The Valued Life Activities Scale (VLAs) : linguistic validation, cultural adaptation and psychometric testing in people with rheumatic and musculoskeletal diseases in the UK

Prior, Y ORCID: https://orcid.org/0000-0001-9831-6254, Tennant, A, Tyson, S and Hammond, A ORCID: https://orcid.org/0000-0002-5266-9991 2020, 'The Valued Life Activities Scale (VLAs) : linguistic validation, cultural adaptation and psychometric testing in people with rheumatic and musculoskeletal diseases in the UK' , BMC Musculoskeletal Disorders, 21 (1) , p. 505.

Preview	PDF - Published Version Available under License Creative Commons Attribution 4.0. Download (804kB) | Preview
	Microsoft Word - Accepted Version Restricted to Repository staff only Download (97kB)

Abstract

Background The Valued Life Activities Scale (VLAs) measures difficulty in daily activities and social participation. With various versions involving a different number of items, we have linguistically and culturally adopted the full VLAs (33-items) and psychometrically tested it in adults with rheumatic and musculoskeletal diseases in the United Kingdom. Methods Participants with Rheumatoid Arthritis, Ankylosing Spondylitis, Chronic Pain/ Fibromyalgia, Chronic Hand/ Upper Limb Conditions, Osteoarthritis, Systemic Lupus, Systemic Sclerosis and Primary Sjogren’s Syndrome were recruited from out-patient clinics in National Health Service Hospitals, General Practice and patient organisations in the UK. Phase1 involved linguistic and cultural adaptation: forward translation to British English; synthesis; expert panel review and cognitive debriefing interviews. In Phase2 participants completed postal questionnaires to assess internal construct validity using (i) Confirmatory Factor Analysis (CFA) (ii) Mokken scaling and (iii) Rasch model. Results Responders (n = 1544) had mean age of 59 years (SD13.3) and 77.2% women. A CFA failed to support a total score from the 33-items (Chi Square 3552:df 464: p < 0.0001). Mokken scaling indicated a strong non-parametric association between items. Fit to the Rasch model indicated that the VLAs was characterised by multidimensionality and item misfit, which may have been influenced by clusters of residual item correlations. An item banking approach resolved a 25-item calibrated set whose application could accommodate the ‘does not apply to me’ response option. Conclusions The UK version of the VLAs failed to satisfy classical and modern psychometric standards for complete item sets. However, as the scale is not usually applied in complete format, an item bank approach calibrated 25 items with fit to the Rasch model. Suitable Computer Adaptive Testing (CAT) software could implement the item set, giving patients the choice of whether an item applies to them, or not.

Actions (login required)

Edit record (repository staff only)