Measuring changes in health over time using the EQ-5D 3L and 5L: a head-to-head comparison of measurement properties and sensitivity to change in a German inpatient rehabilitation sample
- 334 Downloads
To compare measurement properties and sensitivity to change of the standard version of the EQ-5D (3L) with a newly developed 5-level version (5L) in a multicenter sample of German rehabilitation inpatients (n = 230).
Rehabilitation patients (n = 114 orthopedic, n = 54 psychosomatic, n = 62 rheumatic) were asked to complete both versions of the EQ-5D and several other questionnaires at the beginning of, the end of and 3 month after inpatient rehabilitation. 3L and 5L were compared regarding missing values, ceiling effects, redistribution properties, informativity and sensitivity to change.
There were nearly no missing values in both questionnaires. Ceiling effects were 1.6 % points to 16.4 % points lower on average for the 5L. For psychosomatic patients, ceiling effects for 5L were as high as in the general German population. Absolute informativity (mean 5L: 1.76, 3L: 1.06) and relative informativity (5L: 0.76, 3L: 0.67) were both higher for 5L. 5L could better detect both positive and negative health changes in most dimensions and patient samples. Overall, patients made better use of the response levels of the 5L. Average proportion of inconsistent responses between 3L and 5L was 6.1 %.
Cross-sectionally and longitudinally, 5L was associated with an improved ability to detect health changes over time, reduced ceiling effects, and improved discriminatory power. Overall, these findings were in line with previous study outcomes, although differing in magnitude. Since the sample size is moderate and generalizability of the reported results is unclear, further comparisons in other patient populations will be informative and should be encouraged.
KeywordsQuality of life EQ-5D Psychometric comparison Head-to-head comparison Sensitivity to change German inpatient sample
We want to thank the two participating hospitals, Median Klinik Heiligendamm and Moorbad Bad Doberan, for data collection. This study was supported by the research funding vffr (Association for the Advancement of Rehabilitation Research).
- 1.Brazier, J. (2005). Current state of the art in preference-based measures of health and avenues for further research. Discussion paper series 05/5. MPRA paper no. 29762. http://mpra.ub.uni-muenchen.de/29762/.
- 13.Quality Metric. (2014). http://www.sf-36.org/.
- 17.Grissom, R. J., & Kim, J. J. (2012). Effect sizes for research: Univariate and multivariate applications (2nd ed.). New York: Taylor & Francis.Google Scholar
- 18.Leech, N. L., & Onwuegbuzie, A. J. (2002). A call for greater use of nonparametric statistics. Paper presented at the annual meeting of the Mid-South Educational Research Association, Chattanooga, TN, November 6–8.Google Scholar
- 20.Wilcox, R. R. (1997). Introduction to robust estimation and hypothesis testing. San Diego, CA: Academic Press.Google Scholar