Which factors explain the patient global assessment in patients with ankylosing spondylitis? A hierarchical cluster analysis on REGISPONSER-AS.

Abstract Objectives To determine groups of factors (clusters) potentially associated with the patient global assessment (measured with the Bath Ankylosing Spondylitis Patient Global Score (BAS-G)), and to quantify the contribution of each cluster to the patient's well-being. Methods This was a cross-sectional study in patients with a diagnosis of ankylosing spondylitis (AS) from the national, multicentre Spanish REGISPONSER-AS registry. A hierarchical cluster analysis was conducted to group the potential factors (sociodemographic, socioeconomic, patient-reported outcomes, physical exploration variables and depression) associated with the BAS-G. The contribution of each cluster to the variability of the BAS-G was evaluated using a multivariate linear regression model and the determination coefficient (R2) for each cluster. Results A total of 681 patients with complete data were included. Three clusters of variables potentially associated with the BAS-G were found: cluster 1 contained the Maastricht Ankylosing Spondylitis Enthesitis Score (MASES), depression, sex (female) and university studies; cluster 2 included the Graffar scale, age and body mass index; and cluster 3 contained the Bath Ankylosing Spondylitis Functional Index (BASFI), the individual items of the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), pain during the last week, nocturnal pain and the number of swollen joints. A total of 6.8% of the variability of the BAS-G was explained by cluster 1, 0.5% was explained by cluster 2, and 60.8% was explained by cluster 3. Conclusion The BAS-G is mostly explained by pain and function, while demographic and socioeconomic factors are weakly associated with the BAS-G. Depression also has a weak effect on this score.