Direct and indirect mapping of the BATH disease activity and function indices (BASDAI/BASFI) onto the EQ-5D-5L utility index in axial spondyloarthritis

What is the problem?
Clinical studies designed to assess the effectiveness of health technologies would ideally include outcome measures that are directly linked to utility measures of health-related quality of life. Health utilities are preference values that patients attach to their overall health status. This data is required to calculate quality-adjusted life years (QALYs) for use in cost-effectiveness analyses which is an evidence requirement for many decision makers, such as by NICE in the UK. In cases where this does not happen, an alternative solution that is extensively employed is mapping which involves using an additional data set to estimate the (statistical) relationship between outcomes measured in clinical studies and health utility.

What is Axial Spondyloarthritis?
Axial Spondyloarthritis (AxSpA) is a common rheumatic chronic progressive inflammatory disease, leading to joint damage/pain, stiffness, impaired physical function, fatigue and reductions in quality of life. The condition primarily affects the spine and sacroiliac joints. The onset is typically early in adulthood and is more common in men than in women. AxSpA patients incur significant direct NHS costs, which are mainly due to costs associated with prescriptions, and outpatient and day unit use. There are also indirect costs to society as axSpA affects young patients who have to take time off work and in the worst cases may permanently cease paid employment. The introduction of anti-TNF biologic therapies licensed for use in the management of axSpA (e.g., etanercept, adalimumab) has been associated with significantly improved outcomes. However, many of the pivotal trials for new biologics therapies did not include a generic preference-based HRQoL instrument such as the EQ-5D. Consequently, economic models of Health Technology Assessments of axSpA have been based on health utilities estimated from mapping exercises.

Our mapping study
There are no published mapping algorithms that estimate EQ-5D-5L utilities from the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) or the Bath Ankylosing Spondylitis Functional Index (BASFI) scores. We have recently had a journal article published in which we report on new mapping algorithms to obtain EQ-5D indices when only BASDAI/BASFI data are available. This is the first mapping study using a UK data set from the British Society for Rheumatology Biologics Register in AS (BSRBR-AS) which includes observations that come from England, Scotland and Wales.

What we did
Different statistical models to estimate health utility directly, or responses to individual EQ-5D questions (response mapping) indirectly from BASDAI/BASFI scores, were tested for patient self-report questionnaires. A range of metrics were used to compare and assess models performance.

Main findings
Our main study results showed that direct mapping models, and to a slightly lesser degree, response mapping models provided reliable algorithms for predicting EQ-5D-5L utilities from BASDAI/BASFI scores. These algorithms can be used in applied cost-effectiveness analysis in axSpA where EQ-5D-5L is the target outcome of interest.
A user-friendly accessible Excel tool is provided to assist analysis with the implementation of the best performing mapping algorithms (available in the online article Appendix)

The recommended mapping algorithms produced from our recent work allows researchers for the first time to estimate EQ-5D-5L values from BASDAI/BASFI data, enabling cost-utility analyses using datasets where the BASDAI/BASFI but no utility measures were collected.

Article details – this is an Open Access article and is freely available.
Neilson AR, Jones GT, Macfarlane GJ, Pathan EM, McNamee P. Generating EQ-5D-5L health utility scores from BASDAI and BASFI: a mapping study in patients with axial spondyloarthritis using longitudinal UK registry data. Eur J Health Econ. 2022 Feb 3. doi: 10.1007/s10198-022-01429-x. Epub ahead of print. PMID: 35113270.


Leave a Reply

Your email address will not be published. Required fields are marked *