| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, The University of Arizona Health Sciences Center,
1501 N. Campbell Ave., Tucson, AZ 85724-5067.
2 Radiology Ltd., 3170 E. Ft. Lowell Rd., Tucson, AZ 85716.
3 Department of Radiology, St. Mary's Hospital, 1601 W. St. Mary's Rd., Tucson,
AZ 85745.
Received May 8, 2002;
accepted after revision November 22, 2002.
Address correspondence to W. K. Erly.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We studied the interpretations of 716 consecutive emergency CT scans of the head by a group of 15 board-certified general radiologists practicing in the community (as opposed to an academic setting). The scans were sent via teleradiology, and the preliminary interpretations were made. Three of the general radiologists were functioning as nighthawks, and the remaining 12 were acting as on-call radiologists in addition to their normal daytime duties. Each CT examination was interpreted by one of five neuroradiologists the day after the initial interpretation had been performed. The findings of the final interpretation and the preliminary interpretation were categorized as showing agreement, insignificant disagreement, or significant disagreement. The reports in the two categories indicating disagreement were reviewed and reclassified by a consensus of three university-based neuroradiologists.
RESULTS. Agreement between the initial interpretation by the general radiologist and the final interpretation by the neuroradiologist was found in 95% of the CT scans. The interpretations were judged to show insignificant disagreement in 3% (23/716) of the scans and to show significant disagreement in 2% (16/716). Of the 16 significant errors, five were false-positive findings and 11 were false-negative findings. Forty-seven CT scans depicted significant or active disease, and in 11 (23%) of these scans, the final report differed significantly from the preliminary interpretation. Three patients had pituitary masses, none of which had been described on the preliminary interpretation.
CONCLUSION. The rate of significant discordance between board-certified on-call general radiologists and neuroradiologists in the interpretation of emergency CT scans was 2%, which was comparable to previously published reports of residents' performance. The pituitary gland may be a blind spot, and additional attention should be focused on this area.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All CT scans were sent via teleradiology (Cmax, Ikon Systems, Fremont, CA), and preliminary interpretations were made. Bone window settings were included if the patient's presenting complaint was related to trauma. CT scans were sent as image files so that each reviewer had only the capability of changing the brightness and contrast of the scans; the reviewers were unable to perform true adjustments of window or level settings. Three of the general radiologists were functioning as nighthawks, and the remaining 12 were acting as on-call radiologists in addition to their normal daytime duties. A hard copy of each CT scan was subsequently interpreted by one of five neuroradiologists with a certificate of added qualification.
The findings of the final interpretation and the preliminary interpretation were categorized as showing agreement, insignificant disagreement, or significant disagreement. Preliminary reports and final neuroradiologist's interpretations of the CT scans that fell into one of the two disagreement categories were then reviewed by three university-based neuroradiologists who recategorized the scans in which the interpretations disagreed. A disagreement was considered significant if an adverse patient outcome or a change in patient management (for example, missing early signs of cerebral ischemia, which could have resulted in the inappropriate administration of a thrombolytic agent) could have occurred. A disagreement was insignificant if little or no possibility for adverse outcome (such as missing a remote lacunar infarction) existed.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Overall, of the 39 cases in which the initial and final findings disagreed, 16 were believed to be significant errors, and 23 were insignificant. Of the significant errors, five were false-positive findings (Table 1). Eleven errors were false-negative findings (Table 2). Forty-seven scans showed significant or active disease, and in 11 (23%) of these, the final interpretation significantly differed from the preliminary interpretation.
|
|
For 14 (1.7%) CT scans in which the final interpretation was acute cerebral ischemia, the sensitivity of the initial interpretation was 86%, and the specificity was 99%. Twelve cases of acute cerebral ischemia were correctly identified in the preliminary interpretation (true-positive findings), and two were missed (false-negative findings). There were 700 true-negative interpretations and two false-positive interpretations of acute cerebral ischemia.
For intracranial hemorrhage, the sensitivity of the initial interpretation was 85% (22 true-positive results and four false-negative results), and the specificity was 99.8% (689 true-negative findings and one false-positive finding).
In addition, three pituitary tumors were identified, none of which had been noted in the preliminary interpretation (Fig. 1).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In an address to the 2000 annual meeting of the American College of Radiology, Michael Rapp [1], the president of the American College of Emergency Physicians, explained the perspective of the emergency physician regarding emergency department staffing by radiologists. First, a medical emergency is defined by the patient; regardless of the chronicity of a medical problem, if a patient reports to the emergency department, the complaint should be treated as if it were an emergency. Second, the quality of care that a patient receives should not depend on the time of day that he or she presents. A patient who presents at 2 A.M. should receive the same workup with the same timeliness as the patient who presents at 2 P.M. Third, medical decision-making occurs at the time of the visit, not after the patient has been discharged from the emergency department. Because residents are in training, the position of the American College of Emergency Physicians is that the appropriate care of emergency department patients requires 24-hr coverage by an attending radiologist. Resident coverage alone is inadequate, presumably because of a higher rate of errors in resident interpretation compared with the error rates of board-certified radiologists. Although one may have legitimate disagreements with these assertions, the fact remains that this position is the philosophy of the American College of Emergency Physicians and of many practicing emergency department physicians.
Some medical centers offer full-time staffing of the radiology department with board-certified radiologists, whereas others offer 24-hr CT interpretation either via teleradiology or an in-house radiologist [2]. Many academic centers provide contemporaneous CT interpretation, with residents providing a preliminary report. To our knowledge, error rates have not been established for the interpretation of emergency CT scans of the head by board-certified general radiologists (using the final report of a neuroradiologist as the gold standard). We attempted to determine what the community standard is for an on-call general radiologist interpreting CT scans of the head via teleradiology.
We found significant disagreement between the initial general radiologist's and the final neuroradiologist's interpretation in 16 (2%) of 716 scans. The low rate of disagreement may be due in part to a high rate of negative findings. In this series, 47 CT scans displayed significant or active disease, and in 11 (23%) of these, the final interpretation significantly differed from the preliminary interpretation. However, this rate of detection is comparable to that described for other modalities. In chest radiography, the miss rate in the detection of non–small cell carcinoma of the lung is 19% [3]. Rates of significant error in the interpretation of chest radiographs have been reported to be as high as 26% [4]. These rates may seem high. However, the rate of discordance between initial interpretations by emergency department physicians and final interpretations by radiologists of CT scans of the head is 39% [5].
The rate of disagreement in our study was comparable to many previously described rates of discordance between preliminary interpretations by residents and final interpretations by neuroradiologists. On-call error rates have been well documented and range from 2–8% for evaluations of head CT scans [6, 7, 8, 9]. A prior report described a 14% rate of overall discordance and a 4% rate of significant discordance among general radiologists who interpreted head CT scans via teleradiology [10].
Overall, we found that little risk of adverse outcomes was posed by the preliminary interpretations made by the general radiologists in our study. Of the significant misses, four carried a potential for adverse outcome in the acute setting (a missed subdural hematoma in a child, missed hemorrhagic contusion (Figs. 2A, 2B), possible missed subarachnoid hemorrhage in the setting of trauma, and missed intracranial hemorrhage). All of these misses resulted in a change in management (such as observation and rescanning the patient, rather than discharging the patient from the emergency department). However, subsequent clinical follow-up of these patients revealed no adverse outcome.
|
|
The other significant misses were either false-positive CT findings that may have prompted additional workup but caused the patient no harm or were subacute or chronic conditions in which a delayed interpretation had no effect on acute management (Fig. 3). In the case of the missed middle cerebral artery infarction, the patient was not believed to be a good candidate for thrombolysis because she presented 4 hr after the onset of symptoms (Figs. 4A, 4B). Therefore, even if the CT scan had been correctly interpreted, the patient would not have received thrombolytic therapy.
|
|
|
We found a relatively low rate (7%) of positive findings in this series, which is similar to the 10% rate of positive findings described in a prospective study of patients with minimal head injuries [8].
Regarding specific diagnoses, the sensitivity and specificity rates for the detection of acute stroke and intracranial hemorrhage in this study are remarkably similar to those reported for interpretations by residents. For intracranial hemorrhage, the 85% sensitivity and the 99.8% specificity rates of board-certified radiologists were similar to the figures previously reported for interpretations by residents (93% sensitivity and 99.2% specificity) [11]. For the detection of acute cerebral infarction, the 86% sensitivity and 100% specificity rates were comparable to the rates of sensitivity and specificity achieved by residents—87.5% and 99.8%, respectively [11]. These rates are higher than the 82% mean agreement reported among neuroradiologists in the detection of acute stroke [12]. However, the prevalence of disease in the sample group as well as the duration of stroke symptoms at the time of CT scanning undoubtedly had a substantial impact on the rate of agreement. It is likely that in our population, the patients presented to the emergency department later and therefore had more conspicuous findings than those included in studies for early detection and treatment of stroke.
No pituitary lesions were detected by the general radiologists. The region in which the pituitary glands are located is probably not included in the routine search pattern of the radiologists in the evaluation of emergency CT studies. This area is infrequently involved in trauma or a change in mental status, two of the main indications for emergency cranial CT scanning.
In conclusion, the rate of significant discordance between board-certified general radiologists and neuroradiologists in the interpretation of emergency CT scans of the head is 2%, a rate comparable to previously published reports of performance by residents. For CT scans on which a neuroradiologist identified a positive finding, the rate of discordance was 23%. The pituitary gland may be a blind spot that requires additional attention to be focused in the region of the gland.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  W.M. Strub, J.L. Leach, T. Tomsick, and A. Vagal Overnight Preliminary Head CT Interpretations Provided by Residents: Locations of Misidentified Intracranial Hemorrhage AJNR Am. J. Neuroradiol., October 1, 2007; 28(9): 1679 - 1682. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. A. Campbell, R. P. Berger, L. Ettaro, and M. S. Roberts Cost-effectiveness of Head Computed Tomography in Infants With Possible Inflicted Traumatic Brain Injury Pediatrics, August 1, 2007; 120(2): 295 - 304. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R.D. Zimmerman, J.A. Maldjian, N.C. Brun, B. Horvath, and B.E. Skolnick Radiologic Estimation of Hematoma Volume in Intracerebral Hemorrhage Trial by CT Scan AJNR Am. J. Neuroradiol., March 1, 2006; 27(3): 666 - 670. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Zivian, R. Gershater, W. K. Erly, B. C. Ashdown, and R. W. Lucio General Radiologists' Diagnostic Accuracy: Incomplete Presentation of Data Casts Doubt on Study's Conclusions Am. J. Roentgenol., February 1, 2005; 184(2): 697 - 699. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kalyanpur, V. P. Neklesa, D. T. Pham, H. P. Forman, S. T. Stein, and J. A. Brink Implementation of an International Teleradiology Staffing Model Radiology, August 1, 2004; 232(2): 415 - 419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Emergency Head CT Scans: How Good Are Readings in the Middle of the Night? Journal Watch Emergency Medicine, July 30, 2003; 2003(730): 5 - 5. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
