Commentary. In Praise of Studies That Use More Than One Generic Preference-Based Measure

Generic preference-based (GPB) measures of health-related quality of life (HRQL) are widely used as outcome measures in cost-effectiveness and cost-utility analyses (CEA, CUA). Health technology assessment agencies favor GPB measures because they facilitate comparisons among conditions and because the scoring functions for these measures are based on community preferences. However, there is no gold standard HRQL measure, scores generated by GPB measures may differ importantly, and changes in scores may fail to detect important changes in HRQL. Therefore, to enhance the accumulation of empirical evidence on how well GPB measures perform, we advocate that investigators routinely use two (or more) GPB measures in each study.

We discuss key measurement properties and present examples to illustrate differences in responsiveness for several major GPB measures across a wide variety of health contexts. We highlight the contributions of longitudinal head-to-head studies.

There is substantial evidence that the performance of GPB measures varies importantly among diseases and health conditions. Scores are often not interchangeable. There are numerous examples of studies in which one GPB measure was responsive while another was not.

Investigators should use two (or more) GPB measures. Study protocols should designate one measure as the primary outcome measure; the other measure(s) would be used in secondary analyses. As evidence accumulates it will better inform the relative strengths and weaknesses of alternative GPB measures in various clinical conditions. This will facilitate the selection and interpretation of GPB measures in future studies.