Inter- and intra-observer variation in silicosis radiology.

Abstract

The radiologic criteria are important for the diagnosis of silicosis and must be objective. Initial radiologic changes are nonspecific and sometimes misinterpreted. We asked six Pulmonologists with distinct training in pulmonary disease to interpret 112 chest films according to a simple established protocol. No further information was given. One year later, all the observers analysed the same films over again. Then, we compared the results among the observers and between the two analysis made by the same individual. In 34 cases (30%) there was unanimity in asserting or denying the presence of micronodules and in 74 cases (66%) there was an agreement among the three more experienced observers. The intraobserver variability ranged from 8% to 40% and it was more important in the least trained observers. We concluded that when interpreting early radiologic changes in silicosis inter and intraobserver variability was elevated, more objectivity was achieved by the observers with more experience in epidemiologic studies and that we must be careful in interpreting results of epidemiologic studies concerning this matter.