Progressive joint damage, increasing deformity, and declining function characterize rheumatoid arthritis (RA). The evidence suggests structural joint damage is the predominant cause of functional impairment. Structural changes of joints are evaluated by imaging methods. Plain joint radiographs remain the best method for determining the extent and nature of structural changes at present. Newer technologies such as magnetic resonance imaging may eventually replace them. Clinical studies of antirheumatic drugs involving the assessment of prevention or significant decrease in the rate of progression of structural joint damage in RA should meet several standards. All patients who enter a study, whether they continue medication or not, must be assessed on its completion. Studies require sufficient power to determine realistic differences due to therapy. They should last long enough for a reliable analysis of the effects of joint damage; 1 year would be the minimal period for such a study, and 2 years would be preferable. Evaluation should concentrate on erosions and related structural changes in juxtaarticular bone; mapping osteoporotic areas in early disease may be a sensitive and objective measure. Assessments ought to use changes within the hands and wrists to indicate overall progression, with the feet included in evaluating early disease. There should be different therapeutic aims at various stages of RA, focussing on preventing new erosions developing in early disease (< 2 years from diagnosis), preventing new erosions occurring in established disease (2-5 years from diagnosis), reducing the rate erosions develop in established disease (< 5 years from diagnosis), and reducing the rate of joint destruction in late disease (> 5 years from diagnosis).