Synthetic disease-modifying antirheumatic drug prescribing variability in rheumatoid arthritis: a multilevel analysis of a cross-sectional national study.

The objective of this study was to describe the variability in the prescription of csDMARDs for the treatment of RA between centers in Spain and to explore how this variability relates to demographic, disease, physician, and institutional characteristics. A cross-sectional nationwide study was carried out to examine data from 1352 patients. Multilevel logistic regression with two levels was performed to assess the relationships between individual and disease-related factors, as well as physician and hospital characteristics, vis-à-vis csDMARD prescription. Having three or more comorbidities (OR 0.353 [0.173-0.721]), disease duration (OR 0.321 [0.174-0.595]), and the existence of an early-arthritis unit (OR 0.552 [0.335-0.910]) were negatively associated with the prescription of one csDMARD versus nonprescription; contrary, the presence of rheumatoid factor (OR 1.909, 95 % CI [1.181-3.086]) was positively associated. On the other hand, while corticoid intake (OR 1.561 [1.088-2.240]), the maximum number of painful joints, and the presence of nursing consultation (OR 1.626 [1.078-2.452]) were positively associated with the prescription of multiple csDMARDs versus one csDMARD, patient's age (OR 0.984 [0.974-0.995]) and disease duration (OR 0.669 [0.462-0.968]) were negatively associated. Despite all these, variability in the prescription of csDMARDs between hospitals remained statistically significant after adjusting for these individual and hospital characteristics. Within the emAR II study, there was a marked variation in the number of csDMARDs prescribed between hospitals. The reasons for these variations remain unclear and cannot be solely related to disease or center characteristics.