Fig. 3.

Between-assessor agreement on trial population characteristics important for similarity assessment (dark and light bars represent the first and second pairs of trials, respectively). The majority of characteristics were not viewed as important by most assessing physicians.