Published Mar 31, 2014

Aberdeen, UK & Los Angeles, CA, USA - Health-related quality of life measures can provide useful information to decision makers involved in health care resource allocation.  These decision makers are concerned with the planning of health care services and are often interested in how they can most efficiently use the resources available to them.  Many health-related quality of life measures provide multiple scores (“profiles”) but decision makers need a single preference-based score for economic evaluation.  Methods exist for estimating preference-based scores from profile measures to support cost effectiveness analysis.  In the UK, the National Institute for Health and Care Excellence (NICE) has produced guidelines for how to undertake such mapping approaches. Regression models are frequently used in which health-related quality of life profile scores are used to predict a single index preference score.  In the study, “Should Linking Replace Regression When Mapping from Profile-Based Measures to Preference-Based Measures?” researchers from the Norwegian University of Science and Technology and the University of California at Los Angeles, noted that one particular problem with existing statistical approaches is the fact that “true” scores are actually closer to the mean than observed scores because of random variation (regression to the mean). The researchers then proposed an alternative approach to establishing mapping functions that avoids this problem.  A non-parametric variant of linear scale-aligning was proposed to ensure that the predicted scores are the same number of standard deviations above or below the mean as the predictor variable (x-scores). Dr. Peter Fayers, PhD, from the Institute of Applied Health Sciences at the University of Aberdeen, UK, and lead author on the study states, “We recommend the use of parametric linking functions or equipercentile methods to estimate preference-based scores from profile measures. These alternatives avoid the problems of existing methods and offer us a whole new way to address gaps in the evidence base for decision makers.”

---