An analysis of x-ray reader agreement: do five readers significantly increase reader classification reliability over that of three readers.
Lefante-J; Hughes-J; Jones-R; Weill-H
Proceedings of the VIIth International Pneumoconioses Conference, August 23-26, 1988, Pittsburgh, Pennsylvania, USA. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health, DHHS (NIOSH) Publication No. 90-108, 1990 Sep; (Part I):482-486
In a study of a working population engaged in man made mineral fiber manufacturing, the International Labor Office (ILO) 1980 Classifications of a set of chest X-rays by five readers were available for comparison with all possible subsets of three readers. The study materials consisted of the chest radiographs of 1,755 employees at seven facilities. All five readers attempted to classify all films according to the ILO 1980 Classification. A total of 1,603 films were judged readable and were classified by all five readers. The results indicated that when using three experienced readers, which was the widely accepted minimum for epidemiologic research, the effect of including two additional readers appeared to be negligible. Despite the use of readers with significant inter observer variability, only a small proportion of films were changed by more than one category after adding the two additional readers.
Diagnostic-techniques; Synthetic-fibers; Synthetic-fibers; X-ray-analysis; Radiographic-analysis; X-ray-diagnosis
DHHS (NIOSH) Publication No. 90-108
Proceedings of the VIIth International Pneumoconioses Conference, August 23-26, 1988, Pittsburgh, Pennsylvania, USA