Observer variation in quantitative assessment of rheumatoid arthritis Part II. A simplified scoring system

Abstract

A new and more comprehensive simplified scoring system to provide quantitative assessment of radiographic findings in rheumatoid arthritis was studied by six observers with different levels of expertise. Since the observers, film set, and joints scored were the same as in a prior study of a more detailed scoring method, a precise comparison of the results was made. Interobserver variation with the simplified scoring system was small; there were highly significant correlations of the total radiographic scores for all observers (R = .898-.978, P less than .001). There were highly significant correlations between scores obtained with the simplified scoring system and those obtained with the more detailed method (R = .920-.955, P less than .001). When the same joints were evaluated by the two methods, a significantly greater number of possible observations could be scored with the simplified scoring system than with the more detailed method (X2 = 131.07, P less than .001). The simplified scoring system required 2.3 times less time to use in the hands of experienced observers.