Adding a “Cognition Bolt-on” to EQ-5D-5L: Insights from New Research and What’s Next

As part of EuroQol’s current priority-setting work, the development of bolt-ons for the EQ-5D-5L instrument has taken centre stage. A “bolt-on toolbox” is being developed to identify and test potential additional dimensions that could enhance the instrument’s performance in specific contexts.

At the recent 9th EuroQol Academy meeting in Barcelona, which centred around the theme of bolt-ons, cognition emerged as one of the most important and considered candidate bolt-ons. There is growing recognition that cognitive functioning plays a central role in people’s health-related quality of life. However, exactly how to measure cognition within the EQ-5D framework remains an open question.

As part of a EuroQol-funded project, I (Hannah Hussain, OHE) have been working with international collaborators Benjamin Craig, Fanni Rencz, Bernhard Michalowsky and Nan Luo using a large US general population dataset, to explore whether cognition can be meaningfully captured as a single item, or whether it is better represented by more specific sub-domains like memory and concentration. The full analysis is currently being written up into a report. This work was presented at the EuroQol Academy meeting, and feedback received at that meeting has sparked further research ideas, highlighting the need for additional testing and qualitative exploration.

Research conducted by OHE colleagues Nadine Henderson and Chris Sampson has shown that people sometimes struggle to interpret broad terms like “cognition” or “cognitive impairment”. Their qualitative work has also highlighted that asking multiple detailed questions could create additional burden for respondents, particularly for people more vulnerable to mental fatigue.

What have we learned so far?

Our analysis suggests that simply adding a broad “cognition” item may not be the answer. Psychometric testing from our US general population study indicated that breaking cognition down into more specific symptom components, such as memory and concentration, was potentially more informative than using a single umbrella term for cognition. However, further research is needed to identify which aspects of cognition matter most to people when thinking about their health, and how best to phrase items for consistent interpretation by both respondents and proxies.

Where might a cognition bolt-on add value?

While dementia is the most obvious condition where cognitive function is relevant, there are many other clinical areas where cognition plays a central role in people’s experience of health. Acute illnesses, mental health conditions, neurological disorders, long COVID and recovery from stroke or head injury can all impact memory, attention and reasoning. In such cases, having a cognition bolt-on could improve the sensitivity of EQ-5D to detect meaningful changes, especially during recovery and rehabilitation. This is particularly important in areas where proving cost-effectiveness is challenging due to subtle or non-linear changes in health status.

In my own research on mapping utility scores in dementia, I explored the relationship between EQ-5D and core dementia symptoms: cognition, function and behaviour/mood. The latter two are fairly well represented within the existing EQ-5D dimensions. But cognition? Less so. I conducted a targeted review and found weak, inconsistent evidence that the EQ-5D captures the cognitive symptoms of dementia. This suggests that a cognition bolt-on could help improve measurement sensitivity and better reflect the lived experience of people with dementia, something with major implications for HTA and care planning.

Self vs. proxy assessment

One of the key challenges in measuring cognition is determining who should complete the assessment. Self-report may be appropriate for individuals with mild cognitive difficulties, but as impairment becomes more severe, recall, judgement and self-awareness can be compromised. In such cases, respondents may unintentionally misrepresent their level of functioning, or struggle with the cognitive effort required to complete the questions. This makes proxy assessment increasingly important.

However, proxy reporting brings its own challenges. Unlike more observable dimensions of health status such as mobility, cognitive impairment may not present in consistent or easily recognisable ways. If someone seems withdrawn or unengaged, is this due to cognitive decline, or could it stem from sensory loss, fatigue or emotional distress? This raises a critical question: do self and proxy assessments meaningfully capture cognitive impairment, or would proxy reports relying too heavily on outward signs that may be misleading?

Future research could explore the alignment between self and proxy reports of cognition, including inter-rater agreement and how proxies interpret behaviours in relation to internal cognitive states. This would be key to improving the validity and reliability of proxy responses in settings where cognitive impairment limits self-reporting.

Valuation considerations

Adding new bolt-on dimensions to the EQ-5D has implications that go beyond measurement – it directly impacts how utility values are scored, which influences the use of EQ-5D in health technology assessment (HTA).

Key questions remain. Does the bolt-on overlap with existing dimensions like anxiety/depression or usual activities, or does it add unique explanatory power? How would it impact index score generation, and how should it be integrated into existing value sets? Are large-scale preference studies feasible, and would members of the public recognise and value cognition as a distinct aspect of health?

These are not just technical considerations; they are essential to ensuring that any addition to the EQ-5D reflects real-world health experiences and fits within established HTA frameworks. Further testing and valuation work will be crucial before integration is considered.

Where do we go from here?

We need further testing across diverse populations, conditions and settings. Qualitative research is key to ensure any bolt-on is meaningful, understandable and acceptable to both patients and proxies. A cognition bolt-on has the potential to improve the sensitivity and relevance of EQ-5D, particularly in areas like dementia, rehabilitation and mental health. As our work and others’ have shown, cognition is not a straightforward concept and careful design, testing and valuation work is essential before it can be adopted in practice.

If you’re interested in collaboration or would like to find out more about this topic, please feel free to get in touch: hhussain@ohe.org

Hussain, H., Keetharuth, A., Rowen, D., & Wailoo, A. (2022). Convergent validity of EQ-5D with core outcomes in dementia: a systematic review. Health and Quality of Life Outcomes, 20(1), 152.