Psychometric Properties of the Work Ability Index Among Patients With Arthritis in Singapore

The Work Ability Index (WAI) is recommended for use among individuals with arthritis but has yet to be validated in Singapore. We therefore aimed to evaluate the psychometric properties of WAI among individuals with arthritis in Singapore and whether its first item (WAS) and a version without disease-related item (WAInodis) could be proxies for WAI.

Factor structure, internal consistency, convergent validity, and known-groups validity of WAI were evaluated. Concurrent, convergent, and known-groups validity of WAS and WAInodis were examined to assess their potential as proxies.

A total of 509 patients with arthritis and 499 healthy volunteers were included. A one-factor structure was identified for WAI. Internal consistency was acceptable (Cronbach's alpha = 0.725). Convergent validity was demonstrated by correlation with conceptually related measures including work productivity (rho = 0.586), coping (rho = 0.457), general health (rho = 0.497), and fatigue (rho = 0.527). Known-groups validity was demonstrated by significant difference in WAI score between participants with and without arthritis (37.6 vs. 42.1, p < 0.001). Compared to WAS, WAInodis exhibited stronger potential as a proxy, with stronger correlation with WAI (rho = 0.961 vs. 0.763), smaller difference from WAI (mean = -1.0 vs. 49.5), and stronger correlations with conceptually related measures.

WAI demonstrated relatively good psychometric properties among individuals with arthritis, adding to the literature supporting its role in assessing work ability, although further evaluation is needed. WAInodis may serve as a proxy for WAI in studies where the disclosure of medical conditions is a concern.