2010 CAEP/ACMU Scientific Abstracts - Poster Presentations: 72-96

CAEP Abstracts

CJEM 2010;12(3):229-278

Poster presentations

72 A SURVEY of Canadian emergency physicians on utilization of femoral nerve and fascia iliaca compartment blockade for management of midshaft femoral fractures in children

Angelski CL, Black KJL, Taylor B; PEM IWK Health Centre, Dalhousie University, Halifax, NS

Introduction: Femoral nerve (FNB) and fascia iliaca compartment blockade (FICB) have been shown to be safe and effective alternatives to systemic analgesia for midshaft femoral fractures presenting to the pediatric emergency department (ED), yet appear to be underused in Canada. No studies exist that describe under what circumstances Canadian pediatric emergency physicians would be willing to undertake these techniques. We surveyed physician members of the Pediatric Emergency Research Canada (PERC) database to explore practice variation in the Canadian setting. Methods: English-speaking members of the PERC database were surveyed between October and December 2009 by Web-based questionnaire. The questionnaire was developed and strengthened via content validity exercises before distribution. Results: A response rate of 35% (n = 68) was attained. Of those surveyed, 24% (n = 16) have at least some experience with FNB/FICB; 15% (n = 10) have used FNB/FICB within the last 3 years. No adverse events were reported with either technique. Geographical differences were found when comparing those who had experience with FNB/FICB to those who did not (p = 0.04). Recent use of either block was also geographically dependent (p = 0.001). Barriers to FNB/FICB use included lack of education/training in either FNB/FICB, lack of departmental policy or guidelines and lack of familiarity with current evidence. Conclusion: FNB/FICB use in midshaft femoral fractures in children is infrequent. Canadian practice lags behind other settings and locales despite evidence that these techniques are effective alternatives to systemic analgesia. Regional discrepancies exist within Canada and appear to influence willingness to undertake either technique. Keywords: femoral nerve block, femoral fractures, pediatric emergency medicine

73 COMPLIANCE with CAEP asthma clinical practice guidelines at a tertiary care emergency department

Filiatrault L, Harriman D, Abu-Laban RB, FitzGerald JM, Chahal AM, McKnight RD; Vancouver General Hospital, Vancouver, BC

Introduction: Although evidence-based clinical practice guidelines (CPGs) exist, emergency department (ED) asthma management remains highly variable. Our objective was to comparatively evaluate asthma management at a tertiary care ED with that advised by the Canadian Association of Emergency Physicians (CAEP) asthma CPG. Methods: This retrospective study enrolled patients between the ages of 19 and 60 with a prior diagnosis of asthma, who were seen for an acute asthma exacerbation at the Vancouver General Hospital ED in 2008. Standard methodology guidelines for medical record review were followed, including explicitly defined criteria and determination of interrater reliability. Primary outcomes were the proportion of cases with the following: objective assessment of severity using peak expiratory flow (PEF), use of systemic corticosteroids (SCS) in the ED and at discharge, prescription for any inhaled corticosteroids (ICS) and documentation of outpatient follow-up. Results: A total of 204 patient encounters were enrolled. Compliance with the CAEP Asthma CPG was as follows: measurement of PEF 90.2% (95% CI 85.3%–93.9%), use of SCS in the ED 64.4% (95% CI 57.3%–71.2%), prescription of SCS 59.1% (95% CI 51.2%–66.7%), prescription of any ICS 51.0% (95% CI 40.8%– 61.1%; ICS alone 38.0%, longacting agonist/ICS combination 13.0%), documentation of outpatient follow-up 78.3% (95% CI 71.4%–84.2%). Kappa values for interrater assessment of primary outcomes ranged from 0.93 to 1.00. A logistic regression analysis, performed for secondary purposes, identified no factors associated with ICS prescription and found that moderate severity, any ICS on arrival and number of bronchodilator treatments were significantly associated with SCS prescription. Conclusion: This study indicates that discordance exists between asthma management at a tertiary care ED and the CAEP asthma CPG. Further research is warranted to understand the reasons for this finding. Keywords: asthma, guideline compliance, gap analysis

74 A COMPARISON of emergency physician time in Ontario emergency departments using different methods of physician remuneration

Dreyer JF, McLeod SL; London Health Sciences Centre, London, ON

Introduction: Emergency physicians (EPs) in Ontario are remunerated for their clinical work either by means of an alternate funding arrangement (AFA) or by means of a feeforservice (FFS) remuneration method. The objective of this study was to determine the time required for EPs to assess and treat patients in both AFA and FFS emergency departments (EDs). Methods: Twenty hospital-based EDs agreed to participate in this study. Data were collected by research assistants who directly observed EPs for entire shifts and recorded the activities of the EP on a moment-by-moment basis. The individual times of all physician activities associated with a given patient were summed to derive a directly observed estimate of EP time required to treat a patient. Treatment time by CTAS was estimated using Kaplan–Meier Survival Analysis. Hospital sites were grouped into 4 categories; FFS, nonacademic AFA, academic AFA and pediatric AFA. Results: A total of 17 197 patients were observed in this study over 767 shifts. Mean (95% CI) EP time for FFS, nonacademic AFA, academic AFA and pediatric AFA sites was 54.7 (41.6–66.5), 51.8 (39.6–64.1), 72.2 (61.5–82.8) and 194.7 (117.8– 271.5) minutes for CTAS1; 21.7 (20.3–23.1), 33.0 (29.5–36.6), 38.8 (35.0–42.6) and 47.9 (43.5–52.3) minutes for CTAS2; 14.8 (14.3– 15.2), 23.4 (22.2–24.6), 25.2 (24.1–26.3) and 28.9 (27.5–30.2) minutes for CTAS-3; 10.1 (9.8–10.4), 14.1 (13.5–14.7), 16.2 (15.4–17.0) and 21.3 (20.3–22.3) minutes for CTAS4 and 7.8 (7.1–8.5), 10.6 (9.6–11.7), 12.8 (10.8–14.9) and 19.7 (16.8–22.6) minutes for CTAS5. The mean (95% CI) throughput for EDs operating under different remuneration mechanisms was 4.4 (4.2–4.6), 2.9 (2.8–3.1) and 2.3 (2.2–2.4) patients per hour for the FFS, AFA and pediatric AFA sites. Conclusion: FFS treatment times were faster (p < 0.05) in CTAS categories 2–5, resulting in higher patient throughput than EPs working in AFA hospitals. Patient outcomes and satisfaction were not assessed. Keywords: alternative funding arrangement, fee-for-service, fractile response

75 IN out-of hospital cardiac arrest patients, does the description of any specific symptoms to the 9-1-1 dispatcher improve accuracy of the diagnosis of cardiac arrest: a systematic review conducted for the C2010 ILCOR Resuscitation Guidelines

Vaillancourt C, Charette C; University of Ottawa, Ottawa Health Research Institute, Ottawa, ON

Introduction: We sought to determine if, in adult and pediatric patients with outofhospital cardiac arrest, the description of any specific symptoms to the 9-1-1 dispatcher compared with the absence of any specific description, improve accuracy of the diagnosis of cardiac arrest. Methods: For this systematic review, we searched 5 electronic databases (including MEDLINE and Embase), a clinical trial registry and the grey literature. We reviewed without restriction all peerreviewed prospective and retrospective study designs reporting simulated or real dispatcher interaction with callers. Two independent investigators used standardized forms to review papers for inclusion, quality and data extraction. We used for selection agreement and report consensus on science. Lack of homogeneity precluded metaanalyses. Results: We identified 473 citations; 69 were selected for full evaluation (κ = 0.74) and 21 met all inclusion criteria (κ = 0.57). One before–after trial and 10 descriptive studies using a strategy evaluating consciousness and breathing status to identify cardiac arrest report sensitivity ranging from 38% to 96.9% and specificity ranging from 94.8% to 99.2%. One case–control, 1 before– after and 4 observational studies describe agonal breathing as a significant barrier. Other studies report improved cardiac arrest recognition when inquiring about the following: past history of seizure, victim’s age and emotional status of caller, response to painful stimuli, or level of activity. Information spontaneously provided by callers such as facial colour or describing the victim as dead was highly associated with cardiac arrest in a case–control study. Conclusion: Dispatchers most commonly use unconsciousness and absence of breathing or abnormal breathing to identify cardiac arrest victims with variable success. This review will inform the 2010 resuscitation guidelines, and suggest that other clinical features should be explored to improve cardiac arrest recognition by 9-1-1 dispatchers. Keywords: prehospital dispatch, symptom description, cardiac arrest

76 EMERGENCY department syndromic surveillance for pandemic H1N1influnenza

Bellazzini MA, Minor KD, Svenson JE; Division of Emergency Medicine, University of Wisconsin Hospital, Madison, WI

Introduction: Early detection of disease outbreaks is crucial to emergency department preparedness. This study demonstrates the utility of syndromic surveillance for detection of influenza-like illness (ILI) during the 2009/10 H1N1 influenza pandemic and compares a novel linear regression (LM) aberration detection method to traditional cumulative sum (CUSUM) detection. Methods: International Classification of Disease (ICD9) discharge diagnosis data from our emergency department was analyzed from 11/01/08 to 11/01/09. This data was categorized into ILI syndromes and converted to a proportion of illness by dividing by total ED volume for that day. C1 and C2CUSUM algorithms using a 3 standard deviation threshold (3SD) were used to generate alerts indicating increased ILI activity. Alerts were also generated using LM analysis. This detection algorithms used the last 12 day proportion of ILI before the day of interest. The slope of the line was used to determine an increase in ILI activity by using an alert angle threshold of 18 degrees. Laboratory data was used to estimate the first confirmed cases of influenza during seasonal and pandemic waves. Timeliness to detection was determined by comparing syndromic surveillance alerts to dates of laboratory confirmed influenza. Results: C1 and C2 methods failed to alert before the laboratory confirmation of seasonal influenza. LM alerted 5 days before laboratory confirmed seasonal influenza. C1 and C2 methods alerted 11 days before an increase in counts of laboratory confirmed pandemic influenza in spring while LM alerted 12 days before. C1, C2 and LM methods all alerted 13 days before a rise in laboratory confirmed influenza for the fall pandemic wave. Conclusion: Syndromic surveillance provides early notification of ILI before a rise in laboratory confirmed influenza during the 2009/10 pandemic. LM detection methods performed better at outbreak detection than CUSUM with a 3SD threshold in some cases. Keywords: syndromic surveillance, pandemic influenza, H1N1

77 THE EFFECT of a bolus dose of etomidate on hemodynamics: a systematic review

Kelly-Smith C, Hohl C, Yeung T, Sweet D, Doyle-Waters M, Schultzer M; University of British Columbia, Vancouver, BC

Introduction: Etomidate is a preferred induction agent in the emergency department. Our objective was to synthesize the available evidence on the effect of etomidate on mean arterial pressure (MAP). Methods: We developed a systematic search strategy and applied it to 10 electronic bibliographic databases. We hand searched medical journals, conference proceedings, grey literature and bibliographies of relevant literature, and contacted content experts for studies. Studies were included if they reported adult data comparing the effect of etomidate to another rapidly acting intravenous induction agent on blood pressure postinduction. Retrieved articles were reviewed and data were abstracted in duplicate using standardized methods. Data were pooled using the random effects model if at least 4 clinically homogenous studies of the same design reported the same outcome measure. Results: Of 3083 studies found, 27 met our inclusion criteria, of which 22 were conducted in elective surgical and 5 in critically ill patients. Four studies on elective surgical patients met criteria for pooling. The results from pooling showed that etomidate caused no drop in the pooled mean MAP postinduction. However, barbiturates (–9.2 mm Hg, 95% CI –15.0 to 3.4 mm Hg, p < 0.01) and propofol (–9.6 mm Hg; 95% CI –21.3 to 2.2 mm Hg, p = 0.11) were associated with a lower pooled mean MAP 2 minutes postinduction. The remaining 18 studies in elective surgical patients showed mixed results, overall there was a trend to decrease in postinduction MAP with the use of propofol and benzodiazepines compared with etomidate. Of the 5 studies including critically ill patients, 2 showed a significant increase in hypotension (SBP < 90 mm Hg) with midazolam while the others showed no differences in MAP between barbiturates, midazolam and etomidate. Conclusion: The available evidence suggests that barbiturates and propofol lead to brief reductions in MAP in elective surgical patients. The data on critically ill patients indicate an increased risk of hypotension with the use of midazolam. Keywords: etomidate, mean arterial pressure, procedural sedation

78 OUTCOMES of emergency department patients discharged with nonspecific vertigo or dizziness

Grafstein E, Stenstrom R, Harris D, Hunte G, Innes G, Poureslami I, Scheuermeyer F; Providence Health Care, University of British Columbia, Vancouver, BC

Introduction: Patients discharged from an emergency department (ED) with nonspecific vertigo or dizziness (NVD) form a cohort with a broad range of potential etiologies. We attempted to determine this cohort’s long-term outcomes. Methods: A retrospective administrative database review was undertaken for patients who were discharged from 2 urban EDs with a primary discharge diagnosis of NVD (ICD-10 R42) from Aug. 1, 2005, to Aug, 1, 2007. This cohort was probabilistically linked to both the regional ED administrative database and the provincial vital statistics database. The primary outcome was the combined rate of stroke, significant cardiopulmonary diagnoses (ACS, COPD, CHF, syncope, dyspnea NYD, heart block), vertigo related diagnoses (NVD, fall with fracture) requiring admission, or death at 1 year. The secondary outcome was the rate of ED revisits within 1 year. Results: Data on 1286 patients were collected during the 2-year period. Overall there were 69 index admissions (5.4%) leaving 1217 patients. The mean age was 55.3 (SD 18.5) years; 44.9% were male, and 34.6% arrived via ambulance. A total of 349 discharged patients received head CTs (28.7%); 92 patients (7.6%) had consult orders. Within 1 year of the index visit there were 33 patients with a revisit readmission (RA) to the same hospital or ED visit/ admission (VA) to a different regional site for a cardiopulmonary or neurologic diagnoses (2 CVAs). Overall there were 16 total deaths (D). Two patients had a cardiorespiratory etiology and 1 died of a CVA. Of note, there was an 8-fold increase in composite end point of RA + VA + D in the subset of patients greater than or equal to 75 (22/214) compared with those patients under 75 years old (13/1003). A total of 73 patients (6%) had repeat visits for NVD during the subsequent 12 months. Conclusion: Recidivism rates among discharged with NVD are significant. Elderly patients 75 years of age or older have an 8-fold risk (10%) of serious cardiorespiratory or neurologic problems within 1 year compared with those under 75. Keywords: nonspecific vertigo, recidivism, administrative database research

79 OUTCOMES of acetaminophen overdose in the United States: a population-based study of hospitalizations between 1993 and 2007

Yarema MC, Shaheen AAM, Myers RP; Poison and Drug Information Service, Calgary, AB

Introduction: Population-based data describing the outcomes of acetaminophen overdose (AO) in North America are limited. We sought to examine rates and predictors of acute liver injury (ALI), acute liver failure (ALF) and mortality following AO using a United States (US) nationwide database. Methods: Adult patients hospitalized for AO in the US between 1993 and 2007 were identified using the Nationwide Inpatient Sample (NIS) database and International Classification of Diseases codes. We used validated coding algorithms to identify the presence of ALI (alanine aminotransferase > 1000 IU/L) and ALF (hepatic encephalopathy and INR > 1.5). The circumstance of the overdose (intentional v. unintentional) was identified using Ecodes. Independent predictors of hepatotoxicity and in-hospital mortality were evaluated using multivariate logistic regression models with adjustment for patient and hospital factors including hospital volume. Results: Between 1993 and 2007 there were 75 439 discharges for AO identified in the NIS database that met the inclusion criteria. Overall, 5641 patients (7%) developed ALI; 3062 (4%) developed ALF; 995 (1%) died. Independent predictors of ALI included middle age (41–60 yr v. ≤ 40 yr: adjusted odds ratio [OR] 1.19, 95% CI 1.11–1.28), white race (OR 1.11, CI 1.03– 1.19), unintentional overdose (OR 3.05, CI 2.85–3.26), and an increasing number of comorbid conditions, including underlying liver disease (OR 2.28, CI 1.96–2.66), alcohol abuse (OR 1.18, CI 1.10–1.27) and malnutrition (OR 3.41, CI 2.72–4.28). Patients hospitalized in highvolume centres, those transferred from another institution and admitted during more recent years were more likely to develop ALI. Similar factors were associated with progression to ALF and in-hospital mortality. Conclusion: Independent predictors of ALI, ALF and mortality after acetaminophen overdose include older age, unintentional overdoses, underlying liver disease and malnutrition. Further studies are necessary to identify means of improving outcomes among these highrisk patient subgroups. Keywords: acetaminophen poisoning, acute liver injury, risk prediction

80 DO delayed consultation responses prolong ED and hospital length of stay?

Innes G, Wang D, Mercuur L, Godfrey C, Dowling S; Providence Health Care and St. Paul’s Hospital, Vancouver, BC

Introduction: Delayed consultation response to emergency patients requiring hospitalization may prolong ED boarding time and aggravate ED crowding. Slower consultation response may also delay initiation of care by the most appropriate service, extend time spent in a suboptimal environment and prolong overall hospital length of stay (LOS). Our objective was to determine the impact of delayed consultation on overall hospital LOS. Methods: This administrative database study was conducted at the Rockyview Hospital, an urban Calgary ED with an annual census of 70 000 visits. Eligible patients included all those admitted to the RVH Hospitalist service from Sep. 1, 2007, to Aug. 31, 2008. Time from triage to inpatient admission order (ADM) was used as a proxy for consultation delay. Based on triage-to-ADM time interval, patients were stratified into 5 successive delay groups (0–4 h, 4–8 h, 8–12 h, 12–16 h and > 16 h). The primary outcome was overall mean hospital LOS by delay group. Results: During the 1 year period, 2979 admissions were studied, including 76 (2.6%) in the 0–4 hour delay group, 481 (16.1%) in the 4–8 hour group, 591 (19.8%) in the 8–12 hour group, 422 (14.2%) in the 12–16 hour group and 1409 (47.3%) in the > 16 hour group. Overall mean age was 74.3 years, 56.7% of patients were female, 56.4% arrived by EMS and 53.4% were in CTAS level 3. Mean ED LOS and hospital LOS (SD) for the entire study cohort were 22.8 (12.0) hours and 438.1 (622) hours. Patients in the 0–4 hour (short delay) category were younger, more often male, higher acuity and more likely to arrive by EMS. Consultation delays were associated with prolonged ED LOS (8.1, 11.4, 15.5, 19.9 and 31.4 h by group) and hospital LOS (369, 401, 381, 427 and 481 h). Conclusion: Younger sicker patients experienced shorter consultation delays. Patients with prolonged consultation delays had hospital lengths of stay 80–112 hours (3–5 d) longer. Further analysis is required to determine the impact of complexity covariables on these profound outcome differences. Keywords: consultation delays, emergency crowding, administrative database study

81 EVALUATION of a structured wellness curriculum for emergency medicine residents

Dong KA, Dance E, Blouin D, Yarema M, Williams J, Rowe BH; Department of Emergency Medicine, University of Alberta, Edmonton, AB

Introduction: Residency training is known to be a stressful period on the career path to becoming a physician. Emergency medicine (EM) residents face unique stressors such as shift work, the potential for violence in the workplace and other occupational health risks. The objective of this study was to assess the wellness of Canadian EM residents and to determine whether a structured wellness curriculum resulted in improved wellbeing. Methods: A pilot, controlled trial was performed. EM residents from 3 Canadian programs were approached to complete a baseline questionnaire about their wellness. EM residents involved in the intervention program participated in a structured wellness curriculum that involved 2 oneonone meetings with a staff wellness advisor. These meetings generated common “wellness themes” which were then the subject of 2 staff-planned wellness academic half days. EM residents in the 2 control sites did not receive any structured wellness initiatives. Resident wellness was then reassessed 1 year later. Wellness was measured with 2 standardized tools, the SF8 (which contains a physical [PCS8] and mental health [MCS8] component) and the Brief Resident Wellness Profile (BRWP). Results: A total of 33 EM residents participated (17 intervention, 16 control; response rate 89%). The average age was 30 (SD 2.2) years and 69.7% of the residents were male. The majority of the residents (70.6%) in the intervention group felt they benefited from the wellness curriculum; 12 (70.6%) found the meetings with the wellness advisors useful, whereas only 2 (11.7%) found the academic halfday presentations useful. Physical health (PCS8) improved significantly in the intervention group (p = 0.012) and decreased significantly in the control group (p = 0.033). No significant differences were found in mental health scores or in the BRWP. Conclusion: EM residents appear to benefit from a structured wellness curriculum. The optimal design and evaluation of such programs warrants further study. Keywords: wellness curriculum, residency training, controlled trial

82 EMERGENCY department conditions associated with the number of patients who leave a pediatric emergency department before physician assessment

Stang AS, Ciampi A, Strumpf E, McCusker J; Montreal Children’s Hospital, Montréal, Que.

Introduction: As emergency department (ED) waiting times and volumes increase, substantial numbers of patients leave before physician assessment. Little research, and none in a pediatric setting, has been conducted on ED factors associated with the number of patients who LWBS (patients who leave after first contact with the ED, either registration or triage, but before being seen by a physician). The objective of this study was to identify ED conditions reflecting patient input, throughput and output associated with the number of patients who LWBS in a pediatric setting. Methods: This study was a retrospective, observational study using data from 1 tertiary care pediatric ED. The study population consisted of all patient visits to the ED from April 2005 to March 2007. Multivariate Poisson regression analyses were used to examine the impact of variables describing the timing of patient arrival and ED conditions including patient acuity, volume and waiting times on the number of patients who LWBS. Results: During the study period there were 138 361 patient visits corresponding to 2190 consecutive shifts; 11 055 (7.99%) of patients left before being seen by a physician. In the multivariate analysis, the number of patients who LWBS was highest for the overnight shift as compared with the day shift (rate ratio 0.44 95% CI 0.36–0.53) and the evening shift (rate ratio 0.81 CI 0.69– 0.95). The throughput variables, time from triage to physician assessment (rate ratio 2.11 95% CI 2.01–2.21) and time from registration to triage (rate ratio 1.55 95% CI 1.25–1.90) had the largest impact on the number of patients who LWBS. Conclusion: Throughput variables were the most predictive of the number of patients who LWBS. Interventions designed to decrease the number of patients who LWBS should focus on improving ED throughput, particularly on overnight shifts. Keywords: LWBS, throughput measurements, emergency crowding

83 PREDICTORS of admission in Canadian Emergency Department Triage and Acuity Scale level V patients

Lin D, Worster A; McMaster University, Hamilton, ON

Introduction: Canadian Emergency Department Triage and Acuity Scale (CTAS) level V patients are sometimes admitted to hospital despite being assessed as nonurgent. This study sought to determine the frequency and patient factors known before triage associated with admission. Methods: We reviewed ED records with CTAS-V from Apr. 1, 2003, to Sep. 30, 2009, from the National Ambulatory Care Reporting System and Discharge Abstract Database at 3 tertiary academic hospitals in a single city in Ontario. We conducted backward stepwise logistic regression to determine independent predictors of admission known at time of triage. Results: Over the study period, there were 37 416 CTAS-V patients. The mean age was 36.9 (standard deviation [SD] 20.7) and 55.1% were male. A total of 84.0% of patients were discharged, 13.1% left before discharge, 1.5% were admitted and the rest was transferred. The mean length of stay for admitted patients was 8.2 (SD 11.39) days. Admission to hospital was more likely if the patient was over age 65 (odds ratio [OR], 5.45, 95% confidence interval [CI] 4.57–6.51), and arrived by ambulance (OR 7.41, 95% CI 6.14–8.94). Conclusion: Admission to hospital in CTAS-V patients was associated with age over 65 and arrival by ambulance. Although these findings need to be validated in other populations, they should be considered in future triage and ED flow plans. Keywords: low acuity patients, admission predictors, CTAS

84 MINOR and minimal head injuries in patients on warfarin

Alrajhi K, Perry JJ, Forster A; University of Ottawa, Ottawa, ON

Introduction: There is currently very little evidence to guide physicians on how to manage emergency patients who sustain a minor or minimal head injury (HI) while on warfarin. The objective of this study was to determine the incidence of intracranial bleeding (ICB) in anticoagulated patients with minor (initial GCS of 13–15 with witnessed loss of consciousness [LOC]/confusion or amnesia) and minimal (no LOC/amnesia/confusion) HIs and the association of ICB to clinical features, and international normalized ratio (INR) level. Methods: Historical cohort study of all adult patients who presented to 2 tertiary care EDs over 2 years with minor or minimal HIs, identified through the National Ambulatory Care Reporting System database, while on warfarin with INRs of 1.5 or greater. Exclusions included penetrating injuries, new focal neurologic deficits and previous brain surgery. Our structured data collection tool included 25 variables from history, physical exam, CT, surgical interventions, return visits and death. Our outcome, ICB, included any type of bleed within the cranium, and was determined by either 1) CT during initial ED visit, or 2) CT during a subsequent ED visit at any of the 4 regional hospitals within 2 weeks of the initial visit. Results: We identified 176 patients, of which 60.2% were female, mean age was 79.0 (range 25–100) years and mean INR was 2.5 (range 1.5– 8.2). A total of 157 patients (89.2%) had CT scans and 28 (15.9%) had ICB. Comparing the group with ICB to the group with no ICB there was no significant difference in INR levels (mean 2.53 v. 2.45, p = 0.5), respectively, while LOC correlated with higher incidence of ICB (28.6% v. 10.8%, p < 0.05), respectively. Incidence of ICB was 21.9% in the minor HI group versus 4.8% in the minimal group. Conclusion: The incidence of ICB in patients on warfarin is considerable. LOC was associated with higher rates of ICB. This study supports a low threshold for ordering a CT scan for patients with minor or minimal head injuries who are on warfarin. Keywords: minor head trauma, warfarin, intracranial bleeding

85 PROFILING the burden of night work in an academic emergency department

Arntfield R, McLeod SL, Peddle M, Dreyer J; London Health Sciences Centre, University of Western Ontario, London, ON

Introduction: Night shifts are widely regarded as a drawback of being an emergency physician (EP). The objective of this study was to profile the burden that working traditional night shifts (11 pm–7 am) poses toward physicians working in an academic emergency department (ED). Methods: An online, 37-item survey was distributed to all 43 EPs working in the EDs of an academic centre (annual census of 150 000). Participants rated their level of agreement on a 7-point Likert scale regarding questions related to night work. Results: A total of 43 (100%) questionnaires were completed. Of respondents, 69.8% were male and most commonly had greater than 15 years’ experience (30.2%) and lived with partner and children (67.4%). Recovery to baseline vegetative routine after a single night shift was most common after 1 day (55.8%) and after multiple night shifts was days or greater (95.3%). A total of 33 (76.7%) EPs felt they were more irritable following night shifts compared with other shifts. EPs are involved significantly less in household responsibilities (69.8%), hobbies (62.8%) and academic work (76.8%) following night shifts. The majority (79.1%) stated night shifts are the greatest drawback of their job while the same proportion (79.1%) list shift work as one of the greatest advantages of being an EP. The majority of respondents reported that teaching (74.4%), diagnostic test interpretation (58.1%), quality of handover (60.5%) and decisiveness regarding patient care plans (72.1%) are inferior on night shifts compared with day shifts. Despite receiving 150% of day wages, 29 (67.4%) deny financial motivation to work night shifts and 21 (48.8%) felt that the financial remuneration does not justify the disruption caused by working night shifts. Conclusion: ED physicians experience significant morbidity related to traditional night shift duties. These deleterious effects span the domains of sleep physiology, home life, mood, academic productivity and perceived job performance. Study into the role and effects of differently structured night shifts may be warranted. Keywords: shiftwork, night shifts, survey research

86 PEDIATRIC mental health emergencies: a systematic review of crisis interventions used in the emergency department

Hamm MP,Osmond MH, Curran J, Scott S, Ali S, Hartling L, Gokiert R, Cappelli M, Hnatko G, Amanda S. Newton AS; University of Alberta, Edmonton, AB

Introduction: Mechanisms that promote optimal care of pediatric mental health in the emergency department (ED) need to be developed. This systematic review evaluated the effectiveness of management interventions for pediatric mental health. Methods: We searched electronic databases, references, key journals and proceedings from 1985 to 2009. Studies that evaluated ED-based assessment, treatment or psychosocial management were included. The primary focus was on pediatric-based studies, but adult-based studies were used to make recommendations for future pediatric investigation. One reviewer extracted the data and a second checked for completeness and accuracy. Calculations for primary outcomes included odds ratios (ORs) and mean difference (MD) with 95% confidence intervals (CIs). Results: Twelve observational studies were included with pediatric (n = 3) and adult/unknown (n = 9) aged participants. Pediatric studies supported the use of specialized care models to reduce hospitalization (OR 0.45, 95% CI 0.33 to 0.60), return visits (OR 0.60, 95% CI 0.28 to 1.25) and length of stay (MD –43.1 min; 95% CI –63.088 to –23.11). In adult studies, reduced hospitalization was reported with use of a crisis intervention team (OR 0.59; 95% CI 0.43 to 0.82). Five adult studies assessed triage scales. In a comparison of a mental health scale to a national standard, one study demonstrated reduced wait (MD –7.7 min; 95% CI –12.82 to –2.58) and transit (MD –17.5 min; 95% CI –33.00 to –1.20) times. Several studies reported a shift in triage dependent on the scale or nurse training (psychiatric v. emergency), but linkage to systemor patient-based outcomes was not made, limiting clinical interpretation. Conclusion: A small number of studies demonstrated that specialized care models can reduce hospitalization, return ED visits and length of ED stay. Future research must address the reality of the available health care infrastructure, while also determining the most relevant patientcentred outcomes to complement the existing literature on health-services outcomes. Keywords: pediatric mental health, specialized care models, emergency crowding

87 EXAMINATION of Stethoscope Contamination in the Emergency Department (EXSCITED) pilot study

Tang, PHP, Worster A., Srigley JA, Main C; Division of Emergency Medicine, McMaster University, Hamilton, ON

Introduction: The objective of this study was to determine the prevalence of Staphylococcus-contaminated stethoscopes belonging to emergency department (ED) staff and to identify the proportion of these that are Staphylococcus aureus or methicillinresistant S. aureus (MRSA). Methods: We conducted a prospective observational cohort study of bacterial cultures from 100 ED staff members’ stethoscopes at 3 EDs. Study participants were asked to complete a questionnaire. Results: Fiftyfour specimens grew coagulase negative staphylococci and 1 grew methicillinsusceptible S. aureus. No MRSA was cultured. Only 8% of participants, all of whom were nurses, reported cleaning their stethoscope before or after each patient assessment. Alcohol-based wipes were most commonly used to clean stethoscopes. A lack of time, being too busy and forgetfulness were the most frequently reported reasons for not cleaning one’s stethoscope in the ED. Conclusion: This study suggests that although stethoscope contamination rates in these EDs are high, the prevalence of pathogenic bacteria such as S. aureus or MRSA on stethoscopes is low. Keywords: stethoscope contamination, MRSA, infection control

88 FACTORS prolonging emergency department length of stay for patients undergoing successful DC cardioversion for acute atrial fibrillation: a prospective logistic regression analysis

Scheuermeyer F, Stenstrom R, Grafstein E, Innes G, McPhee J, Poureslami I; University of British Columbia, Vancouver, BC

Introduction: Emergency department (ED) length of stay (LOS) is a key performance indicator and prolonged LOS may contribute to ED crowding and access block. Atrial fibrillation is the most common dysrhythmia seen in the ED. We studied factors predicting increased LOS for patients discharged from the ED after successful DC cardioversion. Methods: Consecutive patients undergoing DC cardioversion at a single centre were prospectively enrolled. Those admitted to hospital or failing DC cardioversion were excluded from the LOS analysis but included for a safety analysis. Fortyeight predictor variables including comorbidities, outpatient medications, adverse events and all ED treatments (e.g., rate control, antiarrhythmics and procedural sedation) were documented. The primary outcome was ED LOS, while safety outcomes included prespecified adverse events and stroke, death or return visits to a regional ED within 7 days. Logistic regression models were used to predict which variables increased ED LOS. Generalized estimating equations were used to handle multiple correlated visits. Results: A total of 233 subjects were enrolled and analyzed for safety. Of those, 29 were admitted or failed cardioversion, leaving 204 for LOS analysis. Median LOS was 3.17 (IQR 2.20–4.23) hours. Multiple linear regression showed that initial ED use of rhythm medication increased LOS by 2.25 hours (95% CI 1.59–2.91, p < 0.001) while initial rate control medications increased LOS by 0.91 hours (95% CI 0.22–1.68, p = 0.01). Consulting a specialty service before DC cardioversion increased LOS by 4.48 hours (95% CI 3.40–5.47, p < 0.001). The regression model explained 47% of the variance in ED LOS (F7, 190 = 23.9; p < 0.0001). No deaths or strokes occurred within 7 days (95% CI 0.0%–1.3%). Conclusion: ED cardioversion and discharge appears safe. Modifiable contributors to prolonged ED length of stay in patients are initial use of rate or rhythm control medications, or cardiology consultation before initiating electrical cardioversion. Keywords: electrical cardioversion, atrial fibrillation, emergency crowding

89 EVIDENCE for the use of biomarkers in diagnosing acute myocardial ischemia: a systematic review for the 2010 guidelines for CPR and ECC

Lin S; University of Toronto, Toronto, ON

Introduction: The revised Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care will be published in 2010. As part of the revision process, this systematic review evaluated the use of biomarkers in diagnosing patients with acute coronary syndrome (ACS). Methods: A literature search in MEDLINE, Embase, Cochrane DSR, ACP Journal Club, DARE, CCTR, CMR, HTA and NHSEED for studies from 2004 to 2008 evaluating the value of biomarkers in diagnosing ACS was performed. Manuscripts were included based on criteria focusing on studies relevant to the diagnosis of ACS in patients presenting with symptoms of cardiac ischemia. The level of evidence and quality of each study were evaluated based on established criteria by the International Liaison Committee on Resuscitation. Studies “supported” the use of its respective biomarker if it had a sensitivity of greater than 95% or a specificity of greater than 92% combined with a sensitivity of greater than 90%. Results: A total of 359 studies were identified, of which 71 were included for final review. Eight studies supported cardiac troponin (Tn) testing alone in the diagnosis of acute myocardial infarction (AMI), but not unstable angina, when serum testing was drawn at least 6 hours after symptom onset, emergency department (ED) presentation or serially. No studies supported Tn testing outside of the ED or a short stay cardiac unit. Newergeneration sensitive Tn assays were more sensitive compared with conventional Tn assays when diagnosing AMI. Nine studies showed improved sensitivity when Tn was combined with an early marker (CKMB, ischemia-modified albumin or myoglobin). No studies supported the use of “novel” biomarkers in isolation of Tn. Conclusion: Cardiac Tn was more sensitive than other biomarkers when diagnosing AMI when testing occurred more than 6 hours after symptom onset, ED presentation or serially. Tn testing combined with an early marker of injury showed improved sensitivity. Newergeneration sensitive Tn assays were more sensitive than previous Tn assays in the diagnosis of AMI. Keywords: cardiac biomarkers, troponin, systematic review. Disclaimer: This review includes information on resuscitation questions developed through the C2010 Consensus on Science and Treatment Recommendations process, managed by the International Liaison Committee on Resuscitation (www.americanheart.org/ILCOR). The questions were developed by ILCOR task forces, using strict conflict of interest guidelines. In general, each question was assigned to 2 experts to complete a detailed structured review of the literature and complete a detailed worksheet. Worksheets are discussed at ILCOR meetings to reach consensus and will be published in 2010 as the Consensus on Science and Treatment Recommendations (CoSTR). The conclusions published in the final CoSTR consensus document may differ from the conclusions in this review because the CoSTR consensus will reflect input from other worksheet authors and discussants at the conference, and will take into consideration implementation and feasibility issues as well as new relevant research.

90 THE EFFECT of a brief emergency medicine educational intervention on the number of shifts worked by physicians in the emergency department

Vaillancourt S, Schultz SE, Leaver CA, Stukel T, Schull MJ; University of Toronto, Toronto, ON

Introduction: Some emergency departments (EDs) in Canada have increasing difficulty maintaining full physician staffing, and temporary ED closures sometimes occur. The Ministry of Health funded a brief Emergency Medicine Primer (EMP) course for physicians to upgrade or refresh skills so that they might increase their work in the ED. We sought to determine the effect of the EMP on the ED workload of physicians. Methods: A retrospective longitudinal study of ED work of 239 physicians in the 2 years before and 2 years after completing an EMP course in 2006–2008, compared with ED work of non-EMP physicians. Outcomes were the number of ED shifts per month, and the number of ED patients seen per month. We conducted 2 analyses: a before and after comparison of all EMP physicians, and a matchedcohort analysis matching each EMP physician to 4 non-EMP ED physicians based on sex, year of medical school graduation, rurality and pre-EMP ED workload. Results: Postcourse, EMPphysicians worked 0.7 (p = 0.0032) more ED shifts per month (13% increase) and saw 15 (p = 0.0008) more patients per month (17% increase) as compared with matched non-EMP physicians. The greatest increases were among EMP physicians who were younger, urban, with ED experience or in a high volume ED. The impact of the EMP on physicians with no prior ED experience, and those working in rural areas, was negligible. Conclusion: The EMP was associated with modest increases in ED workload in some physicians, especially younger ones in urban areas. Little effect was seen among rural physicians, or those working in rural areas. Keywords: emergency medicine primer, physician workload, physician productivity

91 VARIABILITY in computed tomography utilization by emergency physicians in 3 urban hospitals

Greenfield G, Dowling S, Wang D, Lang ES, Innes G; University of Calgary, Calgary, AB

Introduction: CT utilization by emergency physicians is variable. CT imaging adds cost, places demands on limited diagnostic imaging resources, prolongs ED length of stay and exposes patients to radiation injury or contrast reactions. If there is substantial practice variation, this has implications regarding diagnostic cost-effectiveness, patient safety and satisfaction, departmental efficiency and modeling behaviour for trainees. Our objective is to describe physician variability in the use of CT imaging in 3 urban Calgary EDs. Methods: This retrospective cohort study was conducted at 3 EDs in Calgary, Alta., over a 1-year period. Every patient visit and CT image was linked to a hospital site and ED physician. Our primary outcome was the physicianspecific CT imaging rate in CTAS level 3, defined as the proportion of these patients who had a CT scan during their ED visit. Secondary outcomes included the CT imaging rate for all patients (aggregate and stratified by physician), and the CT imaging rate stratified by hospital site and triage level. Results: During a 1-year evaluation period, 115 ED physicians treated 188 699 patients, including 93 510 (49.6%) in CTAS level 3. The overall CT rate for all patients was 12.8% (24 070/188 649) and the rate within CTAS-3 was 12.1% (11 313/93 510). Imaging rates varied by site and ranged from 0.9% in CTAS level 5 to 24.7% in CTAS level 1. The median CT rate for our physicians, within CTAS level 3, was 12.0% (IQR 10.0%–14.6%). In the CTAS-3 cohort, the lowest using physician scanned 1.8% of patients (19/1069) while the highest user scanned 25% of patients (229/915). Conclusion: There is profound variability in the utilization of CT scans among physicians working in similar settings with comparable patient groups. Efforts are required to understand this practice diversity and develop strategies to converge on a more rational costeffective approach to diagnostic utilization. Keywords: CT utilization, practice variation, administrative database research

92 CAN I SCRAP the CT chest? The internal validation of a CDR to ruleout major thoracic injury in major blunt trauma

Payrastre J, Upadhye S, Worster A, Sanaee L, Clayden R, Patterson H, Lin D; McMaster University, Hamilton, ON

Introduction: To validate the results of the SCRAP clinical decision rule derivation study. The derivation study found that in major blunt trauma patients, when the 5 scrap variables were normal there were no major thoracic injuries. The 5 variables used were abnormal chest XR, abnormal thorax palpation, abnormal chest auscultation, abnormal oxygen saturation (defined as < 95% on RA or < 98% on oxygen), and respiratory rate greater than 24. Methods: Data were retrospectively obtained from a chart review of all trauma patients presenting to a Canadian tertiary trauma care centre from April 2006 to March 2007. Patients were included if they had a CT chest at admission or had a documented major thoracic injury noted in the trauma database. Patients with penetrating injury or GCS less than 9 were excluded. Three data collectors (κ = 0.83) collected 28 data points, which were then analyzed for their predictive value. The data was analyzed using basic statistics and logistic regression analysis. Results: There were 180 patients that met inclusion criteria. The average ISS of these patients was 23. Once again, the SCRAP variables were found to be highly predictive of injury, and when none of these variables were present, no patients had major thoracic injury (Sens 100% [95.6%–100%], Spec 45% [33.9%–57.2%], NPV 100% [87.3%–100%]). Conclusion: This study confirms the result of the derivation study and retrospectively shows that the SCRAP rule is effective in predicting major thoracic injury in this population. The use of the SCRAP rule would have prevented 45.3% (34/75) of the negative CT chests without missing a major thoracic injury. The next step will be to validate the results of this study in a prospective, multicentre study. Keywords: blunt trauma, clinical decision rule, internal validation

93 SAFETY of assessment of patients with potential ischemic chest pain in an emergency department waiting room: a prospective comparative cohort study

Scheuermeyer F, Christenson J, Innes G, Grafstein E, Boychuk B, Yu E; University of British Columbia, Vancouver, BC

Introduction: Emergency department (ED) crowding has been associated with a variety of adverse outcomes. Current guidelines suggest that patients with potentially ischemic chest pain should undergo rapid assessment and treatment in a monitored setting to optimize the diagnosis of acute coronary syndrome (ACS). These patients may be at high risk of misdiagnosis and adverse events when evaluation is delayed because of ED crowding. In order to mitigate crowdingrelated delays, we developed processes that enabled emergency physicians to evaluate chest pain patients in the waiting room (WR) when all nursestaffed stretchers are occupied. The objective of this study was to investigate the safety of WR chest pain evaluation. Methods: This prospective comparative cohort study was conducted in a tertiary care ED. Explicit triage and waiting room evaluation processes were introduced. A total of 1107 patients with chest pain of potential cardiac origin were triaged either to a monitored bed (MB) or a WR chair, depending upon bed availability and triage judgment. After diagnostic evaluation, patients were followed for 30 days to identify the proportion of missed cases of ACS and other prespecified adverse events. Analysis was based on intention-to-treat. Results: A total of 804 patients were triaged to MB and 303 to WR evaluation. The WR had less cardiovascular risk factors than the MB group, although initial vital signs were similar. The rate of ACS, was 11.7% in the MB group and 7.6% in WR patients. There were no missed ACS cases in either the MB group (0%, 95% CI 0.0%–0.4%) or the WR group (0%, 95% CI 0.0%–1.0%). There were 32 adverse events in the MB group (4.0%, 95% CI 2.6%–5.3%) and 2 in the WR group (0.7%, 95% CI 0.0%–1.6%). Conclusion: An organized approach to triage and waiting room evaluation for stable chest pain patients is safe. Although waiting room evaluation is not ideal, it may be a viable contingency strategy for periods when ED crowding compromises access to care. Keywords: waiting room medicine, chest pain, patient safety

94 DEVELOPMENT of a CanMEDS-based multisource feedback evaluation tool for emergency physicians

Crowder K, Lalani M, Stempien J, Woods R; University of Saskatchewan, Saskatoon, SK

Introduction: Quality and patient safety initiatives are gaining momentum in health care. Additionally, Canada has been a leader in the development of outcome-based medical education with the CanMEDS competency framework, widely used for the assessment of resident trainees. However, there is a paucity of research applying CanMEDS competencies to practising physicians. The use of multisource feedback (MSF) has recently been adopted for physician evaluation. A MSF tool based on the CanMEDS competencies is a logical next step to ensure quality of patient care by practising emergency physicians (EPs), and to ensure faculty are proficient in the competencies they are expected to emulate to their trainees. Methods: The authors reviewed the literature and emergency medicine (EM) training requirements to develop a comprehensive list of evaluation criteria for each of the CanMEDS competencies. These criteria were subjected to 2 rounds of consensus. A 5-point Likert scale was constructed for Round 1 items, surveying all stakeholders (EPs, EM residents, offservice residents, consultants, emergency department [ED] nurses, ED pharmacists, ED unit clerks/aides, social workers, bed managers and emergency medical services personnel) with a cut-off median value of 3.5 for inclusion of criteria. Round 2 surveyed EPs to indicate which health care providers would be best able to evaluate each criterion, with a minimum 50% consensus cut-off value for evaluator selection. Results: The final tool included all valid evaluation criteria and the health care professionals considered to be qualified to evaluate each criterion. Conclusion: It is possible to develop a validated, specific, CanMEDS-based evaluation tool for practising EPs. This method can be used as a model by other physician groups for development of their own MSF tool, or the tool can simply be used by other EP groups. Further research should assess physician satisfaction with the MSF process and the impact of MSF tools on EP performance. Keywords: CanMEDS, multisource feedback, consensus methodology

95 UTILIZATION of 7 clinical decision units in Ontario’s pilot program

Vermeulen MJ, Leaver CA, Guttmann A, Rowe BH, Schull MJ; Institute for Clinical Evaluative Sciences, Toronto, ON

Introduction: Clinical decision units (CDUs) within emergency departments (EDs) have been proposed as a means of improving patient flow by reducing length of stay (LOS) and/or the need for admission in selected patient groups. In 2008, the Ontario Ministry of Health introduced physician funding for CDUs in 7 EDs. We undertook a preliminary evaluation of this pilot program. Methods: We conducted a retrospective analysis of unscheduled ED visits at all 7 CDU pilot sites in the first 11 months of implementation (October 2008 to September 2009) using routinely collected administrative data and key informant interviews, examining trends in CDU utilization and occupancy (measured as the percentage of designated CDU bedhours occupied in each 24-hour period) as well as ED visit characteristics of CDU and non-CDU patients. Results: CDU admissions comprised 3.2% of ED visits (range across sites 1.8%–4.7%); the average daily number of CDU patients was 4.2 (2.2–5.7); occupancy was 47.6% (22.4%–72.8%). CDU patients were older (mean [SD] age 56.4 [21.6] v. 44.5 [24.5]) and were typically triaged as higher acuity (43.5% v. 21.3% resuscitation/ emergent) than non-CDU patients. For CDU patients, the median (IQR) and 90th percentile ED LOS, including CDU LOS, were 12.9 (8.8–19.4) and 26.3 hours, respectively; median (IQR) CDU LOS was 8.2 (4.8–14.1) hours. CDU patients were more likely to be admitted than non-CDU patients (19.8% v. 15.2%). The most common indications for CDU admission were chest pain (17.2%) and abdominal pain (9.0%). Among non-CDU patients, up to 6.9% were potentially eligible for admission to the CDU (based on a predefined indication for CDU admission and an ED LOS of > 6 hours). Conclusion: CDU utilization varied substantially and there appears to be potential for increasing use. In general, CDU patients were older and more acute/complex than non-CDU patients. The next phase of this study will examine the overall impact of the pilot program with respect to ED crowding, ED revisit and short-term admission rates. Keywords: clinical decision units, practice variation, emergency crowding

96 PEDIATRIC pain practice variation among emergency doctors: a Canadian survey

Ali S, Curtis SJ, Johnson DW, Logue EP, Vandermeer B, Newton AS; University of Alberta, Edmonton, AB

Introduction: Eighty percent of all emergency department (ED) visits involve pain, yet pain treatment is suboptimal, especially in children. This study aimed to describe (a) pediatric pain management practices across Canada and (b) the presence of policies that may facilitate or hinder pain management. Methods: The study tool design was based on expert opinions and measurement literature. The Pediatric Emergency Research of Canada database was used to contact physicians (MDs), using a modified version of Dillman’s Total Design Survey Method. Results: The response rate was 68% (139/206). Greater than three-quarters of responding MDs are 31–50 years old, 56% have a pediatric emergency fellowship and more than one-half have 10 years or less of practice. Almost all pain screening in EDs occur at triage. Less than one-half of MDs note mandatory pain score documentation, and one-third indicate absence of standing orders for pain treatment. Ibuprofen and acetaminophen are almost equally prescribed in the pediatric ED for mild–moderate pain. Codeine is prescribed for otitis media by 23% of MDs, and by 75% for burns. Close to one-half of urinary catheterizations and IV starts are done without any analgesia. Ketamine is the most common agent for procedural sedation, with less than 10% using propofol. Pacifier use is the most common nonpharmacologic intervention. Training background and increasing age of MDs affects likelihood of using various nonpharmacologic therapies (p = 0.01) or topical anesthetics (p = 0.02 for training, p = 0.04 for age). Female MDs are more likely to allow breastfeeding during painful procedures (p = 0.0007). Greater than 90% of MDs allowed parents to remain present for painful procedures; however, parents often opt out. Conclusion: Despite good evidence, many pharmacological and nonpharmacological interventions are underused. Ibuprofen is used as frequently as acetaminophen for mild–moderate pain. There is a substantial evidence–practice gap in the management of children’s pain by pediatric emergency physicians. Keywords: pediatric pain management, practice variation, survey research