2007 CAEP/ACMU Scientific Abstracts - Poster Presentations: 191-216

2007 Scientific Abstracts

CJEM 2007;9(3):183-226

Abstracts: 1-30, 31-44, 46-48, 100-130, 131-159, 161-189, 191-216

Poster Presentations

191 RURAL emergency department use by CTAS IV and V patients

Steele S, Anstett D, Milne WK. South Huron Hospital and University of Western Ontario, London, ON

Introduction: It is well known that many Emergency Department (ED) visits are non-urgent as classified by the Canadian Triage and Acuity Scale (CTAS). Many reasons have been suggested for such ED use including a lack of family physicians. A survey done by Field and Lantz (2006) quantified some reasons patients with non-urgent problems use a tertiary care ED. The purpose of our study was to determine if these same reasons applied to patients presenting with non-urgent problems to a low volume rural ED. Methods: The same patient survey used in the aforementioned urban study was administered to CTAS Level IV and V patients, who attended the South Huron Hospital ED in Exeter, Ontario for two weeks in December, 2006. Results: The survey completion rate was 97% (137/141) of non-urgent patients presenting to the ED. Eighty-nine percent (122/137) of patients reported having a family physician. More than half (70/136) had their problem for more than 48 hours while only 24% (33/137) had an acute medical problem of less than 48 hours. However, 39% (53/136) felt they had a problem which needed treatment as soon as possible. Forty-three percent (53/122) reported using the ED because of limited access to their family physician. One third (45/137) presented to the ED because of special services that they thought they would require (xray, sutures, casting and IV medication). Only 4% (6/137) used the ED because they did not have a family physician. Conclusions: In this low volume rural setting, most non-urgent ED visits were for medical problems of more than 48 hours. Many patients used the ED for specialized services; most of which are available at the walk-in clinic. The lack of a family physician was not associated with non-urgent ED use. In fact, many patients had already seen a doctor about their problem. The most common reason cited for coming to the ED was inability to obtain timely access to their family physician. Key Words: Non- urgent, Low acuity, CTAS, Rural ED

192 WHAT strategies are the most efficient for the identification of medical errors in the emergency department?

Dankoff J, Afilalo M, Soucy N, Lang E, Guttman A, Léger R, Xue X, Colacone A, Le Sage N, Guimont C. Division of Emergency Medicine, McGill University, Montréal, QC

Introduction: Medical errors [MEs] and patient safety have become important topics on the national health care agenda. Emergency departments [EDs] are recognized as an environment at high risk for MEs due to their clientele and the nature of the work. However, little is known concerning the various methods for detecting and investigating of MEs in the ED. This project aims to provide a comprehensive method of detecting and investigating MEs in the ED.Methods: A seven month prospective comparative study of methods for detecting MEs was conducted in a tertiary care hospital ED with an annual census of 70 000 visits. Four identification processes were implemented to flag potential MEs in the ED: Doctor's (MD) incident reports; nursing incident reports; database queries 3 consultations, length of stay ≥ (death in ED, > 48 hrs, ICU/CCU admission); and pharmacy reports on specific medications. For each potential ME identified through electronic database queries, an email was sent to the MD taking care of the admitted patient asking them to assess "patient care" within the ED. Potential MEs were classified as case (ME with moderate to severe adverse event [AE]), control (ME without or with minor AE) or normal (No ME). Results: The sample size consisted of 1117 flagged potential MEs: male = 54%; mean age = 65.8; triage code I = 1%; II = 33%; III = 49%; IV = 16%; V = 1%. Six percent of potential MEs were reported by MDs, 11% by nurses, 67% through database queries and 16% through pharmacy reports. Among the 1117 potentials MEs, 14% were identified as MEs. From these 14%, 18% resulted in moderate to severe AEs and 82% in minor or no AEs. Moreover, sensitivity of the various methods to detect MEs is as follows: MD reports = 34%, nursing reports = 56%, database queries = 9% and pharmacy reports = <1%. Conclusions: Although database queries and pharmacy reports flagged 83% of the potential MEs, the MD and nursing incident reports were the more sensitive in identifying the presence of ME (90%). Key Words: Medical error, Patient safety, Incident reports

193 AN analysis of emergency department patient safety culture

Hunte GS, Brubacher JR, Abu-Laban RB, Zed PJ, Shepherd J, Doyle D, Griffin M, Sheps SB. Department of Emergency Medicine, St. Paul's Hospital and Department of Health Care and Epidemiology, University of British Columbia, Vancouver, BC

Introduction: Safety is an emergent phenomenon of sociotechnical system interactions. Therefore, we sought to explore how healthcare providers create patient safety within the culture of an emergency department. Methods: In-depth, semi-structured interviews were conducted with eighteen front-line emergency department (ED) healthcare professionals from two urban Canadian tertiary care EDs (six physicians, nine registered nurses, one resident physician, one ED technician, and one social worker). Statements from the Hospital Survey on Patient Safety Culture were used to prompt detailed narrative exploration of the factors that contribute to patient safety culture. Interviews were recorded and transcribed, and data were analyzed iteratively for patterns and themes. Results: Interviews averaged ninety minutes in length. Six broad areas were identified as issues to address in order to better foster a culture of safety in the ED: 1) Flexibility and resilience with respect to flow, capacity, resources, and recovery when errors occur; 2) Transparency and voice; blame or shame was felt by some, and speaking up was rarely rewarded; 3) Safety and security of person and department, including physical and psychological safety; 4) Reporting of patient safety events was often perceived as ineffectual, with a sense of "no harm, no foul" when a near miss occurs; 5) Feedback and learning, if present, were experienced as ad hoc and local rather than systematic and organizational; 6) Teamwork, communication and responsibility themes endorse the risks associated with transitions in patient care. Conclusions: These findings highlight organizational features that influence the ability of healthcare providers to create safety for patients in busy, urban EDs. Further research and evaluation of interventions to promote patient safety culture in the ED are warranted. Funding: Canadian Medical Protective Association Key Words: Patient safety, Qualitative, Safety culture

194 CLINICAL utility of emergency department targeted ultrasonography in symptomatic first-trimester pregnancy: a systematic review of current evidence

McRae AD, Edmonds M, Murray H. Department of Emergency Medicine, Queen's University, Kingston, ON

Introduction: This systematic review examined the effect of Emergency Department Targeted Ultrasonography (EDTU) on time-to-diagnosis and time-to-surgery for ectopic pregnancy, emergency department length-of-stay, and costs in the evaluation of emergency patients with symptoms of pain or bleeding in the first trimester of pregnancy. Methods: The literature was systematically searched. We included articles that examined the outcomes listed above, and that compared EDTU to formal ultrasonography performed by diagnostic imaging departments or consultant gynaecologists. No constraints on methodological quality were used. Rather, the quality of each study was critically appraised and described in detail in the review. Results: Nine studies were included in this systematic review. Three retrospective studies examined time-to-diagnosis and time-to-surgery in patients evaluated using EDTU compared to historical controls evaluated with formal ultrasonography. Mean time-to- diagnosis was reduced by over 2 hours in one study, while time-to-surgery was reduced by 2.5-3.5 hours. EDTU reduced the proportion of patients with ectopic pregnancy who were discharged from the ED. One randomized, controlled trial and one prospective cohort study demonstrated that EDTU reduced length of stay by approximately two hours for patients with live intrauterine pregnancies. Three retrospective studies had similar findings. One randomized, controlled trial and one retrospective study demonstrated lower health-care costs associated with EDTU compared to formal ultrasonography. No pooled analyses were performed because of heterogeneity of the methods of the included studies. Conclusions: The use of EDTU in symptomatic first-trimester pregnancy reduced ED length of stay for patients with viable pregnancies, and reduced health care costs. There is also evidence that EDTU use reduces time-to-diagnosis and time-to-treatment of ectopic pregnancies. Key Words:Ultrasonography, EDTU, Pregnancy, Ectopic

195 WHAT is the impact of sexual assault nurse examiners on emergency department care?

Sampsel K, Szobota L, Pickett W, Joyce D, Graham K. Department of Emergency Medicine, Queen's University, Kingston, ON

Introduction: Examination and management of the sexually assaulted patient is a complex task. On-call nurses with advanced training are used in some hospitals but their impact on patient care and appropriate forensic examination is largely unknown. We evaluated the impact of the introduction of a Sexual Assault/Domestic Violence Program (SADVP) on emergency department (ED) flow, comprehensive patient care and collection of forensic evidence. Methods:Patients presenting to the Kingston area (pop 185000) EDs following sexual assault were compared for two time periods: 1) pre-SADVP (2001 through August 2004) 2) post-SADVP (September 2004 to August 2006). ED, hospital discharge, SADVP and police records were reviewed. Data abstraction included patient demographics, assault characteristics, forensic examination results and treatment protocols. Results: More patients were seen by the SADVP in a shorter time frame (n=61 pre-SADVP; n=92 post-SADVP). SADVP numbers approximate number of police reports. Median times to initial clinical evaluation were lower in the SADVP group (20 vs. 33 minutes; P=.04). Patients in the post-SADVP group reported less vaginal/anal penetration (77% vs. 98%; P<.001) and experienced fewer genital injuries (13% vs. 39%; P=.007); other sexual assault characteristics were similar between the two study periods. Forensic kits were completed more often in the post-SADVP group (77% vs. 66%; P=.18). Pregnancy and STD prophylaxis was provided more consistently post-SADVP (98% vs. 85%; P=.007) as was counseling (100% vs. 95%; P=.06). Conclusions: The profile of patients observed post-SADVP changed to include less stereotypical sexual assaults with less discrepancy seen between police and hospital estimates of sexual assault. Introduction of the SADVP decreased wait times for sexually assaulted patients, despite the need for the on-call nurses to attend the ED. This program also showed higher completion on a number of important indicators of quality of care: forensic kits, counseling and pregnancy and STD prophylaxis. Key Words: Sexual assault, Specialized nurses, Program impact, Wait times

196 DIAGNOSTIC imaging for renal colic in Ontario emergency departments

Dreyer JF, Edmonds ML, McLeod SL. Division of Emergency Medicine, The University of Western Ontario, London, ON

Introduction: Over the past decade, computed tomography (CT) has become a common imaging modality for patients with suspected renal colic. However, there is ongoing debate as to the most appropriate initial investigation for these patients, and in particular, concern about the amount of radiation exposure from CT for younger women and children. This study examined the inter-hospital variation in investigations performed in patients diagnosed with renal colic, and variation with age and gender. Methods: Ten hospital-based emergency departments (EDs) comprised of a mix of paediatric, teaching, small and large volume community hospitals from across Ontario agreed to participate in this study. Data for all ED visits was collected by each hospital, and submitted to the Canadian Institute for Health Information (CIHI). This data set, for the period from September 1, 2005 to August 31, 2006 was reviewed by the investigators. An analysis was conducted of all patients discharged with a confirmed diagnosis of urolithiasis or renal colic (ICD-10 codes N20-N23). Results:Of 518,488 ED patients treated during the study period, 3,264 (0.6%) had a discharge diagnosis of renal colic. There was marked variation between hospitals in diagnostic modality chosen. 61.8% underwent imaging. Of these patients that had investigations, 16.1% (range 1.8-71.4%) underwent ultrasound (US), 20.9% (range 6.3-64.0%) had only plain x-ray, 61.6% (range 0.0-89.9%) underwent CT and 1.5% (range 0.0-6.3%) had both US and CT. Women <45 and children were less likely to receive a CT scan than adult males and females > 45 years (p=0.02). Conclusions: CT scan appears to be the imaging modality of choice at most centres in patients with a final diagnosis of renal colic, although there was marked variation in use of ultrasound and plain radiographs between hospital sites. In this study there was less use of CT scan in children and young women. Further research is needed to clarify the most appropriate use of CT in the diagnosis of renal colic. Key Words: Renal colic, CT, Imaging, Urolithiasis

197 IMMEDIATE complications of central venous catheterization

Cormier RJ, Woo MY. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

Introduction: Central venous catheterization (CVC) is associated with serious patient morbidity and mortality. The objective of this study was to determine the prevalence of CVC mechanical complications. Methods: A formal health records review was performed for the period August 2005 to December 2005 in a university tertiary referral centre. Research ethics approval was obtained. Inclusion criteria included any patients who received a percutaneously inserted CVC, anywhere in the hospital. Patients with peripherally inserted central catheters were excluded. A single reviewer identified cases using procedure codes for CVC and abstracted data into a standardized electronic database. Descriptive statistics were used on the collected data which included operator, training, procedural approach, and complications. Results: 412 medical charts were identified and reviewed. 242 charts were excluded because of lack of documentation and unavailability of records. 170 charts were analyzed and the mean patient age was 51.0. 60.2% were male. Main indications for CVC were IV access (53.5%), dialysis (25.3%) and OR procedure (7.1%). Most procedures were done in the ICU (48.8%) followed by diagnostic imaging (19.4%), the ward (15.9%), and the emergency department (9.4%). 64.9% of CVC were performed by residents. Most residents were in postgraduate year two (19.4%) and three (9.4%), although 22.9% did not document their level of training in the medical record. There was an overall complication rate of 9.9%. Arterial puncture was the most common complication (4.7%), followed by hematoma (4.1%) and pneumothorax (1.2%). Veins most commonly used were internal jugular (51.8%), subclavian (26.5%) and femoral (21.8%) and each were associated with a complication rate of 7.9%, 9.4%, and 16.1%, respectively. Conclusions: Accurate documentation regarding CVC is quite limited. Strategies to improve documentation are needed to accurately monitor educational interventions to decrease complications associated with CVC. Key Words: Complications, Documentation, Central line, Central venous catheterization

198 A comparison of the care profile of physicians and nurse practitioners in a community hospital emergency department

Dreyer JF, McLeod SL. Division of Emergency Medicine, The University of Western Ontario, London, ON

Introduction: Physician shortages in Ontario have lead to the introduction of alternate care providers to care for lower acuity patients in Emergency Departments (EDs). Nurse practitioners (NPs) are independent primary healthcare providers with a defined scope of practice who work in cooperation with physicians in this setting. We examined the activities of emergency physicians (EPs) and NPs as well as the profile of patients seen by both provider groups in a community hospital ED. Methods: Data was collected at a single community hospital ED by research assistants who directly observed EPs and NPs for entire shifts and recorded their activities on a moment-by-moment basis for two separate one-week periods in 2006. Patients were categorized according to the five-point Canadian Triage and Acuity Scale (CTAS). The individual times of all activities associated with a given patient were summed to derive a directly observed estimate of the amount of provider time required to treat a patient. Results: 1,539 patients were observed. NPs only treated patients in the three lowest acuity categories. Data analysis was therefore confined to CTAS 3, 4 and 5 patients so that a direct comparison of EPs and NPs could be performed. Median (interquartile range) treatment time for CTAS 3, 4 and 5 patients cared-for by EPs was 18.5 min (12.4-28.9), 9.3 min (5.9-14.0) and 8.1 min (5.5-12.5), respectively. Treatment time for CTAS 3, 4 and 5 patients cared for by NPs was 25.1 min (19.5-40.2), 20.4 min (15.2-27.4) and 11.7 min (9.8-27.2), respectively. EPs treated an average of 3.2 patients/hour while the NPs treated 1.9 patients/hour. Total time spent in activities unrelated to patient care was not different between EPs (13.2%) and NPs (14.8%). Conclusions: NPs were able to successfully assess and treat lower acuity patients in a community ED. Treatment times for patients cared for by NPs were longer (p<0.05) in all CTAS categories, resulting in lower patient throughput than EPs. Patient outcomes and satisfaction were not assessed. Key Words: Nurse practitioners, Throughput, Treatment times

199 CRITICAL care in Canadian emergency departments: a survey of CAEP members

MacIntyre J, Green RS. Department of Emergency Medicine, Dalhousie University, Halifax, NS

Introduction: Early and aggressive management has been demonstrated to improve outcomes in critically ill patients. However, little data is available on the management of critically ill patients in the ED. The goal of this survey is to characterize the provision of critical care in Canadian ED's. Methods: An internet-based survey was distributed to members of CAEP. Participants were asked to provide data on the management of critically ill patients in their ED including demographic data, and comfort level in the performance of various invasive procedures and vasopressor medications. In addition, barriers to the provision of critical care in the ED were explored. Pairwise and multivariate Wilcoxon rank-sum tests were used to compare categorical variables. Results: Of the 360 respondents (response rate 44.9%), 79.6% indicated that they managed >6 critically ill patients per month. Critical care consultants and respiratory therapists were commonly available (78.4%; 77.2% respectively). Emergency physicians were responsible for the management of critically ill patients awaiting transfer to an ICU either alone (21.9%) or jointly with the ICU service (50.0%). The majority of invasive procedures performed on critically ill patients were by emergency physicians. Physician comfort with invasive procedures and choice of vasopressor were closely related to their frequency of use in the previous year (p<0.05). Laryngoscopy and lumbar punctures were associated with the highest comfort levels; trans-tracheal jet ventilation and brachial arterial catheters the lowest. 93.2% of respondents indicated that EM physicians should provide critical care. However, barriers were common (77.6%), with the most frequent being "balancing ED flow" and "nursing skill at critical care procedures". Conclusions: Canadian EM physicians provide substantial critical care for patients in the ED. Barriers to the provision of critical care in the ED may influence management, and require further delineation to optimize patient outcomes. Key Words: Critical care, Resuscitation, Physician practice

200 A consensus-established set of important indicators of pediatric emergency department performance

Hung GR, Chalut D. Division of Emergency Medicine, Department of Pediatrics, BC Children's Hospital, University of British Columbia, Vancouver, BC (GH); Division of Emergency Medicine, Department of Pediatrics, McGill University, Montréal, QC

Introduction: Quality assurance (QA) is a new and important area of research in pediatric emergency medicine (PEM). There are few studies that describe which performance indicators best represent PEM practice. This primary study objective is to construct a set performance indicators that have been selected by current and former pediatric emergency department (PED) medical directors as most useful in assessing PED performance. The secondary objective is to assess which indicators are measured to assess performance in PEDs. Methods: Current and former directors of accredited Canadian PEM programs were considered eligible participants. A list of indicators was generated by survey (item pool generation); this list was refined by clarifying unclear terms or eliminating redundant and unquantifiable performance indicators (item scaling); PED directors were asked to rate each item of this refined list to indicate which indicators were more useful in assessing PED performance (item prioritization). A ranking formula was used to prioritize those items considered most useful by a larger proportion of respondents. Results: 14 current and former medical directors were considered eligible participants. Indicators related to patient morbidity and mortality, adverse outcomes, return visits, patient length of stay (LOS), and waiting times were considered to be more useful. Less useful indicators included the number of deaths, daily census, number of incident reports, and individual physicians' admission rates. The most commonly measured PED performance indicators included the left without being seen rate, patient LOS, and the waiting time until being seen by a physician by triage category. Conclusions: The top quartile of performance indicators considered most useful by participants included indicators that reflected clinical outcomes, LOS, and waiting times. A dichotomy may exist between those performance indicators that PED directors are considered more useful, and those indicators that are currently measured. Key Words: Quality assurance, Performance indicators, Pediatric

201 CHILDREN admitted to the hospital after returning to the emergency department within 72 hours

Goldman RD, Kapoor A, Mehta S. Pediatric Research in Emergency Therapeutics (PRETx) Program, Division of Pediatric Emergency Medicine, The Hospital for Sick Children, Department of Pediatrics, University of Toronto, Toronto, ON

Introduction: Children returning to the Emergency Department (ED) within 72 hours after their visit may increase overcrowding and healthcare costs. Some of the returning children need admission. Identifying characteristics of these children may help distinguish who might need admission on their first visit. The objective of this study was to comparing characteristics of children who returned to the ED and needed admission to those discharged. Methods: A retrospective chart review of children under age 19 years visiting our tertiary pediatric ED over a one year period. We excluded patients who left without being seen and against medical advice. We determined the rate of return visits, and then performed Chi square and Student's t-test analyses. The main outcome measures were returning to an emergency department and needing admission to the hospital. Results: Of 47,655 eligible children, 2115 (4.4%) had return visits to the ED within 72 hours. The admission rate for the second visit was 353 (16.7%). There was no significant difference in age, gender, language spoken at home or time elapsing from the first visit to the presentation again in the ED between children admitted on the first versus second visit. The acuity was significantly lower among children discharged after returning (p<0.001), but not among those admitted (p<0.220). Conclusions: Four percent of our pediatric ED visits are for children returning within 72 hours. Progression of illness resulting in higher acuity, not age, gender, time from previous visit or change in chief complaint category, was directly associated with admission on the second visit. Key Words: Return visit, Pediatric

202 AGREEMENT of a computerized triage tool using written case scenarios

Dong SL, Bullard MJ, Meurer DP, Akhmetshin E, Holroyd BR, Rowe BH. Department of Emergency Medicine, University of Alberta, Edmonton, AB

Introduction: Emergency department (ED) triage prioritizes patients based on urgency of care. A web-based triage tool (eTRIAGE©) has been developed with Canadian Triage and Acuity Scale templates and studied in an active clinical setting. Some authors also advocate the use of written case scenarios to evaluate triage. This study examined the agreement between nurses using eTRIAGE© with written patient case scenarios, and agreement between these nurses and a consensus standard triage score. Methods: Volunteer experienced triage nurses were recruited for this study. Each nurse was provided 50 written case scenarios and used eTRIAGE© to arrive at a triage score. The cases were representatively balanced between CTAS 1 through 5, and the nurses evaluated triage notes and vital signs but were blinded to identifying characteristics or the other nurses' triage. These scores were compared to each other and to a consensus score using the weighted and unweighted kappa (κ) statistics. Results: Eight triage nurses volunteered for this project; all completed the 50 cases. An experienced emergency nurse and two emergency physicians generated the consensus scores. Agreement was very good among the eight nurses (overall κ = 0.73; 95%CI: 0.69, 0.77). When compared to each other, the agreement was excellent (mean quadratic weighted κ = 0.88; 95%CI: 0.61-1.00). When compared to the consensus scores agreement was excellent (mean quadratic weighted κ = 0.91; 95%CI: 0.77, 1.00).Conclusions: Nurses using eTRIAGE© to evaluate written patient scenarios had good to excellent inter-rate agreement and excellent agreement when compared to a triage score generated by consensus. Further research to create this degree of agreement in the live setting is needed. It is also unclear what amount of training and experience with eTRIAGE© is required to become competent or to maintain competence. Key Words: Triage, Reliability, Computerized

203 DEVELOPMENT of a standardized diagnosis list for use in Canadian emergency departments

Unger B, Afilalo M, Boivin JF, Bullard M, Grafstein E, Schull M, Guttman A, Lang E, Rosenthal S, Vandal A, Xue X, Colacone A, Leger R. McGill University SMBD-Jewish General Hospital, Montréal, QC

Introduction: Monitoring and managing Emergency Departments (EDs) would benefit from reliable data using a standardized ED specific diagnosis classification list (DCL). EDs across the country use a variety of heterogeneous lists. This study's objective was to use a national consensus process to develop a standardized DCL that could be used by ED personnel to prospectively code patient diagnoses (Dx) in Canadian EDs. Methods: A modified Delphi method was employed to develop the DCL from the ICD10-CA, including over 17,000 Dx. Emergency physicians from across Canada were provided with a randomly selected chapters and were asked to rate the importance of including each Dx using a Likert scale. All Dx with a reviewer consensus of importance of 70% or more were retained in round 1. In round 2 of the Delphi, reviewers assessed only the Dx which did not achieve a strong general consensus (50% to 70%) and those that had achieved a consensus only among paediatric emergency based reviewers. The research team reviewed the results for inconsistencies. Results: The Delphi Round 1 was completed by 83 participating MDs. The reviewers averaged 12 years of experience in emergency medicine, with 63% working in a tertiary care hospital and 26% serving as chiefs of departments. Delphi Round 2 was completed by 69% of the reviewers and resulted in the addition of 26 items for a total of 1416 Dx for the detailed DCL. The chapter with the largest number of retained items was injury and poisoning. This was followed by the gastrointestinal, musculoskeletal and infectious disease chapters. Conclusions: We report the creation of a DCL tailored for EDs developed through a consensus mechanism involving 83 ED physicians from across Canada. This should allow tertiary and academic hospitals to prospectively code ED visits without the need for retrospective coding by nosologists. A shorter and less detailed list applicable to most EDs is being finalized. Key Words: Discharge diagnosis, ICD-10

204 ED overcrowding: a comparison of urban vs rural EDs

Rowe BH, Tam SL, Holroyd BR, Bullard M, Yoon P. Department of Emergency Medicine, University of Alberta, Edmonton, AB

Introduction: The prevalence and severity of emergency department (ED) overcrowding in larger urban centers in Canada has recently been reported. Its role in smaller and rural EDs is less clear. The objective of this study was to explore the degree of severe or major overcrowding in rural EDs. Methods: ED Medical Directors at 102 Alberta sites, located in rural or urban communities were surveyed with a 29-item paper-based questionnaire in the fall of 2006. Data relevant to ED administration were obtained from Alberta Health and Wellness registries. Data are compared using chi-square and T-tests, where appropriate. Results: Overall, 83.3% directors from 17 urban and 85 rural sites responded. The median number of patient visits to urban/teaching hospitals was 17,500 (IQR: 13,000, 30,186), significantly (p = <0.00001) higher than 11,488 (IQR: 5756, 21,974) of rural/non-academic hospitals. Factors such as community size (rural/non-academic: <50,000 vs. urban/teaching: >50,000) and access block to admitted beds (26% vs. 65%; p = <0.00001) were similarly different. Among them, 67% from urban and 19% from rural hospitals (p = <0.00001) reported overcrowding as a major or severe problem during the past year. A similar difference was reported by 57% and 19% directors from teaching and non-academic hospitals (p = 0.00255), respectively. While directors from rural areas commented a lack of alternatives to the ED (after-hours clinics, office availability) to be a significant contributing factor to ED overcrowding, urban ED directors attributed the ED overcrowding problem to a lack of admitting beds. Conclusions: In urban and teaching EDs in Alberta, overcrowding is a serious problem; however, it is much less of an issue for EDs in rural areas. Moreover, the perceived contributing factors to the problem were differed between rural and urban sites. Hence, solutions for resolving ED overcrowding in rural and urban areas may require different policies and interventions. Key Words: Overcrowding, Rural

205 BILLING comparisons between fee-for-service and alternative-funding-arrangement emergency physicians in Ontario

Upadhye S, Cleve P. Division of Emergency Medicine, McMaster University, Hamilton, ON

Introduction: This study compared the billing patterns of Emergency Department (ED) physicians working in both fee-for-service (FFS) and alternative-funding-arrangement (AFA) practices. The study hypothesis was that AFA physicians bill less efficiently than do their FFS counterparts. Methods: A series of hypothetical paper case sets were generated with variations in chief complaints, acuity, time of day, management plans and disposition. Case sets were given to two groups (one FFS, one AFA) of EM physicians in an academic teaching center. Participants in each group billed the cases based on the information documented, using the current Ontario Schedule of Benefits codes. Demographic variables for correlation testing included total years in practice, and years elapsed since last working in an FFS environment. Analyses included descriptive statistics of case billings between groups, and frequency of billing errors. Results: A total of 27 cases were completed by 17 FFS physicians, and 37 cases by 27 AFA physicians. In both groups, there was considerable variation in use of billing codes and mean billings per case. AFA physicians were considerably lower in average billings per case compared to FFS physicians, especially in the use of resuscitation vs. standard assessment codes. AFA physicians were also less likely to utilize appropriate specialized procedure codes, and made more mistakes using inappropriate billing codes and premiums. There was a correlation of optimal use of billing codes and time elapsed from practice in an FFS environment. Conclusions: AFA EM physicians are less efficient than their FFS counterparts in optimal use of OHIP billing codes in this paper case exercise. AFA physicians are less likely to bill appropriate codes in resuscitative/procedural situations, and use of premiums. Since the majority of Ontario ED's are under AFA remuneration, these results illustrate potentially serious problems in billing practices and shadow-billing revenues generated.Key Words: Billing, Fee-for-service, Alternate payment

206 PATIENT flow and bed capacity as causative factors in ED overcrowding

Innes G, Mcknight D, Kaloupis P, Johnson M, Tsang S. Providence Health Care and St. Paul's Hospital; Vancouver Coastal Health Region, Vancouver, BC

Introduction: ED crowding, better termed access block, occurs when there are more patients requiring care than available ED stretchers. Access block is usually attributed to inadequate hospital capacity (lack of beds), but patient flow may be equally important. The difference between patient demand (number of patients needing care) and available ED stretcher capacity can be described as capacity deficit. Our objective was to quantify ED capacity deficits as a proxy for hospital bed shortfalls, and to determine the potential importance of flow dynamics vs. bed capacity in 3 urban hospitals. Methods: For a one year period, our regional data support group used arrival and discharge times for 166,000 ED patients to determine the number of patients present in each ED every hour of every day. Hourly ED census was then compared to ED stretcher capacity to determine ED capacity deficit. Results: Hourly ED census patterns and capacity deficits were highly predictable and consistent across all sites. At all sites, by mid morning, patient inflow exceeded outflow, causing ED census and capacity deficit to rise. Capacity deficits peaked in early evening, after which outflow began to exceed inflow, allowing census and capacity deficits to fall. Average capacity deficit, the difference between ED census and available care spaces, ranged from -1 (one extra stretcher) to +15, depending on time of day (table). On average, hospitals had small capacity deficits at night and large deficits during the day. Conclusions: Patient census and capacity deficits are highly predictable across diverse sites and differ by time of day, showing that imbalance between inflow and outflow is an important cause of ED crowding. Small capacity deficits at night suggest that ED crowding may be more related to poor flow than to hospital bed shortages. Strategies aimed at enhancing flow will profoundly mitigate ED overcrowding. Key Words: Access block, Capacity deficit, Overcrowding

Table 1, Abstract 206. Average deficit by hour of day
Site	0600	1200	1800	0000	Best	Worst
LGH	0	12	13	5	0	15
VGH	0	11	14	10	0	15
SPH	-1	10	11	4	-1	12

207 UTILIZATION of computed tomography angiography in the diagnosis of acute pulmonary embolus

Costantino M, Randall G, Vegas C, Gosselin M, Brandt M, Spinning C, Su E. Department of Radiology, Oregon Health and Science University, Portland, OR

Introduction: The purpose of this study was to assess appropriate use of Computed Tomography Angiography (CTA) in the diagnosis of acute pulmonary embolism (PE). Methods: Review of 580 inpatient (45%), ED (41%) and outpatient (14%) CTAs evaluating for PE at a large US teaching hospital from Jan 04 through Mar 05. Based on chart review blinded to final diagnoses, PE pretest probability using Wells criteria was retrospectively assigned. D-dimer values (if obtained) were also reviewed. Results: The overall PE rate was 10%; rates by location were 12% inpatient, 8% ED and 1% outpatient. Only 3 patients (<1%) were high probability; 2 of these had PE (67%). Of the remaining 577, 48% were intermediate and 51% were low probability. In these two groups, the PE rate was 14% and 5%, respectively. D-dimer was only ordered on 39% of patients; 17% were negative (<0.5), 47% intermediate (0.6-2.0) and 36% positive (>2.0). Only 1 patient with a negative D-dimer and 3 patients with intermediate D-dimers had PE but 146 CTAs (25% of the total) were obtained in these two groups. Of the ED patients, 21 had PE (9%). Rates by probability group were 50% high, 15% intermediate and 2% low. D-dimer was drawn on 59% of ED cases; 21% were negative, 54% intermediate and 25% positive. The PE rate was 50% in the positive group and 3% in both the negative and intermediate groups. Conclusions: CTA is fast, diagnostic and widely available for evaluation of acute PE. Wells criteria stratify patients and guide the PE workup. Our data show suboptimal use of Wells criteria and subjective overestimation of PE probability prior to CTA. Negative D-dimer also does not deter unnecessary CTA. This represents a paradigm shift in which clinical tools are supplanted by imaging that, while noninvasive, is not without cost or risk. While no definitive acceptable positivity rate for CTA has been established, we feel 10% represents inappropriate use of CTA as a screening rather than diagnostic test, equating to ineffective resource utilization and unnecessary radiation exposure. Key Words: Computed tomography angiography, Pulmonary embolism, D-dimer

208 PERCEIVED barriers and facilitators to the implementation of the Canadian C-spine rule by emergency department nurses

Clement CM, Stiell IG, Danseco E, Davies B, O'Connor A, Behaut JC, Leclair C, Marcantonio R. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

Introduction: Currently, ED nurses are validating use of Canadian C-Spine Rule (CCR) in 6 Ontario hospitals. We sought to determine the potential barriers and facilitators to actual implementation of the CCR by ED nurses to clinically clear the c-spine of minor trauma patients. Methods: As part of a prospective validation study, we conducted a mail survey using a modified Dillman technique. Included were all ED nurses who had been trained to evaluate the CCR at 4 community and 2 teaching hospital EDs. These nurses had evaluated the CCR on alert and stable trauma patients for 18 months prior to the survey. Questions included practice patterns, application of the CCR, aids to applying the CCR, and the opportunity to list additional barriers and facilitators. We calculated descriptive and univariate analyses as appropriate for the data. Results: The 86 respondents, representing a 46.9% response rate, had these characteristics: female 77.9%, mean years in nursing 20.0, mean years in ED 10.8, community hospital nurses 37.2%, and full time 62.8%. Among the nurses, 90.7% (95% CI 82-95%) responded that they were comfortable applying the CCR and 88.4% (95% CI 79-93%) indicated that the CCR was easy to use. Other responses included:

Useful to my practice 94.2%
Easy to remember 87.2%
Efficient use of my time 95.4%
Not safe for patients 14.0%
Not nurses role to apply CCR 19.8%

Examples of potential facilitators included: "improved overall triaging", "good MD support", "speeds patient removal from boards and discharge", and "good peer support". Barriers included "fear of making a mistake", "perceived lack of time", and "lawsuits". Conclusions: The majority of ED nurses are comfortable applying the rule and feel it is useful to their practice suggesting great potential for nurses to clinically clear the C-spine. Prior to successful widespread implementation, potential barriers will need to be addressed. Key Words: Decision rules, C-Spine, Implementation

209 CLINICAL predictors of high risk transient ischemic attack

Kerr J, Perry JJ. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

Introduction: Patients with transient ischemic attack (TIA) are at increased risk for a stroke. It is not clear which ED patients with TIA require expedited investigations and management. This study assessed the incidence of patients returning to the ED with a TIA/stroke within 6 months of index visit.Methods: We conducted a one-year historical cohort study at a university-affiliated tertiary care ED (census 60,000 visits/year). All patients with weakness, TIA or stroke were screened. We enrolled all patients with an ED diagnosis of TIA. Data was extracted from paper/electronic records to data extraction forms. Clinical findings, medications and tests were recorded. Data was inputted into a database using Statistical Analysis System (SAS) software. Descriptive analyses were conducted for the primary outcome of recurrent TIA/stroke within 6 months of the index ED visit. Results: 211 patients were enrolled in the study. The patients had the following characteristics: mean age 71.2 years (SD 13.8), 56.9% female, 53.1% hypertension, 26.5% ischemic heart disease, 17.1% stroke, 16.6% previous TIA, 15.2% diabetes mellitus, and 6.2% smokers. 41.2% of patients had unilateral arm weakness. 94.3% of patients underwent CT head, of which 16.1% demonstrated an acute/previous infarct. 32.7% of patients were treated with ASA, 23.7% with dipyridamole/ASA, 8.1% with clopidogrel and 6.6 % with warfarin. 34.1% were on a statin lipid-lowering agent and 59.2% were on antihypertensives. TIA was confirmed by neurology in 76.9% of referred patients. Subsequent TIA/stroke within 6 months occurred in 16.1% of patients, of which 47.1% were completed strokes. Conclusions: This study established that a relatively large number of TIA patients have subsequent TIA/stroke within 6 months of initial diagnosis. A clinical decision rule is needed for ED patients with TIA to identify patients at high risk of an impending subsequent event to ensure prompt testing and optimal risk reduction. Key Words: Transient ischemic attack, Stroke, Outcomes

210 OPTIMAL use of diagnostic imaging for ED patients with cough

AlHamdan T, Stiell IG, AlRajhi A. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

Introduction: There are no reliable criteria to determine which ED patients with cough require diagnostic imaging. We sought to identify criteria that would allow ED patients with cough to be exempt from diagnostic imaging. Methods: We conducted a health records review of adult patients presenting with cough to a tertiary care hospital ED during a 12-month period. We abstracted data from the history and physical examination findings and compared to chest x-ray results as determined by both the ED physicians and by staff radiologists. Data analyses included chi-square, Student's t-test, and kappa statistic. Results: Of 528 patients with cough, we enrolled 354 who had a chest x-ray: mean age 52.9 years, male 48.3%, abnormal chest x-ray 45.8%, admission 15.5%, return to ED within 14 days 4.2%. Variables significantly associated (P<0.05) with abnormal x-rays included history of fever, abnormal chest exam, age ≥ 65, male gender, oxygen saturation ≤ 90%, and temperature ≥ 38. Variables not associated with abnormal x-ray were green sputum, bloody sputum, shortness of breath, heart rate. A combination rule of age < 65, temp < 38, oxygen saturation > 90%, and normal chest exam identified abnormal chest x-ray with 86.4% sensitivity and 44.3% specificity and would have decreased the need for imaging in 30.2% of cases. We found an excellent agreement between ED physicians and radiologists (kappa = 0.86) in their interpretation of chest x-rays. Conclusions: We identified predictive factors for an abnormal chest x-ray among patients with acute cough and could limit the use of imaging to patients with age > 65, temperature > 38, oxygen saturation < 90%, and abnormal chest exam. We also found excellent agreement between ED physicians and radiologists in chest x-ray interpretation. This acute cough rule must be supported by a prospective validation study prior to clinical use. Key Words: Decision rule, Cough, Chest x-ray

211 DOES a clinical decision rule guide physician decision-making? a vignette-based study

Brehaut JC, Wigton R, Tape T, Stiell IG. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

Introduction: To examine if ankle injury radiography decisions are governed by the Ottawa Ankle Rules (OAR). Some rules are reportedly widely used, yet physician self-report on issues like this can be unreliable. Our vignette-based study examined whether radiography decisions for realistic ankle injury scenarios are based on clinical findings that are part of the OAR (Rule-based findings) or others (Non-Rule-based findings). We tested two hypotheses: 1) Physicians should put greater weight on Rule-based findings than Non-Rule findings; 2) Findings used in decision making will differ between older and younger subgroups of physicians. Methods: Our postal survey to 240 Canadian emergency physicians included 20 case vignettes of patients with ankle pain. Each vignette described 8 clinical findings known to affect ankle radiography ordering. Respondents estimated the likelihood they would order an ankle x-ray for each case. Using these data, we inferred - for each physician - the weight they placed on each of the 8 findings in making their decision. Two findings were part of the OAR, 3 were predictive of fracture but not in the rule, and 3 were non-predictive. Results: There were 116 responses (48% response rate). 1) Weights for Rule-based findings differed from 0 for 91.6% of respondents; for Non-Rule predictive findings, 12.7%, and for Non-Rule, Non-Predictive findings, 5% (Rule vs. others: p < .001). 2) Younger physicians (<7 years post-graduation) differed from the oldest physicians (19+ years) in that they placed less emphasis on one Rule indicator (p=.013), and greater emphasis on one Non-rule predictive indicator (p=.015). Conclusions: Physicians placed greatest emphasis on the two clinical findings that are part of the OAR. Younger physicians differed from more experienced physicians in their use of specific clinical findings, suggesting areas for targeted training. This methodology is useful for understanding how CDRs fit into the decision making processes of individual physicians. Key Words: Decision rule, Ankle x-ray

212 DELAYED transfer times of injured patients within a trauma system

Svenson JE. University of Wisconsin, Madison, WI

Introduction: Regionalization of trauma services is based on the idea that injured persons at non-tertiary facilities will be stabilized and rapidly transported to a more definitive center. Trauma systems seem to improve outcomes for urban patients, but this same benefit has not been shown for rural patients. There are many factors associated with the decision to transfer injured patients including referral hospital and patient age for example. The purpose of this study is to examine factors that influence the timing of transfer of trauma patients and specifically to determine if establishing specific trauma systems has led to any changes over time. Methods: All trauma patients at the University of Wisconsin between 7/1/99 and 6/30/05 with an ISS>9 who had been transferred to UW after evaluation at an outside hospital were included. The variables considered were age, referring hospital, emergency department time at referring hospital, ISS, the presence of a head injury, performance of a head CT, mode of transport, and the date of ED evaluation. Results: There were 1,656 patients. The average ED time was 153±82 minutes. ED time was significantly shorter for those with ISS scores >25 and for those transported by helicopter. 30% of patients had a head CT performed before transfer, of which 44% were repeated at the trauma center. The average ED time for those in whom a CT was performed was longer than those without (179±81 minutes vs. 142±84). The ED times were slightly longer for level 3 hospitals (158±82 minutes) than for level 4 hospitals (137±74 minutes). ED times were longer for older patients. The times in the ED showed an upward trend. After controlling for all other variables, ED times were not significantly different over the time period studied. Conclusions: Development of a statewide trauma system and outreach education has not affected transfer times from nontrauma centers in our system. Outreach education should focus on systematic trauma evaluation, prompt transfer and limitation of nontherapeutic testing. Key Words: Trauma services, Transfer times

213 CHILDREN are not crash dummies - a surveillance of road traffic injuries in children in Singapore

Tyebally A, Ang A, Wong HB, Chen J. Children's Emergency, KK Women's and Children's Hospital, Singapore

Introduction: To study the patterns of injuries sustained in children involved in road traffic accidents in Singapore. Methods: 522 children aged 0-16 years, who attended the Emergency Departments in the Singapore Health Services network for injuries caused by road traffic accidents from Feb 2002 to Jan 2004 were surveyed as part of the Childhood Injury Surveillance Project. Data on type of road user, mechanism of injury, injuries sustained and injury severity were collected via the use of questionnaire forms, review of in-patient records and coroner's reports. Data was recorded using the International Classification of External Causes of Injury Codes. Results: There were 522 children who sustained injuries from road traffic accidents in Singapore during the study period. Most of the children were pedestrians (46.6%), car passengers (25.2%) and on bicycles (13%). The commonest serious injuries sustained were limb fractures (10.1%), skull fractures (3%) and intracranial haemorrhage (2.9%). There were 13 deaths (2.5%) and 6.1% of the children required admission to the High Dependency or Intensive Care Unit. 61.5% of the deaths were among pedestrians. Pedestrians sustained the most severe injuries with the highest mean Injury Severity Score (ISS) of 7.55 followed by cyclists with a mean ISS of 6.59. In children with an ISS greater than 20, 44.4% had skull fractures and 18.5% had intracranial hemorrhage. At least 36.6% of car passengers did not have any form of safety restraint. Conclusions: Road traffic injuries involving children can be severe and life threatening. The cost of healthy lives lost is immeasurable and it is important to implement injury prevention strategies focusing on road and bicycle safety to reduce unnecessary mortality and morbidity caused by injuries on the roads. Key Words: Trauma, Traffic accidents, Pediatric

214 PREGNANCY testing in female trauma patients: a retrospective chart review

Fortier CJ, von der Porten F. Division of Emergency Medicine, McMaster University, Hamilton, ON

Introduction: In female trauma patients of child bearing age, it is imperative to quickly determine pregnancy status. Trauma is the leading cause of non-obstetrical morbidity and mortality in pregnant females. Identifying a patient as pregnant may have implications for interpretation of vital signs, resuscitative maneuvers, investigation and management of conditions unique to the pregnant female such as placental abruption. The Hamilton Health Sciences (HHS) Trauma program is referred patients from 25 hospitals and keeps a database on patients with an Injury Severity Score greater than 12. This study will determine the number of female trauma patients being tested for pregnancy. Methods: A retrospective chart review of all female trauma patients 16 to 50 years old in the HHS Trauma database for January 1, 2004 to December 31, 2005 was completed. These patients were manually cross referenced with MediTech, the hospital lab system, for urine and serum BHCG tests within 24 hours of triage and during admission. Charts with no BHCG test were examined by 1 reviewer for a BHCG done elsewhere or documented inability to be pregnant. The results were also compared to a similar analysis at the same centre performed in 2004 for January 1, 2002 to December 31, 2003 using a Chi-square test of independence. This study was approved by the HHS Research Ethics Board. Results: In 2004 and 2005, 118 patients met the inclusion criteria. Sixty-eight (58%) were found to have BHCG tests or documented inability to be pregnant within 24 hours of triage. The proportion of patients being tested for pregnancy did not improve from the 2002 to 2003 period compared to the 2004 to 2005 period (49% versus 54%, p = .14). Three (4%) of the 73 patients with BHCG tests during their admission were pregnant. Based on this 4% incidence of pregnancy, 2 to 3 pregnancies may have been missed during this 2 year period. Conclusions: Further efforts are required to improve the detection of pregnant trauma patients to optimize their care. Key Words: Trauma, Pregnancy, bHCG

215 INJURIES sustained in a MVC do not prevent subsequent impaired driving

Brown D, Purssell RA, Brubacher JR, Fang M, Edwin M, Schulzer M, Abu-Laban RB. Division of Emergency Medicine, University of British Columbia, Vancouver, BC

Introduction: Motor vehicle crashes (MVCs) kill over 2,700 Canadians annually. Drivers with an illegal blood alcohol content (BAC) cause 34% of all crashes involving fatalities. The objective of this observational study was to determine the proportion of injured drivers treated in hospital, categorized by BAC, who subsequently engage in impaired driving activity (IDA). Methods: We retrospectively identified all drivers injured in a MVC who presented to our tertiary care, urban Emergency Department (1999-2003) or were registered in our provincial trauma registry (1992-2005) and had a BAC measured. Injured drivers were categorized into three groups (Gp) according to their BAC: Gp 1: BAC=0, Gp 2: 0<BAC<=17.3 Gp 3: BAC > 17.3 (BC legal limit >17.3 mmol/L (80 mg/dl)). IDA was determined from police records (1989-2005), and defined as any of the following: a conviction for impaired driving; a 24 hour or 90 day license suspension for impaired driving; involvement in a MVC where police listed alcohol as a factor; or presentation to a hospital following an MVC with a BAC above the legal limit. Results: 3366 drivers met inclusion criteria: 189 in Gp 1; 310 in Gp 2; and 1074 in Gp 3. IDA following the index hospital visit was identified in 189 drivers in Gp 1 (9.5%, 95%CI 8.3%-10.9%); 70 drivers in Gp 2 (22.6%, 95%CI 18.2%-27.5%, p<0.001 vs group 1); and 344 drivers in Gp 3 (32.0%, 95%CI 29.3%-34.9%, p=0.001 vs Gp 2). Only 8.7% of the drivers in Gp 3 were convicted of impaired driving as a result of their index crash. Many drivers had engaged in IDA prior to their index crash: 15.2% in Gp 1 (95%CI 13.7%-16.9%), 32.9% in Gp 2 (95%CI 27.8%-38.3%); and 54.2% in Gp 3 (95%CI 51.2%-57.2%). Conclusions: A significant proportion of injured drivers treated in hospital following an MVC engage in subsequent IDA. This is particularly true for those with a BAC above the legal limit. These findings support the need for rehabilitation programs and increased legal efforts to target high risk drivers treated in hospital after a MVC. Key Words: Blood alcohol levels, Motor vehicle collisions

216 DO non-helmeted cyclists use more emergency department and hospital resources?

Barbic D, Brison RJ. School of Medicine, Queen's University, Kingston, ON

Introduction: This study of injured cyclists presenting to the emergency department (ED) assessed whether ED and in-hospital resource utilization varied by helmet use. Methods: A retrospective case series of all patients presenting to the ED of Kingston General and Hotel Dieu Hospitals in Kingston, Ontario with bicycle related trauma in the 2002 and 2003 fiscal years. Cases were identified using the Canadian Hospitals Injury Research and Prevention Program (CHIRPP) database. Information on each patient's management within the ED was obtained by linking the CHIRPP data with the National Ambulatory Care Reporting System database. Inpatient data for all inpatient days, critical care/ICU days, ventilation and discharge disposition were obtained from the hospitals' inpatient databases. Results: 885 cycling injury events were treated during the study period. 36.8% reported wearing helmets at the time of injury. Non-helmeted cyclists were more likely to arrive via ambulance than helmeted cyclists (16.1% vs. 9.8%; p=0.005). Non-use of helmets was associated with more severe head injuries (37.0% vs. 19.8%; p=0.014). There was also greater use of ED-based procedures (p<0.001), and specialty consultation (p=0.014) in patients not wearing a helmet. There were no deaths in this sample. 35 persons required hospital admission and this varied little by helmet use. Yet important differences were apparent for in-hospital resource utilization. Length of stay was longer for non-helmeted cyclists (4.0 days vs. 2.5 days; p=0.002). All seven cyclists requiring ICU care received mechanical ventilation, and were non-helmeted. Conclusions: Utilization of ED-based and hospital resources is greater in cyclists who were not wearing helmets at the time of injury. This study suggests that successful injury prevention strategies targeted at increased helmet use would reduce overall costs of medical care for injured cyclists. Key Words: Bicycle injuries, Helmets, ICU