Factors associated with the decision of the rheumatologist to order sacroiliac joints magnetic resonance imaging (SI-MRI) or HLA-B27 testing in the diagnostic work-up of patients with spondyloarthritis in clinical practice

To evaluate the patients' characteristics associated with the clinical decision to request SI-MRI and/or HLA-B27 in patients with SpA in daily practice. Patients referred to a rheumatology outpatient-clinic in a national referral-centre were selected. Patients with a clinical diagnosis of SpA according to the rheumatologist were included. SI-MRI and HLA-B27 was available for patients in whom the rheumatologists had ordered these tests. Characteristics associated with ordering SI-MRI or HLA-B27 were identified with univariable analyses. Variables with p-value 80% completeness were selected for further analysis. A multivariable logistic regression analysis was used to evaluate the determinants related with the decision to perform SI-MRI and/or HLA-B27 and odds ratios with 95% confidence intervals were calculated. In total, 581 patients with SpA were included in the cohort, 72% were men, mean age 34.6±12.1 and disease duration 7.3±9.7 years. Of these patients, 24% (n=137) had SI-MRI and 77% (n=441) had HLA-B27 tests ordered. Independently predictive factors for ordering a SI-MRI were the presence of IBP (OR=1.81), enthesitis (OR=1.57) and the number of initial-symptoms at presentation (OR=1.27 per additional symptom present). Independently predictive factors of HLA-B27 testing were the number of initial-symptoms (OR=1.45 per symptom) and uveitis (OR=3.19). This study strongly suggests that rheumatologists use certain clinical clues to decide if they order expensive and scarce tests in the diagnostic work-up of SpA patients. These manifestations may increase the efficiency of these tests in clinical practice and suggest that clinical reasoning follows principles of Bayesian theory.