CAEP/ACMU 2001 Scientific Abstracts - Oral Presentations: 15-29

CAEP Abstracts

CJEM 2001;3(2):121-134

| 001-014 | 015-029 | 030-044 | 045-059 | 060-074 | 075-093 |

015 How important is mechanism of injury in predicting the risk of cervical spine injury?
Stiell IG, Clement C, De Maio VJ, Wells GA, Battram E, Morrison L, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: To assist emergency care personnel by measuring the cervical-spine (C-spine) injury risk associated with specific blunt trauma injury mechanisms. The importance of injury mechanisms is a controversial issue in the trauma literature. METHODS: This prospective cohort study was conducted in 10 Canadian EDs and involved adult trauma patients at risk for neck injury, with stable vital signs and a Glasgow Coma Scale (GCS) score of 15. Physicians completed a 20-item data form for all patients who then underwent radiography to determine the outcome, important C-spine injury. Study nurses reviewed ambulance reports, ED records, and in-hospital records to classify each case according to 21 injury mechanisms and 8 dangerous motor vehicle collision (MVC) factors. Variables correlated with the outcome on univariate analysis were then assessed by forward stepwise logistic regression analysis, using SAS. RESULTS: Among 8,924 patients enrolled over 30 months, 151 (1.7%) had clinically important injury, including fracture (143), dislocation (23), and ligamentous instability (9). The most common mechanisms of injury were: MVC 67.0%, fall <3 feet 8.3%, fall 3-10 feet 4.4%, sports 2.9%, assault fist 2.6%. After adjustment for demographic and clinical characteristics, analysis found the following mechanisms to be independently associated with increased risk of C-spine injury (odds ratios with 95% confidence intervals [CIs]): axial load (e.g., diving) 14.2 (8.0-25.2), bicycle collision 9.6 (2.2-35.5), motorized recreational vehicle 8.9 (2.4-32.7), fall 3-10 feet 3.6 (2.0-6.4), fall >10 feet 3.1 (1.2-8.1), MVC >100 km/h 7.6 (3.4-17.0), MVC 60-100 km/h 7.5 (4.6-12.3), MVC rollover 3.8 (2.2-6.6), rear-end MVC 0.06 (0.01-0.5). For this model, Hosmer-Lemeshow goodness-of-fit statistic 0.63 for 7 degree of freedom (DF); area under the receiver operating characteristic (ROC) curve 0.93. CONCLUSIONS: Specific injury mechanisms put patients at much higher risk for C-spine injury and emergency care personnel should carefully ascertain details of the injury situation. Prospective studies will evaluate the accuracy and reliability of these mechanisms.
Key words: cervical spine, injury, diagnosis

016 Pediatric injuries in organized hockey: Does checking make a difference?
Parker MJ, Salter KL, Lipskie TL, Joubert GI. Department of Pediatrics, University of Western Ontario, London, Ontario.

OBJECTIVES: Ice hockey is a sport involving skills such as body checking. Recently, the minimum allowable checking age has been decreased in certain jurisdictions. The impact of checking on injury rates has been hotly debated. This study sought to determine injury rates in organized ice hockey, as reported in the Canadian Hospital Reporting and Prevention Program (CHIRPP) database, and what impact checking has on these rates. METHODS: A retrospective study was conducted using the CHIRPP database, a Canadian injury surveillance tool incorporating data from all 10 pediatric hospitals and 4 general hospitals since 1990. Hockey injury reports between January 1, 1994, to August 31, 1999, were analyzed. Age ranged from <1 to >19 years of age. The database was culled to include only organized ice hockey injury records. Statistical analysis was performed on these records comparing injury trends and relationship to checking by age and year of injury. RESULTS: Data from 12,879 records was analyzed with 4,118 involving checking. The 10-14 age group had the most injuries, with 62% checked and 58% non-checked. Over time, the number of injuries reported per year in the database has been consistent. Through the time period the number of non-checked injuries remained unchanged, while there was a trend toward fewer checking-related injuries (4.2% decrease, p > 0.05). Using hospital admissions as an indicator of injury severity, significantly more checking-related injuries were hospitalized (p < 0.01). CONCLUSIONS: Using the CHIRPP surveillance tool, organized ice hockey injuries have remained stable. There is a trend to decreased checking-related injuries over time. Despite this, injury severity as measured by hospital admission is significantly greater in the checked group. We conclude that checking in organized hockey does not increase injuries. However, increased injury severity in the checked group supports using caution in decreasing the minimum allowable checking age in organized ice hockey.
Key words: injury prevention, hockey, pediatric

017 The role of clinical assessment in the evaluation of patients with potential cervical-spine injury.
Stiell IG, Wells GA, Clement C, De Maio VJ, MacPhail I, Rowe
BH, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: Emergency department (ED) physicians frequently face the decision of ordering cervical-spine (C-spine) radiography for blunt trauma patients. This study measured the accuracy of clinical examination in identifying the likelihood of C-spine injury. METHODS: This prospective cohort study was conducted in 10 tertiary care EDs and involved alert (Glasgow Coma Scale [GCS] score of 15) and stable adult trauma patients at risk for neck injury. Physicians completed a 20-item standardized clinical data form for all patients who then underwent radiography to determine the outcome, important C-spine injury. Variables having univariate correlation with the outcome were assessed by forward stepwise logistic regression (LR) analysis (SAS). RESULTS: The 8,924 patients had these characteristics: important C-spine injury 1.7%; mean age 36.8 years; history: ambulatory 67.5%, delayed onset neck pain 42.1%, paresthesias 9.7%, extremity weakness 3.1%; physical: able to rotate neck 55.3%, absence of neck tenderness 42.2%, upright position 37.3%, visible head injury 20.8%, facial injury 19.2%, distracting painful injuries 7.8%, intoxicated 4.2%, sensory deficit 2.0%, motor deficit 1.2%. After adjusting for other factors, LR analysis found the following clinical findings to be associated with risk of C-spine injury (odds ratios with 95% confidence intervals [CIs]): age >65 3.3 (2.1-5.1), paresthesias 2.5 (1.7-3.9), facial injury 1.8 (1.1-2.7), visible head injury 1.6 (1.0-2.4), delayed neck pain 0.5 (0.3-0.7), absence of neck tenderness 0.4 (0.2-0.6), able to rotate neck 0.07 (0.03-0.16). The following were not independently associated with C-spine injury: ambulatory at any time, extremity weakness, upright position, intoxication, distracting painful injuries, motor deficit, sensory deficit. Hosmer-Lemeshow goodness-of-fit statistic 0.78 for 8 DF; area under receiver operating characteristic (ROC) curve 0.91. CONCLUSIONS: Many clinical findings are significantly associated with either increased or decreased risk of C-spine injury and ED physicians should focus on these findings during assessment of alert trauma patients.
Key words: cervical spine, injury, diagnosis

018 The contribution of blood cultures to the clinical management of adult patients with community-acquired pneumonia: a prospective observational study.
Campbell SG, Anstee R, Ackroyd S, Dickenson G. Dalhousie University, Halifax, Nova Scotia.

OBJECTIVES: To assess the clinical utility of blood cultures (BC) in the management of community-acquired pneumonia (CAP). METHODS: A prospective observational study using the capital study population to investigate how the performance of blood cultures would affect the management and outcomes of adult patients presenting with CAP. The course of therapy of each patients with a positive BC result was examined to assess the influence of the results on clinical management. RESULTS: BC were drawn in 1,049 patients (60.2%), 289 (28%) outpatients and 760 (72%) inpatients. 50 "significant" organisms (from 49 [4.67%] patients) were identified. The yield of BC was 5.7% in admitted patients, and 2.1% in outpatients. Patients with CAP who had blood cultures performed had a 1.5% (16/1049) chance of having a change of therapy directed by the results of the culture. Patients in whom blood cultures were positive had 32.7% (16/49) chance of having a change in therapy determined by the result, and 59.2% (29/49) chance of having a course of therapy in contradiction with BC findings. Severity of illness, as measured by the Pneumonia severity score, compared poorly with the yield of BCs. In admitted patients, BCs were positive in 7.2% of patients in groups I and II, 5.0% in group III, 4.9% in group IV, and 6.9% in group V. CONCLUSIONS: Blood cultures have limited utility in the routine management of uncomplicated community-acquired pneumonia and should not form part of the routine management of CAP.
Key words: blood culture, pneumonia, cost-effectiveness

019 Influenza virus as a determinant of emergency department overcrowding.
Schull MJ, Mamdani M. Institute of Clinical Evaluative Sciences, Toronto, Ontario.

OBJECTIVES: The impact of influenza virus on emergency department (ED) utilization is controversial, and has not previously been studied. We sought to determine the relationship between circulating influenza virus and ED overcrowding. METHODS: Weekly totals of laboratory-confirmed cases of influenza (types A and B) and other respiratory viruses in Toronto, Ontario, from 1996 to 1999 (weeks = 208) were obtained, along with data on weekly ED overcrowding, total ED visits, average patient age and sex distribution. Overcrowding was defined as episodes when EDs diverted all ambulances (Critical Care Bypass), and the average proportion of overcrowding time each week was calculated. Time series analysis using a multivariate autoregressive integrated moving average (ARIMA) model was conducted to longitudinally assess relationships between ED overcrowding and influenza cases, other respiratory virus cases, age, gender, and total number of ED visits. RESULTS: A mean of 10,957 ED visits occurred weekly. Patient age and percent female averaged 39.5 years and 51%, respectively. EDs were overcrowded an average of 0.2% to 23% of the time each week. Weekly influenza cases ranged from 0 to 236 (mean = 19), while other viruses ranged from 0 to 91 (mean = 22) per week. Surprisingly, in the multivariate ARIMA model, circulating influenza virus was a highly significant predictor of increased overcrowding (p < 0.0001). However, increases in other respiratory viruses was not a significant determinant (p = 0.49). As expected, total ED visits was a significant predictor of overcrowding (p = 0.047), while average patient age (p = 0.93) and percentage of female patients (p = 0.28) were not. ARIMA diagnostics showed good fit, and the model R2 was 0.64. CONCLUSIONS: Increases in circulating influenza virus are a highly significant predictor of worsened ED overcrowding, even after controlling for ED utilization. Other respiratory viruses showed no association, despite being more prevalent in the community.
Key words: influenza, emergency department overcrowding

020 Predicting complications from renal colic after emergency department evaluation.
Papa L, Stiell IG, Wells GA, Battram E, Mahoney J. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: To prospectively assess clinical predictors of complications from urinary calculi after discharge from the emergency department (ED). METHODS: A cohort of patients with suspected renal colic from 2 tertiary care EDs were assessed prospectively. Emergency physicians completed a 20-question data form with the clinical findings of each patient. In addition, physician's estimated each patient's pre-test probability of developing complications. Once urinary calculi were confirmed with imaging, patients were discharged and followed via telephone at 1 and 4 weeks to assess development of complications. Complications were defined by: i) persistent pain by 4 weeks, or ii) development of fever (>38°C) or iii) elevation of creatinine >150 mmol/L by 4 weeks. Univariate analysis tested which clinical variables correlated with complications. Fisher's exact test and independent sample t-test were conducted for nominal and continuous data respectively. RESULTS: Over a 6-month period, 150 patients with confirmed urolithiasis were followed, 72 (48%) (95% CI, 40%-56%) patients developed complications; 17 (24%) developed fever or renal failure; and 55 (76%) had persistent pain at 4 weeks. Clinical variables correlated with complications included pain at discharge >= 2 cm on a Visual Analog Scale (VAS) (p = 0.022), length of stay in ED >= 6.5 h (p = 0.018) and temperature >= 37.0°C (p = 0.026) in the ED. Emergency physician judgement had a sensitivity of 31% and a specificity of 66%. They would have misclassified 50% of patients who developed complications. CONCLUSIONS: Physicians have difficulty predicting complications from calculi on initial ED presentation. Patients with elevations of temperature, pain at discharge or have a prolonged stay in the ED, should be considered for early urology referral. Future studies will validate the validity of these predictive criteria.
Key words: renal colic, diagnosis

021 The utility of blood cultures in the management of upper
urinary tract infection.
McInnes JG, Murray HE, Worrall J. Queen's University, Kingston, Ontario.

OBJECTIVES: Studies challenge the use of routine blood cultures in acute pyelonephritis; however, guidelines do not exist for their appropriate use. An examination of the value of blood cultures in the routine workup of acute pyelonephritis and urosepsis was conducted in order to identify clinical and laboratory predictors of bacteremia in these patients. METHODS: All charts for patients seen at a tertiary care referral centre emergency department between October 1996 and March 1998 with a primary diagnosis of acute pyelonephritis or urosepsis were reviewed. Patients under 16 years of age, or with duplicate visits were excluded. Historical features, physical and laboratory findings were recorded. Analyses included descriptive statistics, bivariate and multivariate associations. RESULTS: Of the 158 patients who met inclusion criteria, 110 (69.6%) had blood cultures performed. Growth was observed in 28 (25.4%) blood cultures. Antibiotics were not changed due to culture results. Elevated urea (>7.5 mmol/L), temperature >38.4°C and an abnormal band count (>0.7 * 10⁹/L) were all statistically significant predictors of bacteremia on bivariate analysis (p < 0.05). Multivariate analysis identified only temperature >38.4°C (odds ratio [OR] 2.65, 95% CI 1.03-6.83) and abnormal band count (OR 2.00, 95% CI 1.12-3.55) as predictors of bacteremia. CONCLUSIONS: No change in antibiotics was observed based on blood culture results. Fever and an abnormal band count appear to predict bacteremia. Prospective validation of clinical and lab predictors of bacteremia is required.
Key words: blood cultures, pyelonephritis

022 The epidemiology of acute allergy presentations in Alberta.
Rowe BH, Johnson D, Yiannakoulias N, Spooner CH, Holroyd B, Bullard M, et al. University of Alberta, Edmonton, Alberta.

OBJECTIVES: Many patients with allergic reactions and anaphylaxis present to the emergency department (ED) and despite potentially serious consequences limited information exists about this group. This study examines the epidemiology of acute allergy presentations to the ED using a unique data set. METHODS: All cases of acute allergy and anaphylaxis from all ages presenting to Alberta EDs were eligible for inclusion. Data were derived from a cohort of patients treated at Alberta EDs in 17 health regions over 1 year (fiscal 98/99). Data were extracted from the Ambulatory Care Classification System (ACCS) database, which contains computerized abstracts coded similarly across regions. Allergy and anaphylaxis was coded using ICD-9 coding by medical record nosologists in each hospital and represented the primary physician discharge diagnosis. Descriptive statistics and crude presentation rates are presented. RESULTS: Acute allergic reactions accounted for 8,733 (0.6%) of the 1.494 million annual visits to EDs in Alberta. Males (4,469; 51%) and females were similarly represented; patients <19 years old were involved in 2,856 (33%) of all cases. Most presentations (6,462 [74%]) occurred in the summer months (June-Sept.). Multiple visits were uncommon. The provincial ED presentation rate was 3/1000 persons (range 1-8/1000). The lowest ED presentation rates occurred in the 2 largest urban areas (population >500,000). Most (8,567; 98%) patients were discharged from the ED, with no variation among age groups. Overall, 161 (2%) were admitted; no deaths in the ED setting were noted. CONCLUSIONS: Allergy is a common ED problem with very low acuity. Given the marked variations in presentation rates, it is important to further examine these findings and determine the etiology and treatment of allergy and anaphylaxis in the ED.
Key words: allergy, anaphylaxis

023 Accuracy and reliability of the Object Recall Test for patients with minor head injury.
Stiell IG, McKnight RD, Lesiuk H, Brison RJ, Clement C, Wells GA, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: Patients with minor head injury may harbour serious and sometimes life-threatening brain injuries. This study evaluated the accuracy and reliability of the Object Recall Test for the ED assessment of these patients. METHODS: This prospective cohort study enrolled adults with loss of consciousness, amnesia, or confusion and a Glasgow Coma Scale (GCS) score of 13-15. Physicians in 10 tertiary care EDs assessed each patient for the Object Recall Test prior to computed tomography (CT) scan and, where feasible, 2nd physicians performed interobserver assessments. The 2-minute Object Recall Test consisted of showing patients 3 objects (pen, watch, nose) and asking them to recall the objects after an interval of 2 minutes, and was scored from 0 to 3. The outcome standards were "need for neurological intervention," "clinically important brain injury," and "any brain injury on CT." Analyses included kappa coefficient, chi-square, receiver operating characteristic (ROC) curve areas, sensitivity, and specificity. RESULTS: Among 2,672 patients enrolled over 36 months, the mean age was 38.4 (range 16-97), 68.8% were male, 32 (1.2%) underwent neurological intervention, 206 (7.7%) had important brain injury, and 283 (10.6%) had some acute brain injury on CT. The weighted kappa for interobserver agreement in 86 patients was 0.64 (95% CI 0.50-0.78). Measures of accuracy (95% CIs) for the 3 outcomes were:

CONCLUSIONS: Despite good reliability and strong association with clinical outcomes, the Object Recall Test exhibits only fair discrimination and sensitivity. This test is not accurate enough to be used alone but may prove useful as one component of an algorithm for assessing minor head injury patients.
Key words: brain injury, clinical prediction rule, diagnosis

024 The relationship between out-of-hospital cardiac arrest survival and community bystander cardiopulmonary resuscitation rates.
De Maio VJ, Stiell IG, Wells GA, Martin MT, Spaite DW, Nichol G, et al, for the OPALS Study Group. University of Calgary, Calgary, Alberta.

OBJECTIVES: While recent data have re-emphasized the role of bystander cardiopulmonary resuscitation (CPR) in the chain of survival for out-of-hospital cardiac arrest, its importance has not been clearly quantified. The objective of this study was to measure survival as a function of community bystander CPR rates. METHODS: This prospective cohort study included all adult, cardiac etiology, out-of-hospital cardiac arrests from Phases I and II of the Ontario Prehospital Advanced Life Support (OPALS) Study. Patients in the 20 study communities received a BLS-D (Basic Life Support) level of care by emergency medical services (EMS), but no Advanced Life Support (ALS). Logistic regression (LR) identified adjusted CPR rates using covariates found to be independently associated with bystander CPR. Survival was then modeled using the adjusted CPR rates. The logistic equation was used to estimate community survival for incremental bystander CPR rates. RESULTS: From 1991-97, there were 343 (3.7%) survivors among 9,218 cases treated. The overall bystander CPR rate was 15.2%. LR analysis found the following factors associated with bystander CPR (odds ratios with 95% confidence intervals [CIs]): male sex 1.21 (1.05-1.39), age/10yrs 0.83 (0.80-0.87), non-winter season 1.20 (1.04-1.38), witnessed arrest 2.39 (2.10-2.73), ventricular fibrillation/ventricular tachycardia (VF/VT) rhythm 2.13 (1.88-2.43). Hosmer-Lemeshow goodness-of-fit statistic 0.86 for 8 DF; area under receiver operating characteristic (ROC) curve 0.70. The survival function, using the adjusted CPR rate and the defibrillation response interval, indicates the odds of survival with increasing CPR rate were 1.71 (1.61-1.81) per 5% increase. The survival function predicts, for successive 5% increments: i) survival rates, and ii) additional lives saved per year in the OPALS Study communities: 25% (6.3%; 41 lives), 30% (10.4%; 107 lives), 35% (16.7%; 208 lives), 40% (25.7%; 352 lives). CONCLUSIONS: Improved community bystander CPR rates are associated with dramatically increased out-of-hospital arrest survival in a predictable fashion. EMS and public health directors should focus significant efforts towards improving their community bystander CPR rate.
Key words: cardiac arrest, resuscitation, emergency medical services

025 How accurate is clinical examination in the evaluation of patients with minor head injury?
Stiell IG, De Maio VJ, Clement C, Wells GA, Dreyer JA, McKnight RD, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: Emergency department (ED) physicians frequently evaluate and make management decisions for minor head injury patients. This study assessed the accuracy of common clinical symptoms and signs for predicting brain injury in these patients. METHODS: This prospective cohort study was conducted in 10 Canadian EDs and involved adults with loss of consciousness, amnesia, or confusion and a Glasgow Coma Scale (GCS) score of 13-15. Physicians were trained to complete a 22-item standardized clinical assessment form prior to computed tomography (CT) scan. Variables correlated with the outcome criterion, clinically important brain injury, on univariate analysis were then assessed by forward stepwise logistic regression analysis (SAS). RESULTS: The 3,121 patients enrolled over 36 months had these characteristics: important brain injury 8.1%, required neurological intervention 1.4%, mean age 38.7, amnesia 87.9%, GCS score 15 79.8%, witnessed loss of consciousness 46.5%, intoxicated 12.3%, chronic alcohol abuse 10.9%, repeated vomiting 9.6%, basal skull fracture signs 6.7%, drop in GCS score 2.9%, pupils not equal and reactive 1.2%, lateralizing motor weakness 1.1%. After adjustment for demographic and mechanism factors, LR analysis found the following clinical findings to be associated with risk of brain injury (odds ratios with 95% confidence intervals [CIs]): GCS <15 by 2 hours 7.3 (5.1-10.3), signs of basal skull fracture 5.0 (3.3-7.6), drop in GCS score 4.5 (2.6-7.9), age >65 4.2 (2.9-6.1), repeated vomiting 3.9 (2.6-5.8), possible open skull fracture 3.2 (1.8-5.7), amnesia before injury >30 minutes 1.5 (1.1-2.2). The following findings were not independently associated with brain injury: chronic alcohol abuse, loss of consciousness >5 min, motor weakness, unequal pupils, object recall test, intoxication. The area under the receiver operating characteristic (ROC) curve for this model was 0.895. CONCLUSIONS: In their assessment of minor head injury patients, ED physicians should emphasize the many clinical findings which have been shown to be strongly associated with the likelihood of brain injury.
Key words: brain injury, diagnosis

026 Potential impact of public access defibrillation based upon cardiac arrest locations.
De Maio VJ, Stiell IG, Wells GA, Martin MT, Doherty J, Spaite DW, et al, for the OPALS Study Group. University of Calgary, Calgary, Alberta.

OBJECTIVES: Community public access defibrillation (PAD) programs are becoming commonplace despite the fact that the results of the PAD Trial are not yet known. This study evaluated the potential impact of PAD programs based upon an analysis of out-of-hospital cardiac arrest locations. METHODS: As part of the Ontario Prehospital Advanced Life Support (OPALS) Study, this prospective cohort study included all adult out-of-hospital cardiac arrests in 3 medium-sized cities. Emergency medical services (EMS) response included firefighter defibrillation and Advanced Life Support (ALS) paramedics. Based upon review of ambulance reports and dispatch records, cases were classified into one of 23 pick-up locations and then grouped into 5 categories: small residential, large residential, streets/outdoors, small public buildings, and large public buildings. Data analyses included descriptive statistics and sensitivity analyses to estimate the impact of doubling and tripling survival rates on additional lives saved per year. RESULTS: From 1995-2000, there were 1,373 consecutive cardiac arrests and characteristics included: mean age 68.6, male gender 66.5%, bystander witnessed 47.8%, bystander cardiopulmonary resuscitation (CPR) 16.1%, fire first 74.6%, rhythm ventricular fibrillation/ventricular tachycardia (VF/VT) 33.6%, median defibrillation response interval 5.1 minutes, survival to hospital discharge 3.7%. Cardiac arrest locations were: small residential 56.1%, large residential 27.8%, streets/outdoors 7.9%, small public buildings 3.4%, large public buildings 4.8%. The number of additional lives that would be saved per year in each location if the overall survival rates doubled and tripled, respectively were: small residential 4, 8; large residential 2, 3; streets/outdoors 2, 4; small public buildings 1, 2; large public buildings 1, 3. CONCLUSIONS: Less than 5% of cardiac arrests occur in large public buildings and the potential for lives saved with PAD appears to be modest. EMS and public health directors should recognize the limited potential of PAD and not overlook other methods for improving cardiac arrest survival.
Key words: cardiac arrest, resuscitation, public access defibrillation

027 How well does the Canadian CT Head Rule perform in minor head injury patients with a Glasgow Coma Scale score of 15?
Stiell IG, Wells GA, Clement C, MacPhail I, Rowe BH, Holroyd B, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: The traditional neurosurgical definition of "minor head injury" includes patients with scores of 13-15 on the Glasgow Coma Scale (GCS), and the Canadian CT Head Rule was derived accordingly. Because many ED physicians believe that only GCS scores of 15 truly represent "minor," this study evaluated the accuracy of the Canadian CT Head Rule when restricted to patients with GCS 15. METHODS: This secondary data analysis was based on a prospective cohort study conducted in 10 Canadian EDs and involved adults with loss of consciousness, amnesia, or confusion and a GCS score of 13-15. Physicians completed a 22-item data form for all patients who then underwent CT scan to determine the outcome measure, important brain injury. For the current study, chi-square recursive partitioning analyses (KnowledgeSEEKER) were performed on the sub-set of patients presenting with a GCS score 15. Statistical measures included sensitivity, specificity, and computed tomography (CT) rate. RESULTS: The Canadian C-Spine/CT Head (CCC) Study dataset contained 2,489 minor head injury cases with GCS 15 and these characteristics: male 66.6%, arrived by ambulance 68.4%, fall 30.9%, motor vehicle collision (MVC) 25.7%, assault 9.6%, brain injury 4.8%, neurological intervention 0.8%. The 5 "high-risk" factors from the CT Head Rule predicted need for neurological intervention with sensitivity 100% (95% CI 83%-100%), specificity 80.2% (79%-82%), and required CT rate 20.4%. The additional 2 "medium-risk" factors predicted important brain injury with sensitivity 96.6% (95% CI 91%-98%), specificity 54.5% (52%-56%), and required CT rate 47.9%. Equally efficient for GCS 15 patients would be a rule that substituted "any drop in GCS" for "did not reach GCS 15 within 2 hours." CONCLUSIONS: When applied to minor head injury patients with GCS score 15, the Canadian CT Head Rule performs with a high degree of sensitivity and efficiency and would require only 48% of these patients undergo CT. Ongoing studies will explicitly and prospectively validate these findings.
Key words: brain injury, clinical prediction rule

028 What are the outcomes of patients classified as having clinically unimportant cervical-spine injury?
Stiell IG, Clement C, Lesiuk H, McKnight RD, Wells GA, De Maio VJ, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: The Canadian C-Spine Rule was recently developed to help ED physicians identify which alert and stable blunt trauma patients are at risk for clinically important cervical-spine injury (ICI). This methodological sub-study evaluated the clinical validity of

 the concept of "clinically unimportant cervical-spine injury" (CUCI), as previously endorsed by a survey of academic neurosurgeons and spine surgeons. METHODS: The prospective cohort CCC Study enrolled alert (Glasgow Coma Scale [GCS] score 15) and stable adult trauma patients at risk for neck injury at 10 Canadian EDs. Data collection included physician dataforms, plain and computed tomography (CT) film reviews by study neuroradiologists, in-hospital records, and 14-day telephone follow-up. Patients were considered to have CUCI, and therefore require neither specialized treatment nor follow-up, if they had one of these injuries: isolated osteophyte avulsion, isolated transverse process avulsion, isolated spinous process avulsion, and simple compression <25% of body height. The current study compared the characteristics of ICI and CUCI patients by chi-square and Student's t-test analyses. RESULTS: Among the 8,924 patients, there were 151 (1.7%) ICI cases and 28 (0.3%) CUCI cases (see Table 1). CONCLUSIONS: Patients with CUCI had fewer admissions and follow-up problems. No patient with CUCI had an unstable injury, developed neurological deficit, or required invasive treatment. This study confirms the validity of the concept of CUCI for alert and stable C-spine injury patients.
Key words: cervical spine, injury, diagnosis

029 Emergency department patient willingness to participate in complementary and alternative medicine research.
Abu-Laban RB, van Beek CA, Wu J, Quon J. University of British Columbia, Vancouver, British Columbia.

OBJECTIVES: The emergency department (ED) is a unique potential location for recruitment into studies of complementary and alternative medicine (CAM) therapies. To assess the feasibility of various studies, we sought to determine the stated willingness of ED patients with acute musculoskeletal complaints to participate in CAM research for their presenting problem and the characteristics of this population. METHODS: From Oct. 16, 2000 to Nov. 21, 2000 all adults presenting to a tertiary ED weekdays between 10am and 6pm with acute musculoskeletal complaints involving the spine, upper extremity or lower extremity were approached by a research nurse unless precluded by pain or communication barriers. After being presented background information, consenting patients were asked a series of standardized questions during a 15 minute interview prior to physician assessment. RESULTS: Of 107 eligible patients, 93 participated (87%). Most symptoms began on the day of presentation (44%) or the previous week (41%). The mean age of those studied was 38 years and 56% were male. Most presenting problems involved the ankle/foot (29%), multiple sites (19%), the lumbosacral region (14%) or the wrist/hand (14%). Seventy-six percent of patients had utilized CAM previously and 13% were currently using CAM for their presenting problem. The majority of patients stated an informed hypothetical willingness to enroll in a CAM study: 74% traditional Chinese medicine (69/93: 95% confidence intervals [CI] 64.1%-82.7%); 70% chiropractic (65/93: 95% CI 59.5%-79.0%); and 92% other CAM therapies (86/93: 95% CI 85.1%-96.9%). Of patients asked, 99% stated they would comply with 4-6 weeks of outpatient follow-up and 70% stated they would participate in a placebo-controlled study. Logistic regression modeling was performed for secondary purposes to determine predictors of participation willingness. CONCLUSIONS: ED patients with acute musculoskeletal complaints have a high stated willingness to participate in CAM research, even if this involves outpatient follow-up and/or a placebo-controlled design. ED-based CAM research appears feasible and should be pursued.
Key words: alternative health care

| 001-014 | 015-029 | 030-044 | 045-059 | 060-074 | 075-093 |