[Skip to Navigation]
July 10, 1995

Interobserver Variability in Interpreting Chest Radiographs-Reply

Author Affiliations

Halifax, Nova Scotia

Arch Intern Med. 1995;155(13):1453. doi:10.1001/archinte.1995.00430130149021

in reply  Clark and Conrad raise several points of concern about our recent article.1 The first point is that even though the number and percentage of correct responses for the roentgenographic diagnosis of pneumonia improve with increasing experience (Table 2), we conclude the opposite. Perhaps one of our statements in the conclusions in the abstract was too strong: "This [interobserver] variability does not improve with increasing experience." Instead, the final sentence in the "Comment" section is more appropriate: "This variability only slightly decreases with increasing experience."The next point is that discrepancy in the study design might invalidate the study's conclusions. Consecutive patients admitted to our hospital with pneumonia formed the study cohort; the fact that they were elderly and had a high incidence of concomitant disease reflects reality for patients who are now admitted to hospital with pneumonia. There is no doubt that this represents a severe test