsychotic Trial of Intervention Effectiveness-AD of antipsychotic medication. Spearman rank correlations, multivariate linear regression, and mixed modeling were used to examine the correlates of change in the HUI.

RESULTS

HUI scores decreased by an average of -0.061 over 9 months. Analysis revealed weak bivariate, and largely, nonsignificant multivariate relationships between change in HUI scores and sociodemographic characteristics, psychiatric symptoms, and cognitive performance. There were highly significant associations between decreases in health utilities and change in the AD Cooperative Study for Activities of Daily Living scale (ADCS-ADL) and AD-Related Quality of Life (ADRQoL) (both P < 0.001), even after controlling for other factors. Adjusted R(2) values ranged from 0.14 to 0.20.

CONCLUSION

In AD patients requiring antipsychotic treatment, only weak relationships were found between changes in the HUI and sociodemographic and clinical indicators. While functional capability and quality of life showed more significant associations, less than 20% of the variance in health utility could be explained. Significant cognitive impairment and the need to rely on proxy raters may limit the usefulness of utility measurement in AD patients with serious behavioral symptoms.

Cost-effectiveness of nurse-led disease management for heart failure in an ethnically diverse urban community

BACKGROUND

Randomized, controlled trials have shown that nurse-led disease management for patients with heart failure can reduce hospitalizations. Less is known about the cost-effectiveness of these interventions.

OBJECTIVE

To estimate the cost-effectiveness of a nurse-led disease management intervention over 12 months, implemented in a randomized, controlled effectiveness trial.

DESIGN

Cost-effectiveness analysis conducted alongside a randomized trial.

DATA SOURCES

Medical costs from administrative records, and self-reported quality of life and nonmedical costs from patient surveys.

PARTICIPANTS

Patients with systolic dysfunction recruited from ambulatory clinics in Harlem, New York.

TIME HORIZON

12 months.

PERSPECTIVE

Societal and payer.

INTERVENTION

12-month program that involved 1 face-to-face encounter with a nurse and regular telephone follow-up.

OUTCOME MEASURES

Quality of life as measured by the Health Utilities Index Mark 3 and EuroQol-5D and cost-effectiveness as measured by the incremental cost-effectiveness ratio (ICER).

RESULTS OF BASE-CASE ANALYSIS

Costs and quality of life were higher in the nurse-managed group than the usual care group. The ICERs over 12 months were $17,543 per EuroQol-5D-based quality-adjusted life-year (QALY) and $15,169 per Health Utilities Index Mark 3-based QALY (in 2001 U.S. dollars).

RESULTS OF SENSITIVITY ANALYSIS

From a payer perspective, the ICER ranged from $3673 to $4495 per QALY. Applying national prices in place of New York City prices yielded a societal ICER of $13,460 to $15,556 per QALY. Cost-effectiveness acceptability curves suggest that the intervention was most likely cost-effective for patients with less severe (New York Heart Association classes I to II) heart failure.

LIMITATION

The trial was conducted in an ethnically diverse, inner-city neighborhood; thus, results may not be generalizable to other communities.

CONCLUSION

Over 12 months, the nurse-led disease management program was a reasonably cost-effective way to reduce the burden of heart failure in this community.

Do estimates of cost-utility based on the EQ-5D differ from those based on the mapping of utility scores?

Background

Mapping has been used to convert scores from condition-specific measures into utility scores, and to produce estimates of cost-effectiveness. We sought to compare the QALY gains, and incremental cost per QALY estimates, predicted on the basis of mapping to those based on actual EQ-5D scores.

Methods

In order to compare 4 different interventions 389 individuals were asked to complete both the EQ-5D and the Western Ontartio and McMaster Universities Osteoarthritis Index (WOMAC) at baseline, 6, 12, and 24 months post-intervention. Using baseline data various mapping models were developed, where WOMAC scores were used to predict the EQ-5D scores. The performance of these models was tested by predicting the EQ-5D post-intervention scores. The preferred model (that with the lowest mean absolute error (MAE)) was used to predict the EQ-5D scores, at all time points, for individuals who had complete WOMAC and EQ-5D data. The mean QALY gain associated with each intervention was calculated, using both actual and predicted EQ-5D scores. These QALY gains, along with previously estimated changes in cost, were also used to estimate the actual and predicted incremental cost per QALY associated with each of the four interventions.

Results

The EQ-5D and the WOMAC were completed at baseline by 348 individuals, and at all time points by 259 individuals. The MAE in the preferred model was 0.129, and the mean QALY gains for each of the four interventions was predicted to be 0.006, 0.058, 0.058, and 0.136 respectively, compared to the actual mean QALY gains of 0.087, 0.081, 0.120, and 0.149. The most effective intervention was estimated to be associated with an incremental cost per QALY of £6,068, according to our preferred model, compared to £13,154 when actual data was used.

Conclusion

We found that actual QALY gains, and incremental cost per QALY estimates, differed from those predicted on the basis of mapping. This suggests that though mapping may be of value in predicting the cost-effectiveness of interventions which have not been evaluated using a utility measure, future studies should be encouraged to include a method of actual utility measurement.

Trial registration

Current Controlled Trials ISRCTN93206785

Do estimates of cost-utility based on the EQ-5D differ from those based on the mapping of utility scores?

BACKGROUND

Mapping has been used to convert scores from condition-specific measures into utility scores, and to produce estimates of cost-effectiveness. We sought to compare the QALY gains, and incremental cost per QALY estimates, predicted on the basis of mapping to those based on actual EQ-5D scores.

METHODS

In order to compare 4 different interventions 389 individuals were asked to complete both the EQ-5D and the Western Ontartio and McMaster Universities Osteoarthritis Index (WOMAC) at baseline, 6, 12, and 24 months post-intervention. Using baseline data various mapping models were developed, where WOMAC scores were used to predict the EQ-5D scores. The performance of these models was tested by predicting the EQ-5D post-intervention scores. The preferred model (that with the lowest mean absolute error (MAE)) was used to predict the EQ-5D scores, at all time points, for individuals who had complete WOMAC and EQ-5D data. The mean QALY gain associated with each intervention was calculated, using both actual and predicted EQ-5D scores. These QALY gains, along with previously estimated changes in cost, were also used to estimate the actual and predicted incremental cost per QALY associated with each of the four interventions.

RESULTS

The EQ-5D and the WOMAC were completed at baseline by 348 individuals, and at all time points by 259 individuals. The MAE in the preferred model was 0.129, and the mean QALY gains for each of the four interventions was predicted to be 0.006, 0.058, 0.058, and 0.136 respectively, compared to the actual mean QALY gains of 0.087, 0.081, 0.120, and 0.149. The most effective intervention was estimated to be associated with an incremental cost per QALY of pound6,068, according to our preferred model, compared to pound13,154 when actual data was used.

CONCLUSION

We found that actual QALY gains, and incremental cost per QALY estimates, differed from those predicted on the basis of mapping. This suggests that though mapping may be of value in predicting the cost-effectiveness of interventions which have not been evaluated using a utility measure, future studies should be encouraged to include a method of actual utility measurement.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN93206785.

Item-based versus subscale-based mappings from the SF-36 to a preference-based quality of life measure

OBJECTIVES

Algorithms for mapping descriptive measures of health status into preference-based measures are now widely available and their application in economic evaluation is increasingly commonplace. Existing algorithms make use of scale, subscale, or item scores on descriptive measures. Item-based algorithms entail fewer restrictions than their scale or subscale-based equivalents but are subject to problems in estimation and application. The objective of the present study is to quantify any loss of predictive validity associated with using subscale or scale scores (rather than item scores) to derive conversion algorithms.

METHODS

Multiple linear regression methods to derive item-based, subscale-based, and scale-based algorithms for mapping SF-36 data into Assessment of Quality of Life (AQoL) utility scores in a stratified sample of persons aged more than 16 years and resident in Victoria, Australia. The theoretical consistency and predictive validity of competing algorithms is evaluated against criteria reflecting the intended use of predicted utility scores.

RESULTS

Three mappings were suitable for between-group comparisons. There was no discernible increase in error associated with a move from the item-based mapping to either the subscale- or scale-based mapping.

CONCLUSIONS

Our results do not support the hypothesis that fewer restrictions on functional form necessarily result in a lower magnitude of error when predicting between-group differences. Rather, it appears that the subscale-based mapping offers a good compromise--requiring fewer restrictions on the form of the relationship between SF-36 responses and the AQoL utility score than the scale-based mapping and permitting a more efficient use of SF-36 data than the item-based mapping.

Comparing the Incomparable? A Systematic Review of Competing Techniques for Converting Descriptive Measures of Health Status into QALY-Weights

Background . Algorithms for converting descriptive measures of health status into quality-adjusted life year (QALY)—weights are now widely available, and their application in economic evaluation is increasingly commonplace. The objective of this study is to describe and compare existing conversion algorithms and to highlight issues bearing on the derivation and interpretation of the QALY-weights so obtained. Methods . Systematic review of algorithms for converting descriptive measures of health status into QALY-weights. Results . The review identified a substantial body of literature comprising 46 derivation studies and 16 studies that provided evidence or commentary on the validity of conversion algorithms. Conversion algorithms were derived using 1 of 4 techniques: 1) transfer to utility regression, 2) response mapping, 3) effect size translation, and 4) “revaluing” outcome measures using preference-based scaling techniques. Although these techniques differ in their methodological/theoretical tradition, data requirements, and ease of derivation and application, the available evidence suggests that the sensitivity and validity of derived QALY-weights may be more dependent on the coverage and sensitivity of measures and the disease area/patient group under evaluation than on the technique used in derivation. Conclusions . Despite the recent proliferation of conversion algorithms, a number of questions bearing on the derivation and interpretation of derived QALY-weights remain unresolved. These unresolved issues suggest directions for future research in this area. In the meantime, analysts seeking guidance in selecting derived QALY-weights should consider the validity and feasibility of each conversion algorithm in the disease area and patient group under evaluation rather than restricting their choice to weights from a particular derivation technique.

The e-Mental Health Consultation Service: providing enhanced primary-care mental health services through telemedicine

This article describes the University of California, Davis Medical Center eMental Health Consultation Service, a program designed to integrate tele-mental health clinical services, provider-to-provider consultation, and provider distance education. During the first year of operation, consultations were provided for 289 cases. The most common diagnoses among children were for attention-deficit hyperactivity disorder-spectrum problems. Among the adult patients, mood disorders were most common. A convenience sample of 33 adult patients who completed the SF-12 health status measure showed significant improvements in mental health status at 3-6 months of follow-up. This model of comprehensive rural outpatient primary mental health care delivered at a distance shows promise for wider application and deserves further study.

Which health-related quality of life score? A comparison of alternative utility measures in patients with Type 2 diabetes in the ADVANCE trial

Background

Diabetes has a high burden of illness both in life years lost and in disability through related co-morbidities. Accurate assessment of the non-mortality burden requires appropriate health-related quality of life and summary utility measures of which there are several contenders. The study aimed to measure the impact of diabetes on various health-related quality of life domains, and compare several summary utility measures.

Methods

In the ADVANCE (Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation) study, 978 Australian patients with Type 2 diabetes completed two health-related quality of life questionnaires at baseline: the EQ-5D and the SF-36v2, from which nine summary utility measures were calculated, and compared. The algorithms were grouped into four classes: (i) based on the EQ-5D; (ii) using fewer items than those in the SF-12 (iii) using the items in the SF-12; and (iv) using all items of the SF-36.

Results

Overall health-related quality of life of the subjects was good (mean utility ranged from 0.68 (±0.08) to 0.85(±0.14) over the nine utility measures) and comparable to patients without diabetes. Summary indices were well correlated with each other (r = 0.76 to 0.99), and showed lower health-related quality of life in patients with major diabetes-related events such as stroke or myocardial infarction. Despite the smaller number of items used in the scoring of the EQ-5D, it generally performed at least as well as SF-36 based methods. However, all utility measures had some limitation such as limited range or ceiling effects.

Conclusion

The summary utility measures showed good agreement, and showed good discrimination between major and minor health state changes. However, EQ-5D based measures performed as well and are generally simpler to use.

Measuring preferences for cost-utility analysis: how choice of method may influence decision-making

Preferences for health are required when the economic value of healthcare interventions are assessed within the framework of cost-utility analysis. The objective of this paper was to review alternative methods for preference measurement and to evaluate the extent to which the method may affect healthcare decision-making. Two broad approaches to preference measurement that provide societal health state values were considered: (i) direct measurement; and (ii) preference-based health state classification systems. Among studies that compared alternative preference-based systems, the EQ-5D tended to provide larger change scores and more favourable cost-effectiveness ratios than the Health Utilities Index (HUI)-2 and -3, while the SF-6D provided smaller change scores and less favourable ratios than the other systems. However, these patterns may not hold for all applications. Empirical evidence comparing systems and decision-making impact suggests that preferences will have the greatest impact on economic analyses when chronic conditions or long-term sequelae are involved. At present, there is no clearly superior method, and further study of cost-effectiveness ratios from alternative systems is needed to evaluate system performance. Although there is some evidence that incremental cost-effectiveness ratio (ICER) thresholds (e.g. $US50,000 per QALY gained) are used in decision-making, they are not strictly applied. Nonetheless, as ICERs rise, the probability of acceptance of a new therapy is likely to decrease, making the differences in QALYs obtained using alternative methods potentially meaningful. It is imperative that those conducting cost-utility analyses characterise the impact that uncertainty in health state values has on the economic value of the interventions studied. Consistent reporting of such analyses would provide further insight into the policy implications of preference measurement.

Self-reported discrimination and mental health status among African descendants, Mexican Americans, and other Latinos in the New Hampshire REACH 2010 Initiative: the added dimension of immigration

OBJECTIVES

We examined whether self-reported racial discrimination was associated with mental health status and whether this association varied with race/ethnicity or immigration status.

METHODS

We performed secondary analysis of a community intervention conducted in 2002 and 2003 for the New Hampshire Racial and Ethnic Approaches to Community Health 2010 Initiative, surveying African descendants, Mexican Americans, and other Latinos. We assessed mental health status with the Mental Component Summary (MCS12) of the Medical Outcomes Study Short Form 12, and measured discrimination with questions related to respondents' ability to achieve goals, discomfort/anger at treatment by others, and access to quality health care.

RESULTS

Self-reported discrimination was associated with a lower MCS12 score. Additionally, the strength of the association between self-reported health care discrimination and lower MCS12 score was strongest for African descendants, then Mexican Americans, then other Latinos. These patterns may be explained by differences in how long a respondent has lived in the United States. Furthermore, the association of health care discrimination with lower MCS12 was weaker for recent immigrants.

CONCLUSIONS

Discrimination may be an important predictor of poor mental health status among Black and Latino immigrants. Previous findings of decreasing mental health status as immigrants acculturate might partly be related to experiences with racial discrimination.

Relative disutilities of 47 risk factors and conditions assessed with seven preference-based health status measures in a national U.S. sample: toward consistency in cost-effectiveness analyses

BACKGROUND

Preference-based health measures yield summary scores that are compatible with cost-effectiveness analyses. There is limited comparative information, however, about how different measures weight health conditions in the U.S. population.

METHODS

We examined data from 11,421 adults in the 2000 Medical Expenditure Panel Survey, a nationally representative sample of the U.S. general population, using information on sociodemographics (age, gender, race/ethnicity, income, and education), health status (EQ-5D, EQ-VAS, and SF-12), 4 risk factors (smoking, overweight, obesity, and lacking health insurance), and 43 conditions. From the EQ-5D, we derived summary scores using U.K. [EQ(UK)] and U.S. weights. From the SF-12 we derived SF-6D, and regression-predicted EQ-5D (U.S. and U.K. weights) and Health Utility Index scores. Each of the 7 preference measures was regressed on each of the 47 problems (risk factors and conditions) to determine the disutility associated with the problem, adjusting for socio-demographics.

RESULTS

The adjusted disutilities averaged across the 47 problems for the 7 preference measures ranged from 0.059 for the SF-6D to 0.104 for the EQ(UK). Correlations between each of the measures of the adjusted disutilities ranged from 0.85-1.0. Standardization, using linear regression, attenuated between measure differences in disutilities.

CONCLUSIONS

Absolute incremental cost-effectiveness analyses of a given problem would likely vary depending on the measure used, whereas the relative ordering of incremental cost-effectiveness analyses of a series of problems would likely be similar regardless of the measure chosen, as long as the same measure is used in each series of analyses. Absolute consistency across measures may be enhanced by standardization.

Estimating utility values for health states of overweight and obese individuals using the SF-36

OBJECTIVE

To use health-related quality-of-life (HRQoL) data from the Australian 1995 National Health Survey to estimate the impact of obesity (as measured by body mass index or BMI) on utility and quality-adjusted life expectancy (QALE).

METHOD

SF-36 responses from 12,661 individuals in the general population were transformed into utility values using the SF-6D algorithm developed by Brazier and colleagues. Separate regression analyses for males and females were used to examine the impact of BMI and five obesity-related medical conditions (diabetes, coronary heart disease, depression, musculoskeletal disorders, and cancer) on utility. The utility estimates were used to provide indicative estimates of the decrease in QALE associated with being overweight or obese.

RESULTS

There was a statistically significant negative relationship between BMI and utility for males and females. For males (females), the marginal effect of a one-unit increase in BMI was associated with a -0.0024 (-0.0034) decrement in utility. Based on these estimates, a non-smoking male (female) aged 40 years who is obese can expect 7.2 (8.7) years less of QALE over their remaining lifetime.

CONCLUSIONS

Results suggest that BMI is negatively associated with utility. Evaluation of policies designed to prevent or treat obesity should capture HRQoL as an outcome.

Are decisions using cost-utility analyses robust to choice of SF-36/SF-12 preference-based algorithm?

Background

Cost utility analysis (CUA) using SF-36/SF-12 data has been facilitated by the development of several preference-based algorithms. The purpose of this study was to illustrate how decision-making could be affected by the choice of preference-based algorithms for the SF-36 and SF-12, and provide some guidance on selecting an appropriate algorithm.

Methods

Two sets of data were used: (1) a clinical trial of adult asthma patients; and (2) a longitudinal study of post-stroke patients. Incremental costs were assumed to be $2000 per year over standard treatment, and QALY gains realized over a 1-year period. Ten published algorithms were identified, denoted by first author: Brazier (SF-36), Brazier (SF-12), Shmueli, Fryback, Lundberg, Nichol, Franks (3 algorithms), and Lawrence. Incremental cost-utility ratios (ICURs) for each algorithm, stated in dollars per quality-adjusted life year ($/QALY), were ranked and compared between datasets.

Results

In the asthma patients, estimated ICURs ranged from Lawrence's SF-12 algorithm at $30,769/QALY (95% CI: 26,316 to 36,697) to Brazier's SF-36 algorithm at $63,492/QALY (95% CI: 48,780 to 83,333). ICURs for the stroke cohort varied slightly more dramatically. The MEPS-based algorithm by Franks et al. provided the lowest ICUR at $27,972/QALY (95% CI: 20,942 to 41,667). The Fryback and Shmueli algorithms provided ICURs that were greater than $50,000/QALY and did not have confidence intervals that overlapped with most of the other algorithms. The ICUR-based ranking of algorithms was strongly correlated between the asthma and stroke datasets (r = 0.60).

Conclusion

SF-36/SF-12 preference-based algorithms produced a wide range of ICURs that could potentially lead to different reimbursement decisions. Brazier's SF-36 and SF-12 algorithms have a strong methodological and theoretical basis and tended to generate relatively higher ICUR estimates, considerations that support a preference for these algorithms over the alternatives. The "second-generation" algorithms developed from scores mapped from other indirect preference-based measures tended to generate lower ICURs that would promote greater adoption of new technology. There remains a need for an SF-36/SF-12 preference-based algorithm based on the US general population that has strong theoretical and methodological foundations.

Mapping the SF-12 to the HUI3 and VAS in a Managed Care Population

BACKGROUND

Transforming generic health-related quality of life (HRQOL) instruments to a summary utility index is useful for deriving quality-adjusted life years (QALY) in any cost/QALY analysis.

OBJECTIVE

The purpose of this study was to investigate the role of the SF-12 in predicting utility scores derived from Health Utility Index (HUI3) and Visual Analog Scale (VAS).

METHOD

Data were obtained from a survey of 6923 managed care patients in the United States, aged 18 to 93 years, selected by strata of medication usage (at least 1 medication in target year, 5 or more medications, target medications, and both). The SF-12 was used to assess self-reported HRQOL. Utility was measured by the HUI3 and a VAS. The SF-12 items were used to predict HUI3 and VAS scores using ordinary least square regressions, with sociodemographic covariates. A second model entered each SF-12 item as categorized responses. A third model used the Physical Composite and Mental Composite scores to predict HUI3 and VAS scores.

RESULTS

The SF-12 items and sociodemographic covariates accounted for 35% to 55% of the variations in HUI3 and VAS scores, respectively. Age and most SF-12 items were significantly (P < 0.0001) associated with both utility scores in all 3 models.

CONCLUSIONS

This research provides support that an algorithm can be derived from the SF-12 to estimate utility scores based on the HUI3 and VAS for studies in populations where utility has not or cannot be measured directly.