Comparison of Chest Radiograph Interpretations of Community-Acquired Pneumonia by Clinicians in Nakhon Phanom Province and Chest Disease Institute Radiologist Panel-เปรียบเทียบผลการอ่านภาพถ่ายรังสีทรวงอกผู้ป่วยโรคปอดอักเสบในชุมชน ระหว่างแพทย์ประจำโรงพยาบาลในจังหวัดนครพนมกับรังสีแพทย์ผู้เชี่ยวชาญจากสถาบันโรคทรวงอก

Abstract

            The purpose of this study was to compare the agreement levels on chest radiograph interpretation of suspected community-acquired pneumonia (CAP) by a panel of radiologists from chest disease institute and clinicians in Nakhon Phanom. Trained surveillance staff used standard form to collect data from patients, hospitalized with suspected CAP. Chest radiographs were digitized then
sent for panel interpretations. Reference results from radiologists from chest disease institute were finalized by 2 out of 3 opinions. 

             A total of 3,882 clinical pneumonia patients with chest radiograph were included in our study from December 2003 to December 2004. In all, 522 chest films or 13 percent of the total were interpreted by chest disease institute radiologists. Half of the patients were eventually discharged with diagnoses of pneumonia. The proportions of pneumonia evidence as interpreted by local clinicians’ reading were interstitial infiltrate at 77.9 percent, alveolar infiltrate 16 percent, consolidation at 15.7 percent, pleural effusion, cavitation and atelectasis at 6.1, 4.2 and 1 percent respectively. However the proportion of pneumonia evidence as interpreted by the panel of radiologists at chest disease institute were interstitial infiltrate at 67.2 percent, alveolar infiltrate at 21.1 percent, consolidation at 16.6 percent, pleural effusion, cavitation and atelectasis at 4.5, 2.6 and 0.4 percent respectively. Chest radiograph reading agreement levels among the readers ranged from kappa values of 0.14 to 0.6 (Interstitial infiltrate, K = 0.143; Consolidation, K = 0.215; Pleural effusion, K = 0.6;Cavitation, K = 0.368).

           In summary, it was found that a low to moderate agreement rates of chest radiograph interpretations between the two different readers. The study suggested that local clinicians should consult a radiologist for expert interpretation of chest radiograph, whenever possible. The larger scale study
should be conducted for a better generalization of the national situation in Thailand.

Key words: chest radiograph, community acquired pneumonia (CAP)