Variability in the prescription of non-biologic disease-modifying antirheumatic drugs for the treatment of spondyloarthritis in Spain

Objective: To describe the variability in the prescription of non-biologic disease-modifying antirheumatic drugs (nbDMARDs) for the treatment of spondyloarthritis (SpA) in Spain and to explore which factors relating to the disease, patient, physician, and/or center contribute to these variations. Methods: A retrospective medical record review was performed using a probabilistic sample of 1168 patients with SpA from 45 centers distributed in 15/19 regions in Spain. The sociodemographic and clinical features and the use of drugs were recorded following a standardized protocol. Logistic regression, with nbDMARDs prescriptions as the dependent variable, was used for bivariable analysis. A multilevel logistic regression model was used to study variability. Results: The probability of receiving an nbDMARD was higher in female patients [OR ¼ 1.548; 95% confidence interval (CI): 1.208–1.984], in those with elevated C-reactive protein (OR ¼ 1.039; 95% CI: 1.012–1.066) and erythrocyte sedimentation rate (OR ¼ 1.012; 95% CI: 1.003–1.021), in those with a higher number of affected peripheral joints (OR ¼ 12.921; 95% CI: 2.911–57.347), and in patients with extra-articular manifestations like dactylitis (OR ¼ 2.997; 95% CI: 1.868–4.809), psoriasis (OR ¼ 2.601; 95% CI: 1.870–3.617), and enthesitis (OR ¼ 1.717; 95% CI: 1.224–2.410). There was a marked variability in the prescription of nbDMARDs for SpA patients, depending on the center (14.3%; variance 0.549; standard error 0.161; median odds ratio 2.366; p o 0.001). After adjusting for patient and center variables, this variability fell to 3.8%. Conclusion: A number of factors affecting variability in clinical practice, and which are independent of disease characteristics, are associated with the probability of SpA patients receiving nbDMARDs in Spain.