UTILIZATION AND PREDICTIVE VALUE OF LABORATORY TESTS IN PATIENTS REFERRED TO RHEUMATOLOGISTS BY PRIMARY-CARE PHYSICIANS

Objective, Antinuclear antibodies (ANA), rheumatoid factors (RF), and erythrocyte sedimentation rate (ESR) are among the most frequently req uested tests in the diagnosis and investigation of connective tissue d iseases (CTD). We evaluate the utilization patterns and predictive val ue of these tests in patients referred to rheumatologists by primary c are physicians. Methods. We reviewed the records of all new patients r eferred by primary care physicians in 1994 to 2 rheumatologists practi cing at the University of Alberta. Data extracted from the records inc luded diagnostic tests requested by referring primary care physicians, signs and symptoms at the initial rheumatology consult, and followup diagnoses. Results. Seven hundred eleven new patients had been referre d by over 300 primary care physicians: RF had been requested in 25%, A NA in 21%, and ESR in 29%. One hundred nine (15%) of the 711 patients had a CTD, 45 (6%) had rheumatoid arthritis (RA), and 8 (1%) systemic lupus erythematosus (SLE). The predictive values of positive tests for the diagnosis of CTD were low: 49% for RF, 29% for ANA, and 35% for E SR. For RA, the positive predictive values were 44% for RF, 8% for ANA , 17% for ESR; for SLE, 2, 12, and 3%, respectively. Diffuse musculosk eletal pain and fatigue were significantly associated with test utiliz ation, although most patients with these symptoms had fibromyalgia or localized soft tissue rheumatism. Conclusion. Primary care physicians frequently requested autoantibodies in patients referred to rheumatolo gists. Most tests were negative, and were often requested in patients without CTD, resulting in low positive predictive values and questiona ble clinical utility. These findings suggest inappropriate overuse and lack of understanding of the use of autoantibody tests in diagnosing rheumatic diseases. A decrease in inappropriate use could be achieved by emphasizing that fatigue and diffuse musculoskeletal pain are not i ndicative of CTD in the absence of other features such as joint swelli ng, typical rash, or organ involvement.