2008 CAEP/ACMU Scientific Abstracts - Poster Presentations: 74-103

2008 Scientific Abstracts

CJEM 2008;10(3):255-295

Abstracts: 1-4, 5-24, 25-43, 44-73, 74-103, 104-131

74 DOES the ability of ED clinicians to speak limited Portuguese enhance communication with the Portuguese-speaking patient? The design and pilot of a translation aid

Han A, Laranjo H, Friedman SM. Department of Family and Community Medicine, University of Toronto, Toronto, ON

Introduction: To design and pilot an instrument to facilitate history taking with Portuguese-speaking patients in the ED. Methods: An instrument was designed to facilitate history taking with Portuguesespeaking patients (PSPs). A pocket-sized document incorporated bilingual, problem-oriented, closed-ended questions for common ED presentations as well as numbers, measurements of time and anatomy. An accompanying audio CD demonstrated correct pronunciation of each phrase. A 3-month pilot was undertaken in a downtown teaching hospital on a convenience sample of PSPs who indicated the need for a translator at triage. A trained Portuguese-speaking observer monitored clinician–patient pairs using the instrument and scored differential patient comprehension in a standardized manner. Qualitative patient and clinician impressions were assessed. A follow-up survey assessed EP impressions of the instrument. Results: Eight of the 9 eligible clinician–patient pairs were enrolled. The average proportions of questions answered appropriately in English and then using the instrument were 16.7% and 85.5% respectively (p = 0.001), with mean improvement of 68.8% (95% CI 45.6–92.1). 87.5% (n = 7) agreed that the instrument had helped in communication. Fifty percent of clinicians (n = 4) indicated that the tool had helped them communicate, and 87.5% (n = 7) indicated that they would use the instrument in the future. Only 12% (2/17) of physicians utilized the audio guide. Suggested modifications included incorporation of phonetics. Conclusion: Pilot of the instrument was well received by patients and resulted in improved communication. Physician adoption of the audio CD was suboptimal and likely limited usefulness of the instrument. Keywords: multiculturalism, language translation, translation aid

75 DOES duration of fever affect the likelihood of a positive urinalysis in febrile children age 3–36 months?

Bin Salleeh H, McGillivray D, Martin M, Patel H. King Khalid University Hospital, King Saud University, Riyadh, Saudi Arabia; Montréal Children’s Hospital, McGill University, Montréal, QC

Introduction: The optimal time to screen for a UTI in children with fever without source (FWS) is unknown. Objective: To determine the proportion of children with a positive bag urinalysis with duration of fever of < 1, 2, 3, 4 and > 5 days. Methods: Prospective cohort study of infants and children 3–36 months of age presenting to a tertiary care emergency department with FWS. Patients with antibiotic use in the past 2 weeks, a history of renal pathology, recurrent UTI, immunocompromised state or proceeding directly to catheterization were excluded. Bag specimens were collected for urinalysis after perineal cleaning. A positive urinalysis was defined as a dipstick positive for leukocyte esterase or nitrites or > 5 white blood cells per high power field on a centrifuged urine. The primary outcome was the proportion of positive urinalyses on each day of fever. We also determined the proportion of positive catheter urine cultures on those infants with a positive bag urinalysis. Results: Eight hundred eighteen infants and children were enrolled in the study. The proportion of children with a positive bag urinalysis if the fever was < 2 days was 14.6% versus 23.2% if the fever was > 3 days, p = 0.002 (RR 1.59, 95% CI 1.2–2.1). Results were not affected by age, sex, white race or circumcision status. Temperatures on each day were similar with a median temperature varying between 39.0°C and 39.3°C. The proportion of positive catheter cultures done on infants with a positive bag urinalysis increased nonsignificantly over time: lowest on day 1 (44%, n = 11/25) and highest on day 4 (62.5%, 10/16). Conclusion: The proportion of positive bag urinalyses increased significantly with longer duration of fever in children 3–36 months of age. In order to increase the usefulness of the screening urinalysis in nontoxic appearing infants consideration of duration of fever (> 3 d of fever) may significantly increase detection of a true UTI. Early testing (< 2 d) may have the potential to miss a significant number of urinary tract infections. Keywords:fever, urinalysis, pediatrics

76 MANAGEMENT of acute exacerbations of chronic obstructive pulmonary disease (COPD) in emergency departments in British Columbia

Stenstrom R, Harris DR, Innes GD, Grafstein E, Holmes A. St Paul’s Hospital, Vancouver, BC Objective: Exacerbation of COPD is a common reason for emergency department (ED) visits in Canada and there is considerable practice variability. The objective of this study was to describe specific performance measures in the management of COPD in 48 EDs across British Columbia. Methods: Twenty-two performance indicators for COPD emergency department care were identified by a panel of experts. Four broad areas were included patient history (e.g., number of COPD-related ED visits in past 2 years), patient physical exam (e.g., oxygen saturation level at triage), process measures (e.g., proportion of patients prescribed oral corticosteroids on discharge) and outcome measures (e.g., referred for COPD education). Fourteen data abstractors were given standardized training and were blinded to study hypotheses. Adherence to 22 performance indicators was measured using standardized data abstraction forms. Random sampling of hospitals (stratified by health region) lead to 1452 patient charts from 48 hospitals being assessed.Results: Of 1452 patient visits audited, the median ED length of stay was 3 hours and 18 minutes (interquartile range [IQR] 1 hour 51 minutes to 5 hours 3 minutes). Oxygen saturation was documented for 97.6% of patient visits (95% CI 96.8%–98.4%) and respiratory rate was documented for 98% of patient visits (95% CI 97.3%–98.7%). When stratified by number of ED visits per year, prescription for oral corticosteroids ranged from 55% to 100% of patient visits (p < 0.01) and for antibiotics in 54% to 86% of patient visits (p < 0.05). Conclusion: There is considerable variability in the care of patients with acute exacerbations of COPD in emergency departments in British Columbia. Various knowledge translation and implementation strategies should be utilized to improve care for COPD patients in the emergency department. Keywords: COPD exacerbations, performance indicators, knowledge translation

Winner of the CAEP Resident Research Abstract Award

77 SAFETY of prehospital nitroglycerin use in suspected ischemic chest pain

Hall C, Welsford M. McMaster University, Hamilton, ON

Introduction: Current guidelines from the American Heart Association recommend that sublingual nitroglycerin (NTG) should not be administered to patients with ischemic chest pain and pulse rates > 100 beats/min. We sought to determine if patients treated according to our regional prehospital ischemic chest pain protocol with pulse rates > 100 beats/min were at greater risk of adverse events than patients with pulse rates ≤ 100 beats/min. Methods: A retrospective analysis of data abstracted from ambulance call records from Jan. 1, 2005, to Apr. 30, 2005, was performed. All patients experiencing suspected ischemic chest pain who were treated with NTG were included in the study. The primary outcome of interest was the difference in systolic blood pressure (SBP) drops experienced by tachycardic (> 100 beats/min) and nontachycardic patients after NTG administration. Secondary outcomes included the rate of significant drops in SBP (> 25% drop from the initial value) and the rate of significant hypotension (SBP < 90 mm Hg). Results: There were 467 patients identified who met the inclusion criteria during the study period; 94 (20%) of these were tachycardic (HR > 100 beats/min). The mean drop in SBP after NTG was 6.9 mm Hg greater in the tachycardic group than in the nontachycardic group (p < 0.01). Tachycardic patients were more likely to experience a significant drop in SBP (RR 2.27 95% CI 1.28–4.01). Tachycardic patients were also more likely to experience significant hypotension, but this relationship failed to reach statistical significance (RR 2.65, 95% CI 0.45–15.60). No occurrences of serious adverse cardiac events (e.g., death, dysrhythmia) were noted in either group. Conclusion:Caution is warranted when administering NTG in a prehospital setting for suspected ischemic chest pain in patients with heart rates > 100 beats/min due to a higher risk of significant drops in systolic blood pressure. Further study is needed to determine if this risk is associated with adverse clinical outcomes. Keywords: prehospital, nitroglycerin, patient safety

78 FREQUENCY and outcomes of syncope in the emergency department

Thiruganasambandamoorthy V, Williams K, Stiell IG, Wells GA. University of Ottawa, Ottawa, ON

Introduction: Syncope is a common presenting complaint in the ED, consumes substantial health care resources and has little evidence to guide optimal care. The aim of this study was to evaluate the frequency, management and outcomes of ED syncope patients.Methods: In this health records review we screened consecutive patient visits to a large tertiary care ED over a 13-month period. We included adults with a primary complaint of syncope (sudden transient loss of consciousness with complete recovery). We excluded patients with ongoing altered mental status, alcohol/illicit drug use, seizure, head and severe trauma. We extracted characteristics, ED management, disposition and serious outcomes. A 30-day follow-up at all local adult hospitals and chief coroner’s office for outcomes was done. We conducted descriptive analyses with 95% CIs.Results: Of the 721 visits screened, 357 patients (361 visits) were included with these characteristics: mean age 58.9 years (range 17–96), 50.9% males, past syncope 41%, heart disease 32.1% and arrival by ambulance 68.4%. 96.1% were investigated (EKG 93.1%, CT head 24.9%) and 13% (95%CI 9.5%–16.5%) were admitted. Most common final diagnoses were syncope NYD 57.6%, vasovagal syncope 25.2% and cardiac syncope 4.9%. Of patients, 4.7% had return ED visits within 30 days. Overall 9.9% had 1 of the composite outcomes, including death 0.5%, myocardial infarction 0.3%, arrhythmia 4.7%, pulmonary embolism 0.6%, major bleeding 1.4%, procedural intervention to treat cause of syncope 5.5% (mostly pacer insertion 3.9%) condition causing/likely to cause return ED visit 0.6% and hospitalization for a related event 0.3%. Of the outcomes, 38.9% occurred after discharge from ED. Conclusion: This is the first study to evaluate ED syncope in Canada. Admission rates were much lower than in US studies. A significant proportion of patients suffer adverse outcomes after discharge and there is a need for an accurate clinical decision rule to guide admission for syncope patients.Keywords: syncope, admission rates, prognosis

79 EXTERNAL validation of the San Francisco Syncope Rule (SFSR) in the Canadian setting

Thiruganasambandamoorthy V, Stiell IG, Wells GA. University of Ottawa, Ottawa, ON

Introduction: Syncope is a common presenting complaint in the ED and poses a large challenge to emergency physicians (EPs). The aim of this study was to evaluate the performance of the SFSR in Canadian setting. Methods: Over an 18-month period, consecutive patient visits to a large tertiary care ED were screened in a health records review. Adults with syncope (sudden transient loss of consciousness with complete recovery) as primary complaint were included. We excluded patients with ongoing altered mental status, alcohol/ illicit drug use, seizure, head and severe trauma. Patient characteristics, SFSR variables and outcomes were extracted. A 30- day follow-up for outcomes was done at all local adult hospitals and coroner’s office. All outcomes were confirmed by another EP. We calculated predictive parameters for the rule with 95% CIs. Results:Of 907 visits screened, 436 patients were included (50.8% males, mean age 58.2 years, range 16–101). Of patients, 62.4% had at least 1 abnormal SFSR variable (congestive heart failure 6.0%, shortness of breath 10.2%, triage systolic BP < 90 4.9%, hematocrit < 0.3 2.7%, abnormal EKG 55.3%); 9.8% sustained adverse outcomes (death 0.4%, myocardial infarction 0.2%, arrhythmia 4.0%, pulmonary embolism 0.4%, significant bleeding 1.3%, procedural intervention to treat etiology of syncope 4.4%, condition causing return ED visit 0.7% and hospitalization for related event 0.7%); 12.7% were admitted; 43.1% had outcomes after ED discharge. Sensitivity and specificity of the SFSR were 90.5% and 40.2% (95% CI 78.5%–96.2% and 39%–40.8%). Missed cases were 1 pneumothorax and 3 cases needing pacers with normal EKG but ED monitor abnormalities. Inclusion of monitor abnormalities in SFSR EKG variable would improve sensitivity to 97.6% (95% CI 88%–99.6%). Conclusion:This is the first Canadian study to evaluate the SFSR and found the sensitivity and specificity lower than previously reported. Further research should attempt to refine the SFSR to improve sensitivity and specificity. Keywords: San Francisco Syncope Rule, validation, clinical prediction guide

80 INTERRATER agreement of CTAS triage scores assigned by base hospital and ED nurses

Poitras J, Dallaire C, Aubin K, Lavoie A, Moore L. CHAU Hôtel- Dieu de Lévis, Université Laval, Québec City, QC

Introduction: Chaudière-Appalaches is the first regional healthcare system in the province of Quebec where a base hospital has been set up. One of its intended uses is to streamline processes and assign an ED stretcher in advance to patients coming in by ambulance. This study aimed at measuring the agreement between the triage score attributed by base hospital nurses and the scoring of ED nurses triaging the same patient on arrival, both using the Canadian Emergency Department Triage and Acuity Scale (CTAS). Methods: A prospective study design was used in our academic urban emergency department with a census of 50 000 visits annually. A CTAS score was assigned by the base hospital nurse from the data transmitted by EMS (n = 100). On arrival, the same patient got triaged again by the ED triage nurse, blinded to the first score. The agreement between base hospital and ED nurses triage was measured using the κ statistic.Results: The weighted agreement between base hospital and ED nurses triage was poor, the κ being 0.3564 (95% CI 0.243–0.470). The score assigned on arrival to the ED was the same in 49% of the cases, lower in 39% and higher in 12%. Conclusion: The low interrater agreement obtained supports our own previous data showing poor agreement between experienced nurses triaging from standardized written scenarios. These results might reflect the dynamic condition of the patient, differences in users or differences in application of CTAS among users. These results might even question the validity of the CTAS scale, which has not been supported by high-quality evidence and they warrant caution before implementing triage by base hospital nurses. Further exploration and research are also warranted to determine if computerized triage can allow better treatment of signs and symptoms of presenting patients and more reliable CTAS scores. Keywords: CTAS, interobserver agreement, computerized triage

81 DROWNING and the influence of hot weather: a case-crossover analysis

Denny CJ, Stefan A, Taha N, Redelmeier DA. Sunnybrook Health Sciences Centre, Ornge Transport Medicine, The University of Toronto, Toronto, ON

Introduction: Drowning is the leading cause of death through injury for recreational activities. Public perception often regards cold, stormy weather as a risk factor. We analyzed how weather was a risk factor for drowning deaths. Methods: We reviewed all drowning deaths in Ontario from 1999 to 2004 (6 years) by linking Coroner records with National Climate Archives. We excluded intentional deaths (e.g., suicide and homicide). A case-crossover analysis used a self-matching design to control for multiple confounders including public safety campaigns, seasonal recreational activities, local alcohol sales and population demographics. The interval period was 1 week prior to the time of death. Indoor drowning deaths were separated from outdoor drowning deaths to serve as a control group for the effect of weather. Results: There were 717 drowning deaths. Linked records were possible for 635 cases, which were then analyzed. The average age was 38.7 years (range 0.5–97 years, SD 22.6). Males accounted for 82% of drowning deaths. Alcohol consumption was considered to be a factor in 45.7% of cases. Personal flotation devices were worn in only 3.9% of cases; in 13.6% of cases they were present but not worn. Cardiac disease was a contributing factor in 7.9% of cases, epilepsy in 6.8% of cases. Of drowning deaths, 87.8% occurred outdoors. The risk of drowning death tripled on days when the maximum temperature exceeded 30°C (86°F) (OR 3.04, 95%CI 1.48–6.24, p = 0.004). For indoor drowning deaths, there was no significant association with hot weather (p = 0.157).Conclusion: Hot weather is associated with three times the risk of death by drowning outdoors. Many drowning deaths are associated with alcohol consumption and lack of protective equipment. An awareness of these findings might promote future injury prevention strategies. Keywords: drowning, environmental emergencies, injury prevention

82 BACTERIAL infection and drug resistant patterns in sputum of ambulatory COPD patients

Lee RS, Borgundvaag B, Campbell SG, Worster A, Froh J, Emond M, Willis V, Rowe BH for the ACE Investigators. University of Alberta, Edmonton, AB Objective: There are debates regarding the utility of sputum samples for patients with COPD exacerbations. This study examined the sputum cultures collected from COPD patients enrolled in a clinical trial and describes the microbiological profile of these samples, the sensitivities and the effect of culture on outcomes. Methods: A sputum collection protocol was followed by research staff in 159 enrolled subjects with acute COPD who were discharged from the ED. All patients received corticosteroids and antibiotics at discharge. Sputum cultures were sent for microscopic examination including gram stain and routine bacterial culture with susceptibility testing. Data were extracted from sputum collection forms completed by dedicated research nurses at each site and laboratory reports using a standardized data extraction form. The success and culture results were compared to clinical characteristics and outcomes. A survey on sputum handling was also administered to study sites across Canada.Results: From 156 COPD patients, 60 patients (38.5%) expectorated a sample that could be submitted for analyses. Subjects did not expectorate primarily because of failure to generate sufficient sputum or refusal. From the 60 patients whose samples were processed, 44 patients (28.2% of total sample; 73.3% of those who expectorated) produced a sufficiently high-quality sputum sample to be sent for culture. Of the 44 adequate sputum samples cultured, 25 samples (56.8%) isolated mixed oropharyngeal flora and 19 samples (43.2%) grew identifiable bacterial pathogens. From the susceptibility test, only 1 sample of the 19 identified a resistant pathogen and these results neither influenced management nor impacted outcomes. Survey results from enrolling sites suggest that sputum collection across Canada is variably performed and analyses of samples are not standardized.Conclusion: This study confirms that sputum collection should not be routinely performed on outpatients with exacerbations of COPD. Keywords: COPD, acute exacerbations of chronic bronchitis, bacteriology

83 CHARACTERISTICS associated with admission and longer length of stay due to painful vaso-occlusive crisis in children with sickle cell disease

Rogovik AL, Li Y, Friedman JN, Goldman RD. Pediatric Research in Emergency Therapeutics (PRETx) Program, Division of Pediatric Emergency Medicine, BC Children’s Hospital, and Division of General Pediatrics, The Hospital for Sick Children, Toronto, ON

Background: Sickle cell disease (SCD), characterized by chronic hemolysis and vaso-occlusive crises (VOC) carries a significant risk of morbidity and mortality in children. The objective of the study was to identify demographic, clinical and laboratory characteristics associated with admission and a longer length of stay (LOS) due to VOC in children with sickle cell disease. Methods: Retrospective chart review at a large tertiary pediatric center. All patients under 18 years with VOC due to SCD presenting to the emergency department (ED) were included. We performed univariate and multivariate regression analyses to predict characteristics associated with admission and LOS 4 days. Results: A total of 428 visits for VOC were documented in 2005/06. Children with fever, higher systolic blood pressure, higher leukocyte and monocyte count, and combined extremity and nonextremity location of pain were more likely to be admitted; homozygous genotype, fever, higher mean cell hemoglobin, increased leucocytes, platelets and percentage of reticulocytes were associated with a longer LOS for admitted patients. Higher pain score at triage (p < 0.001), older age (p = 0.04) and increased systolic blood pressure (p = 0.02) were predictors of admission in a multivariate regression analysis. Higher pain score at triage (p = 0.046), older age (p = 0.002), increased polymorphonuclear count (p = 0.02) and homozygous SCD type (p = 0.03) were associated with prolonged hospital LOS in a multivariate analysis.Conclusion: A higher pain score at triage and older age predict both admission and a longer LOS for children with painful VOC. Furthermore, increased systolic blood pressure is associated with admission and increased polymorphonuclear count and homozygous SCD type predict longer LOS. These parameters will help healthcare providers predict and plan admission and management of these children. Keywords: sickle cell disease, vaso-occlusive crises, pediatrics

84 DOES gender influence the risk for subsequent events following an acute ischemic stroke?

Yassa AS, Bellolio M, Hoff A, Gilmore R, Vaidyanathan L, Decker W, Stead L. Mayo Clinic College of Medicine, Rochester, MN

Introduction: To assess whether gender influences the likelihood of a subsequent vascular events acute ischemic stroke (AIS), transient ischemic attack (TIA) and myocardial infarction (MI) in the year following an acute ischemic stroke. Methods: The study was conducted in the emergency department (ED) of a tertiary academic medical center with an annual ED census of 79 000. The study population consisted of 1038 consecutive consenting patients who presented to the ED or were admitted directly to the stroke inpatient service between December 2001 and December 2005. Patients were contacted via telephone at 3, 6 and 12 months to capture all subsequent events that may have occurred following their first admission for stroke. Follow-up was updated using the date of the last service or dismissal available from registration databases. This data was collected in our institutions prospective observational acute brain ischemia registry. Results: There were 547 men (52.7%). The mean age for women was 74.4 (SD 15.0) years and 70.2 ( SD 14.2) years for men. A total of 83 subsequent events were seen: 38 strokes, 20 TIAs, 25 MIs. Using a logistic regression model after adjusting for age, there was no statistical difference (p = 0.168) between men and women in the frequency of and time to subsequent event after stroke; the subsequent events occur in a similar median time from the initial stroke presentation in both genders. In the univariate analysis we found that age was a predictor of subsequent MI in women but not in men. The mean age of women having subsequent MI was 82.7 (7.7) years, versus 74 (SD 15.5) years in those not having MI (p = 0.020). Age was not a predictor of subsequent TIA or stroke after stratification by gender.Conclusion: Both women and men have an equal likelihood of having a subsequent event following a stroke; this is despite women on average being older than men at the time of their initial presentation. Gender does not play a protective role in subsequent event following stroke. Keywords: acute ischemic stroke, transient ischemic attack, gender imbalances

85 INTERRATER agreement of CTAS triage scores assigned by experienced nurses

Poitras J, Dallaire C, Aubin K, Lavoie A. CHAU Hôtel-Dieu de Lévis, Université Laval, Québec City, QC

Introduction: The reproducibility of Canadian Emergency Department Triage and Acuity Scale (CTAS) scores has only been studied in newly trained nurses. At our institution, triage has been performed according to CTAS standards for the last 10 years by all regular nursing staff. Scores are being attributed according to initial training to the use of CTAS and to nurses’ experience and judgment. We aimed at measuring interrater agreement between experienced nurses triaging written scenarios based on real ED cases.Methods: A prospective study design was used in our academic urban emergency department (ED) with a census of 50 000 visits. One hundred patients were selected and triage data collected using ED software and medical records. Five experienced nurses were recruited to blindly assign a triage level to 100 ED written case summaries presenting, in a standardized fashion, initial complaint, vital signs, past medical history and current illness. The agreement among nurses was measured using the κ and AC2 statistics.Results: The κ was 0.16 (95% CI 0.12–0.21) and the AC2 was 0.20 (95% CI 0.03–0.36). The agreement between experienced nurses was very poor. The agreement for any specific score or any subgroup of nurses remained poor and a retest done 6 weeks later with two nurses gave similar results. Conclusion: A low interrater agreement has important implications for ED triage. It suggests that CTAS scores vary greatly among experienced nurses, even when triaging from written standardized scenarios. It suggests also that triage on a daily basis might rely on other factors than the information provided in scenarios, like some more subjective patient characteristics or even actual ED crowding. Further research is required to ascertain if triage scoring would be more consistent if signs and symptoms were analyzed with the help of computerized systems.Keywords: CTAS, triage, interrater reliability

86 OUTCOMES of a regional implementation of a multidisciplinary asthma protocol in emergency departments in British Columbia

Grunfeld A, Yoshitomi K. Fraser Health Authority, British Columbia

Introduction: Considerable variation exists in management of patients with acute asthma in the emergency department (ED). The British Columbia Provincial Emergency Services Project has introduced a province-wide Acute Asthma Management Protocol in early 2006. The protocol, with algorithms for pediatric and adult patients, adjusted for acuity, allowed nurses or respiratory therapists to initiate bronchodilator treatment at triage, continue bronchodilator treatments and administer corticosteroids to patients before being seen by a physician. The objective of the study was to measure the impact of this protocol in the 12 EDs in the Fraser Health Authority (FHA) of British Columbia. Methods: We reviewed consecutive medical records of patients over the age of 2 years presenting with asthma in EDs in the FHA just prior to and 2 months following the implementation. Results: Medical records of 433 patients seen before and 493 patients seen after the implementation were reviewed. Use of an asthma protocol increased in the post period (4% v. 22%, p = 0.005). The average time to first bronchodilator decreased post implementation (66 min v. 46 min, p = 0.04); however, peak flow measurements (43% v. 46%, p = 0.56), mean length of stay in the ED (2 h 36 min v. 2 h 25 min, p = 0.79), use of oral corticosteroids in the ED (35% v. 37%, p = 0.41) and prescription of oral (35% v. 41%, p = 0.2) and inhaled corticosteroids (35% v. 27%, p = 0.12) at discharge remained similar pre and post. ED visits within 7 days following the index visit were 8% and hospitalization within 7 days following the index visit was 2% in the postimplementation group.Conclusion: The study provides evidence that a regional multidisciplinary protocol for evaluation and treatment of patients with acute asthma significantly decreased the time to initiation of treatment with inhaled bronchodilators. However, the use of corticosteroids in the ED and prescription of them at discharge did not change. More research is required to evaluate strategies to improve asthma care by physicians. Keywords: asthma, knowledge translation, care pathways

Winner of the CAEP Resident Research Abstract Award

87 TELEHEALTH Ontario detection of gastrointestinal illness: an early warning system for bioterrorism

Caudle JM, van Dijk A, Rolland E, Moore KM. Queen’s University, Kingston, ON

Introduction: Prompt detection of bioterrorism and rapid introduction of mitigation strategies is a primary concern for public health, emergency and security management organizations. Traditional surveillance methods rely on astute clinical detection and reporting of disease or laboratory confirmation of a pathogen. Although effective these methods are slow, dependent on physician compliance and delay timely effective intervention. To address these issues, syndromic surveillance programs have been integrated into the health care system at the earliest points of access; in Ontario these points are primary care providers, emergency departments and Telehealth Ontario. This study explores the role of Telehealth Ontario, a telephone helpline, as an early warning system for detection of gastrointestinal illness. Methods: Retrospective time-series analysis of the National Ambulatory Care Reporting System (NACRS) emergency department discharge and Telehealth Ontario data for gastrointestinal illness from June 1, 2004, to Mar. 31, 2006. Results: Telehealth Ontario recorded 184 904 calls and the NACRS registered 34 499 emergency department visits for gastrointestinal illness. The Spearman rank correlation coefficient was calculated to be 0.90 (p < 0.001). Time-series analysis resulted in significant correlation at lag (weekly) 0 indicating that increases in Telehealth Ontario call volume correlate with increases in NACRS data for gastrointestinal illness. Conclusion:Telehealth Ontario call volume fluctuation reflects directly on emergency department gastrointestinal visit data on a provincial basis. Telehealth Ontario gastrointestinal call complaints are a timely, novel and representative data stream that show promise for integration into a real-time syndromic surveillance system for detection of bioterrorism events. Keywords: Telehealth, syndromic surveillance, bioterrorism

88 FACILITATION of the management of suspected deep vein thrombosis (DVT) by emergency department based paramedics

Campbell SG, MacKinley RP, Froese P, MacDonald MA, Carr B, Anderson DR, Cairns SL. Dalhousie University, Halifax, NS

Introduction: The management of patients referred to the ED with suspicion of DVT can be conducted according to an evidencebased algorithm, with each step in the process dictated by the result of the previous step. We describe a process whereby DVT investigation and initial treatment is facilitated by an advanced care paramedic (ACP) in the ED without the patient needing to wait for an ED bed. Methods: Patients referred to the ED by their family physician (FP) with suspected DVT were met in the waiting room by an ACP, who calculated their clinical probability of DVT, using Wells criterion, and ordered tests and treatment according to a standard algorithm. Patients and the care plan were reviewed by an emergency physician (EP) before discharge. Consented patients were followed up at 3 months. Satisfaction of ACPs, FPs and EPs was evaluated. Results: ED length of stay for patients unlikely to have DVT according to Wells score decreased by 94.57 min (from 379.31 min before pathway introduction, n = 68, to 284.74 min after introduction, n = 73). There was no change in the percentage of ultrasounds positive for DVT (14.0% v. 13.9%). Of 55 consented patients, 38 were contacted. Of these, anticoagulation to treat DVT had been prescribed in 9 (24%). Regarding the efficiency of the process, 30 (79%) felt that it was efficient, 4 inefficient and 6 no opinion. Regarding their satisfaction with the process, 36 (95%) reported being very satisfied or satisfied. A telephone survey of 30 FPs showed that 70% had used the process. The average satisfaction rating on a 10-point scale was 8.99 (range 8.75–10). Of 23 EPs surveyed, 22 (96%) had used the pathway, 21/22 (95%) felt that the process had improved patient care and was worthwhile. Of 9 ACPs, 7 (78%) felt that the process improved patient care and 8 (89%) that it was worthwhile. Conclusion: DVT investigation and management can effectively be conducted by ACPs resulting in decreased length of stay and satisfaction of patients and the approval of FPs EPs and ACPs. Keywords: DVT, prehospital, clinical prediction guide

89 IMPACT of a homeless curriculum on the attitudes of medical trainees towards homeless people who present for care in the emergency department

Spence JM, Bandiera GW, Taback N, Balu-Kakkar P, Hwang S. St. Michael’s Hospital, University of Toronto, Toronto, ON

Background: Many homeless people (HP) use emergency departments (EDs) for care owing to barriers to access. Negative attitudes of health care workers have been identified as a major theme when HP are asked about difficult aspects of being homeless. Exposure alone to homeless patients may negatively impact attitudes of medical trainees (MTs). Objective: To describe the attitudes of MTs towards HPs pre- and postcompletion of an ED rotation before (Study Yr1) and after (Study Yr2) the inclusion of a formal curriculum on homelessness. Methods: The study was conducted in an urban Canadian ED (55 000 visits/yr; 15% HP). MTs were surveyed using a validated 11-question survey about attitudes towards homelessness. Included were 4 additional questions regarding comfort level providing care. Results: Overall, 228 MTs responded: 113 in Study Yr1 and 115 in Study Yr2. Mean age was 27 years (p = 0.2) and 40% were male. Prerotation data (Yr1 & Yr2): most were comfortable with HP — 96% working, 86% socially. Cutbacks in housing assistance, low minimum wage and welfare were identified as social causes by 91%, 50% and 74%, respectively. Substance abuse was identified with homelessness by 68%. Seventy-seven percent were comfortable dealing with homeless mental health patients. However, 55% felt overwhelmed by problem complexity. Overall scores did not change before or after curriculum implementation. Study Yr1 differences were found for comfort meeting HP (67% v. 74%; p = 0.05), and the impact of low wages on homelessness (49% v. 39%; p = 0.01). Study YR2 differences were found for impact on low wages (63% v. 77%; p = 0.01), little can be done (89% v. 97%; p = 0.01), ability for normal lifestyle (84% v. 93%; p = 0.02) and comfort providing care/mental health care (76% v. 84%; p = 0.03) Conclusion: MTs were comfortable with HP. Exposure to HP alone during an ED rotation did not impact overall scores. Addition of a curriculum improved understanding of causes and solutions for homelessness.Keywords: homelessness, EM curriculum, survey

90 NURSE practitioners in Canadian pediatric emergency departments: current roles and physician attitudes towards those roles

Pitters C, Plint AC, English S, Tetzlaff J, Clifford T. Department of Pediatrics, University of Ottawa, Ottawa, ON

Introduction: The Canadian Association of Emergency Physicians and the National Emergency Nurses Affiliation suggested the use of nurse practitioners (NPs) as a strategy to address emergency department (ED) overcrowding. The objectives of this study were to determine the utilization of NPs among Canadian PEDs and to examine pediatric emergency medicine (PEM) physicians’ views in NPs assuming traditional physician duties. Methods: Self-administered, mailed survey of PEM physicians at 10 of the 12 Canadian PEDs. The 10 PEDs are members of a collaborative research group. Directors were surveyed regarding PED demographics. PEM physicians were surveyed on attitudes towards NPs assuming traditional physician duties. Results: There were 164 physicians surveyed; 68% of general surveys and 90% of director surveys were returned. Only 1 PED currently employs NPs and they manage minor illness and follow up lab and x-ray reports. Overall, 66% of physicians saw barriers to NPs. This belief varied by site (p < 0.001) and was associated with whether NPs were working at the respondent’s institution (p < 0.001) but not with length of practice (p = 0.95) or way of physician remuneration (p = 0.34). Among respondents who saw barriers, 67% reported financial, 40% credentialing and 26% hospital policy barriers. The majority of physicians (76%) were comfortable with NPs managing nonurgent patients without supervision but a small minority (6%) were not comfortable with NP managing any class of patient with or without supervision. Overall, physicians were comfortable with NPs ordering investigations such as EKG and x-rays without review but were not comfortable with them ordering more complex/invasive investigations without review. The 3 top procedures that physicians were comfortable with NPs performing were suture removal (91%), splinting (64%) and use of tissue adhesive (55%). Conclusion: NPs are not utilized in most Canadian PEDs despite physician comfort with NPs managing nonurgent patients, ordering some investigations and performing several procedures.Keywords: nurse practitioners, pediatrics, survey

91 DERIVATION of a prediction model for diagnosing acute heart failure based on the application of Bayesian Theorem to multiple NT-ProBNP cut points

Steinhart B, Mariano Z, Mazer D, Thorpe K, Moe G. Saint Michael’s Hospital, Toronto, ON

Introduction: Acute heart failure (aHF) is often challenging to diagnose. N-terminal prohormone brain-type natriuretic peptide (NTProBNP) is a biomarker of modest specificity for aHF which national guidelines suggest can be clinically helpful. Bayesian Theorem is a statistical concept that correlates posttest probability with pretest probability and likelihood ratios (LRs). This study analyzed data from a blinded clinical trial to assess the performance of NT-ProBNP at multiple cut points applying Bayesian Theorem, using this to derive a prediction model for diagnosing aHF. Methods:A prospective study of 485 patients presenting to one of 7 Canadian emergency departments with undifferentiated shortness of breath were enrolled. The emergency physician estimated the probability of aHF (1%–99%) without knowledge of the drawn NT-ProBNP value; blinded adjudication for aHF was subsequently determined. The performance characteristics of NT-ProBNP were analyzed by applying Bayesian Theorem to its ranges of less than 300 ng/L to greater than 8000 ng/L after calculating LRs. Multiple logistic regression analysis of clinical variables revealed significant correlates for aHF; these were incorporated into a post test probability model for aHF by expressing the regression model on the probability scale by the inverse logit transformation. Bootsrap analysis and calibration were performed for internal validation. Results: The LRs for aHF with NT-ProBNP ranged from 0.11 (95% CI.06–0.19) for values < 300 ng/L to 13.03 (95% CI 5.31–32.01) for values > 8000 ng/L. Significant variables for aHF were age, pretest probability and the logNT-ProBNP value. A posttest probability prediction model was then created with a c statistic of 0.905, being validated internally.Conclusion: A model using NT-ProBNP, age and pretest probability has been derived which appears to be useful in improving the clinical diagnosis of aHF in the ED. External validation of the model will help to establish its clinical utility. Keywords:NT-ProBNP, heart failure, diagnosis

92 METHODOLOGIC quality of American College of Emergency Physicians Clinical Practice Guidelines for emergency medicine

Upadhye S, Ghuman J, Rauchwerger D, Fan J, Kapur A. McMaster University, Hamilton, ON

Introduction: Clinical practice guidelines (CPGs) represent a higher order of clinical evidence, combining literature reviews and expert stakeholder assessments to formulate recommendations for practicing clinicians treating numerous conditions. It is important that CPGs meet high methodologic standards in order to provide reliable information to consumers. This study examined the methodological quality of the American College of Emergency Physicians (ACEP) CPGs using an accepted critical appraisal instrument.Methods: Currently active CPGs (n = 20) were obtained from the ACEP website list and reviewed using the previously validated Appraisal of Guidelines for Research and Evaluation (AGREE) instrument. CPGs were scored using multiple AGREE domains: scope and purpose, stakeholder involvement, creation methodology, clarity of presentation, applicability, bias and overall assessment. CPGs were reviewed in a chronologically random fashion by 4 blinded reviewers, and aggregate scores descriptively summarized. Aggregate methodological scores for domains were ranked as strong, moderate, weak or undetermined. Results: CPG scores were summarized as follows: strong for scope and purpose, creation methodology and presentation clarity domains, moderate for stakeholder involvement and bias domains, and weak for applicability domains. Overall assessments also yielded mixed recommendations, although there was a trend to favorable endorsement with more recent CPGs. Recurrent flaws noted were the lack of patient involvement in creation, lack of pilot testing, identifying organizational barriers and costs to use, supplying monitoring and audit tools, and potential conflicts of interest. Conclusion: ACEP CPGs are methodologically strong in domains of scope and purpose, creation methodology and clarity of presentation. They are weaker in areas of stakeholder involvement, bias and application. CPG creators should consider strengthening these particular domains, so as to make their use more generalizable. Keywords:AGREE, ACEP, clinical practice guidelines

93 THE impact of emergency department targeted ultrasound on management of patients with early pregnancy complaints

Tang JC, McLeod SL, Skoretz TG. London Health Sciences Centre, Schulich School of Medicine & Dentistry, The University of Western Ontario, London, ON

Introduction: Emergency department (ED) targeted ultrasound (EDTU) is a useful clinical tool in the management of the patient with early pregnancy bleeding or pain. EDTU has recently been introduced at our institution and approximately half the emergency physicians in our group currently use EDTU in the management of patients with early pregnancy complaints. The objective of this study was to describe the impact of EDTU on patient outcomes and ED length of stay (LOS). Methods: A convenience sample of 123 (n = 60 EDTU, n = 63 non-EDTU) adult female patients (age > 17 yr) less than 20 weeks pregnant complaining of pelvic/abdominal pain, vaginal bleeding, syncope or shock were included in this prospective, cohort study. Data was gathered from 2 tertiary care EDs with a combined annual census of 95 000 during a 3-month period (July–September 2007). Results: Mean age was 28.3 (SD 6.2) years old and average LOS was 4.1 (SD 2.2) hours. LOS times were significantly less (p = 0.05) in the EDTU group (3.7 hr) compared to the non-EDTU group (4.5 hr). Referral rates for obstetrical follow-up care were also similar in the EDTU (51.7%) and the non-EDTU (60.3%) groups. There was no difference in the proportion of patients sent for formal radiologic ultrasound in the EDTU and non-EDTU groups (53.3% v. 66.7%, respectively).Conclusion: ED LOS was reduced when EDTU was used to manage patients with early pregnancy complaints. A trend to lower rates in both obstetrical follow-up and formal radiologic ultrasound was noted, but a significant difference was not detected between the EDTU and non-EDTU groups. EDTU is a valuable clinical tool that may impact patient outcome and improve patient flow in the ED. Keywords: EDTU, first trimester emergencies, length of stay

94 THE effectiveness of an innovative EM curriculum for offservice junior residents

Frank JR, Nath A, Lee AC, Elder B. University of Ottawa, Ottawa, ON Objectives: Hundreds of non-EM residents rotate through teaching EDs each year, but their experience has been shown to be variable. We evaluated the effectiveness of an innovative EM curriculum for off-service junior residents. Methods: Using data from a previous systematic needs assessment of non-EM residents and their program directors (Wolpert 2007), we implemented a new PGY1 curriculum that included: 1) decreased shift length from 10 to 9 hours; 2) EM skills day involving simulation and ED orientation; 3) ACLS, ATLS, suturing, central line and airway workshops; and 4) an updated tutorial series. We then conducted a cross-sectional online survey using Dillman methods of all junior non-EM residents in the first 10 EM rotations at 2 teaching hospitals in 2007. We obtained ethics approval and used descriptive statistics and t tests to analyze the data versus those obtained by Wolpert et al the year before (n = 86). Results: There were 100 eligible residents and 61% completed the survey. Respondents were 53% male and 97% in PGY1. Their specialties were 47% family medicine, 22% medicine, 6% surgical and 22% other. On a 5-point Likert scale residents in the new curriculum reported greater confidence with the skills of suturing (3.7 v. 3.0, p < 0.001), casting (2.8 v. 2.1, p < 0.001), central line insertion (2.0 v. 1.8, p = 0.35), defibrillation (2.9 v. 2.3, p < 0.001), intubation (2.4 v. 2.2, p = 0.180) and trauma management (2.7 v. 2.1, p = 0.002). This cohort was more satisfied with shift length (3.5 v. 2.3, p < 0.001) and orientation to the ED (4.2 v. 3.8, p < 0.05). Ratings of shift number, teaching, supervision, exposure to procedures, trauma, and resuscitation were all similar before and after the new curriculum. All seminars were positively rated (mean scores 3.6–4.0/5.0). Conclusion: This innovative EM rotation for off-service junior residents effectively addressed the needs of PGY1s and their program directors. Others should consider adopting these features to enhance education and learner satisfaction.Keywords: EM curriculum, graduate medical education, offservice residents

95 IMPACT of a province-wide head injury guideline on outcomes and process of care

Arbeau RP, Tallon JM, Clarke DB. Dalhousie Medical School, Dalhousie University, Halifax, NS

Introduction: A head injury (HI) management guideline (HIG) including specific treatment, imaging and transport recommendations for moderate-to-severe head injury was implemented on a provincewide basis in Nova Scotia (NS) on July, 1 2002. The guideline includes suggestions for intubation, spine immobilization, 2-minute neurological assessment and accessing the provincial trauma hotline for potential transport. The HIG was implemented owing to the regionalization of tertiary trauma services with a single provincial neurosurgical/ adult tertiary trauma centre (TTC). It was hypothesized this guideline would significantly decrease the time from injury and arrival at first hospital to reaching the TTC and that improved outcomes would follow. Methods: All major HI patients (GCS < 12) (> 15 y/o) in NS with at least 1 intermediate hospital before TTC transfer, from July 2001 until December 2003 were included in this study and categorized into 3 time periods: pre-HI guidelines implementation (July 1, 2001–June 30, 2002), implementation period/maturation phase (July 1, 2002–Dec. 31, 2002) and post-implementation (Jan. 1, 2003–Dec. 31, 2003). Time to TTC (TTTC) was measured and descriptive statistics were generated with 95% CIs. Results:There were 214 patients that met inclusion criteria: 71 preimplementation, 44 implementation and 69 postimplementation major HI patients. Overall, there was a 41 minute (p < 0.05) decrease in TTTC from the pre- to postimplementation periods. Final outcomes analysis using mortality and Glasgow Outcomes Scores are pending at this time. Conclusion: We have demonstrated a statistical decrease in TTC for moderate-to-severe HIs for the province of NS following implementation of a specific HIG. This process of care effect was demonstrable in the major HI category where rapid transport within a regionalized trauma system with a single tertiary care neurosurgical site could be of considerable importance to patient care and outcomes. Final analysis on patient outcomes is pending. Keywords:head injury, clinical practice guideline, implementation research

96 ARE triage criteria for prioritizing ICU admission feasible and useful during an influenza pandemic?

Gray SH, Stone A, Toma A. St. Michael’s Hospital, Toronto, ON

Introduction: Future pandemics will likely overwhelm the resources available in emergency departments (EDs), particularly for those patients requiring intensive care. Triage guidelines were developed by expert consensus in Ontario (2006) to determine who would benefit most from intensive care unit (ICU) admission during a pandemic. However, these criteria have not been tested clinically. During a large mock pandemic exercise, we piloted the use of these guidelines to assess their feasibility and ease of use during a simulated crisis. Methods: The triage criteria were used during a mock pandemic exercise involving over 100 simulated patients and including both a full-scale ED and ICU. Use of the guidelines during the exercise was optional, at the discretion of the multidisciplinary team. All health care workers (HCWs) participating in the exercise were asked to complete a voluntary pre- and post-survey regarding the triage criteria. Results: Sixty-five HCWs completed at least 1 of the surveys. None of the participants had used these guidelines previously, and most (99%) could not name all of the triage criteria prior to the exercise itself, despite the information being precirculated electronically. Of respondents, 93% thought that triage criteria to guide ICU admission would be useful during a pandemic. During the exercise, 90% of the respondents chose to use the criteria to triage ICU patients. Of the HCWs, 79% reported that they would use the criteria during a real pandemic to guide patient care. Conclusion:During a large pandemic exercise, most HCWs chose to use triage guidelines to assess patients for appropriateness of ICU admission and ongoing care. Despite very little knowledge of the criteria before the exercise, physicians and nurses were able to rapidly learn and apply the criteria. Further training could maximize the efficacy of these guidelines and ensure that they were applied fairly and consistently.Keywords: influenza pandemic, ICU, disaster plan

97 WHAT are the resources emergency physicians use to maintain clinical competence? A pilot study

Sherbino J, Upadhye S, Worster A. McMaster University, Hamilton, ON

Introduction: This study sought to determine the resources that practising emergency physicians utilize to maintain clinical competence.Methods: After obtaining institutional Research Ethics Board approval and informed consent, an electronic survey was distributed to all practising emergency physicians at 2 university academic centres using a modified Dillman method. Results: Of 54 emergency physicians, 37 (68.5%) responded. All were board certified (59.5% CCFP[EM]). Only 38.9% attended grand rounds 3 or more times per year. The top reasons for nonattendance were conflicting clinical responsibilities (80.6%), conflicting personal/family responsibilities (38.9%) and conflicting nonclinical professional responsibilities (25.0%). Of respondents, 89.2% read journals at least monthly; 61.1% regularly read review synopses of original articles with 63.9% receiving these updates by email; 61.1% used websites to receive updates on clinical practice; 31.4% subscribed to an audio journal or podcast that provided review lectures. Although 91.7% attended a medical conference at least once per year, only 34.3% regularly attended the original research tracks. Conclusion: This survey reveals 3 themes about the resources used to maintain clinical competence by practising emergency physicians. First, traditional group learning is used by a minority with clinical/professional and family scheduling conflicts preventing attendance. Second, most access the medical literature to maintain competence, with the majority using electronic media. Finally, only a minority of respondents is routinely accessing original research. Most practising emergency physicians are relying on review synopses of original findings in order to maintain clinical competence. Further research is required to determine the generalizability of these findings. Keywords: continuing professional development, clinical competence, survey

Winner of the CAEP Resident Research Abstract Award

98 EFFECTS of an emergency department sepsis protocol on time to antibiotics in severe sepsis

Francis M, Rich T, Peterson D. University of Calgary, Calgary, AB

Introduction: Our objective was to evaluate the time to antibiotics in our emergency department (ED) in patients meeting criteria for severe sepsis pre- and postimplementation of an ED sepsis protocol. Compliance with published guidelines for time to antibiotics and initial empiric therapy in sepsis was also assessed. Methods: A retrospective chart review of ED encounters with ICD codes related to severe infections were screened over a 3-month period pre- and postimplementation of a sepsis protocol. Encounters meeting criteria for severe sepsis were further assessed. The time to initiation of antibiotics was determined as well as the initial choice of antimicrobial therapy based on the presumed source of infection. Results: Two hundred and tweleve unique ED patient encounters meeting criteria for severe sepsis were reviewed. Initial analysis showed a median time from meeting severe sepsis criteria to delivery of antibiotics of 163 minutes. Subsequent analysis postimplementation of the protocol revealed a median time of 79 minutes, representing an overall reduction of 52% (p < 0.05). Initially only 23% of patients meeting criteria for severe sepsis received antibiotics within 1 hour of recognition as recommended by the Surviving Sepsis Campaign Guidelines for management of severe sepsis and septic shock. Postimplementation this number improved to 39%. Prior to the protocol, 47% of patients received correct antibiotic coverage for the presumed source of infection in compliance with locally published guidelines. After initiating the protocol, 73% received appropriate initial antibiotics, an overall improvement of 26%. Conclusion: Prompt initiation of appropriate antibiotic therapy in severe sepsis and septic shock has been shown to be a very important predictor of outcome. Our data reveals that a guideline based ED sepsis protocol for the evaluation and treatment of the septic patient appears to improve the time to administration of antibiotics as well as the appropriateness of initial antibiotic therapy in patients with severe sepsis. Keywords:sepsis, time to antibiotics, implementation research

Winner of the CAEP Resident Research Abstract Award

99 AMBULATORY visits to a Canadian emergency department

Woolfrey K, Chu J. McMaster University, Hamilton, ON

Introduction: According to the Canadian Institute for Health Information, between 2003 and 2004, 57% of patient visits to Canadian EDs were for nonurgent conditions. In 1996, Burnett et al. published a survey, characterizing 200 nonurgent patients presenting to a Montréal ED, during regular business hours. Afilalo, in 2004, compared characteristics of nonurgent with urgent and semi-urgent patients, who visited 5 Quebec tertiary hospitals. To date, there has been no Ontario study done on this topic. The purpose of this survey is to determine the reasons why patients choose the ED as their health care facility, for nonurgent medical care. We hypothesized that these factors are affected by recent changes in provincial health care, namely the introduction of Telehealth and the shortage of primary care physicians in Canada. Methods: A written survey was administered to CTAS IV and V patients presenting to the St. Joseph Healthcare Emergency Department in Hamilton, Ontario. The survey was comprised of questions on patients’ baseline characteristics, accessibility to primary care, perceptions of the urgency of their medical problems, and satisfaction with care received in the ED.Results: We collected 245 completed surveys. Musculoskeletal complaints were the most common reason for low-acuity ED visits. Of patients surveyed, 11% had no GP. Among those who did have GPs, 1e in every 3 patients was unable to access their doctor; 26% did not know of any alternative health care options (e.g. urgent care centres, walk-in clinics, etc.). More than one-half of patients felt their problem required immediate medical attention, while 9% were advised by Telehealth to go to the ED. Conclusion: Although most patients do have GPs, accessibility to primary care remains a major factor in nonurgent ED visits. There is a need to educate the public on alternative health care options for low acuity problems. We hope to repeat this study in other EDs across Ontario and to use this information to improve the efficiency of emergency care in the province.Keywords: nonurgent users, access to care, Telehealth

100 PATIENT safety event reporting and response in British Columbia emergency departments

Brubacher JR, Hunte G, Hamilton L, Kopetsky D. University of British Columbia, Vancouver, BC

Introduction: The primary purpose of reporting patient safety events (PSEs) is to learn from experience to guide changes to prevent similar future events. Hence, reporting, analysis, feedback and system change are required for learning to occur. Methods: We interviewed ED nurse managers (EDNM) from urban and rural EDs in BC to describe the current content, structure and process of ED PSE reporting. Fifty-nine semistructured telephone interviews were conducted between September and December 2007. Data were iteratively analyzed for emergent patterns and themes. Results: Five broad themes emerged with open coding: 1) Off the side of the desk: EDNMs reported that they did not have sufficient time, resources or specific training to conduct system level analyses of PSEs. Yet, despite these limitations, few wanted to give the task of analyzing reports to an outsider. 2) Into the void: Many EDNMs neither uniformly gave nor received feedback on PSE reports. Some received summary statistics on PSEs from their site, whereas others did not even know where to send reports. 3) Stats or stories: In many cases the purpose of PSE reporting was not clear. At sites receiving statistical summaries, reports were often viewed as statistics to be monitored rather than stories to be learned from. 4) Threshold of harm: Incidents resulting in significant harm were usually reported, but near misses and Prehospital incidents were generally not reported. 5) Lack of physician involvement: EDNMs reported that physicians rarely filled out reports, but might verbally report to the EDNM or ask a nurse to report for them.Conclusion: Learning from patient safety events in urban and rural EDs in BC is limited by underreporting, lack of system level analysis and scant feedback. Patient safety stories are left untold,and physicians are rarely involved in the process. PSE reports are managed off the side of the desk and often go into the void. Keywords: patient safety, root cause analysis, qualitative research

101 DEMOGRAPHIC characteristics of emergency department patients with suspected renal colic

Edmonds ML, McLeod SL, Yan JW, Theakston KD, Sedran RJ. Division of Emergency Medicine, Faculty of Medicine, University of Western Ontario, London, ON

Introduction: Southwestern Ontario is a known stone belt with a high incidence of renal colic, and our centre is somewhat unusual in its high use of renal ultrasound (US) in the diagnosis of renal colic. The purpose of this study was to describe the demographic characteristics, imaging results and outcomes of patients who had a renal US for suspected renal colic over a 1-year period. Methods: A retrospective chart review was completed for all adult patients who had an emergency department (ED)-ordered US for suspected renal colic. Data was gathered from 2 tertiary care EDs with a combined annual census of 95 000 during a 1-year period (Jan. 1 to Dec. 31, 2006). Independent, double data extraction was performed for all imaging reports and results were categorized as normal, suggestive, stone seen or nonrenal disease. Results: Of the 857 renal US ordered during the study period, 373 (43.5%) were classified as normal, 182 (21.2%) were classified as suggestive, 241 (28.2%) were classified as stone seen and 61 (7.1%) were classified as nonrenal disease. Of the 857 renal US, 160 (18.7) patients had a computed tomography scan and 29 patients (3.4%) required urologic intervention. Mean age of all patients was 44 (SD 16) years and 53% were male. The mean length of stay in the ED was 5.5 (SD 3.4) hours, and 6.4% of patients were brought to the ED by ambulance. There was no seasonal variation in ED visits, and 4.2% of patients were admitted.Conclusion: To our knowledge, this is the largest Canadian renal colic research study. Although patients presenting to the ED with suspected renal colic frequently will be shown to have urolithiasis, there is a relatively low need for intervention or admission in this population. Future research to define low-risk patient characteristics may help decrease the need for imaging in these patients. Keywords:renal colic, ultrasound, utility

102 PATIENT safety event reporting forms used in British Columbia emergency departments

Brubacher JR, Hunte G, Hamilton L, Kopetsky D. University of British Columbia, Vancouver, BC

Introduction: Patient Safety Event (PSE) reporting forms should be easy to use. Reporters need space to tell what happened, why it happened, and how they think it could have been prevented, whereas managers need to make sense of system level factors that contribute to events. Methods: We obtained 14 PSE reporting forms used in 61 BC EDs. Forms were analyzed to describe their length, layout, content, and specificity for ED events. Results: One form was used in 32 EDs, another in 11 and the remainder in 1 to 4. Several patterns were noted: 1) None of the forms were specific to the ED. 2) Most forms were also not specific to PSEs and also allowed reporting of events involving staff, visitors or property. Some EDs had multiple separate forms for occupational health events, blood and body fluid exposure, medication incidents, falls and good catches. 3) Forms ranged in length from 1 to 3 pages, but had limited space for a narrative description of the events. Most were highly structured with tick boxes to classify events and identify risk factors. 4) All forms had large sections focused on falls and medication events. 5) All but one form asked about injury severity, and usually focused on physical injuries sustained in a fall. 6) The reporters name was always requested and most forms asked for a witness. 7) All but 2 forms asked if a physician had been notified, but only 3 forms had a physicians section. 8) All but 2 forms included a section for the manager investigating the report, which ranged from half a page of free text to highly structured tick boxes with specific follow-up options. Conclusion: PSE reporting forms used in BC EDs are highly structured with a focus on falls and medication events, and are neither anonymous nor specific for ED events. Tick boxes may facilitate monitoring of PSE statistics, but scant narrative space limits understanding of the context of PSEs, as well as those events that do not fit the tick box classification. Multiple forms further complicate reporting in some organizations.Keywords: patient safety, error reporting, reporting forms

103 MALARIA in the Ottawa Hospital emergency department

Vayalumkal P, McCarthy A. The Ottawa Hospital, Ottawa, ON

Introduction: Malaria is a very common and potentially fatal tropical disease. In Canada, malaria has proven to be a diagnostic challenge over the years. The majority of cases first present through the emergency room. Its sporadic and nonspecific presentation, however, has lead to misdiagnosis and delays in treatment. This study sought to review the diagnosis and treatment of malaria cases presenting to the Ottawa Hospital emergency department (ED). It examined the time from initial presentation to diagnosis and time to treatment in the Ottawa Hospital. Methods: The study was conducted using a retrospective chart analysis. All cases of malaria smear positive patients evaluated by the ED at the Ottawa Hospital from January 2000 to December 2006 were included. Cases were excluded if diagnosis or treatment were initiated at another centre, if records were not available or if the cases were referred directly to a consultant service. Charts were reviewed and data entered into a chart review form.Results: Thirty-five cases were reviewed. A majority of the cases were travelers from endemic regions in Africa (57.1%). No chemoprophylaxis was used in 71.4% of cases. Falciparum malaria accounted for 57.1% of cases. The average time in the ED until diagnosis was 4.53 hours, with the average time from diagnosis to treatment being 2.54 hours. The ED initiated treatment in 20.1% of cases. There were no deaths among the 35 cases, with 1 ICU admission in a patient discharged on oral treatment. Conclusion: Although there was no mortality amongst the cases reviewed, changes need to be made in the diagnosis and treatment of malaria. Firstly, patients must not be sent home with malaria smear results pending. Secondly, initial treatment should be observed in the ED to ensure tolerance for those being discharged. Lastly, the use of a malaria treatment protocol may help ED physicians to initiate treatment earlier to avoid wait times for consultant services. Keywords: malaria, tropical diseases, chart review

Winner of the CAEP Resident Research Abstract Award

Abstracts: 1-4, 5-24, 25-43, 44-73, 74-103, 104-131