2008 CAEP/ACMU Scientific Abstracts - Oral Presentations: 5-24

2008 Scientific Abstracts

CJEM 2008;10(3):255-295

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

Oral Presentations

5 A novel method to use routinely collected ED clinical data to measure quality of care

Yip A, Schull MJ. Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON

Introduction: Measuring quality of care is important, but developing efficient methods of capturing valid data is challenging. We sought to develop a novel automated method to use routinely collected emergency department (ED) clinical data to measure evidence- based quality indicators. Methods: In a tertiary care ED, we linked 1) the ED information system (EDIS) with the medication dispensing information system (McKesson) to calculate time to antibiotics administration for pneumonia patients and 2) the EDIS with the hospital electronic patient record (EPR) electrocardiogram data to calculate time to ECG for chest pain patients. Patients were identified using specific EDIS algorithms. Data were linked using Microsoft Access. Chart reviews were carried out on all selected records; the accuracy of diagnosis, time of antibiotic administration and time of ECG were compared with those generated from electronic data linkage. Results: When compared with chart data, 1) the EDIS output correctly identified 40/40 (100%, 95% CI 92–100) pneumonia patients and 65/65 (100%, 95% CI 96–100) suspected ischemic chest pain patients; 2) median difference between time of antibiotic dispensing using McKesson versus chart data was 6 min (SD 14.0); 3) median difference between time of ECG in EPR and in the chart was 0 min (SD 70). Excellent linkage of EDIS and McKesson data (100% linked) and EDIS and EPR (95% linked) was achieved based on a hospital unique identifier. The percent of ED patients meeting performance criteria when calculated from chart versus electronic data was 44% versus 47% for time to antibiotics in pneumonia, and 11% versus 8% for time to ECG in chest pain. Conclusion:A simple and valid method for merging routinely gathered ED electronic clinical data for the purpose of quality of care measurement has been successfully piloted. Replication for other measures and in other settings may enable more comprehensive ED quality measurement in a cost-effective manner. Keywords: quality of care, emergency department information system (EDIS), electronic patient record (EPR)

6 SHIFT length and emergency physician productivity dynamics

Innes GD, Grafstein E, Stenstrom R. St. Paul’s Hospital and the University of British Columbia, Vancouver, BC

Introduction: Managing patient flow in an emergency department (ED) is a stressful activity with high uncertainty and decision density. Stress and progressive shift exhaustion may be associated with declining efficiency and cognitive function during the latter hours of a shift. An emergency physician’s (EP’s) intake capacity (ability to see new patients) may vary based on individual characteristics and on number of shift hours worked. Understanding EP intake dynamics is critical in establishing optimal shift length, shift scheduling and end-of-shift etiquette (i.e., patient handovers or not). Our hypothesis was that EP intake capacity would be maximal during hour 1 and decline progressively, reaching a state at hour X when they no longer contribute meaningfully to seeing new patients and should be replaced. Methods: At our inner city teaching hospital, ED physicians electronically sign on to each patient at the time of initial assessment (time seen). We captured electronic time stamps for every EP–patient encounter during a 6 month period in 2007 to determine the number of new patients seen for every hour of the ED shift.Results: The data incorporated 1561 eight-hour day and evening shifts (n = 24 916 patients), as well as 212 five-hour night shifts (n = 2848 patients). Average new patients/hr was 2.0 during 8-hour shifts and 2.6 during 5-hour shifts (p < 0.001). The difference in hourly productivity was explained by profound intake reductions during hours 7 and 8 (mean 1.3 and 0.3 patients/hr, respectively) and in part by the custom of patient handovers after night shifts. Mean patients seen were 2.8, 2.5, 2.3, 2.6, 2.3, 2.0, 1.3 and 0.3 for hours 1 to 8, respectively. Patient intake was characterized by an hour 1 peak, a 25% fall to plateau (hours 2–6), then rapid decline to negligible intake in hours 7 and 8. Hourly intake volume differed significantly by physician, with the highest scoring EP seeing 41% more patients/hr.Conclusion: In this setting, EP productivity fell dramatically after hour 6. Shorter ED shifts may be associated with enhanced productivity.Keywords: physician productivity, shift-work, time-series analysis

7 ESTIMATING physician workload in the pediatric emergency department

Millar KR, Tough S, Stewart B, Nijssen-Jordan C. Alberta Children’s Hospital, Calgary, AB

Introduction: The factors which contribute to emergency department (ED) physician workload are complex and likely vary from one ED setting to another. Studies designed to quantify ED physician workload have been sparse and have focused on general ED settings. The goal of this study was to develop a model which included demographic and clinical variables that would enable the prediction of physician time needed to treat patients in the pediatric ED. Methods:Pediatric ED physician shifts were shadowed and data was collected on 205 patient visits. Data collected included patient variables and physician time spent on patient care, educational and administrative activities. Multivariate regression was used to determine which variables had the strongest influence on physician time needed to treat patients. For model derivation, physician time was defined as the sum of direct and indirect patient care for each patient. Results:Physicians spent 80% of their time on patient care and 20% on educational and administrative tasks. The variables which were most strongly predictive of physician time needed to treat patients in a pediatric ED were CTAS score, arrival by ambulance, performance of a procedure by a physician, performance of a laboratory test and need for admission. Employing these 5 key variables in a regression model enabled the prediction of 40% of the variance in physician time needed to treat patients. The derived multivariate model was LN (physician time) = 3.6095 – 0.3694 (CTAS) + 0.3353(arrival by ambulance) + 0.3300 (procedure) + 0.7584 (laboratory) + 0.4083 (admitted) + 0.4075 (CTAS × laboratory). Patient age, trainee involvement, language barrier, referral by another physician and need for radiographs were not found to influence physician time needed to treat patients. Conclusion: The model derived in this study could be used for pediatric ED man power planning, derivation of equitable funding formulas for pediatric ED physicians and physician performance evaluation. Keywords: physician workload, pediatric emergency medicine, workflow analysis

8 IMPACT of a rapid-assessment pod on ED overcrowding measures: a randomized trial

Bullard MJ, Lo A, Willis V, Latoszek K, Holroyd BR, Rowe BH. Department of Emergency Medicine, University of Alberta, Edmonton, AB

Introduction: Emergency department (ED) boarding of admitted patients has delayed access to care, especially for triage (CTAS) level 3 patients in academic EDs. This study evaluated the impact of transforming acute care spaces into a rapid-assessment pod (RAP).Methods: Over a 6-week period from June 2007 to July 2007 block randomization was used to allocate RAP and non-RAP study days. The RAP was operational from 0900 to 2300 with 3 standard care spaces converted to 2 stretchers for clinician assessment and procedures; and 7 chairs for patients to be transferred to receive medications, IV therapy, be observed or wait for results. Outcomes included ED lengths of stay (LOS) and proportion of patients who left without being seen (LWBS). Mean values are reported and comparisons made using χ2 and Wilcoxon rank sum tests. Daily nurse–physician questionnaires were also collected to assess satisfaction. Results:Patient volume, sex, age, CTAS, mode of arrival and admitted patient boarding times were similar between RAP and non-RAP days. Despite cancelling the RAP 3 times due to severe overcrowding the mean time from triage to bed location was 24 minutes shorter for CTAS level 3 patients (99 v. 123 min, p = 0.04), with no change in time from bed to physician (50 v. 52 min); this was maintained for overall LOS (544 v. 567 min). On RAP days there was also a nonsignificant decrease in LWBS for all (7.1 to 6.6%, p = 0.42) and CTAS 3 (3.3%–2.6%) patients. Questionnaires revealed problems with increased nursing workload, excessive patient complexity for chairs, disease transmission concerns, and concerns with eliminating 3 stretchers during excessive boarding of patients. Conclusion: Despite evidence that the CDU decreased waiting room time for CTAS level 3 patients, concerns about workload, patient safety, and patient complexity led staff to conclude these benefits were outweighed by the loss of treatment spaces during severe overcrowding. Keywords:rapid-assessment pod, ED crowding, randomized trial

9 IMPACT of an overcapacity care protocol on emergency department and hospital access block

Innes GD, Grafstein E, Scheuermeyer F, Stenstrom R, Hunte G. St. Paul’s Hospital and the University of British Columbia, Vancouver, BC

Introduction: A hospital overcapacity protocol (OCP) was implemented at our urban teaching centre in February 2006 because of waiting room (WR) adverse events related to overcrowding (access block). OCP dictates that arriving level I–III patients are placed in overcapacity ED care spaces rather than WR areas. When the ED is overcapacity by 2 patients, boarded patients move from the ED to inpatient overcapacity care spaces. Prior data showed that OCP reduced hospital length of stay (LOS) and ED LOS for boarded patients. The objective of this study is to assess the impact of OCP on ED access block (number of WR patients) and hospital access block (number of boarded patients held in the ED). Methods: Our ED information system tracks real-time patient location and status using electronic bed location and electronic consult and admission orders. We used these data systems to prospectively record patient status snapshots every hour during a 6-month pre-OCP period (March through August, 2005) and a corresponding 6-month post-OCP period in 2006. Outcomes include the mean number of WR and admitted patients, stratified by hour of the day. Results: During the post- OCP period, ED volume increased 1.2% (30 483 to 30 846), CTAS I–III volume rose 5.7% (13 078 to 13 828) and ambulance arrivals rose 1% (46.1–46.6/d). Despite this, the mean number of WR patients fell from 16.7 to 12, a 28% reduction in ED access block (p < 0.001); and the mean number of boarded patients fell from 12.5 to 7.9, a 37% reduction in hospital access block (p < 0.001). Reductions were consistent across all hours of the day, suggesting access block improvement during busy and nonbusy periods. Improvements equate to 110.4 ED stretcher hours made available daily and a 113 hour reduction in daily WR time for sick patients. During the post- OCP period, no critical events were reported in ED waiting areas or inpatient OCP care spaces. Conclusion: The OCP has led to significant reductions in ED and hospital access block, and reduces the number of patients left at risk without ED evaluation. Keywords:overcapacity protocol, ED crowding, access block

10 CAEP research grants competition: an evaluation of outcomes, dissemination and perceptions over ten years

Bawden J, Manouchehri N, Grafstein E, Villa-Roel C, Rowe BH. University of Alberta, Edmonton, AB

Introduction: Dedicated emergency research funding has been uncommon in the past and specific funding competitions have been proposed. This study examines the scholarly outcomes of the projects receiving Canadian Association of Emergency Physicians (CAEP) research grants over the first 10 years of national funding.Methods: Email surveys were sent to 58 ED researchers funded by CAEP between 1996 and 2005. Data collection focused on grant deliverables (completion, presentations and publications) and opinions regarding the value of the award (1–7 Likert scale). Email surveys and up to 3 reminders were sent; publications were confirmed through literature searches. Results: Grants were commonly awarded to residents (37%), followed by senior (26%) and junior (23%) emergency staff. Ontario (44%) and Quebec (19%) received the majority of the grants. A total of 56 (97%) survey responses were received; most (88%) projects were completed at the time of contact. Overall, 38 (76%) manuscripts had been published (or were in press) from the completed projects. Abstracts were more common, with a median of 2 presented per completed project. Local (46%), national (80%) and international (74%) abstract presentations for the completed projects were documented. Overall, 19 (33%) projects received additional funding. Despite medial funding of $5000, respondents felt CAEP funding was critical (median 6.25, interquartile range [IQR] 5–7) to completing the project and respondents felt strongly (median 7, IQR 7–7) that dedicated funding for EM research should be continued to stimulate research productivity. Conclusion:Overall, the CAEP research grants competition has produced impressive results, despite the small sums available for funding. This source of funding was important to ensuring study completion and in many cases assisted in securing additional funding. CAEP and similar ED organizations need to develop a more robust funding approach so that larger grant awards and more researchers can be supported on an annual basis. Keywords: CAEP research grants, research productivity, young investigators

11 INTERRATER agreement between nurses using a computerized version of the Paediatric Canadian Triage and Acuity Scale in a pediatric emergency department

Gravel J, Gouin S, Manzano S, Arsenault M, Amre D. CHU Sainte- Justine, Montréal, QC

Introduction: The Paediatric Canadian Triage and Acuity Scale (PaedCTAS) is a 5-level triage tool constructed from a consensus of experts. Until now, there has been no evaluation of the reliability of the tool for children presenting to emergency departments (ED) in a clinical context. Objective: To measure the interrater agreement for nurses who assigned triage levels to children visiting a pediatric ED using a computerized version of the PaedCTAS (Staturg from Statdev).Methods: This was a prospective cohort study evaluating all children (0–18 yr old) initially triaged from level II (emergent) to V (non urgent) on the PaedCTAS. Recruitment occurred on a convenience sample of patients triaged during 40 shifts from June to September 2007 in a tertiary care pediatric ED. All patients were initially triaged by a regular triage nurse using a computerized version of the PaedCTAS (Staturg). Those not requiring immediate care were eligible to participate. Research nurses recruited eligible patients immediately after their first triage to perform a second triage (within 15 min) using the same triage tool. Both triages were performed blinded to each other. The research nurses were regular ED nurses performing extra hours for research. The primary outcome measure was the interrater agreement between the two nurses measured by the quadratic weighted κ score. Results: There were 499 patients recruited to participate in the study and they were all evaluated by 2 nurses. The overall interrater agreement between the nurses was good as demonstrated by a quadratic weighted κ score of 0.61 (95% CI 0.52–0.69). There was a discrepancy of more than 1 level in only 10 patients (0.02). Conclusion: The Staturg tool is a computerized version of the PaedCTAS that demonstrates good interrater agreement between nurses assigning triage level to children presenting to a pediatric ED. Keywords: paediatric CTAS, interrater agreement, computerized triage

12 VALIDITY of a computerized version of the Paediatric Canadian Triage and Acuity Scale

Gravel J, Manzano S, Arsenault M. CHU Sainte-Justine, Montréal, QC

Background: The Paediatric Canadian Triage and Acuity Scale (PaedCTAS) is a 5-level triage tool for the emergency department (ED). Staturg (from Statdev, Montréal) is a computerized version of the PaedCTAS. Objective: Evaluate the validity of a computerized version of the PaedCTAS (Staturg) for children visiting a pediatric ED. Methods: This was a retrospective cohort study evaluating all children presenting to a pediatric university-affiliated ED for a 1-year period. Data were retrieved from the ED computerized database. Information regarding triage and disposition was prospectively registered by a clerk in the ED database during patients’ management. In the absence of a gold standard for triage, hospitalization, admission to pediatric intensive care unit (PICU), length of stay (LOS) in the ED and patients that left without being seen by a physician (LWBS) were used as surrogate markers of severity. The primary outcome measure was the correlation between triage level (from I to V) and hospitalization. It was estimated that the evaluation of all patients visiting the ED for 1 year would provide more than 500 patients in each triage level. Results:From Nov. 11, 2006, to Nov. 10, 2007, 58 569 patients were triaged in the ED. Hospitalization proportion was 63% for resuscitation (level I), 37% for emergent, 14% for urgent, 2% for semi urgent and 1% for nonurgent. There was also a good correlation between triage levels and admission to PICU, LOS and proportion of LWBS. Conclusion: Staturg is a computerized version of the PaedCTAS that shows good ability to predict hospitalization, admission to PICU, LOS in the ED and proportion of patient that left without being seen by a physician. These results suggest that Paed- CTAS has a good validity for triage of children in a pediatric ED.Keywords: paediatric CTAS, computerized triage, resource use modelling

13 PREVALENCE of CA-MRSA in purulent skin and soft tissue infections in patients presenting at emergency departments in the greater Toronto area

Borgundvaag B, Katz K, Allen V, Adam H, Currie A, Lee JS, Ellis P, Somers A, Rutledge T, Rizos J, Goldman RD; and The EMERGENT Working Group. Mount Sinai Hospital, Toronto, ON; North York General Hospital, North York, ON; Sunnybrook Health Sciences Centre, Toronto, ON; the Toronto Hospital, Toronto, ON; Credit Valley Hospital, Mississauga, ON; the Hospital for Sick Children, Toronto, ON

Introduction: Community associated methicillin resistant Staphylococcus aureus (CA-MRSA) caused primarily by the CMRSA-10 (USA 300) clone has replaced methicillin-sensitive Staphylococcus aureus (MSSA) as the primary cause of skin and soft tissue infections (SSTIs) in the US. The prevalence of CAMRSA causing SSTIs in Canada, and the optimal treatment of such patients is unknown. We document the prevalence and antimicrobial susceptibility patterns of CA-MRSA in SSTIs in the GTA.Methods: Prospective observational study in 7 GTA hospitals. Patients presenting with SSTIs were swabbed as per current guidelines. Patients whose swabs grew Staph aureus were eligible to participate and contacted by study staff for consent to participate. Patients answered a standardized health and lifestyle questionnaire by phone regarding risk factors for CA-MRSA and outcomes of treatment. MRSA isolates underwent reference susceptibility testing and molecular testing. Results: Two hundred and ninety-eight eligible patients were identified. One hundred and sixty-one (54%) consented to participate, and 135 (55%) of the 244 MSSA patients, and 26 (48%) of the 54 MRSA patients completed interviews. Patients with SSTIs due to typical community-associated strains of MRSA were younger (p < 0.01) and less likely to report healthcare exposures (p < 0.01) than patients with typically hospital-associated strains. Two hundred and forty three (82%) of Staph aureus isolates were methicillin sensitive, while 54 (18%) were methicillin resistant. Twenty-eight of 54 MRSA isolates (52%) were the CMRSA- 10 clone (all were Panton-Valentine Leukocidin [PVL] positive and carried SCCmec IV). All CA-MRSA isolates were clindamycin susceptible and all were susceptible to tetracycline, doxycycline, minocycline, septra and tigecycline; 19/28 CMRSA-10 isolates were susceptible to ciprofloxacin. Only 13 MRSA isolates were CMRSA-2 (the most common healthcare associated clone). Conclusion:The most common isolate of CA-MRSA found in SSTIs in this population was the CMRSA-10 clone typical of communityassociated MRSA. Keywords: community acquired MRSA, skin and soft tissue infection, antibiotic resistance

14 IS a prolonged stay in the ED associated with adverse events in older patients?

Ackroyd-Stolarz S, Guernsey J, MacKinnon N, Kovacs G. Dalhousie University, Halifax, NS

Introduction: Older adults are frequent users of the ED. International patient safety studies have identified this as a high-risk group for adverse events (AEs). The purpose of the study was to determine whether a prolonged stay in the ED is associated with an increased risk in the occurrence of AEs for older patients admitted to hospital.Methods: This retrospective cohort study was conducted at an adult tertiary care facility in Atlantic Canada between July 1, 2005, and Mar. 31, 2006. All community-dwelling people 65 years and older admitted to an acute care inpatient unit from the ED were eligible for inclusion. Using the 2007 CAEP Position Statement Guidelines, the exposed group was defined as those patients that had a total length of stay (LOS) in the ED of greater than 6 hours for those triaged to CTAS levels I–III and greater than 4 hours for patients triaged to levels IV–V. The unexposed group had an ED LOS less than the benchmark times. Outcomes were determined using the previously validated Wisconsin Medical Injury Prevention Program screening criteria. The criteria were applied to diagnostic and clinical data from administrative data sources to identify AEs. The results were compared between the exposed and unexposed groups. Results: A total of 982 patient records were analyzed. The average age was 77.8 years (SD 7.8). The majority of patients (75.0%) experienced a prolonged ED LOS as defined by the Canadian benchmark. Of the 982 records, 140 (14.3%) had evidence of an AE. The most common AEs were procedure-related (47.2%), device-related (31.2%) and medication-related (20.8%). After adjustment, the total LOS (hr) in the ED was associated with an increased risk of an AE (OR 1.03, 95% CI 1.004–1.05). Those with an AE had twice the hospital LOS (20.2 v. 9.8 d, p < 0.001). Conclusion: A prolonged stay in the ED for older admitted patients is associated with an increased risk of an AE. The excess hospital LOS associated with AEs further reduces the availability of acute care beds thus exacerbating ED overcrowding.Keywords: geriatrics, ED crowding, patient safety

15 ARE patient and crowding factors associated with adverse outcomes in the emergency department?

Dankoff J, Afilalo M, Xue X, Soucy N, Colacone A. Emergency Multidisciplinary Research Unit, SMBD–Jewish General Hospital, McGill University, Montréal, QC

Introduction: The emergency department (ED) has been identified as a high-risk setting for adverse outcomes (AO) because of the nature of the work (high volume, rapid intervention, etc.) and the clientele (seriously sick, anxious, etc.). This study aims to identify patient and crowding factors that are associated with AO in the ED.Methods: A case–control study with 20-month prospective data collection (March 2006 to November 2007) was conducted at 2 university teaching tertiary EDs in Quebec. Cases were patients who had AO due to a medical error; controls were randomly selected patients from those who had no AOs. Cases were identified by physicians, nurses, pharmacists and through database queries. Patient factors (e.g., age, triage, past admissions) were collected through administrative database and chart review. Crowding factors (e.g., inflow, occupancy) were calculated within 2 and 4 hours before the incident for cases, and before the midpoint of the emergency visit for the controls. Univariate and multivariate (GEE) approach were used for the analysis with adjustment for hospitals. Odds ratios and their 95% confidence intervals (CIs) are presented. Results: The study sample consisted of 292 patient visits, among which 135 were cases. Cases were mainly reported by physicians (52%) and nurses (25%). Multivariate analysis showed that 5 factors were significantly associated with AO: being older (1.2, 1.1–1.3), ambulance arrival (2.0, 1.04–4.0), urgent triage code (2.4, 1.2–4.9), admission within past 2 years (2.3, 1.1-4.8) and presented with different chief complaints (CCs) (i.e., compared with others, patients presenting with gastro (2.4, 1.2-4.7), cardiac [1.8, 1.7-2.0] and pulmonary [7.4, 1.4-39.3] problems have a higher risk of suffering AOs). None of the crowding factors were significantly associated with AOs. Conclusion:Older patients, patients with more urgent conditions, increased morbidity and those with specific CCs were found to be more at risk for AOs in the ED. ED crowding was not found to be associated with AOs. Keywords: patient safety, ED crowding, case–control study

16 DOES emergency department length of stay impact the mortality and morbidity of patients in the medical–surgical intensive care unit?

Gray SH, Boone A, Vainshtein A. St. Michael’s Hospital, Toronto, ON

Introduction: Mortality in the intensive care unit (ICU) often reaches 30%. Many patients presenting to the emergency department (ED) have to wait for an ICU bed to become available. However, patients waiting longer in the ED may not receive optimal ICU care during a critical physiological window, and therefore may experience worse outcomes. This study hypothesized that increased ED length of stay would worsen in-hospital mortality for patients admitted to the medical–surgical ICU. Methods: We examined a 5-year retrospective cohort at a university hospital, including all adult patients in the ED who were subsequently admitted to the medical– surgical ICU. Relevant data was extracted from computerized databases by trained personnel. Double data entry was performed on 20% of the sample to ensure reliability. Results: 926 patients were identified, of which 571 were male (61.7%) with a mean age of 58 years (SD 17 yr). Median ED length of stay was 6.0 hours (IQR 6.6 hr). The in-hospital mortality for the entire sample was 27.4%. Our primary outcome was in-hospital mortality: ED length of stay was a significant independent predictor of in-hospital mortality for ICU patients (univariate logistic regression χ2 5.7, p = 0.017). This relationship remained statistically significant in multivariate analysis that controlled for the effects of possible confounders, including age, sex and hospital length of stay (χ2 8.2, p = 0.004). The κ coefficient for our double-data entry was 94% indicating excellent interobserver agreement. Conclusion: This is the first Canadian study to demonstrate that longer ED lengths of stay increase mortality for patients admitted to the ICU. Further research is underway to assess the impact of ED length of stay on ICU length of stay and other morbidity outcomes. This data, if confirmed prospectively, should be used to facilitate rapid ICU admission for patients presenting to the ED.Keywords: intensive care unit, ED crowding, morbidity and mortality

17 30 DAY outcomes in an ED population undergoing DC cardioversion for atrial fibrillation or flutter

Grafstein E, Innes GD, Scheuermeyer F, Stenstrom R, Harris DR. Providence Health Care, St. Paul’s Hospital, University of British Columbia, Vancouver, BC

Introduction: Despite the general belief that emergency DC cardioversion is a safe practice, there is a lack of outcomes data on atrial fibrillation/flutter patients who are cardioverted in an ED setting. The study objective was to characterize the 30-day hospitalization, CVA and mortality rates for this patient population. Methods:Using an emergency administrative database, we reviewed all patients with a discharge diagnosis of atrial fibrillation or flutter from 2000 to 2005 who had an ED procedure code for DC cardioversion. We linked these patient records with provincial vital statistics data and the regional emergency visit database from the same time period to ascertain which patients presenting with atrial fibrillation or atrial flutter and who received DC cardioversion died within the subsequent 30 days and/or if they subsequently presented at another of the 5 regional EDs. We excluded patients if they had an address that was out of province. Results: There were 1163 patients who made 1731 visits to the ED related to atrial fibrillation or flutter. Patients underwent DC cardioversion a total of 192 times (145 unique patients). Of these, 87.5% were cardioverted for atrial fibrillation, the rest for flutter. The admission rate was 12% (23/192). Average age was 53.8 (SD 14.4). 76% (146/192) were male. All patients were triaged as CTAS level II (53%) or CTAS III (47%). The 30 day mortality for this cohort was 0% (95% CI 0%–2.1%). There were 3 visits by ED patients to other regional EDs in the subsequent 30 days. One patient was admitted and died of lung cancer during the following 30 days. The other 2 patients were treated and released. Eleven of 145 patients (7.6%; 95% CI 3.3%–11.9%) had a 30-day recurrence of atrial fibrillation. There were no associated CVAs in this cohort.Conclusion: This represents one of the few Canadian outcomes studies of DC cardioversion in an ED population. Mortality and CVA after cardioversion are rare events in patients with atrial fibrillation or flutter who undergo DC cardioversion in the ED. Keywords:atrial fibrillation, cardioversion, administrative database

18 A systematic review and meta-analysis of prehospital triage and diversion for patients with ST-elevation myocardial infarction directly to percutaneous coronary intervention centres: Are emergency medical services jumping the gun?

Brooks SC, Allan KS, Welsford M, Verbeek PR, Arntz HR, Morrison LJ. Prehospital and Transport Medicine Research Program, Division of Emergency Medicine, Department of Medicine, University of Toronto, Toronto, ON

Introduction: The optimal prehospital strategy for facilitating rapid coronary reperfusion in ST-elevation myocardial infarction (STEMI) patients is unclear. Our objective was to determine whether emergency medical services diagnosis and diversion of adult STEMI patients to primary percutaneous coronary intervention (PCI) centres improves 30-day all-cause mortality when compared to a strategy of transportation to the closest hospital. Methods: MEDLINE (1980– July 2007), EMBASE (1980–July 2007), Cochrane CENTRAL database (1980–July 2007), Web of Science (1980–July 2007), CINAHL (1980–September 2006), HealthStar (1980–September 2006), Proquest Digital Dissertations (1980–September 2006), the NIH CRISP database (1980–September 2006) and clinicaltrials.gov (July 18, 2007) were searched. Two reviewers independently assessed citations for inclusion and then abstracted data. Studies with diversion to a PCI-capable centre for primary PCI, a control group transported to the closest hospital, and outcomes of treatment time intervals, allcause mortality, reinfarction rate, stroke rate or the frequency of cardiogenic shock were included. A random effects model was used to provide pooled estimates of relative risk where data allowed. Results:The search identified 2264 citations. Five studies were included in the final review. There was clinical heterogeneity across studies and variable quality. The pooled relative risk (RR) for shortterm mortality showed a trend towards increased survival with diversion to PCI but was not statistically significant (RR 0.51, 95% CI 0.24–1.10). One study reported a non-significant trend towards reduced reinfarction (RR 0.43 95% CI 0.11–1.60) and stroke (RR 0.33 95% CI 0.01–8.06) with diversion for PCI. Conclusion: There is insufficient evidence to support the use of prehospital diversion for primary PCI as compared with transportation to the closest hospital in patients with STEMI. Keywords: percutaneous coronary intervention, emergency medical services, systematic review

19 THE Ottawa aggressive protocol leads to rapid discharge of ED patients with acute atrial fibrillation

Stiell IG, Clement C, Vaillancourt C, Perry JJ, Symington C, Dickinson G. University of Ottawa, Ottawa, ON

Introduction: We previously described the safety and efficacy of the Ottawa Aggressive Protocol for rapid ED rhythm control for patients with acute (< 48 hr) atrial fibrillation (AAF) or atrial flutter (AAFL). Our objective was to examine the impact of the protocol on admission rates and lengths of stay. Methods: This 5-year cohort study included consecutive visits to a university hospital ED for adults who presented with acute-onset AAF/AAFL and who were managed with the Ottawa Aggressive Protocol. Patients were identified from the National Ambulatory Care Reporting System database. The Aggressive Protocol was overseen by attending emergency physicians and included: 1) IV procainamide infusion; 2) sedation and electrical cardioversion if necessary, by ED staff; and 3) discharge from the ED with outpatient cardiology follow-up. Outcomes included admission rates and lengths of stay in ED. We conducted descriptive data analyses with 95% CIs. Results: Characteristics for 660 eligible patient visits were: mean age 64.5 years (range 19–92), male 55.6%, mean heart rate 113.4 beats/min (range 45–220), mean duration of symptoms 8.9 hours, AAF 95.1%, AAFL 4.9%, prior AAF 82.1%, warfarin use 33.8%. Of cases, 39.6% received IV rate control drugs prior to rhythm control. All patients received procainamide with a conversion rate of 58.3% (AAF 59.9%, AAFL 28.1%). Electrical cardioversion was attempted in 36.8% of visits with a success rate of 91.7% (AAF 91.0%, AAFL 100%). Overall, only 3.2% of patients required admission. Of those patients discharged, 93.3% were in sinus rhythm and had a mean heart rate of 68.9 beats/min. The median lengths of stay from ED arrival to discharge were, for all patients 4.9 hours (range 1.6–19.7), for those successfully converted with procainamide 3.9 hours, and for those requiring electrical conversion 6.5 hours. Conclusion: These data demonstrate that use of the Ottawa Aggressive Protocol is associated with very rapid conversion to sinus rhythm and discharge for ED patients with AAF and AAFL. Keywords: Ottawa Aggressive Protocol, atrial fibrillation, length of stay

20 ASTHMA presentations by adults to emergency departments in Alberta, Canada: a population-based study

Rowe BH, Voaklander DC, Wang D, Senthilselvan A, Klassen TP, Marrie TJ, Rosychuk RJ. University of Alberta, Edmonton, AB

Introduction: Asthma is a widespread disease with a prevalence of approximately 7%–10% in adults; exacerbations to the emergency department (ED) are common. The objective of this study was to describe the epidemiology of asthma presentations to EDs made by adults in the province of Alberta, Canada. Methods: The Ambulatory Care Classification System of Alberta and provincial administrative databases were used to obtain all ED encounters for asthma during 6 fiscal years (April 1999 to March 2005). Information extracted included demographics, ED visit timing, and subsequent visits to physicians; all data were coded by trained medical records nosologists. Data analysis included descriptive summaries and directly standardized visit rates (DSR). Results: There were 105 813 ED visits for asthma made by 48 942 distinct adults over the study. The DSRs have declined from 9.7/1000 adults in 1999/2000 to 6.8/1000 in 2004/05. In 2004/05, the Welfare (22.5/1000) and Aboriginal (14.9/1000) groups DSRs were higher than the Other subsidy group (6.0/1000; p < 0.001). Most patients (66%) had only 1 asthma-related ED visit during the study period. Daily, weekly and monthly trends were observed. The ED visits resulted in discharge 91% of the time; however, only 33% had at least 1 follow-up visit with a non-ED physician within 7 days of the ED visit. The median time to the first follow-up visit was 19 days. Conclusion: Asthma is a common presenting problem in EDs and further study of these trends is required to understand the factors associated with the variation in presentations. The important findings identified here include an overall decrease in the rates of presentation over the study period; however, disparities based on age, sex and socioeconomic/cultural status exist. Targeted interventions could be implemented to address specific groups and perhaps reduce asthma-related visits to Alberta EDs.Keywords: asthma, socioeconomic status, economic disparities

21 DOES emergency department crowding affect the quality of care provided to patients with acute asthma exacerbation?

Lang E, Afilalo M, Colacone A, Xue X, Guttman A, Willis V, Ciolofan C, Rowe BH. Emergency Multidisciplinary Research Unit, SMBD–Jewish General Hospital, McGill University, Montréal, QC; Emergency Research Unit, University of Alberta Hospital, Edmonton, AB

Introduction: Emergency department (ED) overcrowding has the potential to delay medical assessment (DMA) and delay to treatment (DT) thereby compromising quality of care (QoC). Our objective was to quantify the relationship between ED crowding and QoC for patients with acute asthma presenting to the ED. Methods: Two databases were combined. The first provided measures of QoC (DMA, DT and relapse within 14 days [R14]) for patients who participated in a study designed to determine major risk factors for asthma relapse (AIR substudy). Five AIR sub-study sites provided administrative data for all ED visits during the study period. Measures of crowding (inflow, percent occupancy and boarding time) were then calculated within one hour of each AIR patients triage time. Mixed model analysis for DMA and DT with log transformation and GEE approach for R14 were used to measure the impact of crowding on QoC. Patient and ED factors were adjusted. Results: A total of 369 AIR patients were drawn from the 5 sites for which 354 (96%) were matched to the administrative data. Patient factors (demographics, previous number ED visits and hospitalizations) were similar across sites. Among the indices of crowding examined, only ED occupancy within 1 hour of asthma patients arrival was found to be positively associated with DMA (multiplier; 95% CI 1.7; 1.3–2.3) and DT (multiplier; 95% CI 1.7; 1.2–2.3), but not with R14. For example, an increase in occupancy from 100% to 150% would contribute to an increase of 30% both in DMA and in DT. Other factors associated with longer delays were lower acuity, previous hospitalizations, older age, presentation during weekday and day shifts; while being a female and lower acuity were associated with increased R14. Conclusion: Patients with asthma who presented to the ED when occupancy was higher experienced modest delays to medical assessment and treatment. Other factors associated with a reduced QoC delivered to asthma patients were age, acuity level, previous hospitalization and arrival time. Keywords: asthma, ED crowding, quality of care

22 PROSPECTIVE evaluation of a revised preschool respiratory assessment measure in children with acute asthma

Gouin S, Gravel J, Guimont C, Amre D. Division of Emergency Medicine and Research Institute of CHU Sainte-Justine, CHUQ pavillon of CHUL, Université de Montréal and Laval, Montréal and Québec City, QC

Introduction: Adequate evaluation of the severity of asthma exacerbations is important. A promising clinical severity score is the Preschool Respiratory Assessment Measure (PRAM); however, it was described as cumbersome to use. The aim of our study is to compare the accuracy and responsiveness of the PRAM (5 variables) and a revised PRAM (rPRAM) (4 variables) during an asthma exacerbation in children. Methods: A prospective cohort study was conducted in an academic pediatric ED from February 2006 to October 2007. All patients between the ages of 18 months to 7 years who presented to the ED for an asthma exacerbation while one of the research assistants was available were approached. The outcomes of interest were the length of stay greater than 6 hours in the ED or admission and the physicians’ judgment of severity. Clinical findings were assessed by respiratory therapists at the start of the visit and after 90 minutes of treatment. Results: During the study period, 3788 patients were seen in the ED for an asthma exacerbation. Of these patients, 291 were approached and 8 refused to participate. The mean age was 3.4 years (SD 1.5) and 63.6% were boys. There was a high correlation between the physicians judgment of severity and PRAM (r = 0.45) and rPRAM (r = 0.42). PRAM and rPRAM were moderate predictors of length of stay greater than 6 hours (AUC= 0.67 and 0.64, respectively) but better predictors of admission (AUC = 0.85 and 0.80, respectively) when calculated at the start of the ED visit. When the scores were recalculated after 90 minutes of treatment, the predictive abilities for length of stay greater than 6 hours and admission were higher for both PRAM (AUC = 0.77 and 0.90) and rPRAM (AUC = 0.75 and 0.90). The scores were shown to be responsive, with a 50% relative improvement in score from start to 90 minutes of treatment for both the PRAM and rPRAM. Conclusion: The rPRAM, which is a simplified version of the PRAM, is a valid measure of asthma severity in children and shows similar discriminative and responsive properties. Keywords: pediatric asthma, clinical severity scores, clinical prediction guide

23 DO statins improve outcomes and reduce the incidence of vasospasm following aneurysmal subarachnoid hemorrhage a meta-analysis

Sillberg VAH, Perry JJ, Wells GA. University of Ottawa, Ottawa, ON

Introduction: Vasospasm following subarachnoid hemorrhage (SAH) causes significant morbidity and mortality. Statins are believed to reduce vasospasm by effects on inflammation, endothelial nitric oxide synthase activity and thrombogenesis. This metaanalysis attempts to determine if statins reduce vasospasm and mortality following aneurysmal SAH. Methods: For this systematic review and meta-analysis, we searched MEDLINE (1950–2007), EMBASE (1980–2007) and the Cochrane Central Register of Controlled Trials (second quarter 2007) for randomized controlled trials (RCTs) relating to statins and patients with SAH. We included foreign language and abstract articles. Two independent reviewers assessed studies for eligibility, data extraction and quality. We assessed study quality and allocation concealment. We calculated interrater agreement for study selection and quality using κ statistics. The primary outcome was radiographicallyconfirmed clinical vasospasm; secondary outcomes were delayed clinical ischemic deficits and morality. We combined eligible studies using Review Manager 4.2.10 to determine relative risk (RR).Results: Our search identified 160 titles. From these, we reviewed 26 abstracts. We exclusdd 23 (duplicates, reviews, letters or were not RCTs). Three studies were analyzed (n = 158). The κ for included studies was 1.0. Therapy was started within 96 hours of SAH with simvastatin 80 mg/day or pravastatin 40 mg/day. Most patients were treated for 14 days or until hospital discharge. There was significant reduction in incidence of radiographic vasospasm (RR 0.73, 95% CI 0.54–0.99), delayed clinical ischemic deficits (RR 0.38, 95% CI 0.17–0.83), and mortality (RR 0.22, 95% CI 0.06–0.82) in the statin group. There was no significant heterogeneity.Conclusion: The initiation of statin therapy following aneurysmal SAH significantly reduces vasospasm, delayed ischemic deficits and mortality. This supports the routine use of statins for patients with aneurysmal SAH. Keywords: subarachnoid hemorrhage (SAH), statins, vasospasm

24 WHAT is a clinically significant nontraumatic subarachnoid hemorrhage?

Perry JJ, Symington C, Guarino M, Goindi R, Stiell IG. University of Ottawa, Ottawa, ON

Introduction: There is a relatively new phenomenon in EDs where patients with nonaneurismal, nontraumatic subarachnoid hemorrhage (SAH) are discharged home from the ED. This study’s objective was to define what a significant SAH is. Methods: We surveyed all neurosurgeons listed in the Canadian Medical Directory (lists > 92% of practicing physicians). We used a modified Dillman technique with up to 3 mailed surveys plus a prenotification letter. Neurosurgeons were asked to answer 22 questions including a series of 13 clinical scenarios of SAH. Scenarios involved typical aneurismal SAH to SAH with a normal CT and only positive xanthochromia in cerebrospinal fluid (CSF) analysis. Each scenario was assessed regarding its clinical significance with a 5 point Likert scale (1 = always, 2 = most often, 3 = sometimes, 4 = almost never, 5 = never). Analysis included calculating means and medians as descriptive results. Results: We received 112 responses from 241 surveys. Mean age was 49 years (SD 11), mean years of clinical practice was 16 (SD 11) and 97% were male. All scenarios with an aneurysm requiring an intervention, arteriovenous malformation, death or any surgical intervention all had a median responses of 1 (IQR 1, 1). Scenarios having a negative CT with xanthochromia and few red blood cells in CSF with a negative angiography had a median response of 3 (IQR 1, 4). Scenario with perimesencephalic pattern on CT with negative angiography had median 3 (IQR 2, 4). Scenario where any patient with nontraumatic SAH is discharged directly from the ED had median response of 4 (IQR 3, 5). Conclusion: This study found that neurosurgeons consider SAH cases discharged from the ED after full investigation with CT +/– CSF analysis and cerebral angiography, are not clinically significant SAHs. Emergency physicians need to recognize that not all nontraumatic SAH are clinically significant. Future study should identify a classification to distinguish clinically significant from nonsignificant SAH. Keywords:subarachnoid hemorrhage, survey, angiography negative SAH