Bolt-ons and blind spots: what we’re missing in health metrics

Bolt-ons are additional dimensions that can be added to the EQ-5D-5L — a generic health-related quality of life (HRQoL) measure — with the aim of improving the accuracy and precision of the instrument for a specific population. And last month, some of the Acaster Lloyd team attended the 9th EuroQol Academy meeting in Barcelona, where the theme was exactly that: bolt-ons.

It was valuable to discuss this ongoing area of research with international researchers, collaborating across countries and disciplines, to support the EuroQol foundation’s goal of ‘helping the world make better health decisions’.

One concern that was raised at the conference was the potential for the use of bolt-ons to undermine the status of the EQ-5D as the preferred HRQoL instrument by some health technology assessment (HTA) agencies. However, when bolt-ons are added, the generic EQ-5D-5L instrument is not replaced, data can still be scored using the 5L responses, but bolt-ons provide the opportunity to collect additional data that can be used in sensitivity analyses to give further information about treatment benefit.

In some conditions, such as psoriasis, research suggests that the EQ-5D-5L may not capture all of the important aspects of an individual’s HRQoL. Therefore, two bolt-ons were developed for use in psoriasis populations to capture the additional impacts of skin irritation (itching) and self-confidence.

Acaster Lloyd  presented a study exploring the content validity of the EQ-5D-5L, with and without bolt-ons, in two other chronic skin conditions: atopic dermatitis and chronic urticaria. To assess this, a qualitative study was conducted with N=15 individuals with each condition. Interviews consisted of concept elicitation questions about the symptoms and impacts experienced, followed by cognitive debriefing of the EQ-5D-5L with and without four bolt-ons. The bolt-ons evaluated measured skin irritation (itching), self-confidence, sleep and social relationships.

The study found evidence supporting the content validity of three of the bolt-on dimensions: skin irritation, self-confidence and sleep, with less clear evidence for the social-relationships dimension. You can see the full study right here.

The conference included other presentations on the development and testing of specific bolt-ons, such as assessments of the psychometric properties of bolt-ons. The final session featured discussions about some of the methods and challenges around valuing the EQ-5D with bolt-ons. 

If bolt-ons are going to improve the sensitivity of the generic instrument, appropriate valuation methods are required. This needs to include the avoidance of ‘double-counting’, where the concept measured by the bolt-on is already captured by an existing item and therefore the HRQoL impact is overestimated. 

Currently the views of HTA bodies towards bolt-ons is largely unknown. Presumably, their confidence in the application of bolt-on data for cost-effectiveness evaluation will be dependent on all the topics discussed: measurement properties, valuation and the risk of double counting.

Acaster Lloyd have two ongoing studies investigating the use of EQ-5D-5L bolt-ons. One is a follow-up study to the work presented at the Academy meeting, exploring the psychometric properties of the same four bolt-ons in a survey of individuals with chronic urticaria. The other is a qualitative study evaluating the content validity of the EQ-5D-5L and a selection of bolt-ons in alopecia areata and vitiligo. We’ll update you once these studies are completed.

We also attended the associated 5th early-careers researcher meeting. This was a great opportunity to learn about recent and ongoing international research in more detail, facilitating collaboration and guidance between peers, as well as from senior members of the EuroQol group through the ‘buddy’ programme.

An interesting discussion at the early career researcher meeting was the frequency at which EQ-5D population norm values as well as country-specific value sets should be updated (i.e., whether value sets have a ‘shelf life’).

Meixia Liao, Zhihao Yang, and Nan Luo reported that the range of values assigned by the Chinese general population to EQ-5D-5L health states has broadened over time (between 2012 and 2023). In other words, people thought that severe health states were more severe and that mild health states were milder in the more recent valuation. The most important aspect of health-related quality of life also switched from mobility to pain or discomfort, reflecting changing general population views.

General population perceptions of aspects of health-related quality of life and their relative importance are widely accepted to differ between countries, hence the use of country-specific value sets in EQ-5D based utility estimation. Changes in a population’s perceptions over time may be harder to pin down. An updated general population valuation study is ongoing in the UK — but how often should these updates to value sets be scheduled for? And should this be flexible to capture changes following key population-wide events such as following the covid-19 pandemic?

This interesting discussion is applicable to the new developments with bolt-ons as well — and work on how these should be valued is ongoing. Once valuation protocols for bolt-ons are designed, these should also consider the question of ‘shelf-life’.