Methods to better identify when preference-based outcome measures may undervalue benefits
Project theme: Methods development
It is common for claims to be made that preference based outcome measures key for the calculation of treatment benefits in terms of QALYs, such as EQ5D, are insufficient or inappropriate in specific settings. Current approaches to assessing the validity of these claims range from simple qualitative judgements to the use of empirical evidence based on methods developed for use in psychometrics.
There is a tendency to assess claims in a binary way. Either the instrument is considered appropriate and continues to be used, or it is not in which case the instrument is entirely rejected and a myriad of alternative approaches then come into play (other generic instruments, the development of values for some disease specific instruments, direct valuation by patients, vignettes, expert opinion).
This project will aim to refine new methods that improve the assessment of health benefits for use in economic evaluation. We will use three case studies to:
i) Apply econometric methods that assess which, if any, relevant aspects of health are either not reflected or under-reflected in EQ5D
ii) Consider the extent of bias associated with such omissions and adjustments that can be made to remove or reduce bias
iii) Contrast findings with standard approaches to the assessment of validity
iv) Reflect on findings to refine methods
Monica Hernández Alava, Steve Pudney and Allan Wailoo