Cluster analysis identifies the differential impact of disease activity and severity on functional status and patient satisfaction in rheumatoid arthritis: the FRANK registry.

Abstract

Objectives: The purpose of the present study was to investigate the differential impact of disease activity and severity on functional status and patient satisfaction in rheumatoid arthritis (RA) using cluster analysis on data from the FRANK registry.

Methods: Data from 3,619 RA patients in the FRANK registry were analysed. Patients were grouped using hierarchical and k-means cluster analyses based on age, physician's global assessment (PhGA), patient's pain assessment (PtPA), and Steinbrocker stage. Clusters were evaluated for differences in functional status (mHAQ), quality of life (EQ5D), and patient satisfaction.

Results: Five distinct patient clusters were identified. In hierarchical cluster analysis, Cluster 1 (n=1195, 33.0%) and 2 (n=641, 17.7%) with lower disease activity and severity demonstrated better functional outcomes (mHAQ: 0.18±0.30 and 0.15±0.26, respectively) and higher satisfaction, with treatment efficacy scores of 1.9±0.7 and 2.0±0.7, respectively (1: very satisfied to 6: very unsatisfied). Cluster 3 (n=1117, 30.9%), characterised by less activity and more severity, showed significant joint damage (Steinbrocker stage III-IV: 95.4%) despite controlled inflammation. Cluster 4 (n=385, 10.6%), characterised by patient-physician discordance in disease activity (mean PhGA: 0.9±0.5; mean PtPA: 5.0±2.1), had a more pronounced negative effect on satisfaction. Cluster 5 (n=281, 7.8%), with more activity and moderate severity, had the poorest outcomes in functional status (mHAQ: 0.87±0.65), quality of life (EQ5D: 0.60±0.17), and satisfaction, with a treatment efficacy score of 2.9±0.9. k-Means clustering produced overall similar clusters to hierarchical clustering, allowing the same labels for Cluster 1 to Cluster 5.

Conclusions: The study highlights the importance of understanding the heterogeneous nature of RA and its impact on patient outcomes. Personalised treatment approaches that address both objective disease measures and subjective patient experiences are essential for optimising RA management. Identification of distinct patient phenotypes, particularly those in Clusters 3, 4, and 5, may guide tailored interventions to improve treatment satisfaction and long-term outcomes in RA.