Abstract Objective To analyze patients with rheumatoid arthritis (RA) for inflammatory activity by physician estimate of global assessment (DOCGL) vs an estimate of inflammatory activity (DOCINF) to explain variation in swollen joint count (SJC). Methods Patients with RA were studied at routine care visits. Patients completed a multidimensional health assessment questionnaire (MDHAQ), and the physician completed a 28-joint count for swollen (SJC), tender (TJC) and deformed (DJC) joints, and a RheuMetric checklist with a 0–10 DOCGL visual numeric scale (VNS), and 0–10 VNS estimates of inflammation (DOCINF), damage (DOCDAM) and patient distress (DOCSTR). Disease activity score 28 (DAS28-ESR), clinical disease activity index (CDAI) and routine assessment of patient index data (RAPID3) were calculated. Individual scores and RA indices were compared according to Spearman correlation coefficients and regression analyses. Results 104 unselected patients were included, with median age and disease duration of 55.0 and 5 years respectively. Median (IQR) DAS28-ESR was 2.89 (1.64), indicating low activity. DOCINF was correlated significantly with DOCGL (rho= 0.775). Both DOCGL and DOCINF were correlated significantly with most other measures; correlations with DOCGL generally were higher than with DOCINF other than for SJC. In regression analyses, DOCINF was more explanatory of variation in SJC than DOCGL and other DAS28-ESR components. Conclusions Variation in SJC is explained more by a 0–10 DOCINF VNS than the traditional DOCGL or any other measure in routine care RA patients. DOCINF on a RheuMetric checklist can provide informative quantitative scores concerning inflammatory activity in RA patients monitored over long periods.