Table 2. BI-RADS density scores: inter-rater agreement and reliability (n = 992).
% (N) | |||
---|---|---|---|
ᴋw (95% CI) | % agreement | % agreement (a+b vs. c+d) | |
Rater 1 vs. 2 | 0.81 (0.78–0.83) | 74.3 (737) | 90.2 (895) |
Rater 1 vs. 3 | 0.84 (0.82–0.86) | 72.1 (715) | 89.0 (883) |
Rater 2 vs. 3 | 0.80 (0.78–0.82) | 67.6 (671) | 89.9 (892) |
Rater 1 vs. majority | 0.93 (0.91–0.94) | 89.0 (883) | 94.7 (939) |
Rater 2 vs. majority | 0.89 (0.87–0.91) | 84.8 (841) | 95.6 (948) |
Rater 3 vs. majority | 0.91 (0.89–0.92) | 82.8 (821) | 94.4 (936) |
ᴋw = weighted kappa scores (Fleiss-Cohen, quadratic weights), CI = confidence interval