Radiographic scoring systems for rheumatoid arthritis (RA) should be based on current understanding of disease pathology. Evidence suggests that there may be at least two intraarticular pathologies that may result in change in different radiographic features. There is therefore a strong argument for devising a radiographic score based on the observation of features rather than broad categorizations of the total radiographic change. Features may subsequently be amalgamated in relation to other criteria such as sensitivity, specificity, and responsiveness to change, and may be related to subsequent developments in understanding the biology of RA. A second challenge is in elucidating the relationship between radiographic change and the longterm consequences of RA for the patient. Current practice is predicated on the assumption that in the longterm radiographic change correlates well with functional loss and possibly noninflammatory, endstage joint pain. Although hand and feet radiographs broadly represent destructive change in all joints, in cross sectional studies they correlate only moderately with late stage functional loss. The issue may be resolved by longterm observational studies of radiographic change and functional loss. It is recommended that specific radiographic features relevant to joint pathophysiology be used to create a radiographic damage index for comparison with current scoring systems and that longterm observational studies specifically address the relationship between radiographic joint damage and functional outcome.