Discordance between patient and physician ratings of rheumatoid arthritis (RA) severity occurs in clinical practice and correlates with pain scores and measurements of joint disease. However, information is lacking on whether discordance impacts patients' ability to work. We evaluated the discordance between patient and physician ratings of RA disease activity before and after treatment with etanercept and investigated the associations between discordance, clinical outcomes, and work productivity. In the PRESERVE clinical trial, patients with moderate RA received open-label etanercept 50 mg once weekly plus methotrexate for 36 weeks. Baseline and week-36 disease characteristics and clinical and work productivity outcomes were categorized according to week-36 concordance category, defined as positive discordance (patient global assessment - physician global assessment ≥2), negative discordance (patient global assessment - physician global assessment ≤ -2), and concordance (absolute difference between the two disease activity assessments = 0 or 1). Correlations between discordance, clinical outcomes, and predictors of discordance were determined. At baseline, 520/762 (68.2 %) patient and physician global assessment scores were concordant, 194 (25.5 %) were positively discordant, and 48 (6.3 %) were negatively discordant. After 36 weeks of therapy, 556/763 (72.9 %) scores were concordant, 189 (24.8 %) were positively discordant, and 18 (2.4 %) were negatively discordant. Patients with week-36 concordance had the best 36-week clinical and patient-reported outcomes, and overall, the greatest improvement between baseline and week 36. Baseline pain, swollen joint count, duration of morning stiffness, fatigue, and patient general health significantly correlated with week-36 discordance, p < 0.0001 to p < 0.05. Additionally, baseline pain, patient general health, and C-reactive protein were predictors of week-36 discordance (odds ratios 1.22, 1.02, and 0.98, respectively). For the employed patients, percent impairment while working and percent overall work impairment were highest (greatest impairment) at baseline and 36 weeks in the group with positive discordance. The percentage of patients with concordance increased after 36 weeks of treatment with etanercept, with concordant patients demonstrating the greatest improvement in clinical and patient-reported outcomes. Discordance correlated with several measures of disease activity and was associated with decreased work productivity. ClinicalTrials.gov identifier: NCT00565409 . Registered 28/11/2007.