Interobserver Variability in the Interpretation of Chest Roentgenograms of Patients With Possible Pneumonia
Michelle Young;
Thomas J. Marrie, MD
Arch Intern Med. 1994;154(23):2729-2732.
Abstract
| |
Background
Previous studies have found considerable interobserver variability in the roentgenographic diagnosis of pneumonia. In this study we determined the impact of experience on interobserver variability in the interpretation of chest roentgenograms of patients with pneumonia and defined the characteristics of chest roentgenograms interpreted as pneumonia by all groups.
Methods
The chest roentgenograms of 15 consecutive patients with a clinical and roentgenographic diagnosis of pneumonia were read by a convenience sample of 10 first-year and 20 fourth-year medical students, 21 medical residents, and 21 staff physicians from the department of medicine. Three board-certified radiologists served as the gold standard. The majority rule decision and logistic regression analysis were used to analyze agreement among the various groups of observers.
Results
Agreement with the majority decision of the gold-standard radiologists on the presence or absence of pneumonia for all observations of the various groups were as follows: original radiologist, 87%; first-year medical students, 59%; fourth-year medical students, 54%; medical residents, 66%; and attending staff, 72%. Chest roentgenograms with dense lobar or segmental opacities were generally interpreted as pneumonia by all observers. Patchy opacities caused major disagreements and were usually read as not being due to pneumonia. Bronchopneumonia was not diagnosed by any of the observers. Air bronchograms, atelectasis, and chronic obstructive lung disease were usually not recognized by the nonradiologists.
Conclusions
There is considerable interobserver variability in the roentgenographic diagnosis of pneumonia. This variability does not improve with increasing experience. Dense lobar or segmental opacities are uniformly recognized as pneumonia while patchy opacities result in a variety of interpretations.
(Arch Intern Med. 1994;154:2729-2732)
Author Affiliations
From the Departments of Medicine (Ms Young and Dr Marrie) and Microbiology (Dr Marrie), Dalhousie University and Victoria General Hospital, Halifax, Nova Scotia.
 CiteULike  Connotea  Del.icio.us  Digg  Reddit  Technorati  Twitter
What's this?
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES
Factors Influencing In-hospital Mortality in Community-Acquired Pneumonia: A Prospective Study of Patients Not Initially Admitted to the ICU
Marrie and Wu
Chest 2005;127:1260-1270.
ABSTRACT
| FULL TEXT
Interobserver Variability in Interpreting Chest Radiographs
Clark and Conrad
Arch Intern Med 1995;155:1453-1453.
ABSTRACT
Interobserver Variability in Interpreting Chest Radiographs-Reply
Marrie
Arch Intern Med 1995;155:1453-1453.
ABSTRACT
|
