Misdiagnosis and prevalence of rheumatological diseases in early inflammatory arthritis clinic: results from a UK single-centre retrospective cohort study.

Objectives: Novel diagnostic tools have been developed to aid diagnosis and reduce misdiagnosis. However, the accuracy of these tools depends on estimates of disease prevalence. This study aimed to determine the prevalence of rheumatological diseases in a real-world cohort of patients and assess the rate of misdiagnosis at first presentation and follow-up. Methods: Patients with symptoms of inflammatory arthritis were referred by primary care. Rheumatologists made initial diagnoses, initiated treatment and revised diagnoses during follow-up. Clinical data including diagnoses were extracted from electronic patient records. Results: Over 2 years, 260 patients were seen. Most patients presented with peripheral joint symptoms (95%), and 23% had axial involvement. Psoriasis was present in 15%, inflammatory bowel disease in 4% and uveitis in 4%. Autoantibodies were positive for cyclic citrullinated peptide in 11% and rheumatoid factor in 12%.Seventeen per cent (n = 45) of patients were misdiagnosed, with 8% (n = 20) initially diagnosed with 1 of the 5 most common rheumatological conditions but later found to have a noninflammatory disorder. Conversely, 7% (n = 17) initially received a diagnosis of an 'other rheumatological disease' but later had their diagnoses revised to 1 of the 5 most common rheumatological conditions. Conclusions: Disease prevalence is context-specific, and prevalence estimates in the intended setting should be determined prior to implementation of novel diagnostic tools. In real-world cohorts, noninflammatory diseases like chronic pain syndromes were more prevalent than in observational studies and registries. Misdiagnosis occurs in about one-fifth of patients, highlighting the benefits of novel diagnostic tools for earlier, more accurate diagnosis.