Development and psychometric testing of EQ-5D-5L bolt-on descriptors for vision and cognition: A study in the UK and Australia

Background The use of 'bolt-on' descriptors has been proposed to improve the sensitivity and responsiveness of the EQ-5D in certain contexts. An Australia/UK collaboration is developing bolt-ons for vision (5L-Vis) and cognition (5L-Cog) for self-and proxy-completion, using qualitative and quantitative methods. The application of qualitative methods to bolt-on development is novel. Focus groups and interviews have the potential to generate numerous candidate descriptors, which need to be rationalised and evaluated. EQ-5D bolt-on descriptors need to serve a particular purpose and thus satisfy certain criteria. To date, no such criteria have been defined. A key outcome to support the progress of the study is the generation of criteria to guide the development, selection and testing of candidate bolt-on descriptors. Objectives The aim of this paper is to report the process used to generate the bolt-on development criteria and guidelines, and to present the initial draft guidelines as well as some draft bolt-on descriptors that have been developed. This will facilitate discussion within the EuroQol group, support further development of the criteria, and ensure their suitability in other contexts (including other bolt-ons, and other measures of health and quality of life). Methods Criteria were developed using an iterative approach. First, existing criteria used for other measures were identified from the literature and from past work by the project team. Second, the features of the three separate phases followed in the study were considered, and development and selection criteria were drafted. These were circulated to the project team and iteratively refined. Notional bolt-ons were developed by the Australian team to test the criteria. Results Criteria used to guide the development of condition specific preference-based measures and the SF-6Dv2 were identified. Separate criteria for the development and selection of candidate bolt-ons were formulated. Development criteria focused on issues relating to i) structure, ii) language, and iii) consistency with the existing EQ-5D-5L dimension structure. Selection criteria focused on face validity, and the results from both classical and modern psychometric methods. The selection and development criteria are presented in a checklist form with examples. Conclusions The criteria generated can be used to guide the development of bolt-ons across different health areas and have wider applicability to other generic and condition specific instruments. The criteria are a work in progress, and feedback from the audience on these, as well as a selection of draft candidate descriptors for the 5L-Vis and 5L-Cog, are welcomed.

OBJECTIVES Evidence suggests that, in specific circumstances, the EQ-5D is insensitive to changes or differences in health status. One possible solution is to add 'bolt-on' dimensions to the core EQ-5D dimensions, to provide additional information about the impact of certain conditions or symptoms on health. In this study, we developed EQ-5D-5L bolt-ons for cognition (5L-Cog) and vision (5L-Vis). METHODS The development process involved three stages. First, systematic literature reviews were conducted for outcome measures used in cognitive impairment and visual impairment. Items were extracted from existing measures and mapped to a set of dimensions. Common terminology was also recorded. Second, we conducted focus groups in the UK with people who have experience of either impairment. Participants were encouraged to describe the impact of their own experience of cognitive impairment or visual impairment, with reference to dimensions and terminologies identified in the literature reviews. The findings were used to develop candidate 5L-Cog and 5L-Vis descriptors. Third, we conducted interviews in the UK and Australia with people who have experience of either impairment. Interviewees were asked to complete the EQ-5D-5L with candidate 5L-Cog or 5L-Vis descriptors, followed by cognitive debriefing. RESULTS Candidate descriptors related to dimensions such as memory (for 5L-Cog), visual acuity (for 5L-Vis), mood (for both), and social functioning (for both). Various terminologies were identified in the review, with 'problems', 'limitations', 'difficulties', and 'feelings' being used in both contexts. Focus group participants expressed preferences for different terms and identified problematic words. CONCLUSIONS Qualitative research methods can be used to develop bolt-on descriptors for the EQ-5D that are meaningful to people with lived experience. However, it is challenging to identify a suitable range of participants where people with relevant lived experience constitute a heterogeneous population. In a subsequent phase of the study, 5L-Cog and 5L-Vis descriptors will be tested quantitatively.