Mapping MOS Sleep Scale scores to SF-6D utility index

An Updated Systematic Review of Studies Mapping (or Cross-Walking) Measures of Health-Related Quality of Life to Generic Preference-Based Measures to Generate Utility Values

BACKGROUND

Mapping is an increasingly common method used to predict instrument-specific preference-based health-state utility values (HSUVs) from data obtained from another health-related quality of life (HRQoL) measure. There have been several methodological developments in this area since a previous review up to 2007.

OBJECTIVE

To provide an updated review of all mapping studies that map from HRQoL measures to target generic preference-based measures (EQ-5D measures, SF-6D, HUI measures, QWB, AQoL measures, 15D/16D/17D, CHU-9D) published from January 2007 to October 2018.

DATA SOURCES

A systematic review of English language articles using a variety of approaches: searching electronic and utilities databases, citation searching, targeted journal and website searches.

STUDY SELECTION

Full papers of studies that mapped from one health measure to a target preference-based measure using formal statistical regression techniques.

DATA EXTRACTION

Undertaken by four authors using predefined data fields including measures, data used, econometric models and assessment of predictive ability.

RESULTS

There were 180 papers with 233 mapping functions in total. Mapping functions were generated to obtain EQ-5D-3L/EQ-5D-5L-EQ-5D-Y (n = 147), SF-6D (n = 45), AQoL-4D/AQoL-8D (n = 12), HUI2/HUI3 (n = 13), 15D (n = 8) CHU-9D (n = 4) and QWB-SA (n = 4) HSUVs. A large number of different regression methods were used with ordinary least squares (OLS) still being the most common approach (used ≥ 75% times within each preference-based measure). The majority of studies assessed the predictive ability of the mapping functions using mean absolute or root mean squared errors (n = 192, 82%), but this was lower when considering errors across different categories of severity (n = 92, 39%) and plots of predictions (n = 120, 52%).

CONCLUSIONS

The last 10 years has seen a substantial increase in the number of mapping studies and some evidence of advancement in methods with consideration of models beyond OLS and greater reporting of predictive ability of mapping functions.

Patient-reported outcomes to initiate a provider-patient dialog for the management of hip and knee osteoarthritis

PURPOSE

Although many treatment guidelines exist for hip and knee osteoarthritis (OA), uptake in clinical practice is typically low. Valid patient-reported outcome measures (PROs) that can be easily used in the clinic could aid implementation and evaluation of treatment recommendations, and the tracking of symptoms and function over time. This project responded to a 2012 Call to Action of the Chronic Osteoarthritis Management Initiative of the United States Bone and Joint Initiative; we aimed to develop a tiered list of recommended PROs that could be feasibly applied in common clinical settings, across four domains of pain, function, fatigue, and sleep.

METHODS

PROs were identified through a focused literature review. Clinicians and researchers with OA expertise evaluated each measure' feasibility for use in routine clinical practice, followed by meaningfulness in assessing OA outcomes. Eligible PROs were categorized by domain and ranked into Tiers One (very brief measures for initial use in clinical settings), Two (brief measures with more in-depth assessment), and Three (most detailed assessment).

RESULTS

Total PROs identified were 172 for pain, 160 for function, 55 for fatigue, and 60 for sleep. Of these, 9 pain, 7 function, 7 fatigue, and 8 sleep PROs were ranked into one of three tiers.

CONCLUSIONS

This three-tiered list of recommended PROs provides a basis for tools to systematically track outcomes, facilitate provider-patient dialog, and guide treatment for hip or knee OA. Research is needed to test the utility and feasibility of systematic implementation of these measures in primary care and specialty clinical settings.

The cost-effectiveness of duloxetine in chronic low back pain: a US private payer perspective

OBJECTIVE

To assess the cost-effectiveness of duloxetine in the treatment of chronic low back pain (CLBP) from a US private payer perspective.

METHODS

A cost-utility analysis was undertaken for duloxetine and seven oral post-first-line comparators, including nonsteroidal anti-inflammatory drugs (NSAIDs), weak and strong opioids, and an anticonvulsant. We created a Markov model on the basis of the National Institute for Health and Clinical Excellence model documented in its 2008 osteoarthritis clinical guidelines. Health states included treatment, death, and 12 states associated with serious adverse events (AEs). We estimated treatment-specific utilities by carrying out a meta-analysis of pain scores from CLBP clinical trials and developing a transfer-to-utility equation using duloxetine CLBP patient-level data. Probabilities of AEs were taken from the National Institute for Health and Clinical Excellence model or estimated from osteoarthritis clinical trials by using a novel maximum-likelihood simulation technique. Costs were gathered from Red Book, Agency for Healthcare Research and Quality's Healthcare Cost and Utilization Project database, the literature, and, for a limited number of inputs, expert opinion. The model performed one-way and probabilistic sensitivity analyses and generated incremental cost-effectiveness ratios (ICERs) and cost acceptability curves.

RESULTS

The model estimated an ICER of $59,473 for duloxetine over naproxen. ICERs under $30,000 were estimated for duloxetine over non-NSAIDs, with duloxetine dominating all strong opioids. In subpopulations at a higher risk of NSAID-related AEs, the ICER over naproxen was $33,105 or lower.

CONCLUSIONS

Duloxetine appears to be a cost-effective post-first-line treatment for CLBP compared with all but generic NSAIDs. In subpopulations at risk of NSAID-related AEs, it is particularly cost-effective.

Converting condition-specific measures into preference-based outcomes for use in economic evaluation

Quality-adjusted life-years (QALYs) play an important role in reimbursement decisions when one of the criteria is the cost-effectiveness of the health technology. While for many generic QALYs (e.g., based on the EQ-5D) are viewed as the gold standard, there has been a considerable increase in interest in using condition-specific data to generate QALYs. There are two main methods: mapping from the condition-specific data to a generic health-related quality of life measure; and direct valuation of condition-specific health states. Whether one believes condition-specific data are useful even if generic QALY data are available, or simply that condition-specific data are helpful in the absence of generic measures of health-related quality of life, it is timely to review recent research activity directed at making greater use of condition-specific data to inform assessments of cost-effectiveness.

Evaluation of content on EQ-5D as compared to disease-specific utility measures

PURPOSE

The goal of this study was to appraise the extent of unique content on disease-specific preference-based measures (DSPMs) when contrasted with the EQ-5D using published studies and to inform whether EQ-5D could be inadequate as a utility measure in its content coverage for a given disease-specific application.

METHODS

A structured search of published literature was performed using PubMed and EMBASE/Medline database from Jan 1, 1990 to Mar 31, 2011. Articles were eligible for inclusion if algorithms were developed to convert components from disease-specific measures into utility scores.

RESULTS

Of 1,029 articles identified, 50 studies satisfied the inclusion criteria. The most frequent conditions where DSPMs were developed included cancer (12 studies), coronary artery disease (4 studies), osteoarthritis, rheumatoid arthritis (3 studies of each), obesity, and stroke (2 studies of each). Most studies involved mapping items or scores from disease-specific non-preference-based measures onto a preference-based measure of health such as the EQ-5D. A substantial number of DSPMs appeared to include unique content not covered by EQ-5D dimensions.

CONCLUSIONS

Several conditions were identified as potential areas where the richness of the EQ-5D descriptive system could be enhanced. It is yet unclear whether added dimension(s) would contribute unique explained variance to a utility score. Given the resources required to rigorously develop a utility measure, the need for such measures should be carefully vetted.

Mapping FACT-melanoma quality-of-life scores to EQ-5D health utility weights

OBJECTIVES

We sought to develop a mapping function from functional assessment of cancer therapy-melanoma (FACT-M) quality of life scores to the EuroQol-5D (EQ-5D) utility scores.

METHODS

FACT-M and EQ-5D scores were collected during a prospective study of melanoma-related quality of life at a tertiary cancer care center in the United States. The study sample was divided into development and validation datasets with equal distributions by cancer stage and treatment status. Censored Least Absolute Deviation (CLAD) and Ordinary Least Squares (OLS) regression analyses were performed using the developmental dataset to derive mapping functions, and model performance was examined through comparisons of residuals and measures of fit in the validation dataset. Exploratory analyses examined the predictive ability of clinical factors and individual subscales.

RESULTS

Of 273 patients, 75 were undergoing treatment with 198 in follow-up surveillance. Relatively even distributions were observed by melanoma stage: I/II (n = 102), III (n = 100), and IV (n = 71). OLS regression resulted in a mapping function of EQ-5D = 0.0037*FACT-M+0.2238 with an R(2) 0.499. CLAD regression resulted in a mapping function of EQ-5D = 0.0042*FACT-M+0.1648 with pseudo R(2) 0.328. When applied to the validation dataset, correlations between observed and predicted values resulted in identical coefficients (r = 0.824, P < 0.001). Though the mapping functions were similar, residuals were smaller at the 20th, 40th, and 60th percentiles using the OLS model. The CLAD derived mapping function resulted in smaller residuals only for patients whose EQ-5D = 1.

CONCLUSIONS

The OLS mapping function demonstrated better predictive ability and will facilitate the derivation of utilities when direct population preference measures are not available.

Mapping onto Eq-5 D for patients in poor health

BACKGROUND

An increasing amount of studies report mapping algorithms which predict EQ-5 D utility values using disease specific non-preference-based measures. Yet many mapping algorithms have been found to systematically overpredict EQ-5 D utility values for patients in poor health. Currently there are no guidelines on how to deal with this problem. This paper is concerned with the question of why overestimation of EQ-5 D utility values occurs for patients in poor health, and explores possible solutions.

METHOD

Three existing datasets are used to estimate mapping algorithms and assess existing mapping algorithms from the literature mapping the cancer-specific EORTC-QLQ C-30 and the arthritis-specific Health Assessment Questionnaire (HAQ) onto the EQ-5 D. Separate mapping algorithms are estimated for poor health states. Poor health states are defined using a cut-off point for QLQ-C30 and HAQ, which is determined using association with EQ-5 D values.

RESULTS

All mapping algorithms suffer from overprediction of utility values for patients in poor health. The large decrement of reporting 'extreme problems' in the EQ-5 D tariff, few observations with the most severe level in any EQ-5 D dimension and many observations at the least severe level in any EQ-5 D dimension led to a bimodal distribution of EQ-5 D index values, which is related to the overprediction of utility values for patients in poor health. Separate algorithms are here proposed to predict utility values for patients in poor health, where these are selected using cut-off points for HAQ-DI (> 2.0) and QLQ C-30 (< 45 average of QLQ C-30 functioning scales). The QLQ-C30 separate algorithm performed better than existing mapping algorithms for predicting utility values for patients in poor health, but still did not accurately predict mean utility values. A HAQ separate algorithm could not be estimated due to data restrictions.

CONCLUSION

Mapping algorithms overpredict utility values for patients in poor health but are used in cost-effectiveness analyses nonetheless. Guidelines can be developed on when the use of a mapping algorithms is inappropriate, for instance through the identification of cut-off points. Cut-off points on a disease specific questionnaire can be identified through association with the causes of overprediction. The cut-off points found in this study represent severely impaired health. Specifying a separate mapping algorithm to predict utility values for individuals in poor health greatly reduces overprediction, but does not fully solve the problem.

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.

Modeling the economic and health consequences of managing chronic osteoarthritis pain with opioids in Germany: comparison of extended-release oxycodone and OROS hydromorphone

OBJECTIVE

The Osmotic controlled-Release Oral delivery System (OROS) hydromorphone ensures continuous release of hydromorphone over 24 hours. It is anticipated that this will facilitate optimal pain relief, improve quality of sleep and compliance. This simulation compared managing chronic osteoarthritis pain with once-daily OROS hydromorphone with an equianalgesic dose of extended-release (ER) oxycodone administered two or three times a day.

METHODS

This discrete event simulation follows patients for a year after initiating opioid treatment. Pairs of identical patients are created; one receives OROS hydromorphone the other ER oxycodone; undergo dose adjustments and after titration can be dissatisfied or satisfied, suffer adverse events, pain recurrence, or discontinue the opioid. Each is assigned an initial sleep problems score, and an improved score from a treatment dependent distribution at the end of titration; these are translated to a utility value. Utilities are assigned pre-treatment, updated until the patient reaches the optimal dose or is non-compliant or dissatisfied. The OROS hydromorphone and ER oxycodone doses are converted to equianalgesic morphine doses using the following ratios: hydromorphone to morphine ratio; 1:5, oxycodone to morphine ratio; 1:2. Sensitivity analyses explored uncertainty in the conversion ratios and other key parameters. Direct medical costs are in 2005 euros.

RESULTS

Over 1 year on a mean daily morphine-equivalent dose of 90 mg, 14% were estimated to be dissatisfied with each opioid. OROS hydromorphone was predicted to yield 0.017 additional quality-adjusted life years (QALYs)/patient for a small additional annual cost (E141/patient), yielding an incremental cost-effectiveness ratio (ICER) of E8343/QALY gained. Changing the assumed conversion ratio for oxycodone:morphine to 1:1.5 led to lower net costs of E68 per patient, E3979/QALY, and for hydromorphone to 1:7.5 to savings.

CONCLUSION

Based on these analyses, OROS hydromorphone is expected to yield health benefits at reasonable cost in Germany.