Background/Purpose: Prompt identification and referral of patients with suspected inflammatory arthritis to a rheumatology clinic are crucial for early diagnosis and treatment. In the referral letter to our Early Arthritis Clinic (EAC), the physician should refer, out from seven possible criteria, the ones motivating the referral. The objectives of our study were: 1) to assess the level of agreement between the referring physician and the rheumatologist, regarding the presence of referral criteria; 2) to identify the set of criteria that best predicts the presence of inflammatory arthritis.

Methods: Unicentric retrospective observational study, including patients observed in the EAC until May/2016. Subjects were excluded if they had been referred to the EAC by a rheumatologist and if we couldn´t access the referral letter or the medical records from the first visit to the EAC. Demographic data, provenience, referral criteria and the final diagnosis were retrieved from individual clinical files and the Portuguese Registry of Rheumatic Patients – Reuma.pt. For the referral criteria, the agreement between the referring physician and the rheumatologist in the first EAC visit, was assessed using the Cohen’s Kappa. In the second step, we created four new variables, each one corresponding to a set of criteria, as indicated by the referring physician. ROC curves were drawn by a nonparametric method in order to identify the set that best predicts the presence of inflammatory arthritis.

Results: We included 132 patients, 66% females, mean age 52±17 years; Referred from primary care (72%), emergency department (17%) and other hospital specialties (11%). An inflammatory arthritis was diagnosed in 73% of the cases.  Table 1 shows the level of agreement between the referring physician and the rheumatologist, regarding the presence of each of the referral criteria. Table 2 shows the area under the curve for each set of criteria.

Conclusion: Among the clinical criteria, there was poor agreement regarding the presence of arthritis and no significant agreement regarding the characterization of the arthralgia as inflammatory, the morning stiffness and the squeeze test. The weak performance of all sets of criteria in predicting inflammatory arthritis is probably due to deficient recognition of each criterium. These results suggest the need for improving education among the physicians referring patients to the EAC.

Table 1: Agreement between the referring physician and the rheumatologist, regarding the presence of each of seven criteria.

Table 2: Area under the curve for each set of criteria, as indicated by the referring physician.