2004 CAEP/ACMU Scientific Abstracts - Poster Presentations (#76 to #92)

CAEP Abstracts

CJEM 2004;6(3):173-212

abstracts:  001-016 | 017-032 | 033-052 | 053-075 | 076-092 | 093-111 | 112-132

076 Inhaled anticholinergics in the treatment of acute asthma.

Spooner CH, Spooner GR, Jones A, Rowe BH. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: Acute asthma in the emergency department (ED) is initially treated with systemic corticosteroids and inhaled beta-agonists (B2). Inhaled anticholinergics e.g. ipratropium bromide (IB) may offer additional bronchodilation. This systematic review was conducted to determine the effect(s) of adding inhaled IB. METHODS: Comprehensive systematic searches were conducted of EMBASE, MEDLINE, CINAHL, and 20 leading respiratory care journals. Studies were included if ED patients with acute asthma were randomized to receive inhaled IB+B2 versus B2 alone. Two reviewers independently performed selection, quality assessments, and data extraction. The relative risk (RR) with 95% confidence intervals (CI) was calculated for dichotomous variables; the weighted (WMD) or standardized mean difference (SMD) with 95% CI for continuous variables using a random effects model. RESULTS: 263 articles were reviewed. Fifteen of 42 potentially relevant citations met inclusion criteria. Results are based on 1992 adults (63% female), age 40.5 yr. SD 8.0) in trials conducted between 1987-2001. Overall, inhaled IB+B2 significantly reduced admission to hospital (RR 0.68; 95% CI: 0.55 to 0.84). The effect size for single or multiple dose protocols showed significant change in pulmonary function at < 1 hour when compared with B2 alone (SMD single dose 0.32; 95% CI: 0.18 to 0.46; SMD multiple dose 0.78; 95% CI: 0.53 to 1.03). This effect continued 2 hrs and beyond. There was a trend to a greater response in the more severe sub-group on multiple dose IB regimens (placebo admission >25%): SMD severe 0.61; 95% CI: 0.29 to 0.94; SMD moderate 0.40; 95% CI: 0.13 to 0.55. There were no differences in other clinical responses and no serious adverse events were reported. CONCLUSIONS: Adding inhaled IB to a beta-agonist early in ED treatment of acute asthma appears to provide moderate benefits to the patient in improved lung function but significantly reduces risk of hospital admission by 32%.

Key words: asthma, bronchodilator, anticholinergic

077 Comparison of two methods of measuring quality of life in acute asthma.

Rowe BH, Wong E, Spooner C, Diner B, Ross S, Mackey D, Tyler L, Senthilselvan A. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: Research on acute asthma is often limited by small study sample sizes and rare outcomes (e.g., death, admission). Quality of life (QOL) measurements have been proposed as an alternative ED outcome, especially after discharge; however, many questionnaires are complex, long and few have been validated in the acute setting. This study compares measures of QOL using a validated QOL measure and a shortened version in asthma patients presenting to the ED. METHODS: Multicentre, prospective study of patients, age 18-55, presenting to the ED with an exacerbation of asthma. Data collection focused on severity assessment and quality of life assessment using the asthma quality of life questionnaire (AQLQ) long and short (mini) forms. Patients received both in staggered fashion, as well as pulmonary functions and other assessments. RESULTS: A total of 137 patients were enrolled in this study; all received both measurements. Most patients were young (mean age: 30.5), 59% were women, and 59% were non-smokers. Most (79%) had a regular physician; however, despite many markers of severity on past history, only 48% were receiving inhaled corticosteroids prior to their ED visit. Long-form AQLQ (1 = worst, 7 = best) scores were low for total (3.3), activity (3.7), symptom (2.6) and emotional (3.0) limitations; environmental (4.7) limitations were less impressive. All improved in follow-up. AQLQ and mini-AQLQ were similar for symptoms (2.5; p = 0.84); however, they differed on a priori environmental (4.7 vs 4.4; p = 0.04), emotional (2.6 vs 3.0; p = 0.002) and activity (3.0 vs 3.7; p < 0.0001) limitation comparisons. CONCLUSION: The Mini-AQLQ is easier to use and provides valuable information on the impressive QOL impairments in patients with acute asthma; however, its psychometric properties do not mimic the validated, longer form of the AQLQ. These early results suggest that further evaluation of the mini-AQLQ should be conducted prior to its widespread use.

Key words: asthma, Quality of life, outcomes

078 Increasing the use of anti-inflammatory agents for acute asthma in the emergency department: experience with an asthma care map.

Rowe BH, Chahal A, Spooner C, Blitz S, Senthilselvan A, Wilson D, Holroyd BR, Bullard M. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: Patients commonly present to the with asthma and considerable treatment variation has been documented. This study examined the use of an asthma care map (ACM) in one Canadian ED to improve asthma guideline adherence emphasizing the use of systemic and inhaled corticosteroids. METHODS: Three time periods were studied: 15 months before introduction (PRE); the 15 months following (P-1), then 18-36 months later (P-2). Randomly selected patient charts from each period were included from patients who were 18-60 years old and presenting with a primary diagnosis of asthma. Primary outcomes included: documentation, use of ED corticosteroids (CS) and prescribing of CS and inhaled corticosteroids (ICS) at discharge from the ED. RESULTS: 387 patient charts were included (150 PRE; 150 P-1; 87 P-2). Patient characteristics in the three groups were similar; however, patients in POST periods showed higher use of newer treatment agents. Overall, more females (209; 54%) were seen than males; the mean age was 32.4 years. The care map was used in 67% of cases during P-1 and 70% during P-2. The use of PEFR monitoring was high during the three periods (91%, 89%, 91%), however, documentation of other markers of severity (e.g., prior exacerbations) increased. The length of stay in the ED increased over the study periods (181, 209, 265 minutes; p<0.01). Use of CS occurred earlier (p < 0.01) and more often (57, 68, 75%; p < 0.01) in the post periods. There was a significant increase in the prescriptions for CS (55, 66, 69%; p < 0.05) and ICS (24, 45, 61%; p < 0.001) at discharge in the post periods. Discharge without any corticosteroids decreased over the three periods (32, 21, 17%; p < 0.05). CONCLUSION: This study provides evidence that a standardized care map was widely accepted, improved chart documentation, improved some aspects of in-ED asthma care, and increased prescribing of preventive medications on discharge.

Key words: asthma, corticosteroids

079 Effect of fentanyl pretreatment on sympathetic response in patients with cerebrovascular accident undergoing rapid sequence intubation in the emergency department.

Zed PJ, Abu-Laban RB, Harrison DW. CSU Pharmaceutical Sciences, Vancouver General Hospital, Faculty of Pharmaceutical Sciences & Division of Surgery, Faculty of Medicine, University of British Columbia, Vancouver, BC

INTRODUCTION: Fentanyl pretreatment is recommended for stroke patients requiring emergency intubation to blunt the sympathetic response to laryngoscopy. The purpose of this study was to evaluate the effect of fentanyl on sympathetic response in patients with cerebrovascular accident (CVA) undergoing rapid sequence intubation (RSI) in the emergency department (ED). METHODS: A prospective, observational study was performed on all patients requiring RSI for CVA in the ED between 1/11/01 and 31/12/03. Patient receiving fentanyl was compared to those not receiving fentanyl in the primary analysis. A dose-related response was also explored comparing patients within the fentanyl group that received <3 mcg/kg to those that received 3-5 mcg/kg. Data on systolic blood pressure (SBP) pre- and immediately post-RSI was collected and between group comparisons were obtained using a t-test. RESULTS: 48 patients were included in the final analysis. In patients receiving no fentanyl (n = 28) pre- and post-SBP were found to be 160.9 ± 39.7 mmHg and 168.9 ± 46.0 mmHg, respectively. In patients receiving fentanyl (n = 20) pre- and post-SBP were found to be 155.7 ± 32.8 mmHg and 154.6 ± 34.7 mmHg, respectively. A statistically significant difference was found for the mean pre-post difference between the two groups (p < 0.0001). Among fentanyl-treated patients, no dose-related response was identified. CONCLUSIONS: The use of fentanyl pretreatment in patients experiencing a CVA in the ED requiring RSI appears to blunt the sympathetic response resulting in SBP elevation. We could not find any evidence for a dose-related effect between higher (3-5 mcg/kg) and moderate (<3 mcg/kg) dosing of fentanyl. A small statistically significant elevation in SBP occurs in patients receiving no fentanyl pretreatment; however, the clinical relevance of this is unknown.

Key words: airway management, rapid sequence intubation, fentanyl

080 The impact of a clinical practice guideline on the management of cellulitis in the emergency department.

Choi D, Bullard MJ, Holroyd BR, Meurer D, Rowe BH. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: Cellulitis is a common infection treated in the ED. Considerable practice variation has been previously documented. A clinical practice guideline (CPG) may reduce this variation. We hypothesized that the introduction of a standardized printed order sheet and subsequent electronic CPG with decision support would decrease practice differences. METHODS: Cellulitis cases were randomly selected for review at a University tertiary referral center, and physicians were unaware of the study at the time of patient contact. Charts were excluded if the primary diagnosis was not cellulitis or if the case was complicated by an abscess. A standardized audit form was used to perform a chart review. We compared three different time periods: before the introduction of the CPG (PRE), after introduction of printed orders (pCPG), and after implementation of an electronic CPG (eCPG). Groups were compared using chi-square, ANOVA and K-W tests. RESULTS: A total of 194 charts (77 PRE; 57 pCPG; 60 eCPG) were included in this study. Patient demographics, presenting signs and symptoms, and triage level were similar among the groups. There was a trend towards the use of fewer antibiotic combinations (12 vs 8 vs 6) which did not reach statistical significance (p = 0.08). Fewer patients received anaerobic agents (22% vs 7% vs 13%; p < 0.05), and an increasing trend towards using a single agent with broad coverage (Cefazolin; 65% vs 67% vs 80%; p = 0.09) was observed. Admission at any time was lower in the CPG groups (13% vs. 2% vs 3%; p = 0.003) and ED visits were less frequent (p < 0.05). CONCLUSIONS: These results suggest that utilizing pCPGs and eCPGs for the treatment of ED cellulitis reduced practice variability, increased appropriate antibiotic selection, and may reduce ED congestion. Future research is needed to examine other eCPGs in the ED setting and to examine the other benefits of using electronic CPG including convenience, ease of access, improved legibility, and reduction in medical errors.

Key words: cellulitis, clinical practice guideline

081 Predicting admission in the absence of pneumonia severity index score.

Blitz S, Rowe BH, Marrie TJ. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: This study was designed to determine predictors of hospital admission for community acquired pneumonia (CAP) in patients for whom there was insufficient data to calculate a pneumonia severity index (PSI) class. METHODS: Between 11/2000 and 11/2002, all CAP patients presenting to 6 regional emergency departments (ED) were assessed according to a specified treatment pathway. ED physicians were encouraged to compute a PSI score on those patients who may have required admission. This analysis considers the patients in whom a PSI score was not determined. Logistic regression was used to determine predictors of admission. Sensitivity and specificity was used to determine cutoff values for continuous variables. Odds ratios (OR) and 95% confidence intervals (CI) are reported. RESULTS: 1362 of 3818 entered on the pathway did not have a PSI. 774 patients with complete data on the variables considered for inclusion in the model were included in the final analysis. Oxygen saturation <92 (OR: 11.6; 95% CI; 7.6, 17.7), age >68 (OR: 4.5; 95% CI: 2.9, 7.0), altered mental status (OR: 4.3; 95% CI: 1.7, 11.0), triage score <4 on a scale of 1 to 5 (OR: 3.1; 95% CI: 1.8, 5.5), pulse >94 (OR: 2.0; 95% CI: 1.2, 3.1) and respiratory rate >25 (OR: 1.7; 95% CI: 1.1, 2.6) were significant predictors of admission. Results were transformed to scores and summed to yield an admission risk score, which ranged from 0 to 24. Risk scores were classified into low (<5), moderate (5-8) and high (>8) probability of admission, based on ROC cut-points. Admissions occurred in 2%, 13%, and 60% (p = 0.001), respectively; death occurred in 0%, 1.4%, and 7.5% (p = 0.001), respectively. CONCLUSIONS: An admission risk score based on common variables, which are easily collected in the ED, can successfully predict the probability of admission for CAP patients. The same score was predictive of death; however, not of relapsing back to the ED (data not shown). This model requires prospective validation prior to implementation.

Key words: pneumonia, clinical prediction rule

082 Rapid serum procalcitonin testing for the diagnosis of infection in the ED.

Juquel B, Abderrahim N, Mezaib K, Simon N. Dept. Accueil Urgences Hopital de Poissy

INTRODUCTION: Procalcitonin has been shown to be an early specific biological marker for infection. The use of PCT in emergency situations could influence clinical assessment of patients. METHODS: all adult patients with at least two criteria of systemic inflammatory response syndrome (SIRS) were evaluated by the emergency physician (EP) in the usual manner based on clinical data and standard biological testing (CBC; CRP, urinalysis; LP and pleural or articular or abdominal puncture as indicated; imaging and US ). All patients were tested with a semi quantitative PCT assay (PCT-Q BRAHMS). Levels under 0.5 ng/ml were considered as low and levels over 2 ng/ml as high and indicative of an infectious process. The Emergency physician blinded to PCT results determined a pre test probability (P1) for infection, a diagnosis, patient disposal and the estimated need for antibiotic treatment. PCT levels were then disclosed to the EP and a second post-test probability for infection is determined (P2) as well as any change of diagnosis, disposal or treatment. RESULTS: 212 patients aged 58 y (+/- 23,7) could be evaluated, 44 (21%; CI 15.7% - 27.2%) had been on antibiotic treatment during the previous 8 days. PCT levels were low (<2 ng/ml) in 122 pts (58.4% CI 51%-65%); indeterminate (>0.5 < 2 ng/ml) in 38 pts (18% CI 13%-24%) and elevated >2 ng/ml in 49 pts (23% CI 18%-30%) sensitivity of PCT for infection was 89% (CI 83% -95%)- and specificity 93% (CI 88%-98%) in this series. P1 and P2 did not differ by more than 10% in 86 pts (41.3% CI 34.6%-48.4%) Compared to P1, P2 was lower in 41 patients (19.7% CI 14.5%-25.8%) and higher in 81 pts (38.9% CI 34.6%-48.4%). Diagnosis was changed in 54 pts (25% CI 17%-33%). Intended antibiotics were suspended in 34 pts and inversely introduced in 5 pts. Hospitalization was no longer deemed necessary for 10 (4.7% CI 2.9%-6.5%) patients. CONCLUSIONS: PCT might help to improve the diagnosis of sepsis in patient with SIRS preventing unnecessary antibiotics and hospital admissions.

Key words: Procalcitonin, sepsis, diagnosis

083 Pediatric emergency department staff perceptions of infection control measures against severe acute respiratory syndrome.

Parker MJ, Goldman RD. Department of Pediatrics, The Hospital for Sick Children; University of Toronto, Toronto, ON

INTRODUCTION: Severe Acute Respiratory Syndrome (SARS) is a respiratory illness recognized on several continents. Our city is one of the first major centres affected by SARS prompting the government to mandate rigorous infection control measures. Health care workers' perceptions of the effectiveness and practice of infection control measures during an evolving outbreak are previously unreported. METHODS: All medical staff of the Pediatric Emergency Department (PED) in a tertiary medical centre completed a written questionnaire over 7 days near the onset of the SARS outbreak. Level of concern regarding SARS and perceptions of the effectiveness and use of infection control measures were assessed on a 5-point scale. Statistical analysis was performed with SPSS using Chi-square and ANOVA tests with significance at p < 0.05. RESULTS: Response rate was 97% (116/120). Using isolation rooms (mean score 4.6/5), wearing a mask when examining patients (4.5/5) and handwashing (4.5/5) were considered most effective. Staff physicians reported handwashing more than nurses and trainees (4.9/5 vs 4.5/5 and 4.5/5, p<0.05) while other measures were reported equally. Those who considered SARS a high public health threat reported higher compliance with handwashing (4.8/5 vs 4.4/5), wearing a mask at all times in the PED (3.9/5 vs 3.2/5) and gloves (3.6/5 vs 2.9/5)(p < 0.05), but not eye protection (3.4/5 vs 3.0/5), gown use (4.9/5 vs 4.7/5) or wearing a mask when examining patients (5.0/5 vs 4.8/5) (p = NS). Staff who considered combined infection control measures effective in protecting patients and PED staff did not report increased compliance. CONCLUSIONS: Eye protection was perceived as only moderately effective in protecting against the spread of SARS and reported compliance was relatively poor among PED staff. Concern of SARS as a public health threat rather than perceived effectiveness of infection control measures appears to have a greater impact on compliance.

Key words: SARS, infection control

084 Overuse of narcotic analgesics in the ED treatment of acute migraine headache.

Rothney A, Colman I, Wright SC, Zilkalns B, Rowe BH; Division of Emergency Medicine, University of Alberta, Edmonton, AB

OBJECTIVE: Acute migraine headache is a common presentation to the emergency department (ED). Treatment with narcotics is potentially ineffective and may lead to abuse. This study examined the treatment practice variation across five linked EDs in one Canadian center, focusing on the use of narcotic analgesics and factors associated with their use. METHODS: Five hundred acute migraine headache patient charts were randomly selected from five Canadian EDs. Charts underwent a structured review to determine medication use. Data were analyzed, comparing those who received narcotics as first-line treatment to those who did not, using Chi-square, t-tests, and logistic regression. RESULTS: The majority of patients (59.6%) received narcotics as first-line treatment. Numerous factors were associated with first-line narcotic treatment, but only four variables remained significant in the final multivariate model (all p < 0.002). Having taken anti-headache medications prior to ED presentation (OR: 2.63; 95% CI: 1.53, 4.51) and hospital of presentation other than Hospital A (e.g., Hospital D - OR: 6.32; 95% CI: 2.76, 14.46) increased the odds of receiving first-line narcotics. Having received a more urgent triage score (OR: 0.4; 95% CI: 0.24, 0.65) or having a longer duration of headache (OR: 0.994; 95% CI: 0.99, 0.99) decreased the odds of receiving first-line narcotics. CONCLUSION: The results suggest that acute migraine management in these EDs does not meet current consensus guidelines. Factors associated with narcotic use are predictable and ED physicians should make a concerted effort to replace narcotics with more evidence-based first-line treatments in this patient group in the future.

Key words: headache, analgesia, opioid, migraine

085 Does income level correlate with admission to hospital from the emergency department?

Spence JM, Murray MJ, Morrison LJ. Division of Emergency Medicine, St. Michael's Hospital, University of Toronto, Toronto, ON

OBJECTIVES: Socioeconomic status (SES) has been shown to be related to health care outcomes for many conditions. This study was designed to evaluate the relationship between SES and admission to hospital. METHODS: A retrospective analysis of emergency department (ED) visits from a community hospital in Ontario for 1999 (N = 68,757) was conducted. Mean household income was estimated using postal code conversion of neighborhood-level data from the 1996 Canadian Census. Where income level was suppressed by Statistics Canada due to low sample size, mean income was imputed. Income quartiles were used for univariate and logistic regression analyses. RESULTS: Income data was missing or suppressed for 10% of visits. The mean income for patients was $53,950 (range $11,110-262,450). Mean household income for each quartile was Q1 $38,680, Q2 $49,880, Q3 $57,230, and Q4 $69,750. Patients in the lowest quartile (Q1) were 41% more likely to be admitted when compared to the highest income (Q4) (Q1 14.0% vs Q4 9.9%; p < 0.001). Odds ratios (OR) for admission in each quartile were: Q1 1.47 (95% CI 1.38-1.58), Q2 1.21 (95% CI 1.12-1.30), Q3 1.22 (1.14-1.30), Q4 1.00. Differences were most marked in the pediatric age group (Q1 OR 1.48 (95% CI 1.25-1.75)). The unadjusted OR for lowest income quartile and admission, with all income levels as reference, was 1.28 (95% CI 1.22-1.35). The relationship between admission and lowest income quartile remained significant after adjusting for age, sex, triage level, diagnosis, and time of visit (adjusted OR 1.07; 95% CI 1.01-1.14). CONCLUSIONS: Even in a setting with government-sponsored healthcare, patients from lower SES neighborhoods appear to have increased odds of admission to hospital. The study has the following limitations. It represents one Canadian community hospital. Imputing mean income for missing data minimized income related differences among groups. Finally, community level income data were used as a surrogate for individual level income data.

Key words: Emergency health services, public health

086 Does triage level (Canadian Triage and Acuity Scale) correlate with other factors associated with emergency department visits?

Murray MJ, Levis G. Royal Victoria Hospital, Barrie ON

INTRODUCTION: The CTAS is used in the majority of Canadian emergency departments. The relationship between CTAS level and various other factors associated with Canadian emergency department visits have not been studied. OBJECTIVE: The purpose of this study is to determine if a relationship exists between triage level and: arrival by ambulance rates, admissions rates to hospital, admission rates to the ICU/CCU, and transfer rates to other institutions for specialized care. METHODS: This was a retrospective study of a database of emergency department visits in a large community hospital for a 26-month period from April 2001 to June 2003. The Z-test for difference in two proportions was used to see if there was a statistically significant difference between the percentage rate for a given triage level and the next lower level. RESULTS: The information was available for all patients (n = 164,812) The distribution of patients over the triage levels from highest acuity to lowest was: CTAS 1 = 0.5%, CTAS 2 = 11.1%, CTAS 3 = 32.7%, CTAS 4 = 42.2 %, CTAS 5 = 13.5%. This is similar to other community hospitals of the same volume. For each of the factors studied: arrival by ambulance rates, admissions rates to hospital, admission rates to the ICU/CCU, and transfer rates to other institutions for specialized care, there was a significant difference (p<.001) in the percentage rate for a given triage level and the next lower level for all CTAS levels. A statistically significant difference (p<.001) for admission rates across all triage levels existed for all age groups studied (0-17 years, 18-34 years, 35-64 years and 65 years and greater. CONCLUSIONS: CTAS level which defines level of acuity correlates with: rates for arrival by ambulance, admission rates to hospital, admission rates to CCU/ICU, and transfer rates to other institutions for specialized care. The correlation exists for admission rates across multiple age groupings. CTAS may be a valid predictor for each of these factors.

Key words: Canadian triage and acuity scale, triage

087 Adverse events identified following an emergency department visit.

Rose NGW, Forster AJ, van Walraven C, Stiell IG. Department of Medicine, University of Ottawa, Ottawa, ON

BACKGROUND: Many studies demonstrate a high rate of treatment related adverse outcomes or adverse events. There have not been any studies evaluating the risk in patients sent home from the emergency department (ED). We wished to identify the incidence and cause of adverse events in patients discharged home from an ED. METHODS: Patients were included in the study if they were to be sent home directly from the ED of an urban, academic teaching hospital in Ottawa, Canada. Patient records were reviewed to identify demographic and medical history information. Two weeks following the ED visit; patients completed a standard telephone interview to record post ED visit outcomes. Two physicians reviewed all outcomes to identify all adverse events and their cause. RESULTS: Follow-up was complete for 399 of 408 enrolled patients (response rate = 98%). 24 patients experienced an adverse event (incidence = 6% [95% CI = 4%-9%]). 17 adverse events were caused by errors (incidence = 4% [95% CI = 3%-7%]). Errors occurred in 5 of 78 chest pain patients (incidence = 6% [95% CI = 3%-14%]). An additional five adverse events were unpreventable medication side effects and two were minor, procedure-related complications. 15 adverse events led to an additional ED visit or a hospitalization. CONCLUSION: We found a small but important rate of adverse events in patients discharged from the ED. Adverse events following an ED visit are less frequent than following hospital discharge.

Key words: error, emergency health services

088 ED discharge instruction comprehension and compliance study.

Clarke C, Friedman SM, Shi K, Culligan C. Division of Emergency Medicine, DFCM, University of Toronto, Toronto, ON

INTRODUCTION: To assess patient comprehension of ED discharge instructions and correlates of patient compliance. METHODS: Prospective cohort study in a downtown teaching hospital. Patients were solicited for a structured interview and reading test after discharge, and follow-up telephone interview two weeks later. Two blinded physicians scored comprehension and compliance. Inter-rater variability was assessed using a kappa-weighted statistic, and testing for correlations using Spearman's Correlation and Fisher's exact test. RESULTS: 88 of 106 patients approached (83 %) were enrolled. The inter-rater reliability of physician rating scores was strong (Kappa = 0.66, P < .0001). Approximately 60% of subjects demonstrated reading ability at or below a Grade 7 level. Comprehension was positively associated with reading ability (R = 0.29, P <.01) and first language (R = .27, P = .01). Reading ability was positively associated with first language (R = .24, P = .03) and years of education (R = 0.43, p < .0001), and inversely associated with age (R = -.21, P = .05). First language not English and need for translator were generally associated with poorer comprehension of discharge instructions but unrelated to compliance. Compliance with discharge instructions was correlated with comprehension (R = 0.31, P = .01). Compliance was not associated with age, language, education, years in anglophone country, reading ability, format of discharge instructions, follow-up modality, or affiliation with FP. CONCLUSIONS: ED subjects demonstrated poor reading skills. Comprehension was correlated with reading and language skills. Compliance was correlated with comprehension. Physicians may anticipate poor comprehension and employ alternate modalities of discharge instruction and follow-up to optimize compliance.

Key words: emergency health services

089 Age and region differences in emergency department users in Winnipeg.

McMaster R, Kozyrksyj AL, Martens R. Faculty of Medicine, University of Manitoba, Winnipeg, MN

INTRODUCTION: Over 200 000 visits were made to Adult Emergency Departments (ED) in Winnipeg during one year; however, without a comprehensive ED data abstraction system, little else beyond weekly counts are reported. METHODS: This descriptive study used systematically collected data from all six adult Emergency Departments in Winnipeg from April 1, 2001 to March 31, 2002. Age, triage, registration date and time, discharge date and time, discharge status, and postal code were available for each person registering for care. RESULTS: During the year, 45% of visits were made by young adults (17-44 years) while just fewer than 5% were made by the oldest old (85+ years). By contrast, hospital admission rates were 45% in the 85+-age group compared to 9% in the 17-44 age group. Four-fold differences in the admission rates for young adults were observed over individual hospitals. Over 40% of visits were for non-urgent care in young adults; in the oldest old this decreased to 23%. Geographically, visit rates overall age groups showed large variation throughout the city (161 to 456 per 1000 population). The older age group showed less variation and higher rates overall (323 to 618 per 1000 population). Visit rates were highest in the inner city, regardless of age. Time spent in the ED showed little variation in the young adult age group (on average, 4 hours), but in the oldest old this time ranged from 10 to 18 hours among hospitals. CONCLUSIONS: This study illustrated age and region differences in the utilization of Winnipeg Emergency Departments. Although the older age groups made proportionately fewer visits than younger groups, their visits were more urgent, resulted in admissions to hospital more often, and accounted for longer lengths of time spent in the emergency department.

Key words: emergency health services

090 Triage of patients in the emergency department: comparing a computerized triage tool to standard triage methods.

Dong SL, Bullard MJ, Meurer DP, Blitz S, Colman I, Rowe BH. Division of Emergency Medicine, University of Alberta, Edmonton, AB

INTRODUCTION: Emergency department (ED) triage prioritizes patients based on severity and urgency of care. The Canadian Triage and Acuity Scale (CTAS) is the nationally recognized standard; however, psychometric testing of CTAS has been rare. A web-based triage decision tool (eTRIAGE) based on CTAS has been developed. We describe agreement between triage staff using eTRIAGE and current triage methods (paper-based). METHODS: This prospective study enrolled consecutive patients presenting to a tertiary care ED in a large urban centre. Patients were assessed by triage staff and assigned a paper-based score. A second nurse independently assessed these same patients using eTRIAGE. Agreement between paper-based and eTRIAGE scores are reported using kappa statistics and compared to patient admission status. RESULTS: Over a five week period, 722 ED patients were enrolled; complete data were available from 693 (96%) score pairs. Patient acuity using paper-based method was as follows: 1%, 3%, 45%, 49%, and 2% for CTAS levels 1 to 5, respectively. Using eTRIAGE the acuity was: 1%, 19%, 41%, 29%, and 9%. Kappa agreement between paper-based triage and eTRIAGE was fair (kappa = 0.20, 95%CI 0.15, 0.25; weighted kappa = 0.36, 95%CI 0.31, 0.42) using exact CTAS score agreement; however, improved to good if agreement was considered to be within one CTAS level (kappa = 0.73, 95%CI 0.64, 0.82). Admission rate was significantly different for CTAS level 2 between the paper-based method (73.7%) and eTRIAGE method (39.2%). CONCLUSION: Previous triage studies consider one-level difference to represent agreement. This study showed significantly different conclusions when using exact-level agreement. The admission rate disparity between paper and electronic triage level 2 underscores the need to reliably assign triage level in EDs, especially if used to determine resource allocation.

Key words: Canadian Triage and Acuity Scale, triage, reliability

091 Quelques défis de la recherche en soins pré-hospitaliers.

Lafleur I, Lortie G, Dallaire C, Ulmer Z, Lapointe J.École de service social, Université Laval, Lévis, PQ

INTRODUCTION : Une recherche portant sur la chaîne d'intervention préhospitalière dans Chaudière-Appalaches est en voie d'être terminée. Un de ses objectifs était d'identifier les indicateurs déjà disponibles qui serviront à évaluer la mise en place de nouveaux services et protocoles. Cet objectif est primordial, car l'évaluation des nouvelles façons de faire dans les différents maillons de la chaîne contribuera directement à la qualité des soins aux usagers. La communication consistera à présenter quelques-uns des défis posés par différents indicateurs. Ces défis seront illustrés à l'aide de deux cas, soit l'infarctus du myocarde et les polytraumatisés de la route. MÉTHODES : Il s'agit d'une recherche exploratoire et descriptive effectuée à partir de l'analyse de documents écrits, de banques de données et d'entrevues auprès d'acteurs-clés. RÉSULTATS : La principale démarche méthodologique privilégiée, soit recourir à des sources de données secondaires (documents écrits et banques de données), a permis d'évaluer les possibilités et les limites qu'offrent ces sources de données pour l'évaluation de projets dans le domaine. Ainsi, plusieurs indicateurs pouvant servir à l'évaluation existent déjà et certains de ces indicateurs se retrouvent dans les différents maillons de la chaîne. Cependant, leur mise en commun soulève des défis importants, car 1) ces indicateurs ont des définitions propres à chacun des maillons de la chaîne et parfois à chacun des acteurs d'un même maillon; 2) ils se retrouvent dans plusieurs banques de données 3) le recours à des diagnostics pour préciser la recherche présente aussi plusieurs embûches, etc. CONCLUSION : Nos conclusions sont à l'effet que ces sources de données sont nombreuses et accessibles. Néanmoins, à partir de ce constat, un long travail reste à faire pour pouvoir utiliser les indicateurs de ces sources de données lors de l'évaluation des projets que la région et la province veulent mettre en place. Mots clés : indicateurs, préhospitalier, protocoles

092 Emergency physician time and motion study.

Elinson R, Friedman SM. Division of Emergency Medicine, DFCM, University of Toronto, Toronto, ON

INTRODUCTION: A time and motion analysis of emergency physician (EP) function. To characterize EP time utilization, patterns of communication and task interruption, and potential areas for optimization of EP function. METHODS: A physician observer shadowed 11 EPs and longitudinally documented physician tasks in a standardized manner. This was a nonrandomized convenience sampling over 16 periods of approximately six hours each, including days, evenings, overnights, weekdays and weekend shifts. Statistical analysis was performed using SPSS and SAS. Correlations were identified using Spearman's Coefficient and verified using Fisher's exact test. RESULTS: EPs ranged in age from 29-55, and reported 1-35 years in practice and diverse certifications. 2,889 tasks were characterized by duration and 13 activity types over 5 507 minutes (approx 91.8 hours). EP time was devoted primarily to patient interaction (27.7 %), chart review and entry (16.7 %) and computerized data retrieval and entry (11.8 %). EPs were interrupted 400 times (approx every 13.8 minutes).10 % of interruptions resulted in a switch of EP task, and 9.75 % of interrupts required a transit of 3 metres or more, often across the department. Most frequent sources of interruptions were nurse (53.7 %), physician (31.8%), and family member (5.8 %). Rate of interruption was correlated with shift intensity, as measured by delay to patient assessment (P = 0.005), and negatively associated with certification in emergency medicine (P = .0002), and years in practice (P = 0.01), but not with MD age or time of shift. CONCLUSIONS: EPs devote considerable time to nonclinical tasks. EP interruption rate increases as shifts become busier. Attention should be devoted to EP task delegation and optimization of communication in the ED.