Multinational Evaluation of Conditional Median Models
 |  |
James W. Shaw PhD, PharmD, MPH (Department of Pharmacy
Administration/UIC), Simon Pickard PhD (Department of Pharmacy
Practice/UIC) and colleagues (Lin Zhan MS and Mark Oppe MSc) presented a
poster entitled “Multinational Evaluation of Conditional Median Models
of EQ-5D Health State Preferences” at the 17th Annual Conference of the
International Society for Quality of Life Research (ISOQOL), which was
held October 27-30, 2010 in London, England. ISOQOL is an international
scientific organization dedicated to advancing the study of
health-related quality of life and other patient-centered outcomes.
There were more than 300 posters that were presented at the conference
and were organized into methodology, clinical research, and policy
research tracks. Dr. Shaw and his colleagues were granted the Best
Policy Research Poster Award for their research.
The
EQ-5D questionnaire is designed to measure an individual’s
point-in-time health status. It uses 5 questions to measure the
severity of problems with mobility, the performance of self-care
activities, the performance of usual activities, pain/discomfort, and
anxiety/depression. Combinations of responses to these questions
describe a total of 243 unique health states. By assigning population
preferences to individuals’ self-classified EQ-5D health states, one can
derive quality-adjusted life years for use in economic evaluations of
pharmaceuticals and other health care technologies.
Until recently, researchers have emphasized the estimation and
application of mean societal preferences for EQ-5D health states. For
most states of health, the mean provides an inadequate measure of the
central tendency of population preferences. Over the past several
years, Dr. Shaw has led several studies involving the characterization
of median population preferences for health states described by the
EQ-5D questionnaire. As compared with the mean, the median provides a
fairer representation of the aggregate preference of all population
members. Both the mean and median are amenable for use in decision
analytic applications.
The study presented by Dr. Shaw and colleagues involved an evaluation of
estimators of median population EQ-5D health state preferences using
data collected in four countries. Much of Dr. Shaw’s prior research has
focused on the US population. This study provided evidence regarding
the suitability of median preference estimators for countries outside
the US. Although the authors observed differences among the four
countries, their findings supported the feasibility of median preference
estimation. This work has implications for the methods used to
estimate preferences for the health states described by the EQ-5D and
similar questionnaires.
Click here for the research poster.