Comparison of the Canadian CT Head Rule and the New Orleans Criteria in Patients With Minor Head Injury
- Ian G. Stiell, MD, MSc, FRCPC;
- Catherine M. Clement, RN;
- Brian H. Rowe, MD, MSc;
- Michael J. Schull, MD, MSc;
- Robert Brison, MD, MPH;
- Daniel Cass, MD;
- Mary A. Eisenhauer, MD;
- R. Douglas McKnight, MD;
- Glen Bandiera, MD;
- Brian Holroyd, MD;
- Jacques S. Lee, MD;
- Jonathan Dreyer, MD;
- James R. Worthington, MBBS;
- Mark Reardon, MD;
- Gary Greenberg, MD;
- Howard Lesiuk, MD;
- Iain MacPhail, MD, MHSc;
- George A. Wells, PhD
- Author Affiliations: Department of Emergency Medicine (Drs Stiell, Worthington, Greenberg, and Reardon), Department of Epidemiology and Community Medicine (Dr Wells), Division of Neurosurgery (Dr Lesiuk), and Clinical Epidemiology Program (Ms Clement), University of Ottawa, Ottawa, Department of Emergency Medicine, University of Alberta, Edmonton (Drs Rowe and Holroyd), Division of Emergency Medicine, University of Toronto, Toronto (Drs Schull, Cass, Lee, and Bandiera), Department of Emergency Medicine, Queen’s University, Kingston (Dr Brison), Division of Emergency Medicine, University of Western Ontario, London (Drs Eisenhauer and Dreyer), and Division of Emergency Medicine, University of British Columbia, Vancouver (Drs McKnight and MacPhail), Canada.
Abstract
Context Current use of cranial computed tomography (CT) for minor head injury is increasing rapidly, highly variable, and inefficient. The Canadian CT Head Rule (CCHR) and New Orleans Criteria (NOC) are previously developed clinical decision rules to guide CT use for patients with minor head injury and with Glasgow Coma Scale (GCS) scores of 13 to 15 for the CCHR and a score of 15 for the NOC. However, uncertainty about the clinical performance of these rules exists.
Objective To compare the clinical performance of these 2 decision rules for detecting the need for neurosurgical intervention and clinically important brain injury.
Design, Setting, and Patients In a prospective cohort study (June 2000-December 2002) that included 9 emergency departments in large Canadian community and university hospitals, the CCHR was evaluated in a convenience sample of 2707 adults who presented to the emergency department with blunt head trauma resulting in witnessed loss of consciousness, disorientation, or definite amnesia and a GCS score of 13 to 15. The CCHR and NOC were compared in a subgroup of 1822 adults with minor head injury and GCS score of 15.
Main Outcome Measures Neurosurgical intervention and clinically important brain injury evaluated by CT and a structured follow-up telephone interview.
Results Among 1822 patients with GCS score of 15, 8 (0.4%) required neurosurgical intervention and 97 (5.3%) had clinically important brain injury. The NOC and the CCHR both had 100% sensitivity but the CCHR was more specific (76.3% vs 12.1%, P<.001) for predicting need for neurosurgical intervention. For clinically important brain injury, the CCHR and the NOC had similar sensitivity (100% vs 100%; 95% confidence interval [CI], 96%-100%) but the CCHR was more specific (50.6% vs 12.7%, P<.001), and would result in lower CT rates (52.1% vs 88.0%, P<.001). The κ values for physician interpretation of the rules, CCHR vs NOC, were 0.85 vs 0.47. Physicians misinterpreted the rules as not requiring imaging for 4.0% of patients according to CCHR and 5.5% according to NOC (P = .04). Among all 2707 patients with a GCS score of 13 to 15, the CCHR had sensitivities of 100% (95% CI, 91%-100%) for 41 patients requiring neurosurgical intervention and 100% (95% CI, 98%-100%) for 231 patients with clinically important brain injury.
Conclusion For patients with minor head injury and GCS score of 15, the CCHR and the NOC have equivalent high sensitivities for need for neurosurgical intervention and clinically important brain injury, but the CCHR has higher specificity for important clinical outcomes than does the NOC, and its use may result in reduced imaging rates.
