2005 CAEP / ACMU Scientific Abstracts - Oral Presentations: 1-14

Faster ALS response intervals may improve cardiac arrest survival.

De Maio VJ, Stiell IG, Nesbitt L, Wells GA, for the OPALS Study Group. Department of Emergency Medicine, University of North Carolina, Chapel Hill, NC

INTRODUCTION: To date, no evidence exists to support a survival advantage of advanced life support (ALS) measures in prehospital cardiac arrest. The objective of this study was to analyze survival as a function of ALS response intervals (ALSRI). METHODS: This prospective cohort study included all adult, cardiac etiology, prehospital cardiac arrests occurring prior to EMS arrival in the 17 communities of Phase III of the Ontario Prehospital Advanced Life Support (OPALS) study. All centers provided a tiered EMS response of basic life support (BLS) with defibrillation by ambulance and firefighters and ALS by medics trained to intubate and administer cardiac drugs. All case and survival definitions followed the Utstein style. Survival to hospital discharge was plotted as a function of the ALSRI. Chi square and logistic regression analyses identified the association between incremental ALSRIs and survival. RESULTS: From 1998-2002, there were 3545 arrests managed by ALS paramedics. There were 138 (3.9%) survivors overall. Mean response intervals for ALS and BLS were 7.7 and 5.9 minutes with ALS arriving first in 32% of arrests. The survival curve began to plateau after 6 minutes. Patients responded to earlier had incrementally better survival. Univariate analysis indicated that faster ALSRIs were significantly associated with survival (%; 95% CI): 6 min (5.0%; 3.8-6.3); 5 min (6.4% 4.6-8.2); 4 min (8.1%; 5.1-11.1); 3 min (8.2%; 3.1-13.3). Controlling for known covariates, multivariate analysis provided an adjusted odds ratio of survival for an ALSRI of ≤6 min of 1.6 (95% CI 1.1-2.3). CONCLUSIONS: Faster ALS response may increase the chances of survival despite no improvement on overall cardiac arrest survival through the addition of ALS to a system of rapid defibrillation. This association is similar to that previously identified for BLS and most likely results from quicker times to basic CPR and defibrillation. However, further study is required to identify if other factors may contribute. Key words: cardiac arrest; emergency medical services; advanced life support

OPALS pediatrics study: What is the impact of advanced life support on out-of-hospital cardiac arrest?

Stiell IG, Nesbitt LP, Osmond MH, Campbell S, Gerein R, Munkley DP, Luinstra LG, Maloney JP, Wells GA, for the OPALS Study Group. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

INTRODUCTION: The Ontario Prehospital Advanced Life Support (OPALS) Study is designed to evaluate EMS interventions for critically ill and injured patients. The OPALS Pediatric Study tested the impact on children with out-of-hospital cardiac arrest of adding a full ALS program to existing BLS-D EMS systems. METHODS: This multi-center before-after controlled clinical trial was conducted in 17 communities (population 20,000 to 750,000) and enrolled all children (<16 years of age) with out-of-hospital cardiac arrest during the 36-month BLS-D phase and the subsequent 36-month ALS phase. Paramedics were fully trained to ALS standards including endotracheal intubation and administration of IV drugs. The primary outcome was survival to hospital discharge. Standard univariate chi-square and t-test analyses were performed. RESULTS: The 163 children enrolled during the BLS-D (N = 91) and ALS (N = 72) phases were well matched and had these characteristics: mean age 3.3 (range 0-15), male 52.87%, respiratory etiology 65.6%, unwitnessed 73.6%, bystander CPR 18.4%, initial rhythm not recorded 86.5%, defibrillator at scene <8 minutes 85.0%, defibrillated 1.8%. During the ALS phase, intubation was only attempted for 8 children with 87.5% success; intravenous access was only attempted for 14 children with 50% success. From the BLS-D to the ALS phase, there was no increase in any outcome including hospital discharge (3.3% vs. 2.8%; P = 0.85), hospital admission (11.0% vs. 12.5%; P = 0.61), or return of spontaneous circulation (11.0% vs. 13.0%; P = 0.69). There was no survival improvement for any subgroup including cases witnessed by bystander (10.5% vs. 0%; P = 0.25) or by EMS (20.0% vs. 0%; P = 0.34); BLS vs. ALS paramedic (4.1% vs. 1.5%; P = 0.36). CONCLUSIONS: The OPALS Pediatric Study is the first controlled trial to evaluate full ALS programs for out-of-hospital cardiac arrest in children. The addition of a system-wide EMS ALS program did not improve pediatric survival although few children actually received ALS interventions. Key words: cardiac arrest; pediatric; advanced life support

The clinical reliability and validity of CTAS among British Columbia ambulance service paramedics.

Andrusiek DL, Sheps S, Grafstein E, Sanderson M. British Columbia Ambulance Service, Vancouver, BC

INTRODUCTION: The Canadian Triage Acuity Scale (CTAS) is a five-point triage ordinal scale used in Canadian EDs. The British Columbia Ambulance Service CTAS propose using CTAS as a tool to identify patients who could be left unsupervised in the ED prior to admission. We sought to evaluate the clinical inter-rater reliability and criterion validity of CTAS used by paramedics. METHODS: Driving and attending paramedics who received CTAS training and worked in the Greater Vancouver Regional District (GVRD) from March 8 to 25th, 2004, independently recorded CTAS scores for patients transported to hospital. Kappa statistics were calculated to measure agreement between the driver and attendant. A subset of these data, patients transported to St Paul's Hospital (SPH), were linked using a unique identifier to a computer derived CTAS score. Spearman's correlation coefficient measured criterion validity. Sample size was determined a priori. RESULTS: 2710 patients (69% of all transports) had paired CTAS scores. BLS paramedics evaluated 92% of the patients. Overall probability of agreement between the driver and attendant was 0.82 (0.81 - 0.84). Unweighted kappa was 0.75 (0.73 - 0.77) and quadratically weighted kappa was 0.87 (0.85 - 0.88). Agreement among the BLS paramedics was similar to the agreement within the ALS paramedics: unweighted kappa BLS = 0.74 (0.72-0.76) and ALS = 0.81 (0.74-0.88). Probability of exact agreement for the least acute category (CTAS 5) was 0.64. Agreement between driver and attendant was excellent (unweighted kappa score 0.77) for the 419 patients transported to SPH. However, correlation between paramedic and computer score was 0.44 (0.36-0.52). CONCLUSION: The clinical inter-rater reliability of CTAS is very good between paramedics, but criterion validity is only moderate suggesting that paramedic generated CTAS scores do not adequately distinguish between patients who do and do not require paramedic supervision prior to ED admission. Key words: triage; CTAS; emergency medical services

Is the cerebral performance category score a valid measure of functional outcome after out-of-hospital cardiac arrest?

Nesbitt LP, Stiell IG, Cousineau D, Beaudoin T, Brisson D, Blackburn J, Bordeleau C, Wells GA, for the OPALS Study Group. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

INTRODUCTION: The Cerebral Performance Category (CPC) score measures functional outcome following cardiac arrest but there are few data regarding its validity. We studied the accuracy of the CPC score in predicting health-related quality of life (HRQL). METHODS: This prospective substudy of the Ontario Prehospital Advanced Life Support (OPALS) Study included adult out-of-hospital cardiac arrest patients treated in 20 cities with a mixed BLS-D/ALS EMS system. Survivors were interviewed at 12 months for both CPC and the Health Utilities Index Mark 3 (HUI3), a validated measure of HRQL. CPC is a simple measure of function that ranges from 1 (best) to 5 (brain death) and can be obtained from the hospital chart. HUI3 evaluates quality of life from 0 (dead) to 1.0 (perfect health) and requires a detailed interview and calculation. Data were also collected from ambulance, dispatch, ED, and hospital records. Data analyses included descriptive statistics, 95% CIs, sensitivity, specificity, and kappa statistics. RESULTS: Of 20,867 cardiac arrest patients (1994-2002), 1056 (5.1%) survived to discharge, and 305 (1.5%) completed the HRQL interview and had CPC scored at 12 months. Of the 305 patients: mean age 63.9; male 78.0%; EMS witnessed arrest 25.6%; bystander CPR 32.1%; initial rhythm VF/VT 86.9%. Overall, the median scores were CPC 1 (IQR 1-1) and HUI3 0.84 (IQR 0.61-0.97). For each CPC score, the median HUI3s were (see Table 1, Abstract 4):

Sensitivity of CPC score 1 for good HRQL (>0.80) was 61.8%, specificity 78.9%. Sensitivity of CPC score 3 for poor HRQL (<0.40) was 77.8%, specificity 100%. The weighted Kappa was 0.29. CONCLUSIONS: This is the first study to compare the CPC score with the well-validated HUI3 and demonstrates that most survivors have CPC score 1. An important problem for cardiac arrest research is the limited ability of CPC to discriminate at the high end of quality of life. Key words: cardiac arrest; outcomes; quality of life

CAEP research grants competition: an evaluation of effect and growth over seven years.

Manouchehri N, Grafstein E, Rowe BH. Department of emergency Medicine, University of Alberta, Edmonton, AB

OBJECTIVE: Research funding for Emergency Physicians has been problematic in the past and separate research funding has been proposed by several organizations. This study examines the scholarly outcomes of the projects receiving CAEP Research Grants over its first 7 years of funding. METHODS: From 07/02-10/02, E-mail surveys were sent to 40 lead investigators funded by the national research program between 1997 and 2003. Data collection focused on grant deliverables (project completion, presentations and publications) and opinions regarding the value of the award (1-7 Likert scale). E-mail contact was made with the investigator and up to three reminders were sent. RESULTS: A total of 39 (97%) of 40 principal investigators responded to the survey; 33 projects were completed at the time of contact. Overall, 23 (70%) of completed projects were published or in press as manuscripts; 3 projects produced 2 publications. Presentations were common (n = 97); 14 (14%) local, 35 (36%) national and 48 (50%) international presentations were reported. Award winners commonly (28 {70%}) presented at CAEP meetings. The mean grant amount received from CAEP was $4,189 and 32% of respondents received additional funding; the median total award was $5,000 (IQR: 4,546-9,133). Most respondents felt CAEP funding was important in accomplishing the project (median: 6.5 or 7; IQR: 6, 7); however, not in securing additional funding (3.0 of 7; IQR: 1, 4). Most respondents felt designated grants for emergency researchers was a continuing need (median 7; IQR 7,7). CONCLUSION: Overall, the CAEP Research Grants competition has produced impressive scholarly results, despite the small sums available for funding. The funding is important in ensuring completion of the study and results are widely disseminated in the emergency field. Funding organizations need to consider dedicated funding for EM research so that larger funds and more researchers can be supported on an annual basis. Key words: research grant; CAEP

Communication between emergency departments and community physicians: a survey of Ontario's emergency department chiefs.

Stiell AP, Forster AJ, Stiell IG, van Walraven C. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

INTRODUCTION: Continuity of patient information is required when patient care is transferred between physicians to maintain continuity of care. This survey determined how, and how well, emergency departments (EDs) communicate patient information to community physicians. METHODS: We surveyed each provincial ED chief to determine the most common media and methods of disseminating information used. We employed a modified Dillman approach to the survey, using mail, email, fax, and telephone to contact respondents. We also measured the perceived quality of their system, which was regressed against the hospital teaching status and community size using generalized logits modelling. Finally, we elicited the components of an ideal communication system for the ED. RESULTS: 143 (85.6%) of ED chiefs participated. The ED record of treatment was the most commonly used medium, being the most common medium in 95% of EDs. Regular mail was the most common method of disseminating information, being the most common in 55% of EDs. 33 (23%) chiefs perceived the quality of communicating patient information from their ED as unsatisfactory or inadequate. This perception was significantly more prevalent in larger communities (excellent vs. unsatisfactory OR 44.9 [95% CI 13.9-140] and satisfactory vs. unsatisfactory OR 2.9 [1.6-5.1]) and teaching hospitals (satisfactory vs. unsatisfactory OR 9.7 [4.7-20.3]). 78% of responding physicians felt that patient information should be disseminated using electronic means. Other issues raised were the need for confidentiality, problems with handwriting, and inadequate telephone messaging systems. CONCLUSIONS: To communicate patient information to community physicians, EDs in this province most commonly send a copy of the ED chart by regular mail. More than one fifth of ED chiefs perceived communication from their department as unsatisfactory or inadequate. Practicing physicians believe that the dissemination of patient information would be improved with new technology. Key words: emergency health services; communication

Can emergency department utilization be predicted?

Soucy N, Xue X, Unger B, Lang E, Colacone A, Léger R, Rosenthal S, Afilalo M. Sir Mortimer B Davis Jewish General Hospital- McGill University, Montreal, PQ

INTRODUCTION: Crowding and diminished resources in EDs are growing concerns among the public, health care professionals and government policy makers continent-wide. Managers face difficulties in the daily management of their EDs as the number of visits continue to rise, while hospitals downsize, consolidate or close altogether. Accurate forecasting of ED visits would be beneficial for better planning for upcoming trends and required resources. The aim of this study is to examine the potential of using administrative databases and time-series models to aid in the prediction of patient ED visits. METHODS: This study used provincial ministry database of financial years 1995 to 2004 including demographic, medical and administrative data. Based on only administrative data (i.e. arrival date and time), time-series analytic approach was used to model and forecast monthly ED visits. Monthly data on patient volume for the first 8 years (April 1995-March 2003) were used for model derivation while the remaining one-year data (April 2003-March 2004) were used for model validation. The best forecasting model was automatically selected by the SAS/ETS system according to the smallest mean absolute percent error (MAPE). RESULTS: Twenty-seven EDs (out of 85) who reported complete data for 9 years were included. The MAPE between predicted vs. actual number of ED visits over the validation period is reported. For 27 EDs, the MAPE is 2.13% (Winters method-additive). The MAPEs for 8 tertiary care EDs (full range of specialized services), 14 secondary care EDs (some specialized services) and 4 primary care EDs (no specialized service) are respectively 1.35% (Linear Trend with autoregressive errors), 1.07% (Linear Trend with autoregressive errors) and 2.35% (Seasonal Exponential Smoothing). Finally for a single ED (tertiary teaching hospital), the MAPE is 1.53% (ARIMA). CONCLUSIONS: Time-series methods applied to past ED data showed that accurate forecasting of ED utilization patterns is achievable. Key words: emergency health services; utilization; crowding

Equity of emergency care access in an inner city hospital.

Innes GD, Grafstein E, Harris D, Hunte G, Christenson J. St. Paul's Hospital, Vancouver, BC

INTRODUCTION: Equity is a key quality domain. Our objective was to determine whether marginalized patients receive comparable access to emergency care. METHODS: This cohort study was performed using the ED database in an inner city hospital. Patients with no primary care provider and no fixed address were considered marginalized while patients with stable housing and a primary care provider were controls. Triage nurses trained in CTAS elicited each patient's presenting complaint and assigned a subjective triage level. Without the nurse's knowledge, the ED information system simultaneously linked the patient's presenting complaint to a CTAS-defined (complaint-linked) triage level. As a result, all patients had two triage levels recorded: a standard complaint-linked level and a subjective nurse-assigned level. Our equity marker was whether patients with similar complaint-linked triage levels fell into lower subjective triage levels. Our primary outcome was the proportion of CTAS 1-3 patients who were down-triaged (ie. subjective triage level lower than complaint-linked triage level). RESULTS: See Table 1 for Abstract 8. Over a three-year period, 133696 patients (99.3%) had both triage levels recorded, including 5116 in the marginalized group and 128580 in the control group. Marginalized patients were 81% male (mean, 32.8 years) while controls were 58.7% male (mean, 46.7 years). Overall, 20.3% of marginalized patients and 19.5% of controls were down triaged, but in CTAS levels 1-3, these rates were 42% vs. 34.8% (p < 0.001). Mental health complaints accounted for the largest proportion of down-triages (27% vs. 20% respectively). When MH complaints were excluded from the analysis, down-triage rates converged to 34.5% vs. 32.9% respectively (p = 0.2). CONCLUSIONS: Complaint-linked and subjective triage levels were substantially different. Marginalized patients with mental health complaints may be seen as lower priority. Key words: emergency health services; access; CTAS

Do low-acuity emergency eepartment patients worsen ED crowding? Results from the CROWDED study.

Schull MJ, Kiss A, Katic M. Institute for Clinical Evaluative Sciences, Toronto, ON

INTRODUCTION: It is frequently stated that reducing the number of low-acuity patients presenting to Emergency Departments (ED) would help improve crowding. We studied the impact of low-acuity ED patients on waiting times for other patients. METHODS: We obtained records on all visits to Ontario EDs from April 2002-March 2003. Variables for each ED were computed for consecutive 8hr intervals. The primary outcome was the mean ED length-of-stay (LOS) per 8hr interval for non-low acuity patients; secondary outcome was LOS for all patients. The main predictor was the number of new low acuity ED patients (defined as ambulatory arrival and CTAS 4 or 5 and discharged) in each interval. Covariates were the number of new high acuity (defined as admitted) and medium acuity (defined as neither high or low) patients, patient age, sex, hospital teaching status, time of day and day of week, and total patient-hours during the interval. Auto-regressive modeling was used given correlation in the data. RESULTS: 1095 consecutive 8hr intervals were analysed, during which 4.8 million patient visits occurred at 110 EDs. 49% of patients were male and mean age was 37.7 years. Low, medium and high acuity patients represented 54%, 35% and 11% of all patients. Mean LOS for low, medium and high acuity patients was 2.2, 2.7 and 3.3 hours in teaching EDs and 1.6, 2.1 and 2.8 hours in non-teaching EDs. In multivariable analyses, every 10 additional low acuity patients arriving per 8hrs was associated with a 5.7-minute (p < 0.0001) increase in mean LOS for medium and high acuity patients, and a -3.8 minute (p < 0.0001) reduction in mean LOS for all ED patients. Ten additional high acuity patients was associated with a 71.2-minute increase (p < 0.0001) in mean LOS for medium and high acuity patients, and a 65.0-minute (p < 0.0001) increase in mean LOS for all ED patients. CONCLUSION: Low acuity ED patients have a negligible impact on the LOS of other patients. Reducing the number presenting to EDs is unlikely to reduce waiting times or improve crowding. Key words: emergency health services; crowding; CTAS

Access to care prior to the emergency department visit.

Han A, Russell B, Blitz S, Strome T, Rowe BH. Department of Emergency Medicine, University of Alberta, Edmonton, AB

OBJECTIVE: It is unclear if Emergency Department (ED) patients seek primary or alternative care before ED visits. This study attempted to understand the variety of actions that patients take before presenting to the ED. METHODS: Patients > 17 years were randomly selected from computerized ED records at 2 urban ED sites in Edmonton, AB (UAH; RAH). Following initial triage, stabilization, and informed consent, patients were asked to complete an on-line or paper survey. Survey data were cross-referenced to a minimal patient dataset. The questionnaire asked simple demographic questions, questions regarding reasons for presentation, primary care visit history and preventive health practice. RESULTS: Of the 1425 patients approached, 904 (63%) surveys were completed (UAH: n = 393, RAH: n = 514). Of these 904 patients, 32.7% of patients came to the ED for a problem for which they were already receiving medical treatment. Overall, 34.7% of patients visited a physician before the ED visit (family physician = 14.4%, specialist = 6.5%, walk-in clinic = 10.4%); 13.9% of patients visited an alternative health care professional (e.g., chiropractor, physiotherapist, nurse, etc). Telephone contact also occurred with physicians' offices (29%) or the regional health information line (8.7%) before the ED visit. Finally, 26.5% sought another source of care (e.g., pharmacist, optometrist, etc; self-treatment or researching on the internet) before the ED visit. Some patients (21%) sought no alternate sources of care before coming to the ED and this was not dependent on triage level (p > 0.1). The majority of patients (89.9%) perceived the ED as the best option for their problems. CONCLUSIONS: Most patients, irrespective of triage level, make an attempt to abort an ED visit prior to presenting. While a third seek traditional physician care, a wide variety of alternatives are also used. Despite this access to alternative care, patients perceive the ED as their best option for care. Key words: emergency health services; access; public health

Assessing the quality of reports of systematic reviews in emergency medicine: Are they improving?

Ospina MB, Kelly K, Klassen TP, Rowe BH. Department of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: To describe the scientific quality of systematic reviews (SR) published in major emergency medicine (EM) journals from 2000 to March 2004 compared to the quality and number of SRs published during the 1988-1998 period before the launch of the QUOROM guidelines. To identify factors associated with methodological quality of systematic reviews in EM. METHODS: Retrospective study of all SRs published from January 2000 to March 2004 compared to those published from January 1988 to December 1998 in major EM journals. MEDLINE and EMBASE searches and hand searches of 7 major EM journals were conducted. Two investigators independently selected potential reviews for inclusion and assessed their methodological quality using a validated 10-point scale. RESULTS: From 190 references identified, 45 reviews published from January 2000 to March 2004 met the inclusion criteria (inter-rater kappa = 0.64) compared to 29 from 1988-1998. The scientific quality of the 2000 - 2004 cohort of SRs was low (OQAQ mean score: 2.96; 95% CI: 2.44, 3.48). When compared to the 1988-1998 group of SR (OQAQ mean score: 2.66; 95% CI: 2.09, 3.22), no statistically or clinically significant differences in the overall methodological quality were found (F = 1.39, df = 69; p = 0.24). No change in the methodological quality of the reviews after the introduction of the QUOROM statement was found (Δ: -0.30; 95% CI: -1.08, 0.48). Results of the multiple regression analysis demonstrated that the type of analysis was significantly associated with the quality of the reviews. The OQAQ total score increased by 2 points when the review included a meta-analysis. CONCLUSIONS: While the number of SRs published in the EM literature has increased dramatically, the quality remains relatively poor, despite the publication of QUORUM guidelines in 1999. To ensure the EM community maximizes the potential use of the scientific evidence, it is important to ensure that SRs are conducted and reported using the highest possible methodological standards. Key words: systematic reviews; research; methodology

Reliability and validity of retrospective peer review of the quality of care in emergency medical services.

Andrusiek DL, Sheps S, Abu-Laban RB, Levy A, Reid R, Zumbo BD. Department of Health Care and Epidemiology, University of British Columbia, Vancouver, BC

INTRODUCTION: Peer review is common in EMS, however its reliability and validity is not well studied. We sought to determine the reliability and criterion validity of EMS appropriateness and protocol compliance evaluation performed by peer review. METHODS: Six peers retrospectively reviewed 168 patient care reports (PCRs) for severely injured trauma patients as defined by explicit criteria. Care was rated with a tool derived by a sequential derivation process. Emergency physicians (EPs) prospectively evaluated 118 of the patients with the same tool, blind to the PCR. Inter-rater reliability was determined between paramedics for all 168 PCRs. Criterion validity was determined between the EP (gold standard) and peer rating. Intra-rater reliability was determined from a repeated scoring of 50 PCRs after a washout period. The sample size was defined a priori. RESULTS: The criterion validity correlation coefficient was 0.29 (95% CI 0.06-0.53). The inter-rater reliability kappa score was 0.35 (95% CI: 0.17-0.53). The intra-rater reliability kappa score was 0.62 (95% CI: 0.37-0.86). Three of 14 questions had very good inter-rater reliability: was a prehospital intervention required to manage airway, cervical spine, or orthopedic injuries. Two questions had strong criterion validity: was a prehospital intervention required to manage airway or cervical spine. All questions asking if treatments provided were appropriate or compliant with written protocols had poor inter-rater reliability and criterion validity. CONCLUSION: Intra-rater reliability of retrospective peer review is good, however both inter-rater reliability and criterion validity are frequently poor. In particular, questions rating the appropriateness or compliance of treatments have poor reliability and validity. Although more research is required to fully understand this issue, our results call into question the appropriateness of retrospective peer review of the quality of care in EMS. Key words: emergency medical services; peer review; methodology

Randomized controlled trials in emergency medicine: Where do they all go?

Ospina MB, Kelly K, Klassen TP, Rowe BH. Department of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: To determine the publication fate of randomized controlled clinical trials (RCTs) presented at the 1995-2003 Society of Academic Emergency Medicine (SAEM) meetings. The impact of positive outcome bias, time-lag bias, grey literature bias, and place of publication bias were also explored. METHODS: Prospective cohort study of RCT abstracts presented at 9 consecutive SAEM annual scientific meetings. MEDLINE, EMBASE and CINAHL searches (1995-April 2004) were completed to identify publications in peer-reviewed journals resulting from these abstracts. RESULTS: Of 4399 abstracts, 383 (8.7%) were identified as RCTs. The median time to publication was 32 months (95% CI: 22.97, 41.10) and 59% of the RCT abstracts were published in full within 5 years of presentation. Therefore, the fate of 41% of the RCTs that were not published afterwards is unknown. Positive outcome bias was absent: abstracts reporting positive results were not more likely to be published than those with negative/neutral results. No time-lag bias was found: abstracts reporting positive results were not published faster than those reporting negative/neutral results. A grey literature bias was present in this sample of abstracts: author's conclusions in the scientific abstract differed from those in the corresponding publications (OR: 0.26; 95% CI: 0.098, 0.69). No evidence of place of publication bias was found: studies with positive results were not more likely to be published in higher rated journals than studies with negative/neutral results. CONCLUSIONS: The proportion of emergency medicine RCT/CCT abstracts published is slightly lower than for other specialty societies; however, biases reported by others do not appear to be as common or problematic in emergency medicine RCT abstracts. Given the differences between the results reported in abstracts and manuscripts, caution is warranted with respect to employing meeting abstracts as a source of evidence for future research or systematic reviews. Key words: research; methodology; randomized controlled trials

What do triage nurses believe to be the most important features?

Upadhye S, Eva K, Fernandes CMB, Worster A. Division of Emergency Medicine, McMaster University, Hamilton, ON

INTRODUCTION: To determine the features of emergency triage tools deemed important by triage nurses as a preliminary step toward comparing the acceptability of various triage tools. METHODS: ED triage nurses were surveyed to explore all possible influential features in the triage process. Items were generated from validated tools and triage nurse focus groups. The relative importance of each was rated using a 7-point scale and piloted on a random sample of triage nurses at four EDs. RESULTS: 25 items were generated with alpha = 0.75. The inter-rater reliability of the full sample of 30 nurses was kappa = 0.89 (95% CI: 0.86, 0.91) and the test-retest reliability after 14 days kappa = 0.67 (95% CI: 0.61, 0.72). "Very important/Essential" items included vital signs (mean score 6.8; 95% CI = 6.6, 7.0), triage time (6.5; 6.1, 6.9), allergies (6.4; 6.0, 6.8), medical directives started (6.6; 6.3, 6.9), and pediatric data (6.0; 5.6, 6.4). "Less useful" items included a full review of systems (3.3; 2.7, 3.9), and the tool's physical form (3.7; 3.2, 4.2). CONCLUSIONS: A pilot survey designed to explore the relative importance of specific items used in the ED triage process was created, with good reliability. Key words: triage; methodology