CAEP/ACMU 2001 Scientific Abstracts - Oral Presentations: 1-14

CAEP Abstracts

CJEM 2001;3(2):121-134

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

Oral Presentations

001 The epidemiology of emergency department presentations in Alberta.
Holroyd BR, Bullard M, Craig W, Klassen T, Johnson D, Yiannakoulias N, et al. University of Alberta, Edmonton, Alberta.

OBJECTIVES: Few places can accurately evaluate the pattern of emergency department (ED) presentations and valuable population-based administrative information is often unavailable. This study examines the epidemiology of presentations to the ED using an electronic provincial database. METHODS: All patients presenting to Alberta EDs were eligible for inclusion. Data were derived from the population of ED patients treated in the 17 health regions over 1 year (fiscal 98/99). Data were extracted from the Ambulatory Care Classification System (ACCS) database, a computerized database of abstracts coded similarly across all regions. Diagnostic categories were recorded using ICD-9 coding by medical record nosologists in each hospital and represented the primary physician discharge diagnostic code. Descriptive statistics and crude presentation rates are reported. RESULTS: During the 1 year 1,493,659 ED visits were recorded; males (52%) and females were similarly represented. Despite an aging population, patients over 60 account for only 336,365 (17%) of all ED visits. Limited seasonal variation was observed although weekly and hourly variation was marked. Of the approximately 2.3 million residents of Alberta, 687,389 (30%) make at least 1 visit to the ED annually. Multiple visits are common with more than 38,148 (6%) patients visiting EDs more than 5 times annually in this province. Left without being seen (LWBS) cases were also common (18,672 [1.3%]). The lowest ED presentation rates occurred in the 2 largest urban areas (population >500,000). Average rates are 229/1000 population (regional range: 148-526/1000). Most (1.2 million; 89%) patients are discharged from the ED. Overall, 1490 (<1%) deaths occurred in the ED setting.
CONCLUSIONS: On an annual basis, nearly one-third of Alberta citizens present to an emergency department for care; multiple visits and LWBS require further detailed evaluation. These results provide an opportunity to further characterize ED utilization and disposition to determine the role of ED in health care delivery.
Key words: emergency department, utilization

002 Development and evaluation of a chief complaint classification system for comparing emergency department clientele.
Afilalo M, Unger B, Colacone A, Lang E, Guttman A, Robitaille C, et al. McGill University, Montreal, Quebec.

OBJECTIVES: Emergency department (ED) clientele are characterized through indicators that reflect administrative interests (i.e., length of stay [LOS], hospitalization rates, and socio-demographics). These parameters are used in governmental statistics to establish performance outcomes and compare caseloads. Chief complaint (CC) or "reason for ED visit," a parameter not currently part of governmental statistics, may also serve as an index of caseload. At present there is no standardized CC classification method that can describe clienteles across institutions. A CC classification system was developed and evaluated for describing and comparing ED clienteles. METHODS: A classification method for "reason for ED visit" was developed for the adult population using sources including the list of CCs from one hospital since 1993, literature reviews, and expert recommendations. The classification list consists of 143 CCs divided into 12 anatomical categories (head, eyes, ear, nose, face, etc.) and 6 other categories (stat, traumatic, behavioural/social, non-specific, body fluid, and returns). This classification system was applied in a prospective sample of patients in 6 EDs over 9 months. RESULTS: At least 1 CC was recorded in 2,810 of the 2,841 patients recruited (99%). Among all CCs, 96% were represented. Overall, 3 CCs accounted for 22% of total visits (chest pain 8.0%, abdominal pain 7.3%, dyspnea 6.4%). However, these 3 represented 32% of patient-hours (LOS) in the ED, with each CC contributing approximately 10% of the total. All other CCs accounted for less than 4% of total patient-hours. Inter-institution variations between ED clienteles were observed in terms of frequency of specific CCs and associated LOS. For example, for chest pain, percent total visits varied between 4.8 and 15.7 while percent patient-hours ranged from 3.6 to 20.5. CONCLUSIONS: The proposed CC-based classification method is comprehensive and valuable for describing heterogeneous ED clienteles. Adopting its use across Canada would permit comparisons between different EDs.
Key words: triage, presenting complaint

003 The impact of a nurse discharge coordinator on the successful discharge of elderly patients from the emergency department.
Guttman A, Guttman R, Colacone A, Robitaille C, Lang E, Rosenthal S, et al. McGill University, Montreal, Quebec.

BACKGROUND: With the aging of the Canadian population, the elderly make up a progressively larger proportion of emergency department (ED) visits. These patients are particularly vulnerable to inadequate discharge planning after leaving the ED, thereby placing them at a greater risk for return visits. OBJECTIVES: To examine the impact of an ED nurse discharge coordinator on the rates of unscheduled return visits to the ED in elderly patients. METHODS: Patients over 75 years of age discharged from the ED of the Sir Mortimer B. Davis-Jewish General Hospital were recruited in a pre-post study design aimed at comparing return visits to any ED within 14 days after the index discharge. During the pre phase (June to December 1999), study patients (n = 905) received standard discharge care. Patients in the post phase (January to July 2000), (n = 819) received the services of a nurse discharge coordinator in the ED. The nursing intervention included patient education, coordination of appointments, telephone follow-up and access to the discharge nurse for up to 7 days post ED discharge. RESULTS: Preliminary results indicate that patient groups were similar with respect to sex and age, however patients in the post phase were on more medications at baseline (5 vs. 4, p < 0.05) and a greater proportion perceived their presenting complaint as very severe (53% vs. 45%; p < 0.05). Bivariate analyses show a relative risk reduction of 36%, 95% confidence interval (CI) (-7% to 62%) for unscheduled return visits the day following the index discharge, 25%, 95% CI (0% to 44%) for up to 8 days post discharge and 19%, 95% CI (-2% to 36%) for up to 14 days post discharge. CONCLUSIONS: The assignment of a nurse discharge coordinator, dedicated specifically to the ED, reduces the rate of unscheduled return visits to the ED.
Key words: older patients, discharge planning

004 Emergency department overcrowding in Toronto from 1991 to 2000: the effect of systematic hospital restructuring.
Schull MJ, Szalai JP, Schwartz B, Redelmeier DA. University of Toronto, Toronto, Ontario.

OBJECTIVES: To determine the impact of systematic hospital restructuring on emergency department (ED) overcrowding at hospitals in Toronto, Canada. Restructuring was characterized by acute care bed reductions and by the closure of some EDs. METHODS: Data on all ambulance diversions at all EDs (hospitals = 24) from 1991 to 2000 (months = 108) was obtained. Overcrowding was classified as severe (all ambulances diverted) or moderate (most ambulances diverted). For each month, the average proportion of time that EDs requested diversion of ambulances was calculated. Autoregression analyses were conducted separately for each of the 2 severity levels. Models included month, an indicator variable for period (before or after restructuring began in March 1997), and lag terms to assess seasonality. Linear trends were estimated and compared before and following restructuring. Secondary analyses were conducted with three additional variables: monthly ED patient volume, average patient age, and the proportion of females. RESULTS: The average ED experienced overcrowding 10% of each month at the midpoint prior to restructuring, 25% at the midpoint after restructuring, and 40% at the study's conclusion. Before restructuring, neither severe nor moderate overcrowding was increasing (slope of 0.0% per month [p = 0.54] and -0.04% [p = 0.04] respectively]). Following restructuring, both severe and moderate overcrowding began increasing significantly (slope of 0.2% per month [p < 0.0001], and 0.5% per month respectively [p < 0.0001]). The rates of change were significantly different before and following restructuring for both severe (p < 0.0001) and moderate (p < 0.0001) overcrowding. No evidence of seasonality was found. Results were similar after controlling for patient volume, age and sex. CONCLUSIONS: A large increase in ED overcrowding occurred following systematic hospital restructuring that is not explained by changes in patient demand. More effort is needed to understand how increased overcrowding affects patient care.
Key words: overcrowding, emergency department

005 An "ultra-sensitive" version of the Canadian CT Head Rule for US physicians.
Stiell IG, Dreyer JA, Worthington JR, Greenberg GH, Clement C, Wells GA, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: The high sensitivity and efficiency of the Canadian CT Head Rule for minor head injury patients appears to be very acceptable to Canadian and European physicians. This study attempted to derive an even more sensitive rule that might be more acceptable to US physicians. METHODS: This secondary data analysis was based on a prospective cohort study conducted in 10 Canadian EDs and involving adults with loss of consciousness, amnesia, or confusion and Glasgow Coma Scale (GCS) scores of 13-15. Physicians completed a 22-item data form for all patients who then underwent computed tomography (CT). The outcome measures were need for neurological intervention, clinically important brain injury, and any acute brain injury on CT. For this study, chi-square recursive partitioning (RP) analyses (KnowledgeSEEKER) were used to derive a more sensitive model. Statistical measures included comparison of measures by chi-square analysis. RESULTS: The CT Head (CCC) Study data-set contains 3,121 minor head injury cases with mean age 38.7 years, males 68.4%, falls 30.9%, acute brain injury on CT 11.2%, important brain injury 8.1%, and required neurological intervention 1.4%. RP analyses lead to a new "ultra-sensitive" model with 1 additional variable, the object recall test, giving a total of 8 variables. Comparing the "ultra-sensitive" model to the original CT Head Rule:

CONCLUSIONS: This study has derived an "ultra-sensitive" version that is more sensitive but less specific and would require a higher CT rate than the original Canadian CT Head Rule. The accuracy and acceptability of this revised rule for minor head injury patients should be prospectively and explicitly validated in US settings.
Key words: clinical prediction rule, brain injury

006 The validity of a clinical model to predict the presence or absence of deep vein thrombosis in the emergency department.
Scheibel N, Spooner CH, Sukhrani N, Cunningham R, Kelly KD, Holroyd BR, et al. Mayo Clinic, Rochester, Minnesota.

OBJECTIVES: In patients presenting to the emergency department (ED) with suspected deep vein thrombosis (R/O DVT), <25% are assigned a final diagnosis of DVT. A new clinical model (CM) to determine the pretest probability of DVT and guide efficient investigation has been developed. This study examined the validity of this CM in 3 urban North American EDs. METHODS: In a prospective fashion, patients were enrolled in 2 Canadian and 1 US ED. Patients >18 years of age with a R/O DVT were approached for enrollment. A CM was completed by the ED physician prior to obtaining any laboratory or diagnostic evaluation. The clinical model comprises eight examination and historical factors and is converted into risk categories (low, moderate, high). Patients were subsequently followed by telephone and through a database evaluation to confirm the diagnosis of DVT. RESULTS: Overall 1,067 patients presented to the 3 EDs with diagnosis of rule-out DVT; the CM was completed in 489 patients (46%). Patients in the CM group had the following risk classifications: 217 (45%) low, 164 (34%) moderate and 98 (20%) high risk for DVT. After follow-up, 17 (8%) low, 41 (25%) moderate and 39 (48%) high-risk patients were shown to have a proximal DVT (p < 0.001). Physician vs. CM estimates of risk agreed only moderately (kappa = 0.48). In low, moderate and high risk groups, D-dimers were performed 39%, 40% and 33% of the time, ultrasound was first investigation in 79%, 75% and 72%, and venograms were performed 12%, 21% and 25%, respectively. CONCLUSIONS: Clinical models applied in the ED are valuable tools for improving quality of care. However, the DVT CM was used <50% of the time and investigations did not appear to be influenced by risk categorization. Further research is required to determine how to reduce practice variation in the ED.
Key words: deep vein thrombosis, diagnosis

007 Retrospective validation of the "New Orleans" criteria for minor head injury.
Stiell IG, Eisenhauer MA, Dreyer JA, Reardon M, Clement C, Wells GA, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: The "New Orleans" criteria have been recently proposed to guide the need for computed tomography (CT) in minor head injury patients with a score of 15 on the Glasgow Coma Scale (GCS). This study retrospectively validated these criteria on an existing data-set of head injury patients. METHODS: This secondary data analysis was based on a prospective cohort study conducted in 10 Canadian emergency departments (EDs) and involving adults with loss of consciousness, amnesia, or confusion and GCS scores of 13-15. Physicians completed a 22-item standardized data-form prior to CT scan, and the outcome criterion was clinically important brain injury. Data collection also included CT reviews by study neuroradiologists, ambulance call reports and in-hospital records. For the current study, chi-square recursive partitioning analyses (KnowledgeSEEKER) were performed using the 10 variables that corresponded to the 7 New Orleans criteria. Statistical measures included sensitivity, specificity and CT rate. RESULTS: The Canadian C-Spine/CT Head (CCC) Study data-set contains 2,489 minor head injury cases with a GCS score of 15 and these characteristics: injury above clavicle 74.6%, headache 58.8%, vomiting 20.9%, >60 years 14.0%, intoxication 8.0%, persistent amnesia 7.3%, seizure 0%, important brain injury on CT 4.8%, required neurological intervention 0.8%. For the primary outcome, important brain injury, the New Orleans criteria classified patients with a sensitivity of 97.5% (95% confidence interval [CI] 92%-99%) and specificity of 8.4% (95% CI 7%-10%). For the outcome, need for neurological outcome, the criteria had a sensitivity of 90% (95% CI 67%-96%). According to the criteria, 91.9% (95% CI 91%-93%) of patients would require CT. CONCLUSIONS: The criteria performed with good sensitivity for important brain injury and fair sensitivity for neurological intervention and would have required almost all GCS-15 head injury patients undergo CT. The New Orleans criteria should be further explicitly and prospectively evaluated for accuracy, reliability and potential impact prior to widespread clinical use.
Key words: brain injury, clinical prediction rule

008 Urgent imaging for suspected renal colic in the emergency department.
Papa L, Stiell IG, Wells GA, Battram E, Mahoney J. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: To predict which patients with suspected renal colic have severe obstruction and require urgent urinary tract imaging. Several studies suggesting that intravenous pyelogram (IVP) be delayed or eliminated have not suggested criteria. Clinical practice is known to be inaccurate in predicting severe obstruction. METHODS: This prospective cohort study included suspected renal colic cases presenting to 2 tertiary care hospital emergency departments (EDs). Patients had a 20-variable dataform completed by ED physicians and an IVP performed within 24 hours. All IVPs were reviewed by a radiologist and a urologist to identify those cases with severe ureteral obstruction (urine extravasation and/or no visualization of contrast beyond the obstruction after 2 hours). Categorical data were analyzed using Fisher's exact test, continuous variables assessed by independent sample 2-tailed t-test, and ordinal variables tested with Mann-Whitney U test. Those variables found to be associated with the outcome measure of severe obstruction (p < 0.15) were combined using logistic regression analysis. RESULTS: Of the 119 patients included in the analysis, 18 (15%) had severe obstruction identified on IVP. Four clinical variables were significantly correlated with having severe obstruction on IVP: i) residual pain at discharge > = 1 cm on VAS (OR = 7.8 95% confidence interval [CI] 1.6%-82.1%), ii) rebound tenderness on abdominal palpation (OR = 18.5, 95% CI 1.1%-303.0%), iii) vomiting (OR = 7.8 95% CI 1.0%-62.0%) and iv) persistent pain after 6 hours in the ED (OR = 18.6 95% CI 2.2%-159.0%). Hosmer-Lemeshow goodness of fit statistic was 7.4 for 6DF (p = 0.283). If all variables were positive there was a 99.4% probability of severe obstruction. If all variables were negative there was a 0.6% probability of obstruction. CONCLUSIONS: This preliminary study has identified clinical variables that could be used to identify those patients with renal colic who require urgent imaging. Future studies will prospectively validate this model.
Key words: clinical prediction rule, renal colic

009 Long-acting beta-agonists following emergency department discharge: a randomized controlled trial.
Rowe BH, Travers A, Brown J, Tyler L, Folk D, Spooner CH, Kelly KD. University of Alberta, Edmonton, Alberta.

OBJECTIVES: One regimen to reduce relapse following emergency department (ED) therapy for acute asthma includes oral prednisone and inhaled corticosteroid (ICS). We investigated the effect of adding a long-acting beta-agonist to this regimen for patients discharged from 3 Canadian EDs. METHODS: Patients aged 18-60 with acute asthma receiving <2000 mcg of beclomethasone dipropionate or equivalent were eligible. All patients received similar ED treatment and upon discharge received oral prednisone (50 mg * 7 days) and short-acting inhaled beta-agonists. In a double blind fashion, patients were randomly assigned to receive either Advair (ADV; 1000 mcg/d fluticasone and 100 mcg salmeterol) or fluticasone alone (FLU; 1000 mcg/d) via dry powder delivery system. Patients were followed for 21 days or until relapse. The main outcome was the Asthma Quality of Life Questionnaire (AQLQ) score; the minimal clinically important difference for the AQLQ is 0.5/domain. RESULTS: 70 patients were enrolled (35 ADV; 35 FLU); the groups were similar at the outset. Full follow-up was achieved in 63 (90%) patients. The ADV group showed less impairment (ADV vs. FLU) for total (6.4 vs. 6.2) AQLQ, and the activity (6.4 vs. 6.2), environmental (6.5 vs. 6.1), emotional (6.3 vs. 6.1) and symptom (6.3 vs. 6.2) AQLQ domains at 21 days; no differences were clinically or statistically significant (p > 0.1). Relapse (11% vs. 23%; p = 0.23) and beta-agonist puffer use/day (1 vs. 2; p = 0.22) outcomes also favoured ADV at 21 days. Early inhaler compliance was high (66% vs. 83%, p = 0.34) and side effects were similarly rare in both the groups. CONCLUSIONS: Adding a long-acting beta-agonist to prednisone and ICS therapy does not appear to significantly improve quality of life for patients recovering from acute asthma. However, the promising trends that show a reduction in relapses and beta-agonist use suggest further large scale studies are required to verify these preliminary findings.
Key words: asthma, corticosteroids, quality of life

010 Are intubation conditions using rocuronium comparable to succinylcholine? A meta-analysis.
Perry JJ, Lee J, Wells GA. Clinical Epidemiology Unit, University of Ottawa, Ottawa, Ontario.

OBJECTIVES: To complete a systematic review to determine if rocuronium creates excellent intubation conditions as frequently as succinylcholine during rapid sequence induction (RSI). Individually, existing studies lack sufficient power to determine their equivalence. METHODS: Medline, Embase, and the Cochrane Controlled Trials Register databases were searched for controlled or randomized clinical trials (RCT). The search strategy included all generic and trade names for succinylcholine and rocuronium, anesthesia, neuromuscular blockade and a validated RCT filter. Intubation conditions were a required outcome. Foreign language journals were included, and references were hand searched. Two independent reviewers assessed studies for eligibility, data extraction and quality. Intubation conditions were scored using Goldberg's scale (excellent conditions defined as clear vocal cords, easy tube insertion, and no cough). A priori subgroup analysis was conducted using propofol for induction. The effect of narcotics, true versus modified RSI, and the dose of rocuronium were checked by sensitivity analysis. Data was analyzed with Metaview 4.0 for relative risk (RR) with a variable effects model. RESULTS: The search identified 40 articles. 10 articles were excluded for not meeting inclusion criteria, 2 were duplicate publications and 2 were abstracts with insufficient data. For the 26 studies included in the analysis, overall, rocuronium was inferior to succinylcholine, with a RR = 0.87 (95% confidence interval [CI], 0.81,0.94) (n = 1606). Intubation conditions were similar in the propofol subgroup, with a RR = 0.96 (95% CI, 0.87, 1.06) (n = 640). A total sample size of 468 is required for equivalence (alpha = 0.05, beta = 0.10, MCID 12%, % expected with excellent conditions = 80%). Sensitivity analysis of the propofol subgroup demonstrated no effect of narcotics, true versus modified RSI, or the dose of rocuronium. Failed intubations (n = 28) were equivalent in both groups. CONCLUSIONS: Overall, succinylcholine creates excellent intubation conditions more reliably than rocuronium. If an alternative is required, rocuronium used with propofol creates intubation conditions equivalent to succinylcholine.
Key words: rocuronium, succinylcholine, intubation

011 Emergency department presentations of chronic obstructive pulmonary disease in Alberta.
Spooner CH, Yiannakoulias N, Holroyd B, Bullard M, Craig W, Klassen T, et al. University of Alberta, Edmonton, Alberta.

OBJECTIVES: Chronic obstructive pulmonary disease (COPD) is an increasingly common problem in North America. However, its burden on emergency departments (EDs) is virtually unknown. This study examines the epidemiology of COPD presentations to the ED using a provincial database. METHODS: All patients presenting to Alberta EDs were eligible for inclusion. Data were derived from a population of patients treated at Alberta EDs in 17 health regions over 1 year (fiscal 98/99). Data were extracted from computerized abstracts coded similarly across all regional EDs contained within the Ambulatory Care Classification System (ACCS) database. Diagnostic categories are recorded using ICD-9 coding by medical record nosologists in each hospital and represented the primary physician discharge diagnostic code. Patients aged 55 and over with a ICD-9 code compatible with COPD were selected for study; descriptive statistics and crude presentation rates are reported. RESULTS: Over 1 year, 1,493,659 ED visits were recorded; 21,147 (1.4%) patients aged 55 and over presented to the ED during this period with a diagnosis of COPD. Males (10,419; 50%) and females were similarly represented. The highest months for presentation are December-February (36%); daily variation is low however most presentations occur in the morning. Only 5,266 (44%) of patients present to the ED once per year, the rest make multiple presentations annually (up to 45). The provincial presentation rate is 10/1000 persons (regional range: 5-47/1000). The lowest ED presentation rates occurred in the 2 largest urban areas (population >500,000). An important percentage (28%) are admitted to hospital, few leave without medical care (<0.1%) and admission to ICU occurs in 1% of cases. CONCLUSIONS: Repeat ED presentations for COPD in the elderly are common in this jurisdiction. Severity is high and admission is more frequent than for other respiratory diseases. Further research is required to understand the regional variation and examine ED treatments.
Key words: utilization, obstructive lung disease

012 Comparison of Canadian versus US emergency department visits for asthma/COPD (chronic obstructive pulmonary disease) exacerbation among older adults.
Rowe BH, Cydulka RK, Camargo CA Jr, for the MARC Group. University of Alberta, Edmonton, Alberta.

OBJECTIVES: To compare emergency department (ED) visits for asthma/COPD (chronic obstructive pulmonary disease) exacerbations among adults age 55+ in Canada vs. the US. METHODS: Prospective, multicentre, inception cohort study. 23 EDs (4 Canadian and 19 US) enrolled patients (pts) 24 hrs/day for a median of 2 weeks, as part of the Multicenter Airway Research Collaboration. Enrolled pts underwent a structured interview in the ED and follow-up by telephone 2 weeks later. Inclusion criteria were prior MD diagnosis of asthma or COPD, current exacerbation of their obstructive airway disease, and age 55+. Chi-2, t-test, and Kruskal-Wallis (K-W) test were used as appropriate. RESULTS: Of the 371 enrolled pts, 62 (17%) were seen in Canadian EDs. Enrollment was similar in Canada vs. US sites (64% vs. 69% of consecutive pts, p = 0.30). Canadian pts were older (71 vs. 69 years, p = 0.03), more likely to be white (92% vs. 58%, p < 0.001), but less likely to have completed high school (37% vs. 54%, p = 0.02). Prior MD diagnosis of asthma only, asthma and COPD, and COPD only did not differ between Canadian and US pts (p = 0.49). Canadian and US pts also did not differ according to reported breathing difficulty on arrival (p = 0.11). While pts in both countries were equally likely to receive steroids (74% vs. 64%; p = 0.14), Canadian pts were more likely to receive anticholinergics (92% vs. 77%, p = 0.007) and more likely to receive antibiotics (47% vs. 23%, p < 0.001) in the ED. Canadian and US pts were equally likely to be admitted overnight to the hospital (68% vs. 57%, p = 0.11) but Canadians were more likely to keep pts in the ED for 6+ hours (90% vs. 29%, p < 0.001). At 2-week follow-up, Canadian and US pts did not differ according to risk of having an ongoing exacerbation or a "relapse" requiring acute medical care (both p > 0.15). CONCLUSIONS: Among ED pts age 55+ with asthma/COPD exacerbations, compared to US pts Canadian pts were more likely to receive supplemental ED treatments and had longer ED stays, but had similar 2-week outcomes.
Key words: asthma, obstructive lung disease, utilization

013 The role of injury mechanism in the evaluation of patients with minor head injury.
Stiell IG, Lesiuk H, Clement C, Wells GA, De Maio VJ, Battram E, et al, for the CCC Study Group. University of Ottawa, Ottawa, Ontario.

OBJECTIVES: Emergency department (ED) physicians frequently face the decision of ordering computed tomography (CT) for patients with minor head injury. This study measured the risk of brain injury associated with specific injury mechanisms for minor head injury patients. METHODS: This prospective cohort study was conducted as a component of the Canadian C-Spine/CT Head (CCC) Study 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 completed a 22-item assessment form prior to CT scan and the outcome criterion was clinically important brain injury on CT. 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. RESULTS: Among 3,121 patients enrolled over 36 months, 254 (8.1%) had important brain injury and 44 (1.4%) required urgent neurological intervention. The most common injury mechanisms were: falls 30.9%, MVC 25.8%, assault 10.7%, sports 9.9%, bicycle 6.7%, pedestrian struck 5.9%, head struck by object 5.8%, motorcycle 3.5%. After adjustment for demographic and clinical factors, analysis found the following mechanisms to be independently associated with increased risk of brain injury (odds ratios with 95% confidence intervals [CIs]): MVC ejected 15.0 (6.2-36.7), pedestrian struck by motor vehicle 6.2 (3.8-10.0), bicycle collision 3.6 (1.1-12.5), head hit by object 3.1 (1.6-6.3), fall >10 feet 3.1 (1.7-5.7), bull's eye damage to windshield 3.0 (1.2-8.0), fall 3-10 feet 2.2 (1.4-3.6). This model discriminated with an area under the receiver operating characteristic (ROC) curve of 0.90. CONCLUSIONS: Specific injury mechanisms place patients at increased brain injury risk and ED physicians should carefully ascertain details of the injury when making decisions regarding CT for minor head injury patients. The accuracy and reliability of these mechanism risk factors will be prospectively evaluated.
Key words: brain injury, diagnosis

014 Emergency department patients' opinions of screening for intimate partner violence.
Campbell SG, Crooks C, Wallace T, Venugopal R. Dalhousie University, Halifax, Nova Scotia.

OBJECTIVES: Universal screening for intimate partner violence (IPV) in the emergency department (ED) has been advocated by numerous medical institutions. Implementation of policies to screen IPV, however, have met with numerous obstacles, one of which is the perception by emergency staff that patients might be offended by such screening if they had presented to the ED for problems unrelated to trauma. To assess the opinion of patients presenting to the ED, regarding a policy of universal screening of female ED patients for IPV. METHODS: A convenience sample of 250 undifferentiated ED patients were asked whether it was appropriate far all women to be asked if they had experienced violent or threatening behaviour from someone close to them. Patients in significant pain or in extremis were not approached. The Queen Elizabeth II IPV committee prospectively decided by consensus that approval of 85% of patients would be adequate to demonstrate that a significant majority of patients supported universal screening. RESULTS: Of 250 undifferentiated ED patients, ages ranged from 16-95 years, and 151 (60.4%) were female. 213 (85.2%) answered "yes" to the question. 30 (12%) answered "no," 5 (2%) had no opinion, and 2 (0.8%) could not be counted. There were no significant differences between the proportion of "yes" and "no" answers in the male and female groups; "yes" in 84.8% of males and 85.4% of females, and "no" in 10% and 13.2%, respectively. CONCLUSIONS: Universal screening for IPV of female patients presenting to the ED is supported by 85% of patients, and patient objections should not be seen as a reason to withhold questioning on the issue.
Key words: domestic violence, emergency

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