To determine the disease course and outcome in a multicenter cohort of patients with juvenile rheumatoid arthritis (JRA). All patients with JRA seen at 3 pediatric rheumatology centers were identified from databases and/or clinic records. Inclusion criteria were a diagnosis of JRA (1977 American College of Rheumatology criteria), a followup period of at least 5 years since onset, and a minimum age of 8 years. Patients were examined and completed a Childhood Health Assessment Questionnaire (CHAQ). Kaplan-Meier curves were constructed to estimate rates of remission, relapse, and arthroplasty. Remission was defined as absence of active arthritis while off treatment for at least 2 years. Outcome measures were active disease duration, CHAQ scores, pain determined by visual analog scales, physician's global assessments, and Steinbrocker functional classifications. Years of education and employment status were ascertained. We studied 392 patients of 652 (60%) who met the selection criteria. The probabilities of remission at 10 years after onset were 37, 47, 23, and 6% for patients with systemic, pauciarticular, RF- polyarticular, and RF+ polyarticular JRA, respectively. The probability of relapse varied from 30 to 100% at 15 years. The probability of arthroplasty varied from 13 to 57% after 15 years of active disease. We found 2.5% of patients assessed were in Steinbrocker Classes III or IV and 6% were in the highest CHAQ score (> 1.5) group. Compared with national statistics, fewer female patients received post-secondary education and unemployment rates for patients 20 to 24 years of age were higher. Our results indicate that JRA is a disease that often extends into adulthood. Compared to previous decades, functional outcome has improved; however, the estimated rate of arthroplasty remains very high. Patients with JRA may have difficulty entering the workforce.