Methodological issues in radiographic scoring methods in rheumatoid arthritis.

Radiographs are important endpoints in clinical trials in rheumatoid arthritis (RA). Several scoring methods exist. However, many methodological issues are unsolved and require more attention. The following issues will be addressed: the abnormalities that should be scored; joints that should be included in a scoring method; whether both right and left hands and feet should be scored; views that should be used; the order in which radiographs should be scored; how data should be evaluated; how intra/interobserver variation and sensitivity to change should be assessed; the optimum number of readers to assess radiographs; the score to be used if there are multiple readers; quality assurance, international training, a validation set of radiographs, and automated scoring of radiographs.