Grinding the Axe-ray! The Controversy Surrounding Radiograph Interpretation in the ED
ABSTRACT & COMMENTARY
Source: Klein EJ, et al. Discordant radiograph interpretation between emergency physicians and radiologists in a pediatric emergency department. Pediatr Emerg Care 1999;15:245-248.
This prospective study enrolled 2083 radiographs obtained in a university-affiliated children’s hospital emergency department (ED). This represented one-third of the radiographs obtained during the study period, and only the films that had an interpretation recorded by an emergency physician (EP). One of the authors (a radiologist) determined if the ED reading and radiology over-reading were either concordant (83%) or discordant (17%). Klein and associates then eliminated a few films that were not bone, chest, or abdomen x-rays and studied the remaining 324 discordant films. Klein et al then reviewed the ED chart and determined if the discordance was clinically significant. For example, if the ED interpretation and the radiology interpretation differed on location of infiltrate or a "normal" was later interpreted as a "viral pattern", this was considered clinically insignificant. Other examples included splinted extremity fractures when the ED interpretation was "no fracture" and the radiology interpretation was "possible fracture." This left 23 clinically significant discordant interpretations for an overall rate of 1.1%. Of the 23 patients having clinically significant discordance, 12 were patients given antibiotics for an infiltrate when one was not identified on re-interpretation. Other examples of the missed findings include three patients recalled for infiltrate, one patient with pulmonary edema and myocarditis mistakenly treated as pneumonia, and a patient with missed free air on an abdominal plain film who was admitted for vomiting and diarrhea and subsequently had a perforated appendix. Klein et al conclude that EPs could benefit from more rigorous interpretation of chest radiographs to avoid unnecessary treatment with antibiotics. In addition, they note that the small number of significant misinterpretations does not make 24-hour radiology interpretation cost-effective.
Comment by Richard J. Hamilton, MD, FAAEM, ABMT
There is an agenda that drives studies of this nature. A major debate is emerging in this country which has, as its central focus, one essential question: "Who gets paid for interpreting studies in the ED?" With health care costs rising, why pay a radiologist to read a radiograph after the patient has left the ED and is already being treated? In one study, the cost has been reported to be approximately $300 per study. Studies of this nature lay the groundwork for what I believe will be the radiology perspective—1) EPs occasionally miss significant findings that could lead to adverse outcomes; 2) the radiologist’s interpretation is the only correct one, the final impression, and necessary to maintain quality care; and 3) the current practice is of value to the patient. EPs will argue that numerous larger, combined adult/pediatric studies, as well as this one, confirm that radiologists’ re-interpretations change clinical management in less than 1% of all cases. One interesting suggestion is to have the EP submit the radiograph for re-interpretation on a selected basis. This would decrease the cost of patient care and still maintain quality. ECG reading is another example where this quality assurance practice is of questionable value to the patient.
However, while it may appear that I am strident on the issue of cost, I am actually more troubled by the medico-legal implications of the re-interpretation and "final impressions" that can appear to contradict an EP’s clinical decisions. What is particularly troubling about this study and other studies of this sort is that they are often done as a quality assurance project, and thus are not subject to the standard rigors of interobserver and intraobserver reliability such as kappa statistics. Numerous studies demonstrate that even radiologists cannot 1) completely agree with each other, and 2) completely agree with their first interpretation when given the same radiograph at a different time! The truth is found in the axiom "Treat the patient, not the test results." EPs should interpret radiographs with the confidence that they do an excellent job, and that they are clinicians who use images to care for patients, whatever the "final impression" might be.
Important issues regarding research on x-ray interpretation in the ED include all of the following except:
a. clinical significance of radiologist over-reads.
b. interobserver reliability.
c. intraobserver reliability.
d. region of the body that was imaged.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content