Discriminant validity, responsiveness and reliability of the arthritis-specific Work Productivity Survey assessing workplace and household productivity within and outside the home in patients with axial spondyloarthritis, including nonradiographic axial spondyloarthritis and ankylosing spondylitis

INTRODUCTION The arthritis-specific Work Productivity Survey (WPS) was developed to evaluate productivity limitations associated with arthritis within and outside the home. There is an unmet need for an instrument assessing similar productivity limitations in axial spondyloarthritis (axSpA), including nonradiographic axSpA and ankylosing spondylitis. Following its validation in rheumatoid and psoriatic arthritis, we aimed to assess psychometric properties of WPS in adult-onset active axSpA in this analysis. METHODS Psychometric properties were assessed using data from the RAPID-axSpA trial (NCT01087762) in which researchers investigated certolizumab pegol efficacy and safety in axSpA. WPS was completed at baseline and every 4 weeks until week 24. Validity was evaluated at study baseline via known-groups defined by the first and third quartile cutoffs of patient scores to Ankylosing Spondylitis Disease Activity Score (ASDAS), Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), back pain, Bath Ankylosing Spondylitis Functional Index (BASFI), Short Form 36 health survey (SF-36) and Ankylosing Spondylitis Quality of Life Scale (ASQoL). Responsiveness and reliability were assessed by comparing WPS mean changes in ASAS 20% improvement criteria (ASAS20), BASDAI50, ASDAS clinically important improvement/major improvement (CII/MI) and BASFI minimum clinically important difference (MCID) responders versus nonresponders at week 12. All comparisons were conducted on observed cases in the randomized set using a nonparametric bootstrap-t method. RESULTS The results confirmed the psychometric properties of WPS. AxSpA patients with a worse health state had significantly more days of household work lost, household work with reduced productivity, social activities missed and outside help hired, as well as a higher interference rate of arthritis, than patients with a better health state. Similarly, employed patients with a worse health state had significantly more work days lost or with productivity reduced, and a higher interference of arthritis on work productivity. Similar findings were also observed in the nonradiographic (nr) axSpA and AS subpopulations. The WPS was responsive to clinical changes, with responders reporting larger improvements at week 12 in WPS scores versus nonresponders. Effect sizes in responders were generally moderate to large (standardized response mean >0.5). CONCLUSIONS These analyses demonstrate that WPS is a valid, responsive and reliable instrument for the measurement of productivity within and outside the home in adult-onset axSpA, as well as the in subpopulations of AS and nr-axSpA.