Reliability of the articular examination in children with juvenile rheumatoid arthritis: interobserver agreement and sources of disagreement.

OBJECTIVE: To assess the interobserver agreement of articular examination in children with juvenile rheumatoid arthritis (JRA) and identify sources of disagreement.
METHODS: Four rheumatologists graded tenderness/pain on motion, swelling, and limitation of motion in the joints of 10 children with JRA, as recommended by the Pediatric Rheumatology Collaborative Study Group, and 17 different joint indices were computed. Agreement was measured by kappa (kappa) and intraclass correlation coefficients (Ri).
RESULTS: All 4 observers detected tenderness in 15.7% of the joints, but they disagreed (2 vs 2) on 4.2% (kappa = 0.71). They detected swelling in 5.2% but disagreed on 6.2% (kappa = 0.47). They found limitation in 4.9%, but disagreed on 8.1% (kappa = 0.54). The tender joint count, and the American Rheumatism Association cooperating clinics and Hart modified Ritchie indices were the most reliable (Ri > 0.93); the swelling severity index fared the worst (Ri = 0.40). There were differences in examination maneuvers and judgment among examiners. Discrepancies were larger in metacarpophalangeal joints and in patients with many involved joints.
CONCLUSION: There was low agreement in the assessment of joint swelling and limitation of motion. Differences in examiners' techniques, patients with severe disease, and the small hand joints were important sources of disagreement.