2010 CAEP/ACMU Scientific Abstracts - Oral Presentations: 1-23
CAEP Abstracts
CJEM 2010;12(3):229-278
Oral presentations
1 TEST characteristics of ultrasonography for the detection of pneumothorax: a systematic review and meta-analysis
Alrajhi K, Woo MY, Vaillancourt C; Department of Emergency Medicine, The Ottawa Hospital, Ottawa, ON
Introduction: We sought to determine the test characteristics of ultrasound (US) in adult patients clinically suspected of having a pneumothorax (PTX), and how it compares to chest xray (CXR) using CT scan or release of air on chest tube placement as the gold standard. Methods: For this systematic review and meta-analysis, we searched MEDLINE, Embase and reviewed the reference list of the included articles and recent reviews. Two independent investigators used standardized forms to review papers for inclusion, quality (QUADAS tool) and data extraction. We included English language prospective studies comparing US to CT scan or release of air on chest tube placement in adult patients suspected to have PTX. We calculated interobserver agreement for study selection and evaluated clinical and quality homogeneity before meta-analysis of data. Results: Our search strategy identified 570 papers, 21 were selected for full review (κ = 0.89) and 8 papers met all including criteria (κ = 0.81). CXR data was available for 864/1048 patients evaluated with US. US was performed by emergency physicians (EPs) in 4/8 papers (606/1048 patients). PTX was traumatic in 767/1048 patients and iatrogenic in 281/1048 patients. All but 1 study used a combination of lung sliding and comet tail signs for the detection of PTX. Among 1048 patients, US was 90.9% sens (95% CI 0.87–0.94), 98.2% spec (95% CI 0.97–0.99), with 94.4% PPV and 97.0% NPV. CXR was 53.9% sens (95% CI 0.47–0.61), 97.1% spec (95% CI 0.96–0.98), with 85.2% PPV and 87.2% NPV. EPs were 95.3% sens (95% CI 0.90–0.98), 99.0% spec (95% CI 0.97–1.00), with 97.0% PPV and 98.5% NPV. Among 767 trauma patients, US was 90.2% sens (95% CI 0.85–0.94), 98.8% spec (95% CI 0.97–1.00), with 95.9% PPV and 97.0% NPV. Conclusion: The performance of US for the detection of PTX is excellent and is better than CXR. Considering the rapid access to bedside US and the excellent performance of this simple test, this study supports the routine use of US for the detection of PTX. Keywords: emergency ultrasound, pneumothorax, systematic review
2 THE EFFECT of a bolus dose of etomidate on cortisol levels, mortality, and health care services utilization: a systematic review
Hohl CM, Kelly-Smith CH, Titus C, Yeung TC, Sweet DD, Doyle-Waters MM, Schulzer M; University of British Columbia, Vancouver, BC
Introduction: Etomidate is a widely used induction agent for rapid sequence intubation (RSI). Our objective was to synthesize the evidence on the effect of a bolus dose of etomidate on adrenal function, mortality and health services utilization compared with other induction agents. Methods: We developed a systematic search strategy and applied it to 10 electronic bibliographic databases. We hand searched medical journals, conference proceedings, grey literature, bibliographies of relevant literature, and contacted content experts for studies comparing a bolus dose of etomidate with other induction agents. Studies reported in English, French and German were included if they reported adult data comparing the effect of a bolus dose of etomidate to another rapidly acting intravenous induction agent. Retrieved articles were reviewed and data was abstracted in duplicate using standardized forms. Data was pooled using the random effects model if at least 4 clinically homogenous studies of the same design reported the same outcome measure. Results: Of the 3083 articles found, 20 met our inclusion criteria. Pooled mean cortisol levels were lower in patients induced with etomidate compared with those induced with other agents between hours 1 and 4 postinduction. The differences varied from 6.1 µg/dL (95% CI 2.4–9.9 µg/dL, p = 0.001) to 16.4 µg/dL (95% CI 9.7–23.1 µg/dL, p < 0.001). No difference was observed after 4 hours. None of the studies reviewed, nor our pooled estimate (odds ratio 1.14, 95% CI 0.81– 1.60) showed a statistically significant effect on mortality. Only one study reported longer ventilator, intensive care unit (ICU) and hospital lengths of stay (LOS) in patients intubated with etomidate, the other studies reported no difference. Conclusion: The available evidence suggests that etomidate suppresses adrenal function transiently without demonstrating a significant effect on mortality. However, no studies to date have been powered to detect a difference in hospital, ventilator or ICU LOS, or in mortality. Keywords: etomidate, adrenal function, systematic review
3 ST. Michael’s Hospital Basic Life Support Termination of Resuscitation Guideline Implementation Trial (TORIT)
Morrison LJ, Eby D, D’Souza P, Zhan C, Kiss A, Welsford M, Loreto C, Arcieri V, Prowd C, Pilkington M, Dodd T, Scott J, Mooney E, Reichl R, Verdon J, Waite T, Verbeek PR; St. Michael’s Hospital, Toronto, ON
Introduction: This implementation study was designed to evaluate the transport rate of outofhospital cardiac arrest (OHCA) patients when the universal termination of resuscitation (TOR) guideline was applied in 8 services. TOR is recommended for arrests not witnessed by EMS, when there is no return of spontaneous circulation, and no shocks delivered. Secondary aims were to report errors in guideline application and comfort in application. Methods: This prospective multicentre observational trial was conducted from January 2006 to September 2008 across Ontario. Adult OHCA patients of presumed cardiac etiology treated by defibrillator only trained paramedics were eligible. Providers contacted the delegating physician when termination was recommended. Both provider and physician completed a data collection form indicating their comfort on a 5-point Likert scale. Results: Of 2421 OHCA and 953 patients were eligible for TOR guideline application. The TOR guideline was followed in 755 cases resulting in 388 terminations and 367 transports. There were no errors in guideline application. In 198 cases where the TOR guideline was not followed, paramedics cited 241 reasons: family distress (56), short time intervals (54), patient age (13) and public venue (10) accounted for 55%. Paramedics cited discomfort 28 times (11%). In 14 cases they were unable to establish telephone contact and in 30 cases the guideline recommended termination and the physician chose to transport (12.4%). All these 198 TOR eligible patients died in hospital. When the TOR guideline was applied the transport rate was 48.6%, which is significantly different than the previously reported transport rate of 100% when the TOR guideline was not applied (p < 0.001). Both providers and physicians were very comfortable (median [IQR] of 5 [4–5]; p < 0.001). Conclusion: The transport rate is significantly reduced when the TOR guidelines are followed, error rates are minimal and most providers were comfortable with following the guideline recommendations. Keywords: termination of resuscitation, implementation research, prehospital, cardiac arrest
4 THE IMPACT of a triage nurse ordering on ED overcrowding: a systematic review
Rowe BH, Guo XY, Wong L, Villa-Roel C, Schull MJ, Vandermeer B, Ospina M, Holroyd BR, Bullard M Innes G; University of Alberta; Edmonton, AB
Introduction: Emergency department (ED) overcrowding is a crisis for many urban and highacuity hospitals and effective interventions are urgently needed. The role of triage nurse order (TNO) sets has been proposed as a method to reduce ED delays. The aim of this study was to evaluate the evidence for TNO and its influence on ED overcrowding. Methods: Electronic databases (Cochrane Central Register of Controlled Trials, MEDLINE, Embase, Web of Science, HealthSTAR, Dissertation Abstracts, ABI/INFORM Global), controlled trial registry websites, conference proceedings, study references, experts in the field and correspondence with authors were used to identify potentially relevant TNO studies. Interventional studies in which TNO was used to influence ED overcrowding metrics (length of stay [LOS], left without being seen) were included in the review. Two reviewers independently assessed the citation relevance inclusion and study quality. Weighted mean differences (WMD) were calculated and reported with corresponding 95% confidence intervals (CIs). Results: From 14 716 potential relevant studies, 6 were included in the systematic review. All studies were described as trials, 5 were journal publications and 1 was a thesis. The trials were completed in various locations; all but one were single-centred ED studies. All studies were rated as weak due to poorly reported methods. ED LOS was reduced by 19.4 minutes (95% CI 15.7–23.1) for patients with fracture/injury status; however, no improvement was shown for patients without an injury/fracture (WMD 0.78 min; 95% CI 5.22–6.78); there was considerable heterogeneity (I2 = 88%). Conclusion: Overall, triage nurse ordering appears to be an effective intervention to reduce ED LOS for injury/ruleout fracture cases. Medical management of patients using TNO approaches did not seem to reduce LOS. Nurses are capable of employing this strategy; one option would be to limit TNOs to injury cases. Keywords: emergency crowding, systemic review, triage nurse order
5 PROSPECTIVE evaluation of the ABCD and ABCD2 scores in a Canadian ED setting
Hall C, Oczkowski W; Calgary Health Region, Calgary, AB
Introduction: The ABCD and ABCD2 scores were derived to predict stroke risk after TIA. Subsequent prospective validations have demonstrated lukewarm results. We sought to prospectively evaluate these scores in a Canadian ED setting. Methods: Prospective cohort study of all suspected TIAs referred from the ED for outpatient follow-up at the Hamilton Health Sciences Stroke Prevention Clinic (SPC) from Nov. 1, 2006, to Oct. 31, 2007. ED physicians completed a referral form including all elements of the ABCD2 score. Patients were then evaluated in the clinic by stroke neurologists. The primary outcome of interest was completed stroke at 7 days, with a secondary outcome of completed stroke at 30 days. Results: Of the patients, 285 attended the SPC for an outpatient appointment after being referred from the ED; 238 (83.5%) of these consented and were enrolled in the study; 113 (47.5%) patients were diagnosed as having had true TIA events based on the neurologist assessment in the clinic. Three patients suffered a stroke within 7 days of their initial presentation (1.3% of all patients; 2.7% of all patients diagnosed with TIA). One of these patients had a score of less than 5 on both the ABCD and ABCD2 at presentation to the ED. Five patients (2.1% and 4.4%) had a completed stroke within 30 days. Three of these presented with scores less than 5 on both the ABCD and ABCD2. The areas under the ROC curves for the ABCD score for stroke at 7 and 30 days were 0.74 (95% CI 0.40–1.00) and 0.54 (0.24–0.84), respectively. For the ABCD2 score the AUCs for the ROC curves were 0.82 (0.58–1.00) at 7 days and 0.63 (0.35–0.91) at 30 days. Conclusion: More than half of the patients referred to the SPC for evaluation of suspected TIA were ultimately diagnosed with a nonischemic etiology for their symptoms. Of those patients diagnosed with a true TIA, 2.7% experienced a stroke within 7 days and 4.4% within 30 days. The recommended cut-off score of 5 points or greater failed to predict a significant proportion of the strokes which occurred within 7 and 30 days of the index presentation. Keywords: transient ischemic attack, risk stratification, clinical decision rule
6 HYPERTONIC saline in acute traumatic brain injury? A systematic review and meta-analysis
Barbic D, Barbic S, Lang E, Jayaraman D; McGill Emergency Medicine Residency Program, Montréal, QC
Introduction: Recent studies have suggested a benefit with the use of hypertonic saline (HTS) for the treatment of raised intracranial pressure (ICP) secondary to traumatic brain injury (TBI). However, the magnitude of the potential benefit conferred by HTS treatment versus alternative interventions such as mannitol (M) or lactated Ringer’s (LR) is unknown. We examined the benefits and risks of HTS compared with M or LR in the treatment of raised ICP due to TBI. Methods: We searched the MEDLINE, Embase, ISI Web of Science, CINAHL, Scopus, Cochrane Library and Cochrane Injuries Group databases, as well as reference lists of articles and proceedings of major conferences. We contacted trial authors and experts in critical care. Randomized controlled trials were included in which subjects with severe primary TBI were assigned to treatment or control groups (placebocontrolled, no drug or different drug). Trials with a crossover design were excluded. All reviewers agreed on trial eligibility. Two reviewers independently extracted data. Missing and unpublished data were obtained from the trials’ authors. The primary outcome for this review was mean ICP at 120 minutes postinfusion. Secondary outcomes included mortality, and adverse events. Results: We identified 5 trials (n = 143) that had acceptable methodological quality to pool in a meta-analysis, but comparable data for the primary outcome was only available for 3 trials (n = 86). Data for mortality was available for 4 trials (n = 120). The mean ICP at 120 minutes postinfusion for HTS versus controls was –2.604 mm Hg (95% CI –603 to –3.525; p = 0.006; I2 = 13.2% REM). The difference in mortality in patients treated with HTS versus controls was 10.8% versus 30.6% (RR 0.40, 95% CI 0.17 to 0.95; p = 0.04; I2 = 0% REM). Conclusion: HTS reduces mean ICP at 120 minutes postinfusion compared with control solutions, and leads to a marked improvement in mortality. HTS may benefit patients with raised ICP due to acute TBI. Further research is warranted to determine the optimal infusion tonicity and infusion time. Keywords: hypertonic saline, traumatic brain injury, systematic review
7 EMERGENCY department targeted ultrasound for the detection of hydronephrosis
Bruder E, Rang L, Wall B, McKechnie K, Moore K; Queen’s University, Kingston, Ont.
Introduction: CT imaging of patients suspected of ureterolithiasis entails radiation exposure and prolongs ED length of stay. Emergency department targeted ultrasound (EDTU) allows rapid bedside evaluation of hydronephrosis. We compare EDTU to formal imaging for the detection of hydronephrosis. Methods: Patients undergoing CT or formal ultrasound of their abdomen were imaged using EDTU within 60 minutes of their test using a 3.5 MHz curvilinear ultrasound probe at the bedside. Physicians investigating these patients were experienced EDTU operators who underwent additional standardized training before their involvement in this study. Research assistants identified eligible patients, and obtained informed consent. Sample size was determined based on a power calculation. The presence of hydronephrosis was determined in a binary fashion (present or absent) and compared with formal ultrasound or CT. EDTU operators were blinded to results of formal imaging. Results: Of the patients, 196 were imaged with EDTU, 38 of which had hydronephrosis; 152 patients were evaluated with CT, and 44 were evaluated with formal ultrasonography. EDTU demonstrates moderate accuracy when compared with CT or formal ultrasound with a sensitivity of 82% (95% CI 65%–92%), specificity of 89% (95% CI 83%–93%), positive likelihood ratio of 7.7 (95% CI 4.7–12), and a negative likelihood ratio of 0.21 (95% CI 0.1–0.4). EDTU identified all cases of hydronephrosis seen on formal ultrasound, and only missed 7 cases identified on CT. Of 6 EDTU operators, 5 had greater than 90% accuracy compared with formal imaging, and 1 was 79% accurate. Conclusion: EDTU compares favourably to formal ultrasound and CT, supporting its use in the bedside evaluation of hydronephrosis. EDTU is a rapid test without radiation exposure that can aid in the diagnosis of hydronephrosis and the exclusion of other life threatening conditions. Keywords: emergency ultrasound, renal colic, hydronephrosis
8 AN INTEGRATED rapid assessment zone and waiting room care initiative improves throughput for CTAS-III and CTAS-IV patients
Koonar H, Kathy Taylor K, Lang E, Mercuur L, Wang D, Dowling S, Innes G; University of Calgary, Calgary, AB
Introduction: The evaluation of lower acuity patients who present to the emergency department (ED) is often hampered by a lack of physical space and inefficient care. The waiting room care (WRC)/rapid assessment zone (RAZ) initiative is a multidisciplinary reengineering of processes of care for ambulatory patients. The goal of this study was to determine its impact on measures of throughput and quality for CTAS level III and IV patients. Methods: Using administrative databases, measures of efficiency and accessibility were compared during a 5-month period when WRC/RAZ was in place against equal lengths of time immediately prior (IP) to WRC/RAZ implementation and during the same period 1 year earlier (1YE). All registered patients were eligible. Confounding variables including total volume and admission rates were incorporated into the analysis; nursing staffing levels were noted as well. Primary outcomes were length of stay (LOS), unplanned revisits and left without being seen (LWBS). Results: Of all CTAS-III and -IV patients, 29.3% were treated in WRC/RAZ during the intervention period. CTAS-III and -IV patients who presented to the ED while WRC/RAZ was operational experienced a mean LOS of 3.95 hours while the IP and 1YE patients experienced LOS of 4.47 and 4.27 hours, respectively (p < 0.0001 for both comparisons). The improvements in throughput yielded a reduction in LWBS (8.5% during WRC/RAZ, 11.1% in IP and 11.6% in 1YE time periods, p < 0.0001). Unplanned revisits within 72 hours were unaffected by WRC/RAZ (7.4% v. 7.3% and 7.6% for IP and 1YE, respectively, p = NS). Potential confounders did not influence the observed differences, and overall volume of CTAS-III and -IV actually increased during WRC/RAZ as did nurse staffing by 5.6%. Conclusion: WRC/RAZ implementation has led to significant improvements in throughput without compromising quality of care for CTAS-III and -IV patients. The WRC/RAZ model merits study in other centres to determine if its impact is institutionspecific. Keywords: emergency crowding, waiting room care, operations research
9 ATTITUDES and factors associated with successful CPR knowledge transfer in an older population most likely to witness cardiac arrest: a national survey
Vaillancourt C, Kasaboski A, Charette M, Islam R, Brehaut J, Grimshaw J, Osmond MH, Wells GA, Stiell IG; University of Ottawa, Ottawa Health Research Institute, Ottawa, ON
Introduction: Bystander CPR rates are lowest at home, where cardiac arrest is most likely to occur. We sought to identify barriers and facilitators to performing CPR and taking CPR training among older citizens. Methods: We conducted this national randomdigitdial telephone survey among independentliving individuals aged 55 and older from urban and rural settings. We randomly assigned participants to complete the CPR training or performance survey. We developed the interview guides based on the constructs of the Theory of Planned Behaviour, which elicits salient attitudes, social influences, and barriers and facilitators potentially influencing CPR training and performance. The survey instruments were developed after completion of 24 iterative interviews, verified for internal validity, and piloted before implementation. Recruitment took place between December 2009 and January 2010 until reaching 100% of the required sample size. Results: Demographics of the 226 interviewees: 48.2% older than 65, 58.0% women, 53.5% married, 50.9% retired, 19% completed high school, 59.7% CPR trained (most > 10 years ago), 82.7% never performed CPR. Intention to take CPR training in the next 6 months was slight. Most believe that CPR is an important skill to have, but many are afraid to forget or doing it wrong. Leading social influences are friends and spouse. Many believe courses should be free, and that more publicity is necessary. Intention to perform CPR if necessary was high. Most know that CPR is beneficial, but many believe they could harm the victim, catch a disease or get sued. People are unlikely to perform CPR if perceived to be less competent than other bystanders, but could do it if prompted by 9-1-1. Conclusion: This is the first national survey conducted using the Theory of Planned Behaviour to identify key facilitators and barriers to CPR training and performance. These findings will inform the design of interventional trials to improve bystander CPR and survival rates for cardiac arrest. Keywords: knowledge transfer, cardiopulmonary resuscitation, survey research
10 PLANNING and effectiveness of an H1N1 surge protocol and lessons to be learned from implementation delay
Lam TV, Bullard MJ, Holroyd BR, Villa-Roel C, Vester M, Latoszek K, Meurer DP, Rowe BH; University of Alberta, Edmonton, AB
Introduction: For the second wave of the 2009 H1N1 Pandemic a graded provincial surge capacity response plan was developed to trigger appropriate system responses defined by patient volume increases of 10%–20%, 21%–40% or greater than 40%; or staff absenteeism of 10%–20%, 20%–30% or greater than 31%, respectively. Proposed responses included ED influenza-like illness (ILI) screening, opening influenza assessment clinics (IAC), public communication, EMS protocol changes and utilitarian care principles. This study looks at adherence to the plan and lessons learned. Methods: Annual differences in regional ED visit volumes were determined for the period between Sep. 15 and Dec. 15, 2008 and 2009, to identify surge levels and timing of response triggers. Data were gathered from the ED information system (EDIS) database, ED staff sick call records and the IAC records. Descriptive data are presented as proportions for categorical variables and medians with interquartile range (IQR) for continuous variables. Results: ILI positive ED patient volumes began rising Oct. 17, with ED visits 9.3% above normal on Oct. 19 and then returning to baseline over the next 3 days. By Oct. 25 presentations had risen 16% above normal and peaked at 42% on Oct. 28. Volumes remained greater than 10% above baseline until Nov. 11 and were greater than 20% above baseline on 6 of those 15 days. Over the same period, average daily sick leave was 6.3% presurge, 10.1% surge (11 d > 10%), and 5.6% postsurge. The IAC opened Oct. 30 and closed on Nov. 23; from Oct. 30 to Nov. 10 IAC received 248.5 (IQR 223.5–302.5) patients per day. Conclusion: Despite careful planning and use of EDIS data for monitoring and reporting, operational responses could have been more effective, implemented more efficiently and initiated earlier. The early lack of alternative destinations for care and delayed public educational announcements resulted in a significant number of minor ILI positive patients presenting to EDs which was eventually mitigated by the IAC opening. Keywords: H1N1, influenza assessment clinics, pandemic planning
11 UTILITY of the ABCD and ABCD2 scores in identifying true TIA events in the emergency department
Hall C, Oczkowski W; Calgary Health Region, Calgary, AB
Introduction: Distinguishing true transient ischemic attacks (TIAs) from neurologic mimickers can be difficult in the ED. The ABCD and ABCD2 scores have been derived to predict stroke risk after TIA. We sought to determine if they can also help to diagnose true TIAs in the ED. Methods: Prospective cohort study of all suspected TIAs referred from the ED for outpatient follow up at the Hamilton Health Sciences Stroke Prevention Clinic (SPC) from Nov. 1, 2006, to Oct. 31, 2007. ED physicians completed a referral form including all elements of the ABCD2 score. Patients were evaluated in the clinic by stroke neurologists and assigned a diagnosis of TIA versus non-TIA event based on history, examination, and imaging studies performed at the neurologists’ discretion. The sensitivity, specificity and positive and negative likelihood ratios were calculated for the ABCD and ABCD2 scores for a diagnosis of TIA. ROC curves were derived for both scores. Results: A total of 329 patients were referred from the ED for evaluation; 44 (13.4%) did not attend their appointment. Of the remaining 285 patients, 238 (83.5%) consented and were enrolled in the study; 113 (47.5%) patients were diagnosed as having had a TIA. The area under the ROC curves for the ABCD and ABCD2 scores were 0.62 (95% CI 0.54–0.68) and 0.63 (0.55– 0.69), respectively. Using a cut-off point of 5 points or greater for a diagnosis of TIA, this yielded the following measures for the ABCD score: LR (+) 1.90 (1.28–2.80), LR (–) 0.74 (0.62–0.89), sensitivity 42.5% (33.2%–52.1%) and specificity 77.6% (69.3%–84.6%). The performance of the ABCD2 score was nearly identical for each of these measures when using a cut-off score of 5 or greater for a diagnosis of TIA. Conclusion: More than half of the patients referred from the ED to the SPC for evaluation of suspected TIA were ultimately diagnosed with a nonischemic etiology for their symptoms. Neither of the ABCD or ABCD2 scores performed well enough to be recommended as a useful tool to assist ED physicians in distinguishing true TIAs from nonischemic mimickers. Keywords: transient ischemic attack, risk stratification, clinical decision rules
12 ONE-year outcomes of patients undergoing electrical cardioversion for atrial fibrillation in the emergency department: a prospective analysis
Scheuermeyer F, Grafstein E, Innes G, Stenstrom R, McPhee J, Poureslami I; British Columbia
Introduction: While the short-term (< 7 d) safety of electrical cardioversion (DC CV) for emergency department (ED) patients with atrial fibrillation (AF) have been established, the long-term outcomes have not been reported. Methods: A prospective 2-centre cohort of consecutive ED patients undergoing DC CV for atrial fibrillation between Apr. 1, 2006, and Dec. 31, 2008, was enrolled. AF patients with acute underlying illnesses or those deemed hemodynamically unstable were excluded. Information collected included patient demographics, cardiovascular risk factors (in the form of a CHADS 2 score), vital signs, EKGs, medical treatment, consultations, and admissions. Procedural sedation data collected included patient ASA class, medications given, success of cardioversion and prespecified adverse events. This cohort was probabilistically linked with both a regional ED database and the provincial health registry to determine which patients had a subsequent ED visit or hospital admission, stroke or thromboembolic event, or died within 1 year. Data was analyzed by descriptive methods. Results: During the study period, 244 patients were enrolled and 22 excluded. Of the 222 eligible patients, 167 (70%) had a CHADS 2 score of 0 and 201 (84%) had prior AF. All those who converted to NSR were discharged from the ED, but 27 patients failed DCCV (11.3%, 95% CI 7.3%–15.4%) and 14 were admitted (5.9%, 95% CI 2.9%–8.9%). During 1-year follow-up, 1 patient died of unrelated causes (95% CI 0.0%–1.3%) and no patients had a stroke or thromboembolic event (95% CI 0.0%–1.3% for both). 13 patients (5.5%, 95% CI 2.6%–8.4%) had an ED adverse event that did not change disposition. 116 (52%) patients had 270 ED visits and 19 admissions in the following year. Conclusion: In this lowrisk group of ED patients with AF, electrical cardioversion appears to have few long-term complications. Keywords: atrial fibrillation, cardioversion, prognostic study
13 PROGRESSION of EM competence: Delphi method for derivation and validation of milestones
Frank JR, Choi S, Wiesenfeld L, Nussbaum C, Clark E, Weitzman B, Elder B, Yeung M, Kapur A, Greenberg G, Vaillancourt C, Johns C, Stiell I; Department of Emergency Medicine, University of Ottawa, Ottawa, ON
Introduction: Residency education is evolving to be more competency-based and organized around milestones that define progression of specialist ability. However, there are currently no published milestones to guide Canadian emergency medicine (EM) residency teaching, learning or assessment. Methods: We employed a modified Delphi method and the CanMEDS roles as applied to EM to develop a framework of descriptive competency markers for residency education. Starting with the Royal College of Physicians and Surgeons of Canada’s CanMEDS-based EM objectives of training, we engaged our group of EM educators (n = 19) to define the essential milestones for each CanMEDS domain for years 1 to 5 of residency education. These were then validated by the larger group academic EM physicians (n = 51) in serial iterations until consensus. Results: We achieved consensus on a “Progression of EM Competence” milestones framework after 7 iterations of the group process. The final framework divided the 7 CanMEDS roles further into 11 applicable horizontal domains of EM (knowledge and clinical reasoning, procedures, communicator, collaborator, health advocate, manager, teaching, lifelong learning, critical appraisal, research and professional). Postgraduate years 1 and 2 included 43 milestones, PGY3 included 44 milestones, and PGY 4 and 5 included 44 milestones. Conclusion: We developed and validated a novel “Progression of EM Competence” milestones framework suitable to guide teaching, learning, and assessment of Canadian training. All EM training programs should consider adopting a framework like this one. Keywords: clinical competence, CanMEDS, Delphi methodology
14 OUTCOMES of patients who present to the emergency department because of an adverse drug event
Hohl CM, Nosyk B, Zed PJ, Kuramoto L, Brubacher JR, Sheps S, Abu-Laban RB, Peter S, Loewen PS, Sobolev B; University of British Columbia, Vancouver, BC
Introduction: Adverse drug events (ADEs) are a common cause of preventable nonsurgical adverse events. Drug related visits (DRVs) — defined as ED visits due to an ADE — are a leading cause of adult emergency department (ED) visits. Our objectives were to compare mortality, health services utilization and cost of care in patients with and without DRV. Methods: This prospective observational study enrolled adults presenting to a tertiary care ED. Pharmacists evaluated all patients for DRVs using standardized algorithms. Emergency physicians (EPs), blinded to the pharmacist opinion, were interviewed at the end of each shift to determine if they felt the patient’s ED visit was a DRV. An independent committee reviewed and adjudicated cases in which the EP and pharmacist assessments were discordant or uncertain to establish a criterion standard. Data from the index ED visit were linked to vital statistics and administrative health utilization and cost of care data using provincial health numbers. Patients were followed for 6 months after the index ED visit. Results: Of 1000 enrolled patients, 122 (12.2%, 95% CI 10.3%– 14.4%) were diagnosed with a DRV. We found no difference in mortality during the follow-up period among patients who presented with or without a DRV. The adjusted odds of being hospitalized on any day during follow-up were 1.51 (95% CI 1.41–1.61, p < 0.001) greater in the DRV group than the non-DRV group. The adjusted median monthly cost of care was 1.88 times higher (95% CI 1.16– 3.05, p = 0.010) for DRV than non-DRV patients. Conclusion: Patients with a DRV spend more time in hospital, and incur greater health care costs in the 6 months following the index ED visit than patients presenting for other reasons. ADEs warrant further investigation to explore the development of preventive strategies. Keywords: adverse drug events, drug related visits, patient safety
15 FACTORS associated with the successful recognition of agonal breathing and cardiac arrest by 9-1-1 communications officers: a qualitative iterative survey
Jensen JL, Vaillancourt C, Tweedle J, Kasaboski A, Charette M, Grimshaw J, MD, Jamie Brehaut J, Osmond MH, Wells GA, Stiell IG; University of Ottawa, Ottawa Health Research Institute, Ottawa, ON
Introduction: Agonal breathing can delay recognition of cardiac arrest and delivery of CPR instructions by 9-1-1 communication officers (COs). We sought to identify barriers and facilitators to COs ability to recognize cardiac arrest and agonal breathing. Methods: We conducted semistructured qualitative interviews with a purposeful sample of COs from 4 Canadian provinces. We developed an interview guide based on the constructs of the Theory of Planned Behaviour, which elicits salient attitudes, social influences, and behavioural control potentially influencing COs ability to recognize agonal breathing as a symptom of cardiac arrest. Interviews were recorded, transcribed verbatim and analyzed until data saturation was achieved. Two independent reviewers performed inductive analyses to identify emerging categories and themes, and ranked them by way of consensus. Results: We interviewed 24 COs from Ontario, Quebec, New Brunswick and Nova Scotia (67% female, median 9.5 years’ experience, 92% full time employees, 33% had paramedic training). Leading attitudes were as follows: 1) afraid to cause harm by teaching CPR if unsure victim is in cardiac arrest; 2) agonal breathing is an early sign of death; and 3) dispatchassisted CPR instructions can improve survival. Leading social influences came from: 1) management/quality assurance staff; 2) other health care providers; and 3) national cardiovascular organizations. Behavioural control was the construct most associated with COs’ ability to recognize cardiac arrest including: 1) adherence to existing scripted protocol; 2) poor caller description of breathing pattern; and 3) lack of education about agonal breathing. Conclusion: This is the first qualitative study using the Theory of Planned Behaviour to identify key facilitators and barriers for cardiac arrest recognition by 9-1-1 COs. These findings will inform the design of a national survey, and interventional trials to improve the diagnostic accuracy of COs and the efficient delivery of CPR instructions. Keywords: cardiopulmonary resuscitation, agonal breathing, qualitative research
16 CONSENSUS on evidence-based quality of care indicators for Canadian emergency departments
Leaver CA, Hatcher C, Guttmann A, Rowe BH, Vermeulen M, Schull MJ; Institute for Clinical Evaluative Sciences, Toronto, ON
Introduction: Currently, no standard or widely accepted quality of care and patient safety measures for ED care exist in Canada, thus hampering efforts for common measurement, crossjurisdiction comparisons and evaluation of interventions to improve care. The purpose of this project was to select and prioritize from among existing quality of care indicators a parsimonious, evidence-based set of indicators for Canadian EDs. Methods: Nationally representative clinical, administrative and quality experts participated in a modified Delphi panel and nominal review process over 2008/09. Expert panelists reviewed and prioritized candidate indicators identified in a comprehensive literature review and environmental scan; indicators were rated on dimensions of scientific soundness (evidence-based link to outcomes) and relevance/importance to patients or providers. A steering committee selected the final indicator set based on expert panel ratings (though revisions were permitted); the final set of indicators was then prioritized within clinical groupings by expert panelists. A feasibility review was also conducted. Results: A total of 170 candidate indicators were generated from the literature and assessed by panelists in 2 survey rounds. A final set of 48 were selected and prioritized within 8 clinical/ operational categories. Feasibility review determined that 13 indicators out of 48 can currently be captured using existing administrative databases; and a further 9 would be feasible with improved/enhanced data quality in existing data elements. Conclusion: The prioritized indicators have facevalidity representing many of the most serious/acute emergencies seen in EDs. Next steps include specification of technical definitions, developing appropriate and valid data sources for longitudinal and crossjurisdiction measurement, and regularly reviewing and revising the indicators to ensure they remain relevant and accurately reflect current knowledge and practice. Keywords: quality of care, patient safety, Delphi methodology
17 METHODOLOGICAL quality of reports of interventions to improve emergency department overcrowding: a systematic review
Guo XY, Villa-Roel C, Wong L, Schull MJ, Vandermeer B, Ospina M, Holroyd BR, Bullard M, Rowe BH, Innes G; University of Alberta, Edmonton, AB
Introduction: Emergency department (ED) overcrowding is a major concern for most urban and highvolume hospitals, and interventions to reduce overcrowding are urgently required. decision-makers require the highest quality evidence upon which to base decisions. This study evaluated the quality of methods used in ED overcrowding intervention research. Methods: Electronic databases (Cochrane Central Register of Controlled Trials, MEDLINE, Embase, Web of Science, HealthSTAR, Dissertation Abstracts, ABI/INFORM Global), controlled trial registry websites, conference proceedings, study references, experts in the field and correspondence with authors were used to identify potentially relevant studies involving interventions to reduce ED overcrowding. Interventional studies in which any intervention was used to influence ED overcrowding metrics (length of stay, left without being seen, etc.) were included. Two reviewers independently assessed citation relevance, inclusion and study quality. The methodological quality of studies was scored using the modified Effective Public Health Practice Project (EPHPP) tool. Results: From 14 716 potential relevant studies, 354 were potentially relevant and 183 were retained. The most common topics were triage liaison physician (29) and system wide interventions (29). Controlled clinical trials were infrequent (19.5%); most studies employed single-centred before–after designs. The majority of studies were classified as weak (85.7%) using the EPHPP tool. Selection bias, lack of valid and reliable outcomes, failure to compare groups between study periods or adjust for confounders, and inadequate statistical reporting were common limitations. Conclusion: Any efforts to synthesize evidence in ED overcrowding interventions using systematic review methods will produce potentially misleading conclusions due to weak study methodologies. Future efforts should be targeted at improving the quality of research conduct and reporting in this field. Keywords: emergency crowding, research methodology, systematic review
18 FACTORS influencing the intentions of nurses and respiratory therapists to use automated external defibrillators during in-hospital cardiac arrests: a qualitative iterative survey
Tweedle J, Vaillancourt V, Jensen JL, Kasaboski A; University of Ottawa, Ottawa Health Research Institute, Ottawa, ON
Introduction: Nurses and respiratory therapists (RTs) can initiate CPR during in-hospital cardiac arrest, but most of them are not allowed to use automated external defibrillators (AEDs). We sought to identify barriers and facilitators to AED use by nurses and RTs during in-hospital cardiac arrest before implementing a new medical directive. Methods: We conducted semistructured qualitative interviews in person or by telephone with a purposeful sample of nurses and respiratory therapists from various clinical settings, including medical/surgical wards, critical care units, operating rooms and outpatient clinics. We developed an interview guide based on the constructs of the Theory of Planned Behaviour, which elicits salient attitudes, social influences and behavioural control potentially influencing the use of AEDs during in-hospital cardiac arrests. Interviews were recorded, transcribed verbatim and analyzed until data saturation was achieved. Two independent reviewers performed inductive analyses to identify emerging categories and themes, and ranked them by way of consensus. Results: Demographics of the 24 interviewees (19 nurses, 5 RTs) are as follows: mean age 41, female 79.2%, used CPR on victim 87.5%, used AED on victim 29.2%. Leading attitudes were as follows: 1) faster defibrillation may improve patient survival, 2) AEDs are easy to use, 3) AED rhythm analyses may be inaccurate and shock inappropriately. Leading social influences came from the following: 1) physicians, and 2) hospital administrators. Behavioural control was the construct most associated with the ability to use an AED, including the following: 1) training and familiarity with AED use, 2) policy allowing AED use, and 3) unclear role during resuscitation. Conclusion: This is the first qualitative study using the Theory of Planned Behaviour to identify key facilitators and barriers for in-hospital use of AEDs by nurses and RTs. These findings will inform the design of a provincial survey, and education research focused on safety, and efficacy of AED use by nurses and RTs. Keywords: automated external defibrillators, cardiac arrest, qualitative research.
19 EFFECTIVENESS and adverse event profile of different procedural sedation and analgesia regimens in injection drug users: a prospective comparison
Scheuermeyer F, Lange L, deJong D, Andolfatto G, Grafstein E, Christenson J, Innes G, Stenstrom R; University of British Columbia, Vancouver, BC
Introduction: Injection drug users (IDUs) often require procedural sedation and analgesia (PSA) as part of emergency department (ED) treatment. However, the optimal sedation strategy remains unclear. This observational study compares adverse event (AE) rates in different PSA regimens. Methods: As part of quality assurance procedure, IDU status was documented among consecutive patients undergoing PSA at 2 urban Vancouver EDs from Apr. 1, 2006, to Jan. 31, 2009. Structured data describing comorbidities, vital signs, sedation regimens and medical treatments were collected, along with prespecified AE. PSA regimens consisted of propofol (PR), fentanylmidazolam (FM), and ketaminepropofol (KP). The primary outcome was rate of AE in each group. Secondary outcomes included the recovery time and ED length of stay (LOS) for nonadmitted patients, stratified by group. Secondary outcomes were analyzed by nonparametric methods. Results: A total of 198 consecutive patients participated; 110 received PR, 54 FM and 44 KP. Ages, vital signs and presenting complaints were similar in each group. There were 2 AEs in the PR group, both patients with transient apnea (1.8%, 95% CI 0%–4.3%). Two patients in the FM group required reversal agents (3.7%, 95% CI 0%–8.7%) Three emergence reactions were recorded in patients receiving KP (6.8%, 95% CI 0%–14%). (For the primary outcome, p = 0.90 by 1way ANOVA). No AE changed patient disposition. Median recovery times were 25 (IQR 18–36) minutes, 27 (IQR 22– 45) minutes and 30 (IQR 20–40) minutes, respectively (p = 0.88). Median ED LOS for the 165 discharged patients was 3.0 (IQR 2.5– 4.3) hours, 3.5 (IQR 3.0–4.5) hours and 4.0 (IQR 3.4–5.5) hours, respectively (p = 0.67). Conclusion: All 3 medication regimens for PSA in IDU appear safe and resulted in few AE. Recovery times and ED LOS were similar for all groups. Keywords: procedural sedation, injection drug use, adverse event, prospective study
20 TREATING febrile neutropenia: the role of an electronic clinical practice guideline (eCPG)
Lim C, Bawden J, Wing A, Villa-Roel C, Singh M, Meurer DP, Bullard MJ, Rowe BH; University of Alberta, Edmonton, AB
Introduction: Febrile neutropenia (FN) is a potentially life-threatening condition that requires urgent attention and management in the emergency department (ED). Evidence-based clinical guidelines for managing FN have been developed; however, the integration of guidelines into routine practice is often incomplete. We evaluated the impact of implementing an electronic clinical practice guideline (eCPG) on the management and outcomes of patients presenting to the ED with FN. Methods: A retrospective chart review over a 3-year period at 4 hospitals in Edmonton, Alta., was performed. Potentially eligible patient visits were identified by searching the Ambulatory Care Classification System database using ICD-10 codes and ED physician diagnoses of FN. ED patients with fever (> 38°C at home or in ED) and neutropenia (WBC count of < 1000 cells/mm3 or a neutrophil count of < 500 cells/mm3) and receiving an ED diagnosis of FN were included. Bivariable analyses were performed using χ2or Mann–Whitney U tests according to the nature of the variables. Results: From 371 potential cases, 201 unique cases of FN were included. Overall, the eCPG was used in 76 of 201 (37.8%) patient visits; however, there were significant differences in eCPG utilization between hospitals. eCPG usage were highest at the academic health sciences centre where it was developed (57%). The implementation of the eCPG increased the proportion of FN patients receiving an ECG (46.9% v. 31.5%, p = 0.03) and having blood cultures drawn (96.1% v. 93.1%, p = 0.04). The eCPG correlated with a decrease in time from triage to first antibiotic by 1 hour compared with the 3 control hospitals (3.9 v. 4.9 h, p = 0.02). Patients presenting to control hospitals were also more likely to be discharged home (15.1% v. 7.0%, p = 0.04). Conclusion: The eCPG is a safe and useful clinical tool that can improve patient management in the ED, and strategies to increase its utilization in this and other regions should be pursued. Keywords: febrile neutropenia, electronic decision support, clinical practice guidelines.
21 CONSENSUS on paramedic clinical decisions during high acuity emergency calls: results of a Canadian Delphi study
Jensen JL, Croskerry C, Travers AH; Emergency Health Services, Dalhousie University Division of EMS, Halifax, NS
Introduction: Paramedics make decisions which impact clinical outcome and patient safety. This Delphi study sought to establish consensus on the most important clinical decisions paramedics make during high acuity emergency calls. Methods: Canadian paramedics and medical directors participated in this multiround online survey. In Round I, participants listed important clinical decisions. In Round II, participants scored each decision in terms of its importance for patient outcome and safety on a 5-point Likert scale. In Rounds III and IV, participants could revise their scores. Consensus was defined a priori: if 80% or more of the panel scored a decision important or extremely important, it was included. Results: The panel (17 paramedics, 7 medical directors) had a mean 16.5 years experience. Response rates were as follows: Round I: 96%; II: 92%; III: 83%; IV: 96%. Consensus was reached on 42 decisions, grouped into 6 categories: airway management (n = 13); assessment (n = 3); cardiac management (n = 7); drug administration (n = 9); scene management (n = 4); general treatment (n = 6). The highest level of consensus was the assessment category (97% scored assessment decisions important or extremely important). Paramedics scored 4 decisions higher than medical directors: Decide on airway device (p < 0.04); Perform chest decompression (p < 0.01); Begin chest compressions on decompensated child (p < 0.04); Decide when to leave scene versus stay (p < 0.02). Medical directors scored one decision higher than paramedics: Give epinephrine for anaphylaxis (p < 0.04). Conclusion: In a Delphi study of clinical decision-making by paramedics in high acuity emergency calls, consensus was reached on 42 decisions in 6 categories, with the highest level of consensus on assessment decisions. The decisions found to be most important for patient outcome and safety should be a focus of paramedic training, continuing education and clinical auditing. Keywords: paramedic decision-making, prehospital, online survey research.
22 PREDICTORS of hospitalization in patients presenting to the emergency department with recent-onset of atrial fibrillation and flutter
Langhan T, Lang ES, Clement CM, Brison RJ, Rowe BH, Borgundvaag B, Magee K, Stenstrom R, Perry JJ, Birnie D, Wells GA, Xue X, Innes G, Stiell IG; University of Calgary, Calgary, AB
Introduction: The Recent-onset Atrial Fibrillation and Flutter (RAFF) study demonstrated that the decision to treat patients with either a rate or rhythm control dominated strategy is sitedependant, reflecting local practice patterns and an absence of clear guidance from clinical research. We sought to identify the primary drivers of the decision to admit patients with RAFF. Methods: This was a Canadian multicentre health records review with unique case identification and central quality control, conducted at 8 centres over a 12-month period. Eligible patients demonstrated RAFF requiring emergency management. In this secondary analysis of the RAFF database, univariate t test or χ2 analyses were conducted to select factors related to admission decision at a significance level of p < 0.05 and multiple logistic regression was employed to evaluate independent predictors of the decision to admit after adjustment. Results: Overall, 1068 patients were included, 178 of whom were hospitalized (16.7%), range by site (10%–27%). Seven factors emerged as independent predictors of admission. These included the following: associated acute coronary syndrome or congestive heart failure (OR 10.2, 95% CI 4.0–26.3); use of heparin in the ED (OR 4.1; 95% CI 1.7–9.4); atrial flutter (OR 3.2, 95% CI 2.1–4.8); new ischemic changes on ECG (OR 2.7, 95% CI 1.7–4.3) age by decade (OR 1.6, 95% CI 1.4– 1.8); elevated heart rate at time of disposition by 10 beats/min (OR 1.4, 95% CI 1.3–1.6) and conversion to sinus rhythm (OR 0.39, 95% CI 0.26–0.59). The triage level, site and CHADS score were not independent predictors of hospitalization. Conclusion: Multiple variables were positively associated with hospitalization in this population; however, conversion to sinus rhythm was associated with discharge. Admission of RAFF patients is driven largely by rational clinical parameters. Further research should define whether these factors predict clinical outcomes that would benefit from hospital admission. Keywords: atrial fibrillation, admission predictors, health records review
23 CLINICAL experience predicts CT utilization rates in lower acuity patients
Dowling S, Greenfield G, Wang D, Lang E, Innes G; University of Calgary, Calgary, AB
Introduction: CT imaging is increasingly stressing emergency department (ED) resources and a contributor to prolonged length of stay. Recent studies of utilization have demonstrated significant variability among physicians in CT imaging rates. Exploring the nature of this variability can lead to more rational CT usage. Our objective was to identify emergency physician (EP) characteristics associated with both high and low rates of CT use. Methods: This administrative database study was conducted at 3 adult EDs over a 1-year period. Every visit and CT image was linked to a hospital site and an EP. To minimize confounders and provide a more homogenous sample, only CTAS-3 patients were included. Predictor variables included for analysis were EP’s age, gender, years in practice, training path (EM v. FRCP) and number of patients seen per hour. EPs were categorized into 1 of 3 groups based on their CT ordering rates (quartiles): high CT users (top 25%), control (middle 50%) and low CT users (bottom 25%). Results: During the 1-year period, 93 510 CTAS-3 patients were treated by the 115 EPs included in the study. The mean CT usage rate was 12.0%. Comparing the highest CT users (> 15% CT ordering rate) to the control group, none of the following variables were significantly predictive (odds ratio and 95% CI): age = 1.01 (95% CI 0.96–1.07), years in practice = 1.01 (95% CI 0.96–1.07), EM versus FRCP trained = 0.90 (95% CI 0.42–2.66), patients seen per hour = 1.39 (95% CI 0.46–4.21). Comparing low CT users (< 10% ordering rate) to the control group, increased age and years in practice were predictive of reduced use: age = 1.08 (95% CI 1.03–1.15); years in practice = 1.07 (95% CI 1.02–1.13). Conclusion: In this study, physician age and clinical experience were associated with lower CT imaging rates. We did not identify any variables that were significantly predictive of high CT usage. Further research should examine the patient outcomes these extremes in practice pattern and to better understand the rationale employed by more limited CT users. Keywords: CT utilization, low acuity ED patients, administrative database study
