Objective: The aims of this study were to assess the agreement of physicians and nurses performing tender and swollen joint counts (TJCs/SJCs) in rheumatoid arthritis (RA) and identify factors that might influence their examinations including patient age, sex, race, RA disease duration, body mass index, RA disease activity level, comorbid fibromyalgia, comorbid osteoarthritis, and levels of acute-phase reactants.
Methods: Seventy-two RA participants underwent TJCs/SJCs of 28 joints using a standardized protocol by 2 nurses and 2 rheumatologists. Demographic, laboratory, radiographic, and clinical data were obtained to assess the influence of these factors on TJCs/SJCs. Intraclass correlations (ICCs) among examiners were determined for TJCs/SJCs. Nurse-physician differences and agreement of individual joints were evaluated using Cohen κ. Analysis of variance was performed to detect differences in means between examiners for TJCs/SJCs. Intraclass correlation and Fisher Z tests were used to identify factors influencing TJCs/SJCs.
Results: Agreement was strong among these nurses and physicians for total TJCs/SJCs (ICC = 0.84/ICC = 0.79, respectively). κ was best for hand joint tenderness and poorest for shoulder swelling. Some significant differences in mean TJCs/SJCs were found between examiners. Fibromyalgia significantly reduced agreement of both TJCs and SJCs. Agreement of TJC was significantly reduced when patients had lower disease activity, greater work impairment, lower mental health quality of life, and elevated erythrocyte sedimentation rate, whereas female sex, assessor's perception of but not radiographic hand osteoarthritis, and elevated C-reactive protein significantly reduced agreement for SJC.
Conclusions: Strong agreement was found among nurses and physicians for total 28-joint counts, with agreement at individual joints being stronger for tenderness than swelling. Fibromyalgia significantly reduced ICCs of TJCs/SJCs.