OBJECTIVE: To determine which activity indices better correlate with assessor's (AGA) and patient's (PGA) global assessment of disease activity and to compare the improvement with American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) criteria and their association with PGA and AGA of overall improvement. METHODS: Seventy-five patients with rheumatoid arthritis (RA) were studied. Swollen and tender joints, morning stiffness, grip strength, pain, AGA, PGA, Health Assessment Questionnaire (HAQ) score, erythrocyte sedimentation rate (ESR), C-reactive protein (CPR), and hemoglobin were determined before and 6 months after treatment. Several activity indices were calculated: Disease Activity Score (DAS), DAS 3, DAS 28, DAS '28,' ACR > or = 20%, Mallya, Riel, IDA, and a modification of the Stoke index. RESULTS: All indices correlated with PGA and AGA before and after treatment (r > 0.38, p < 0.01), but better results were obtained with AGA than PGA. DAS, DAS 3, DAS 28, and modified Stoke had the best correlation with AGA (r > or = 0.77, p < 0.01). The indices that better detected the differences after treatment for AGA were DAS, DAS 3, DAS 28, and modified Stoke (r > or = -0.42, p < 0.01). The level of agreement between EULAR and ACR improvement classifications with both reduced and extensive joint counts was comparable and its association with PGA and AGA overall improvement was significant (p < 0.01). CONCLUSION: All activity indices correlated with PGA and AGA, although the best results were obtained with AGA. Although indices' correlations were similar, the DAS group and the modified Stoke seemed to be the most useful indices to measure disease activity in RA. The discriminating potential between ACR and EULAR improvement classification was comparable, as was the association with PGA and AGA overall improvement.