I was recently directed towards a new report on the EQ-5D, by the NICE Decision Support Unit at the University of Sheffield. As health economists we can all see the benefits of a measure like the EQ-5D, but it’s also hard to ignore its limitations. Its apparent brevity and imprecision has often (for me at least) made it a hard sell to non-economists.
Claims against the EQ-5D
Insensitive. Unresponsive. Unreliable. Invalid. The EQ-5D has been on the receiving end of much criticism in recent years, yet it has maintained its position as chief measure of health related quality of life. The DSU study investigates the claims made against the EQ-5D, finding that many are unsubstantiated and that criticisms are often aimed at QALYs rather than the EQ-5D itself. By carrying out case studies the report sought to investigate whether these claims could be supported.
In some cases the EQ-5D was found to be less responsive or sensitive than disease specific outcome measures. Surely this was to be expected; proof that the use of a generic measure involves a trade-off. The good news for the status quo is that other generic preference-based measure were not found to systematically perform better than the EQ-5D. Even more reassuring is the fact that the EQ-5D is actually more sensitive to differences between patients than some disease specific measures! The authors also highlight the difficulty in testing the reliability and validity of the EQ-5D, as there is no gold standard with which to compare it. Thus any claims made against it based on this can only be circumstantial.
It seems that the EQ-5D has plenty of life in it yet. Unless you are opposed to the theoretical basis of the EQ-5D then there isn’t much bad to say about it. It works. However, for me the problem never lay within the construct of the EQ-5D but with the limitations of the QALY and in (my opinion) the inappropriateness of using (public) preferences to value health related quality of life. To echo the view of the DSU study, future research should remain focused upon tools that are globally relevant rather than condition specific. For me this may mean the development of alternative tools or alternative methods of valuation, not because the EQ-5D is problematic but because I see room for improvement.
Is there a consensus on the usefulness of the EQ-5D? Do we use it because it is a good tool or because NICE say so? In what ways do we think the EQ-5D could be improved?
A helpful summary. I’d agree what is critical is comparability of health gain for all possible uses or our limited resources. It seems unlikely that QALY and DALY will be dislodged from their prime position, and within that the place of EQ5D falls into the ‘best currently available’ – The price of criticism is a viable alternative, and this has not been evident to date.
It also seems to me that we are only at the beginning of our journey to embed responsive tracking of health gain, and within that health and life utility, into the normal feedback loop of patient’s life experiences. The attrition in benefit from research led to real world outcomes, and its sensitivity to the quality of care delivery, the starting position, and the effects of non-healthcare factors will keep us all busy for many years yet. Having EQ5D as a fixed point to make progress in all of these areas seems no bad thing.