Appraising Ultra-Orphan Drugs: Is Cost-Per-QALY Appropriate? A Review of the Evidence
Four papers presented at the EuroQol Plenary discussed preliminary findings.
As reported earlier, OHE’s Prof Nancy Devlin and Prof Ben van Hout of the University of Sheffield were awarded a £457,000 grant from the Policy Research Programme (PRP) of the UK’s National Institute for Health Research (NIHR) for research on an EQ-5D-5L value set for England.
As reported earlier, OHE’s Prof Nancy Devlin and Prof Ben van Hout of the University of Sheffield were awarded a £457,000 grant from the Policy Research Programme (PRP) of the UK’s National Institute for Health Research (NIHR) for research on an EQ-5D-5L value set for England. The EQ-5D-5L is an expanded version of the EQ-5D, able to capture subtler differences in attitudes about health states. Widely used in the UK and internationally to measure patients’ self-reported health, these instruments are accompanied by “value sets”—numerical summaries of the preferences of the general public about living in certain states of health, along a continuum from “good” to “bad”. One of the applications of these value sets is to produce estimates for quality-adjusted life years (QALYs), which in turn help inform resource allocation decisions.
Although some research and analysis still is in progress, initial findings from the project were described in four papers presented at the 30th EuroQol Plenary Meeting in September[1]. Since results still are preliminary, the papers are not yet available publicly. Further modelling work on the data is underway, with the final value set and related methodological papers to be released by the end of 2013.
The project included careful monitoring of the integrity of the data during and after data collection. The results are important not only as a basis for modelling in this study, but also as guidance on good practice for other studies intended to create EQ-5D-5L data sets. The findings include, for example, that respondents appeared to find the five-level instrument more challenging to use than the three-level—for instance, in differentiating between “serious” and “extreme” problems with health.
Quite apart from technical challenges in eliciting health-state preferences are issues about how to use individuals’ data to represent society’s preferences. A further paper presented at the EuroQol meeting by OHE’s Nancy Devlin and Ken Buckingham (University of Otago) discussed the normative issues in selecting what measure of “average” preferences should be used[2]. The average preference of society (the mean), for example, is not the necessarily the same as the preference of the average person (the median). Which option is chosen has an important impact on the final value set. Finding little guidance on this issue from a review of the literature, this paper develops a set of criteria applicable in evaluating alternative approaches.
In addition to Nancy Devlin and Ben van Hout, the project team for this research includes Koonal Shah and Yan Feng from OHE and Brendan Mulhern from the University of Sheffield. Each co-authored papers presented at the EuroQol Scientific Plenary. As reported in our earlier post, the project has been guided by an advisory group comprising senior staff from the English Department of Health, NICE and external academics.
This is an independent report commissioned and funded by the Policy Research Programme in the Department of Health. The views expressed are not necessarily those of the Department.
For further information, please contact Prof Nancy Devlin at OHE.
Earlier OHE research
This project builds on previous research under an NIHR PRP grant for methodological research that was completed by Prof Nancy Devlin and colleagues at OHE, in collaboration with others: An interim EQ-5D-5L value set for England and development of EQ-5D-5L valuation methods. That research contributed to the development and testing of preference elicitation methods as reported in: Devlin, N., Buckingham, K., Tsuchiya, A., Shah, K., Tilling, C., Wilkinson, G. and van Hout B. (2013) A comparison of alternative variants of the lead and lag time TTO. Health Economics. 22(5), 517-532.
The project also produced data that helped create an interim value set for the EQ-5D-5L based on mapping to EQ-5D value sets. This is reported in: van Hout, B., Janssen, M.F., Feng, Y-S., Kohlman, T., Busschbach, J., Golicki, D., Lloyd, A., Scalone, L., Kind, P. and Pickard, S. (2012) Interim scoring for EQ-5D-5L: Mapping the EQ-5D-5L to EQ-5D-3L value sets. Value in Health. 15(5), 708-715.
Papers reporting on a multi-country pilot study to test the methods used in EQ-5D-5L value set studies was published recently in a special issue of the European Journal of Health Economics,co-edited by Nancy Devlin and Paul Krabbe.
[1] The results to date are based only on interviews conducted in England. Data collection of samples of the general public in Scotland, Northern Ireland and Wales are in progress. These will be combined with the English data to model a UK value set to complement the value set for England.
[2]Devlin, N. and Buckingham, K. (2013) What is the normative bias for selecting the measure of average’ preferences to use in social choices? Paper presented at the 30th EuroQol Plenary Meeting. Montreal. 12-13 September.
Prof Nancy Devlin
