Does linear equating improve prediction in mapping? Crosswalking MacNew onto EQ-5D-5L value sets

Abstract

Purpose Preference-based measures are essential for producing quality-adjusted life years (QALYs) that are widely used for economic evaluations. In the absence of such measures, mapping algorithms can be applied to estimate utilities from disease-specific measures. This paper aims to develop mapping algorithms between the MacNew Heart Disease Quality of Life Questionnaire (MacNew) instrument and the English and the US-based EQ-5D-5L value sets. Methods Individuals with heart disease were recruited from six countries: Australia, Canada, Germany, Norway, UK and the US in 2011/12. Both parametric and non-parametric statistical techniques were applied to estimate mapping algorithms that predict utilities for MacNew scores from EQ-5D-5L value sets. The optimal algorithm for each country-specific value set was primarily selected based on root mean square error (RMSE), mean absolute error (MAE), concordance correlation coefficient (CCC), and r-squared. Leave-one-out cross-validation was conducted to test the generalizability of each model. Results For both the English and the US value sets, the one-inflated beta regression model consistently performed best in terms of all criteria. Similar results were observed for the cross-validation results. The preferred model explained 59 and 60% for the English and the US value set, respectively. Linear equating provided predicted values that were equivalent to observed values. Conclusions The preferred mapping function enables to predict utilities for MacNew data from the EQ-5D-5L value sets recently developed in England and the US with better accuracy. This allows studies, which have included the MacNew to be used in cost-utility analyses and thus, the comparison of services with interventions across the health system.

WITHDRAWN: Mapping the FACT-G to EQ-5D-3L utility index in cancer: data from a cross-sectional study in in Guangxi, China

“…Where and were the mean and standard deviation of the observed EQ-5D-3L utility scores, and and were the mean and standard deviation of the predicted EQ-5D-3L utility scores obtained from the regression model. Then, using scatter plots and the Bland-Altman plot, we recoded predictions (equivalent) scores and observed scores to show the difference between them [23].…”

“…While regressionbased models seek to predict the most likely true preference-based score using the pro le-based score, linking seeks to nd the preference-based score that is equivalent to the pro le-based score by aligning the score distributions of the two scales [22]. In a few mapping studies, linear equating was used to predict utility using a regression-based method and then scale aligning between predicted and observed values to force them to have the same mean and variance [23].…”

Predicting EQ-5D-5L crosswalk from the PROMIS-29 profile for the United Kingdom, France, and Germany

“…The nRMSE is the normalized root of the sum of the squared residuals between observed and predicted scores and the nMAE is the normalized mean absolute error of the absolute residuals. Both are normalized with respect to the different scale ranges of the EQ-5D-5L crosswalk in the UK, France, and Germany [39][40][41]. We…”

Predicting EQ-5D-5L crosswalk from the PROMIS-29 profile for the United Kingdom, France, and Germany

“…The nRMSE is the normalized root of the sum of the squared residuals between observed and predicted scores and the nMAE is the normalized mean absolute error of the absolute residuals. Both are normalized with respect to the different scale ranges of the EQ-5D-5L crosswalk in the UK, France, and Germany (40)(41)(42). We also determined the width between the 95% empirical limits of agreement and compared them to the 95% theoretical limits of agreement (i.e., ± 1.96 * SD(residuals)).…”