Skip to main content
. 2005 Aug 13;331(7513):379–382. doi: 10.1136/bmj.331.7513.379

Table 2.

Intraobserver agreement among Russian clinicans on evaluation of chest radiographs. Values are κ (95% confidence intervals)

Radiographic finding All participants Tuberculosis specialists Radiologists Respiratory specialists
Presence of clinically important abnormality all five categories 0.457 (0.004 to 0.911) 0.456 (−0.036 to 0.949) 0.604 (0.347 to 0.861) 0.292 (−0.141 to 0.725)
Localisation of abnormality* 0.477 (−0.009 to 1.044) 0.473 (−0.081 to 1.026) 0.627 (0.143 to 1.110) 0.277 (−0.280 to 0.834)
Presence of cavity 0.358 (−0.595 to 1.311) 0.443 (−0.202 to 1.089) 0.594 (−0.205 to 1.393) 0.448 (−0.244 to 1.140)
Radiographic findings consistent with tuberculosis 0.493 (−0.03 to 1.016) 0.481 (−0.019 to 0.981) 0.529 (−0.042 to 1.100) 0.483 (−0.078 to 1.044)
Form of tuberculosis 0.488 (−0.073 to 1.050) 0.448 (−0.074 to 0.970) 0.611 (0.021 to 1.220) 0.449 (−0.144 to 1.042)
Tuberculosis process active 0.490 (−0.098 to 1.078) 0.471 (−0.097 to 1.039) 0.546 (−0.158 to 1.250) 0.470 (0.015 to 0.927)
*

Average of κ values for each radiological zone.

According to Russian classification: military, focal, caseous pneumonia, cavernous, cirrhotic, tuberculosis of mediastinal lymph nodes, infiltrative, disseminated, tuberculoma, fibrocavernous, pleuritis.