Global health in axial spondyloarthritis: thresholds for the Assessment of SpondyloArthritis international Society Health Index and the EuroQol score: analysis of the ASAS-PerSpA study.

Abstract

Objectives: In axial spondyloarthritis (axSpA), patient-perceived quality of life/global functioning and health (GH) can be assessed using disease-specific [Assessment of SpondyloArthrit is international Society Health Index (ASAS-HI)] or generic [(3-level EuroQol 5 Dimensions (EQ-5D-3L)] scores. Our objectives were to explore the link between these scores and to define thresholds for good and poor GH.

Method: We conducted a post-hoc analysis of the cross-sectional ASAS-PerSpA study for patients fulfilling ASAS criteria for axSpA. The ASAS-HI and EQ-5D scores were analysed visually (distribution, scatterplot) and through Spearman correlation and agreement (deciles). To determine cut-offs for good and poor GH on EQ-5D based on the validated ≤5 and ≥12 cut-offs for ASAS-HI, respectively, receiver operating characteristics (ROC) curves and distribution-based methods were applied. Validity was assessed using crude concordance and prevalence-adjusted bias-adjusted kappa; discordance between groups was explored.

Results: In 2651 patients (median age 41.0 years, 66.5% men), the correlation between ASAS-HI and EQ-5D was high (r = -0.73) and agreement (between deciles) was moderate (weighted kappa = 0.51). Both ROC areas under the curve were 0.86; thresholds of 0.69 and 0.54 for EQ-5D were chosen for good and poor GH, respectively. Crude concordances and agreement were satisfactory (0.80-0.81 and 0.60-0.61, respectively). The EQ-5D cut-off for good GH performed better than that for poor GH.

Conclusion: ASAS-HI and EQ-5D were highly correlated but did not fully overlap. We propose EQ-5D thresholds corresponding to the ASAS-HI thresholds for good and poor GH; however, caution is needed when assessing poor GH with EQ-5D. These findings will be useful to compare GH when only one of the outcome measures is available.