Referral and diagnosis of common rheumatic diseases by primary care physicians

Objective. To describe primary care patterns of referral and diagnoses of patients with rheumatic diseases referred to rheumatologists. Methods. The medical records of all consecutive patients referred in 1994 by >300 primary care physicians to two rheumatologists at an academic centre were reviewed. The referring physician diagnosis was compared with the rheumatologist's diagnosis. Sensitivity, specificity and predictive values of primary care diagnoses were estimated using the rheumatologist diagnosis as the 'gold standard'. Setting. University-based rheumatology out-patient clinic. Results. Over half of the patients referred had a rheumatologist diagnosis of soft-tissue rheumatism or a spinal pain syndrome. Three hundred and forty-seven patients (49%) had a primary care diagnosis of a defined rheumatic disease. Of these, 142 (41%) of the primary care diagnoses were subsequently modified by the rheumatologist. The highest agreement between primary care physician and rheumatologist was observed for crystal-induced arthritis (kappa = 0.86), and the lowest agreement for polymyalgia rheumatica (kappa = 0.39) and systemic lupus (kappa = 0.46). Sensitivity was lowest for a primary care diagnosis of fibromyalgia (48%) and highest for ankylosing spondylitis (94%). Positive predictive values were generally low, in particular for systemic lupus erythematosus (33%) and polymyalgia rheumatica (30%). Conclusion. Most patients referred to an academic rheumatology centre had soft-tissue rheumatism or other pain syndromes. In general, diagnostic agreement between rheumatologists and primary care physicians was low. Increased emphasis on musculoskeletal disorders should be encouraged in medical education to increase the efficiency of rheumatology referrals.