Table 5 Inter-observer agreement on five-point confidence scale scores among the three radiologists

From: Performance of the deep convolutional neural network based magnetic resonance image scoring algorithm for differentiating between tuberculous and pyogenic spondylitis

Ā 

Kappa value†

95% CI

Readers 1 and 2

0.6728

0.6090–0.7366 (substantial agreement)

Readers 1 and 3

0.6644

0.6027–0.7261 (substantial agreement)

Readers 2 and 3

0.7399

0.6923–0.7876 (substantial agreement)

  1. Abbreviations: CI, confidence interval.
  2. †We used the linear weighted kappa to account for partial agreement because the outcome encompasses ordinal scoring.