Table 6 Reliability of the gold standards for the fracture dataset.

From: The Feature Ambiguity Mitigate Operator model helps improve bone fracture detection on X-ray radiograph

Body part

Cohen’s κ

All Fleiss’ κ

Between physicians 1 and 2

Between physicians 1 and 3

Between physicians 2 and 3

Hand

0.95

0.97

0.92

0.94

Wrist

0.74

0.81

0.73

0.76

Elbow

0.83

0.72

0.67

0.74

Shoulder

0.85

0.95

0.79

0.86

Pelvic

0.85

0.87

0.86

0.86

Knee

0.81

0.88

0.79

0.83

Ankle

0.93

0.95

0.92

0.93

Foot

0.83

0.86

0.77

0.82

All

0.85

0.88

0.81

0.85

  1. The guidelines of Fleiss and colleagues characteristic κ values of more than 0.75 were set as excellent agreement, 0.40–0.75 as fair to good agreement, and less than 0.40 as poor agreement beyond chance.
Back to article page