How Should We Measure `Change' in Utility Measures of Health Staus - or Should We?
2017-06-08T06:20:35Z (GMT) by
The two standard requirements of instruments that measure cross-sectional differences in respect of abstract and subjective phenomena (such as health status or health related quality of life) are reliability and validity. However, for evaluative instruments (those designed to measure longitudinal change) another property is required: the instrument must be sensitive to clinically important changes over time, even if such changes are small. To the extent that economic evaluation is dependent on the equality of the underlying medical evidence; and the assessment of the efficiency becomes (more) "vertically integrated" with the assessment of the efficacy and effectiveness dimensions of health care, as prophesied by Drummond and Stoddart (1984), health economists, too, may have to become concerned with measuring responsiveness. The question of responsiveness also lies at the root of some of the disputation about whether to use disease-specific or generic health status measures and the clinical significance or usefulness of scores on health status instruments. The literature on the measurement of change is often confusing and contradictory. There is controversy on at least two fronts: whether a separate concept in necessary and if it is, how should it be measured. Some researchers advocate the use of change scores as the best approach to the analysis of treatment effects and others maintain that they should be avoided like the plague. We explore the issues surrounding the measurement of change scores and the methodological advantages and disadvantages surrounding the several change measures that have been suggested. We also examine the limited literature on the extent to which utility measures of health status are capable of measuring change. Implications for the practice of economic evaluation are drawn.