Validation of the EQ-5D-5L and psychosocial bolt-ons in a large cohort of people living with multiple sclerosis in Australia

Julie A Campbell; Hasnat Ahmad; Gang Chen; Ingrid van der Mei; Bruce V Taylor; Suzi Claflin; Glen J Henson; Steve Simpson-Yap; Laura L Laslett; Kirsty Hawkes; Carol Hurst; Hilary Waugh; Andrew J Palmer

doi:10.1007/s11136-022-03214-y

Validation of the EQ-5D-5L and psychosocial bolt-ons in a large cohort of people living with multiple sclerosis in Australia

Qual Life Res. 2023 Feb;32(2):553-568. doi: 10.1007/s11136-022-03214-y. Epub 2022 Aug 29.

Authors

Affiliations

¹ University of Tasmania, Menzies Institute for Medical Research, Hobart, TAS, Australia. Julie.Campbell@utas.edu.au.
² University of Tasmania, Menzies Institute for Medical Research, Hobart, TAS, Australia.
³ Centre for Health Economics, Monash University, VIC, Australia.
⁴ School of Population and Global Health, Neuroepidemioloy Unit, The University of Melbourne, Melbourne VIC, Australia.
⁵ School of Population and Global Health, Health Economics Unit, The University of Melbourne, Melbourne VIC, Australia.

^# Contributed equally.

Abstract

Background: Multiple sclerosis (MS) is an inflammatory, neurodegenerative disease of the central nervous system which results in disability over time and reduced quality of life. To increase the sensitivity of the EQ-5D-5L for psychosocial health, four bolt-on items from the AQoL-8D were used to create the nine-item EQ-5D-5L-Psychosocial. We aimed to externally validate the EQ-5D-5L-Psychosocial in a large cohort of people with MS (pwMS) and explore the discriminatory power of the new instrument with EQ-5D-5L/AQoL-8D.

Methods: A large representative sample from the Australian MS Longitudinal Study completed the AQoL-8D and EQ-5D-5L (including EQ VAS) and both instruments health state utilities (HSUs) were scored using Australian tariffs. Sociodemographic/clinical data were also collected. External validity of EQ-5D-5L-Psychosocial scoring algorithm was assessed with mean absolute errors (MAE) and Spearman's correlation coefficient. Discriminatory sensitivity was assessed with an examination of ceiling/floor effects, and disability severity classifications.

Results: Among 1683 participants (mean age: 58.6 years; 80% female), over half (55%) had moderate or severe disability. MAE (0.063) and the distribution of the prediction error were similar to the original development study. Mean (± standard deviation) HSUs were EQ-5D-5L: 0.58 ± 0.32, EQ-5D-5L-Psychosocial 0.62 ± 0.29, and AQoL-8D: 0.63 ± 0.20. N = 157 (10%) scored perfect health (i.e. HSU = 1.0) on the EQ-5D-5L, but reported a mean HSU of 0.90 on the alternative instruments. The Sleep bolt-on dimension was particularly important for pwMS.

Conclusions: The EQ-5D-5L-Psychosocial is more sensitive than the EQ-5D-5L in pwMS whose HSUs approach those reflecting full health. When respondent burden is taken into account, the EQ-5D-5L-Psychosocial is preferential to the AQoL-8D. We suggest a larger confirmatory study comparing all prevalent multi-attribute utility instruments for pwMS.

Keywords: AQoL-8D; Australian MS Longitudinal Study; Cost-utility analysis; EQ-5D-5L; EQ-5D-5L-Psychosocial; Multi-attribute utility instrument; Multiple sclerosis; Sleep.

MeSH terms

Australia
Female
Humans
Longitudinal Studies
Male
Middle Aged
Multiple Sclerosis*
Neurodegenerative Diseases*
Psychometrics / methods
Quality of Life / psychology
Surveys and Questionnaires

Grants and funding

19-0702/Multiple Sclerosis Research Australia