The Canadian C-Spine Rule
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Stiell et al. (Dec. 25 issue)1 report the sensitivity of the National Emergency X-Radiography Utilization Study (NEXUS) low-risk criteria (NLC) for important cervical-spine injury to be 90.7 percent. This figure dramatically contrasts with the 99.6 percent sensitivity we observed among more than 34,000 patients examined by hundreds of clinicians in multiple emergency departments, representing real-world practice.2
There are several reasons to believe that the Canadian researchers underestimated the sensitivity of the NLC. As noted by Yealy and Auble3 in their editorial accompanying the report, the use of surrogate criteria by Stiell et al. introduces the potential for misclassification errors. Misclassification can also occur when clinicians fail to assess or record criteria accurately; the potential for such failure increases with the number and complexity of the required assessments. Whereas we restricted our data form to the five NEXUS criteria, the clinicians in the Canadian study had to complete a long, complex form that required the assessment of a great number of items. The five NEXUS criteria were placed at the end of the form, making misclassification of these items even more likely.
Before the NEXUS study was undertaken, a literature review revealed no credible reports of patients with clinically significant cervical spine injury who would have been mislabeled by the NLC as being at "low risk." Even the two false negative cases we reported in our study proved not to be true failures of the NLC; one involved an old injury, and the other was almost certainly misclassified.2 If the Canadians' estimation of sensitivity were correct, the NLC would be missing more than 1000 of the approximately 11,000 cases of devastating cervical-spine injuries that occur annually in the United States.4 Yet in the three years since our study was published, and the NLC widely adopted, there continue to be no reports of cases of important cervical spine injury with a false negative classification by the NLC.
Ultimately, a decision instrument is of little value if it is so complex that clinicians choose not to use it. NEXUS provides a simple, straightforward instrument that safely allows physicians to select candidates for cervical-spine radiography with a sensitivity that is very close to 100 percent.
William R. Mower, M.D., Ph.D.
UCLA School of Medicine
Los Angeles, CA 90024
Allan B. Wolfson, M.D.
University of Pittsburgh School of Medicine
Pittsburgh, PA 15208
Jerome R. Hoffman, M.D.
UCLA School of Medicine
Los Angeles, CA 90024
Knox H. Todd, M.D., M.P.H.
Emory University School of Medicine
Atlanta, GA 30309
References
Stiell IG, Clement CM, McKnight RD, et al. The Canadian C-Spine Rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003;349:2510-2518.
Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma. N Engl J Med 2000;343:94-99.
Yealy DM, Auble TE. Choosing between clinical prediction rules. N Engl J Med 2003;349:2553-2555.
Spinal cord injury: facts and figures at a glance. Birmingham, Ala.: National Spinal Cord Injury Statistical Center, December 2003.
To the Editor: Stiell et al. note that inappropriate ordering of cervical-spine examination "adds substantially to health care costs because of the high volume of its use. Furthermore, patients are often immobilized on a backboard for many hours while awaiting radiography, leading to considerable discomfort and unnecessary use of space in crowded emergency departments." Even more important, in my experience, are the unanticipated downstream costs of ordering posttraumatic cervical-spine examinations in low-risk patients. The examination is frequently suboptimal, often allowing only incomplete visualization of the seventh cervical vertebra, and in today's medicolegal climate these patients are usually referred for additional computed tomographic scanning.
Ferris M. Hall, M.D.
Beth Israel Deaconess Medical Center
Boston, MA 02215
fhall@bidmc.harvard.edu
The authors reply: We disagree with the comments of Dr. Mower and colleagues from the NEXUS group, as well as with the similar comments of Drs. Yealy and Auble in their editorial. First, we explicitly and accurately validated the NEXUS criteria in accordance with good decision-rule methodology.1,2 We were very aware of the importance of faithfully reproducing the NEXUS criteria and enlisted the help of the NEXUS group for this component of our data form. Both the five basic NEXUS criteria, as well as the lengthy explanatory descriptors attached, were used exactly as they had been prepared for us by a NEXUS investigator. The two rules were placed side by side in large, separate boxes in the middle of our data form.3
Second, we can comment on why others have not found that the sensitivity of the NEXUS criteria is low. In the NEXUS validation study, the specificity was only 12.9 percent, most patients were positive for the criteria, and all underwent radiography.4 In our study, physicians found that the NEXUS criteria were negative for 33.4 percent of patients; among these patients, 16 important injuries would have been missed. There are no data on how widely or how accurately the NEXUS criteria are being used in practice and whether they have affected the use of radiography. If the NEXUS criteria are being applied so liberally that most patients require radiography, then of course no injuries will be missed. Almost all of the 16 patients with a clinically important injury that would have been missed by the NEXUS criteria in our study had a dangerous injury, were over the age of 65 years, or were unable to rotate their neck. We do not agree that these missed cases reflect the design of our data form or the doctors' behavior. Instead, we believe the problem is that the NEXUS criteria do not include these important correlates of injury. We encourage others to conduct prospective evaluations of the accuracy and potential effect of the NEXUS criteria.
Ian G. Stiell, M.D.
University of Ottawa
Ottawa, ON KY1 4E9, Canada
Brian H. Rowe, M.D.
University of Alberta
Edmonton, AB T6G 2B7, Canada
Jacques Lee, M.D.
University of Toronto
Toronto, ON M4N 3M5, Canada
References
Laupacis A, Sekar N, Stiell IG. Clinical prediction rules: a review and suggested modifications of methodological standards. JAMA 1997;277:488-494.
Stiell IG, Wells GA. Methodologic standards for the development of clinical decision rules in emergency medicine. Ann Emerg Med 1999;33:437-447.
The CCC Study Group. Canadian C-Spine Rule study physician data collection form. (Accessed March 11, 2004, at http://www.ohri.ca/programs/clinical_epidemiology/ohdec/clinical.asp#five.)
Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma. N Engl J Med 2000;343:94-99.
There are several reasons to believe that the Canadian researchers underestimated the sensitivity of the NLC. As noted by Yealy and Auble3 in their editorial accompanying the report, the use of surrogate criteria by Stiell et al. introduces the potential for misclassification errors. Misclassification can also occur when clinicians fail to assess or record criteria accurately; the potential for such failure increases with the number and complexity of the required assessments. Whereas we restricted our data form to the five NEXUS criteria, the clinicians in the Canadian study had to complete a long, complex form that required the assessment of a great number of items. The five NEXUS criteria were placed at the end of the form, making misclassification of these items even more likely.
Before the NEXUS study was undertaken, a literature review revealed no credible reports of patients with clinically significant cervical spine injury who would have been mislabeled by the NLC as being at "low risk." Even the two false negative cases we reported in our study proved not to be true failures of the NLC; one involved an old injury, and the other was almost certainly misclassified.2 If the Canadians' estimation of sensitivity were correct, the NLC would be missing more than 1000 of the approximately 11,000 cases of devastating cervical-spine injuries that occur annually in the United States.4 Yet in the three years since our study was published, and the NLC widely adopted, there continue to be no reports of cases of important cervical spine injury with a false negative classification by the NLC.
Ultimately, a decision instrument is of little value if it is so complex that clinicians choose not to use it. NEXUS provides a simple, straightforward instrument that safely allows physicians to select candidates for cervical-spine radiography with a sensitivity that is very close to 100 percent.
William R. Mower, M.D., Ph.D.
UCLA School of Medicine
Los Angeles, CA 90024
Allan B. Wolfson, M.D.
University of Pittsburgh School of Medicine
Pittsburgh, PA 15208
Jerome R. Hoffman, M.D.
UCLA School of Medicine
Los Angeles, CA 90024
Knox H. Todd, M.D., M.P.H.
Emory University School of Medicine
Atlanta, GA 30309
References
Stiell IG, Clement CM, McKnight RD, et al. The Canadian C-Spine Rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003;349:2510-2518.
Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma. N Engl J Med 2000;343:94-99.
Yealy DM, Auble TE. Choosing between clinical prediction rules. N Engl J Med 2003;349:2553-2555.
Spinal cord injury: facts and figures at a glance. Birmingham, Ala.: National Spinal Cord Injury Statistical Center, December 2003.
To the Editor: Stiell et al. note that inappropriate ordering of cervical-spine examination "adds substantially to health care costs because of the high volume of its use. Furthermore, patients are often immobilized on a backboard for many hours while awaiting radiography, leading to considerable discomfort and unnecessary use of space in crowded emergency departments." Even more important, in my experience, are the unanticipated downstream costs of ordering posttraumatic cervical-spine examinations in low-risk patients. The examination is frequently suboptimal, often allowing only incomplete visualization of the seventh cervical vertebra, and in today's medicolegal climate these patients are usually referred for additional computed tomographic scanning.
Ferris M. Hall, M.D.
Beth Israel Deaconess Medical Center
Boston, MA 02215
fhall@bidmc.harvard.edu
The authors reply: We disagree with the comments of Dr. Mower and colleagues from the NEXUS group, as well as with the similar comments of Drs. Yealy and Auble in their editorial. First, we explicitly and accurately validated the NEXUS criteria in accordance with good decision-rule methodology.1,2 We were very aware of the importance of faithfully reproducing the NEXUS criteria and enlisted the help of the NEXUS group for this component of our data form. Both the five basic NEXUS criteria, as well as the lengthy explanatory descriptors attached, were used exactly as they had been prepared for us by a NEXUS investigator. The two rules were placed side by side in large, separate boxes in the middle of our data form.3
Second, we can comment on why others have not found that the sensitivity of the NEXUS criteria is low. In the NEXUS validation study, the specificity was only 12.9 percent, most patients were positive for the criteria, and all underwent radiography.4 In our study, physicians found that the NEXUS criteria were negative for 33.4 percent of patients; among these patients, 16 important injuries would have been missed. There are no data on how widely or how accurately the NEXUS criteria are being used in practice and whether they have affected the use of radiography. If the NEXUS criteria are being applied so liberally that most patients require radiography, then of course no injuries will be missed. Almost all of the 16 patients with a clinically important injury that would have been missed by the NEXUS criteria in our study had a dangerous injury, were over the age of 65 years, or were unable to rotate their neck. We do not agree that these missed cases reflect the design of our data form or the doctors' behavior. Instead, we believe the problem is that the NEXUS criteria do not include these important correlates of injury. We encourage others to conduct prospective evaluations of the accuracy and potential effect of the NEXUS criteria.
Ian G. Stiell, M.D.
University of Ottawa
Ottawa, ON KY1 4E9, Canada
Brian H. Rowe, M.D.
University of Alberta
Edmonton, AB T6G 2B7, Canada
Jacques Lee, M.D.
University of Toronto
Toronto, ON M4N 3M5, Canada
References
Laupacis A, Sekar N, Stiell IG. Clinical prediction rules: a review and suggested modifications of methodological standards. JAMA 1997;277:488-494.
Stiell IG, Wells GA. Methodologic standards for the development of clinical decision rules in emergency medicine. Ann Emerg Med 1999;33:437-447.
The CCC Study Group. Canadian C-Spine Rule study physician data collection form. (Accessed March 11, 2004, at http://www.ohri.ca/programs/clinical_epidemiology/ohdec/clinical.asp#five.)
Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma. N Engl J Med 2000;343:94-99.