Can We Predict the Impact of the Differential Sensitivity of Health-Related Quality of Life (HRQoL) Instruments on the Valuation of Health State Changes?

OBJECTIVES: Cost effectiveness evaluation (CEA) using quality-adjusted life years (QALYs) gained as a measure of health outcomes typically relies on generic HRQoL instruments based on multi-attribute utility (MAU) theory. Yet, different MAU instruments produce different utility values, differ in content validity, and show different sensitivity to changes in HRQoL. Since instrument choices may exert a profound impact on utility values, it is of interest to predict their impact on expected QALY gains as a function of a given health state profile change.

METHODS: The multi-instrument comparison (MIC) database comprises data from 8,022 respondents from six countries (either healthy or suffering from asthma, arthritis, cancer, depression, diabetes, hearing loss, or heart disease), who completed six MAU instruments (EQ-5D, SF-6D, HUI3, 15D, QWB, AQoL) and the generic HRQoL profile instrument SF-36. After mapping the utilities onto the eight dimensions of the SF-36, we assessed the sensitivity of the instruments to changes of each of them using scale invariant beta coefficients, which we derived from OLS regressions, and estimated the relative impact of each dimension on the utility change based on standardized coefficients. In addition we performed pairwise comparisons of instrument sensitivity by dimension.

RESULTS: Our analyses consistently confirm the substantial differences between the utilities produced by the various MAU instruments. While pairwise comparisons assist qualitatively selecting an instrument that is sensitive to the particular health state changes of interest, the standardized beta coefficients derived from our OLS regressions allow for semi-quantitative and quantitative predictions of the impact of instrument choice.

CONCLUSIONS: In addition to documenting the different psychometric properties of the six MAU instruments, our analyses can be used to predict the sensitivity of generic MAU instruments (and the effect of their choice on the computation of QALYs gained) to changes in any of the eight dimensions captured by the SF-36.