CAEP/ACMU 2002 Scientific Abstracts: 30-44

2003 Scientific Abstracts

CJEM 2002;4(2):124-154

030 Evaluation of the Pediatric Risk of Admission (PRISA) Score in a Pediatric Emergency Department.
Gravel J, Gouin S, Bergeron S, Amre D, Lacroix J. Divisions of Emergency Medecine, Research Institute and Intensive Care, Department of Pediatrics, Sainte-Justine Hospital, Montreal University. Montreal, QC.

BACKGROUND: Few studies have evaluated the case severity of patients visiting a pediatric Emergency Department (ED). Recently, a pediatric risk of admission score (PRISA) was developed to predict the risk for children of being hospitalised on the basis of their ED evaluation. It was retrospectively evaluated in 1 centre. OBJECTIVES: To evaluate the predictive value of the PRISA score with respect to the risk of admission in a pediatric ED. METHODS: Prospective cohort study conducted in a pediatric tertiary centre ED with 65,000 patient-visits annually. From November 1st 2000 to October 31st 2001, 3 periods of 8 hours per month, were randomly chosen. During these periods, all patients triaged to the ED were evaluated. Data required for calculation of the PRISA score were collected by a single investigator uninvolved in the treatment of patients before the decision regarding admission or discharge was made. Analysis of the odds ratio of each variable for admission and PRISA score were calculated. The criterion validity of the score was evaluated. RESULTS: During the study periods, 1,930 patients were evaluated. Of these patient-visits, 203 admissions were observed while the PRISA score predicted 231. The overall accuracy of the score was 97%. It had a sensitivity of 0.94, a specificity of 0.98, a positive predictive value of 0.82 and a negative predictive value of 0.99. The Hosmer-Lemeshow goodness-of-fit test demonstrated good agreement in consecutive deciles of admission probability: X2 was 28.15 (P = 0.78). The area under the receiver operating characteristic curve was 0.79. Some variables of the score did not reach statistical significance as risk factor of admission in our population. CONCLUSION: Probability of admission can be predicted with the PRISA score in a pediatric ED. Some variables could be deleted or modified in order to optimise the accuracy of the score.

031 How Accurate and Reliable is Mental Status Evaluation in Minor Head Injury Patients?
Stiell IG, Rowe BH, Clement C, Morrison L, Bandiera G, Dreyer J, MacPhail I, Holroyd B, McKnight RD, Schull M, Greenberg G, Worthington JR, for the CCC Study Group. Division of Emergency Medicine, University of Ottawa. Ottawa, ON.

OBJECTIVES: Many physicians take a casual approach to mental status evaluation in minor head injury patients. This study assessed the accuracy and reliability of the Glasgow Coma Scale and the Object Recall Test. METHODS: This prospective cohort study was conducted in 10 tertiary care EDs and involved adult head injury patients who presented with GCS 13-15 after loss of consciousness, amnesia, or confusion. Physicians made standardized assessments of the Glasgow Coma Scale and the Object Recall Test. 2nd physicians performed interobserver assessments for 101 cases. The 2-minute Object Recall Test consisted of showing patients 3 objects and asking them to recall the objects after 2 minutes and was scored from 0 to 3. Patients then underwent CT to determine the outcome criterion, clinically important brain injury. Data analyses included univariate association, kappa, sensitivity, specificity, odds ratio by multivariate logistic regression. RESULTS: We enrolled 3,121 patients with mean age 38.7 (16-99) and important brain injury 8.1%. This Table shows % of injury (N = 254) and non-injury (N = 2867) patients with findings, p-value, kappa, sensitivity, specificity, adjusted odds ratio.

CONCLUSIONS: The most reliable and accurate mental status findings for identifying patients at risk for brain injury are 'initial GCS 13' and 'GCS <15 2 hours after injury'. Decision rules for the management of minor head injury should incorporate these findings.

032 Evaluation of the Canadian Paediatric Triage and Acuity Scale in an Emergency Department.
Gravel J, Bergeron S, Amre D, Gouin S. Division of Emergency Medecine, Department of Pediatrics, Research Institute, Sainte-Justine Hospital, Montreal University. Montreal, QC.

BACKGROUND: The Canadian Paediatric Triage and Acuity Scale (PaedCTAS) was proposed for the triage of pediatric patients coming to an Emergency Department (ED). The validation of this 5-category triage tool was not previously reported. Admission rate and Pediatric Risk of Admission Score (PRISA) may be used for case severity evaluation. OBJECTIVES: To compare the PaedCTAS to a previous triage tool with respect to admission rate, diagnostic and therapeutic interventions and PRISA score. METHODS: A before-and-after study was conducted in a pediatric tertiary centre ED. Data were prospectively collected for 4 months before (Feb-May 2001) (group pre) and for 4 months following (July-Nov 2001) (group PaedCTAS) the implementation of the PaedCTAS. Three periods of 8 hours were randomly chosen each month. During these periods, all the patients triaged to the ED were included. A single investigator, not involved in the treatment of the patients, collected the variables of interest. RESULTS: 744 patient-visits were analysed in the group pre and 537 in the group PaedCTAS. Both groups had similar characteristics: chief complaints, admission rates and mean PRISA scores. In group PaedCTAS, more patients were triaged in the most urgent categories (64% vs 53%, p < 0.05) but the admission rates for these categories were lower (15% for group PaedCTAS vs 37% for group pre, p < 0.05). The Goodness-of-fit test demonstrated a weaker correlation between triage score and admission rate for the group PaedCTAS vs group pre (p = 0.43 vs p = 0.85). Both triage tools predicted similarly the interventions required (hemoculture, IV bolus, nebulizations) and the PRISA score. CONCLUSION: The PaedCTAS seems to attribute a higher level of severity to patients but is less accurate for the prediction of admission than a previous triage tool. The PaedCTAS does not appear to perform, as a more discriminatory triage tool regarding interventions and treatments needed than a previous triage tool.

033 Plain Gut versus Non-Absorbable Nylon Sutures in Traumatic Pediatric Lacerations: Long-Term Outcomes.
Karounis H, Gouin S, Eisman H, Chalut D, Pusic M, Morin I, Williams B. Divisions Of Emergency Medicine, Biostatistics & Plastic Surgery, The Montreal Children's & Ste-Justine Hospitals. Montreal, QC.

OBJECTIVE: To compare long-term cosmetic outcomes in pediatric traumatic lacerations repaired with absorbable plain gut versus non-absorbable nylon suture material. METHODS: Randomized clinical trial conducted in a pediatric Emergency Department (ED). Patients (1-18 year-old) who presented to the ED with lacerations <12 hours old were recruited between Jan 1999 and Dec 2001. Patients were randomized into 1 of 2 groups: absorbable plain gut sutures (group A) and non-absorbable nylon sutures (group NA). BE/BC pediatric emergency physicians performed laceration repair in a standardized approach. The patients were evaluated at 4 months by a plastic surgeon blinded to the suture material used for wound repair. The wounds were assessed using the visual analog scale (VAS) of cosmesis (0-100 mm) and the wound evaluation score (WES) composed of 6 items: a score of 6/6 was considered optimal. In addition, the need for any surgical revision was noted. To detect a 12 mm difference on the VAS (alpha = .05, beta = .10), 43 patients per group were needed. RESULTS: 137 patients were eligible of which 49 patients declined to participate. Of the 88 patients enrolled, 45 were randomized to group A and 43 to group NA. Both groups had similar demographics (sex, age), wound size (length, width), wound location and mechanism of injury. No significant differences were found in the mean of VAS (±1 SD) between group A and group NA (82 ± 13 mm vs 73 ± 22 mm, P = 0.19). No significant differences were found in the proportion of optimal WES between group A and group NA (45% vs 50%, RR; 95% CI = 0.91; 0.44, 1.88). None of the wounds required surgical revision. CONCLUSIONS: The use of absorbable sutures in the repair of traumatic lacerations in children appears to be an acceptable alternative to non-absorbable sutures as long-term cosmetic outcomes and complication rates are similar.

034 Utilization and the Aging of Emergency Department Patients in Ontario: a Tale of Two Trends.
Chan B, Schull M, Schultz S. Institute for Clinical Evaluative Sciences, Sunnybrook & Women's College Health Sciences Centre, University of Toronto. Toronto, ON.

OBJECTIVE: To determine trends in emergency departments (ED) utilization in Ontario. METHODS: Over the study period (1993 to 2000), most emergency physician services in Ontario were remunerated on a fee-for-service (FFS) basis. We obtained data all FFS physician billing records for ED patient visits. We excluded scheduled visits and those where patients left without being seen, were not assessed, lacked a health insurance number, were workplace accidents, or direct referrals to non-EM specialists. Utilization was analysed by age-sex groups, and rates were calculated using direct standardization techniques and the 1996 census. Statistical tests were not conducted since we included virtually the entire ED patient population. RESULTS: In 2000, there were 3.27 million visits to Ontario EDs. Out of approximately 11.5 million people in the province, 2.25 million, or almost 1 in 20, made at least 1 visit to an ED. Children <5 years and the elderly had the highest rates of use: 34% of children <5 years and 29% of patients >75 years visited an ED at least once, compared to 18% for patients 5-74 years. From 1993 to 2000, utilization rates among all age-sex groups <55 years decreased, but they increased in all groups >55 years. Age-sex specific utilization rate changes ranged from a decline of 24.7% for males 14-19 years of age to an increase of 15.5% for males >90 years. Overall, per capita ED use declined by 10.3%. The average age of ED patients rose by 4.1 years during the study period, as compared to a 1.3 year increase for all Ontarians. CONCLUSION: The overall rate of ED utilization has decreased by 10% since 1993, due to substantially lower rates among individuals <55 years. However, utilization has increased among individuals >55 years. The average ED patient is now older than the average citizen of Ontario.

035 Construct Validity of a Prehospital Acuity Scale: Case Severity Rating Scale.
Morrison LJ, Vermeulen MJ. Sunnybrook & Women's College Health Sciences Centre. Toronto, ON.

OBJECTIVE: To evaluate the construct validity of a 5-point acuity scale (1 = minor [Mi], 2 = moderate [Mo], 3 = severe [S], 4 = life-threatening [LT], 5 = vital signs absent [VSA]) using operational and clinical benchmarks. This scale has inter- and intra-rater reliabilities of 0.98 and 0.99, respectively, and generalizeability of 0.97 for EMS credentials. METHODS: An EMS database of 106,552 (74% of total) emergency ambulance calls in 2000 was used; accuracy of recorded data was 84% and data entry reliability was 0.57. A valid case severity code was entered in 94,299 calls (89%) of which 92,705 (98%) had a value of 1-5 (codes 6 and 7 define death as pronounced in field or obviously dead). The scale was evaluated using dispatch and return priority, patient status change scale, selected drugs or procedures and time intervals. RESULTS: High priority dispatch was associated with a higher severity code (%, 95% CI): severe = 75.0 (74.3, 75.7), life-threatening = 91.0 (89.0, 90.9). A high return priority was found in 58.7% Mi calls, (58.2, 59.2) and increased in all categories to 90.7% in LT (89.7, 91.6). In the status change scale, the majority of patients remained unchanged; improvement rates were 19.2% in Mi (18.7, 19.6), 26.1% in Mo (25.6, 26.6), 30.8% in S (30.0, 31.5), 33.3% in LT (31.7, 34.9), 17.9% in VSA (15.7, 20.2). Severe and LT ranked patients were more likely than Mi and Mo to worsen (10.7 versus 1.7%) or become VSA (2.1 versus 0.13%). The use of drugs and procedures increased incrementally from Mi to LT. Scene interval (90th percentile) increased from Mi = 25.2 to Mo = 27.2, S = 28.1, LT = 32.1, VSA = 41.2; among patients pronounced in field (code 6) it was 46.1. Transport interval (90th percentile) decreased from Mi = 24.1, Mo = 23.9, S = 21.9, LT = 17.9, VSA = 17.3. CONCLUSION: The 5-point case severity scale has validity with respect to operational constructs (dispatch and return priority, scene and transport intervals) and clinical constructs (use of drugs and procedures, patient status change scale).

036 Probability of Acute Coronary Syndrome Stratified by Presentation ECG and Serum Markers.
Christenson J, Innes G, McKnight D, Boychuk B, Grafstein E, Thompson C, Rosenberg F, Gin K, Anis A, Tilley J, Singer J. St. Paul's Hospital, UBC. Vancouver, BC.

The probability of acute coronary syndrome (ACS) in individual patients presenting with chest pain is best estimated using information from subgroups with similar clinical features. OBJECTIVE: To estimate the probability of ACS in patients with chest discomfort and no clear alternate cause prior to and after initial ECG and marker testing. METHODS: Consecutive, consenting patients >24 years old who presented with chest discomfort to 2 urban, tertiary care EDs from May 2000 to April 2001 were enrolled prospectively. Exclusion criteria were inability to communicate or contact, terminal illness, or an obvious traumatic or radiographic cause. Of 1831 patients, 1332 had ECG and serum markers. Patients were assigned explicitly defined, adjudicated 30-day outcomes of AMI, definite unstable angina (UA) or no ACS. Pretest probability of AMI or UA (ACS) was calculated for the whole cohort prior to testing, then recalculated after initial ECG, and after initial ECG plus first marker evaluation. Values are reported for clinically relevant subgroups. RESULTS: ACS proportions (95% CI) in subgroups follow: overall cohort, 27.0% (24.6, 29.4); normal ECG, 13.8% (11.3, 16.7); T-wave flattening only, 20.8% (13.4, 30.6); LBBB, 30.8% (17.5, 47.7); T-wave inversion, 31.1% (22.0, 41.9); ST depression without LBBB, pacing or LVH, 42.8% (34.5, 51.5); ST elevation without LBBB, pacing or LVH, 80.3% (69.2, 88.2); normal ECG + negative marker, 10.1% (7.8, 12.8); T-wave flattening + negative marker, 16.5% (9.8, 26.1); normal ECG + positive marker, 60.0% (45.2, 73.3). CONCLUSION: In this sample of patients with chest pain and a clinical suspicion warranting an ECG and serum markers, the pretest likelihood of ACS was 24.5%. Different initial ECG findings modify likelihood substantially. These tests are strong predictors but imperfect: 11% of patients with a normal initial ECG and normal initial markers had ACS and only 60% of those with an elevated initial marker and a normal initial ECG had ACS.

037 Are Chest Pain Units Cost-Effective Relative to Unstructured Emergency Department Evaluation?
Innes G, Christenson J, Anis A, McKnight D, Boychuk B, Grafstein E, Thompson C, Rosenberg F, Gin K, Tilley J, Singer J. St. Paul's Hospital, UBC, Providence Health Care. Vancouver, BC.

OBJECTIVES: Chest pain units (CPU) are a cost-effective alternative to coronary care units for the evaluation of low risk patients. No studies have compared CPUs to unstructured ED assessment (EDA). Our objective was to compare outcomes and resource utilization for a CPU vs. EDA approach. METHODS: We compared data from a systematic literature review to data from a prospective EDA cohort. Articles eligible for review were those that described their CPU protocol, patient risk factors, index AMI rate, admission rate, length of stay (LOS) and death + AMI rate during a follow-up period. In the EDA cohort, physicians applied an initial diagnosis of AMI, definite unstable angina, possible acute coronary syndrome (ACS) or no ACS. Resource utilization and blinded explicit outcomes were ascertained at 30-day follow-up. Aggregate data from the possible ACS/no ACS groups (low/intermediate risk) were compared to CPU review data. Primary outcomes were admission rate, hospital LOS and 30-day death + AMI rate after discharge. RESULTS: Six CPU studies published after 1994 met review criteria. Over a 12-month period, 1831 consecutive patients were enrolled in the EDA cohort and had 30-day follow-up. EDA patients (n = 1562) were a comparable or higher risk group than historical CPU patients and had similar rates of death + AMI within 30-days of discharge (see Table). The EDA cohort had similar or lower admission rates, shorter hospital LOS and fewer diagnostic tests. CONCLUSION: Unstructured EDA may provide similar outcomes and lower resource utilization for low/intermediate risk patients. The cost effectiveness of CPU vs. EDA for low/intermediate risk patients is unclear.

CPU EDA Index AMI rate (%) 1.2–5 4.7 Median LOS (hr) 9.2–33 4.8 Stress testing (%)* >90 15.5 % admitted 13–45 26.3 % AMI within 30 d 0–0.2 0.7 % death within 30 d 0–0.5 0.1 *exercise treadmill, echocardiogram or nuclear scan

038 Antidote Availability in British Columbia Hospitals.
Gorman SK, Zed PJ, Purssell RA, Brubacher J. CSU Pharmaceutical Sciences and CSU Emergency Medicine, Vancouver General Hospital; Faculty of Pharmaceutical Sciences and Faculty of Surgery, UBC. Vancouver, BC.

INTRODUCTION: Timely administration of the correct antidote is often essential to prevent or minimize morbidity and mortality following a toxic exposure. We sought to determine the availability of 14 important antidotes in acute care hospitals in British Columbia (BC). METHODS: A four-part survey, consisting of hospital demographics, community demographics, hospital laboratory capabilities, and antidote stocking information was distributed to BC hospital pharmacy directors in 1997. The antidotes examined in this survey were chosen based on published guidelines and included atropine, calcium gluconate, cyanide kit, deferoxamine, dimercaprol, digoxin Fab fragments, ethanol, glucagon, methylene blue, N-acetylcysteine, naloxone, pyridoxine, rattlesnake antivenin and sodium bicarbonate. RESULTS: Antidote stocking surveys were completed by 107 of 114 (94%) hospitals. Complete antidote stocking surveys were obtained from 96 of 107 (90%) responding hospitals and partial replies were obtained from the remaining 11 centres (10%). The number of antidotes sufficiently stocked ranged from 0/14 to 13/14 (93%), with an average of 3.2. Teaching hospitals stocked 64% of the antidotes versus 20% for non-teaching facilities (p < 0.001), urban hospitals stocked 45% versus 13% for rural hospitals (p < 0.001), and trauma centres stocked 63% versus 21% for non-trauma centres (p < 0.001). Adequate stocking rates increased with hospital size. Small hospitals (<50 beds) adequately stocked 12%, medium hospitals (50-250 beds) 41%, and large hospitals (>250 beds) 55% of the recommended antidotes. Sodium bicarbonate, N-acetylcysteine, ethanol, and naloxone were the best-stocked antidotes with adequate stocking rates of 64%, 50%, 41%, and 38%, respectively. Digoxin Fab fragments, glucagon, pyridoxine, and rattlesnake antivenin were very poorly stocked with respective rates of 3.7%, 4.9%, 6.6%, and 11.2%. CONCLUSIONS: British Columbia hospitals adequately stock less than 1 quarter of the recommended antidotes. This is concerning because BC encompasses a vast geographic area, with many hospitals being located more than a 2-hour drive from another health care centre.

039 An Emergency Department-Based Outpatient Treatment Program for Deep Vein Thrombosis: a Prospective Evaluation of Safety, Patient Satisfaction and Knowledge.
Lang E, Kahn SR, Josephson T, Tselios C, Colacone A, Pike J, Robitaille C, Goulet M, Afilalo M. Sir Mortimer B. Davis Jewish General Hospital. McGill University. Montreal, QC.

OBJECTIVE: To examine the safety, tolerability and educational aspects of an emergency department (ED)-based outpatient treatment program for deep vein thrombosis (DVT). METHODS: Design: Prospective cohort study and structured chart review. Setting: Tertiary-care academic centre. Participants: From April 2000 to March 2001, consecutive patients diagnosed with DVT and eligible for outpatient treatment with a low molecular weight heparin (dalteparin) and warfarin were approached. Interventions/Observations: Management consisted of an evaluation/education session with an ED-based Discharge Planning Nurse and an ED-based pharmacist. Patients or their caregivers were surveyed by telephone within 10 days of the initiation of treatment and again at 3 months. The primary outcome measures were recurrent venous thromboembolism (VTE), death and bleeding complications. Secondary outcomes were measures of patient satisfaction and knowledge. RESULTS: 62 patients were recruited and completed the 10 day survey; mean age 59 years; IQR: 47-83 years; 51 patients (82%) were reached for the 3-month evaluation. 4 patients developed complications of illness or treatment (8%; 95% CI 0.4 to 15%); there were no unexpected or VTE-related deaths (95% CI 0 to 7%). One patient (2%; 95% CI; 0 to 10%) was diagnosed with a pulmonary embolism (PE) and 3 (6%; 95% CI 2 to 16%) patients experienced complications likely related to treatment (hematuria, vaginal bleeding and thrombocytopenia); 2 required hospitalization. Overall, 54 (89%) patients were either satisfied or very satisfied with home treatment. In terms of education provided in the ED, 22 patients (36%) were unaware of bleeding complications of therapy and 35 patients (57%) did not know that their condition could be complicated by symptoms of PE. CONCLUSIONS: ED-based outpatient treatment of DVT is safe, well tolerated, and patient satisfaction is high. Educational programs directed at knowledge of illness and side effects of treatment can be improved.

040 Comparison of Point-of-Care Cardiac Markers vs. Standard Laboratory Assays in Emergency Department Patients with Chest Pain.
Innes G, Christenson J, Rosenberg F, Boychuk B, Grafstein E, Thompson C, Tilley J, Singer J. St. Paul's Hospital, UBC, Providence Health Care. Vancouver, BC.

OBJECTIVE: To determine the sensitivity and specificity of a quantitative point-of-care (POC) cardiac marker panel and a standard lab troponin I (TnI) assay relative to 30-day clinical outcomes. METHODS: A prospective cohort study in ED patients with chest pain. Markers were drawn at T = 0 and 2 hours. Standard TnI levels were done on a Beckman Access immunoassay system; triple marker POC assays (TnI, myoglobin, CKMB) were done using a Biosite Cardiac Panel. Tests were considered positive if any marker was positive at 0 or 2 hours, based on recommended cut-offs. At 30-days, blinded investigators assigned explicit outcome diagnoses of AMI, definite unstable angina (objective criteria), possible unstable angina (clinical suspicion only) or no ACS. Sensitivity and specificity for AMI and ACS (AMI or definite UA) were determined. TnI values for all paired (standard-POC) 0 and 2-hour assays were correlated, and area under the ROC curve for the 2-hour TnI assays (standard vs. POC) was compared using AMI and ACS as the outcomes. RESULTS: Of 204 patients studied, 22 had a final diagnosis of AMI and 57 had ACS. Pearson's correlation coefficient for 401 paired (standard/POC) assays was 0.80. AUC (95% CI) for the 2-hour standard TnI assay were 0.87 (.77-.97) and 0.66 (.57-.74) for AMI and ACS respectively. Corresponding values for the 2-hour POC TnI assay were 0.81 (.70-.93) and 0.62 (.53-.71). CONCLUSION: The standard TnI assay was more sensitive for AMI; the POC triple marker was more sensitive for ACS (see Table). Specificity was better using a single vs. a triple marker approach. For the 2 TnI assays, correlation was high and AUC were similar.

Standard Point of care   AMI ACS AMI ACS Sens .77 (.54–.91) .37 (.25–.51) .68 (.46–.83) .46 (.33–.60) Spec .95 (.91–.98) .97 (.92–.99) .81 (.74–.86) .84 (.77–.89)

041 Distinguishing Missed Patients with Acute Coronary Syndrome from Those Without Disease.
Christenson J, Innes G, McKnight D, Boychuk B, Grafstein E, Thompson C, Rosenberg F, Gin K, Tilley J, Anis A, Singer J. St. Paul's Hospital. Vancouver, UBC.

OBJECTIVE: It is unclear whether patients discharged inadvertently with acute coronary syndrome (missed ACS) are different from those with ACS. Our objective was to determine whether specific risk criteria distinguish patients with missed ACS from those without disease. METHODS: Consecutive, consenting patients >24 years old who presented with chest discomfort to 2 urban, tertiary care EDs were enrolled prospectively. Age, gender, risk factors (smoking, diabetes, dyslipidemia, hypertension and family history), previous MI or angina, initial vitals signs, initial ECG and initial troponin (TnI) levels were documented. After 30-day follow-up, each patient was assigned an explicitly defined outcome diagnosis of definite ACS or no ACS. Patients discharged with a non-ACS diagnosis and no cardiac investigations arranged but whose outcome diagnosis was definite ACS were classified as missed ACS. We determined the prevalence of predefined risk criteria in patients with missed ACS and in patients with no ACS as well as the relevant odds ratios. RESULTS: There were 383 patients with detected ACS, 22 with missed ACS (AMI or unstable angina), and 1426 without ACS. Odds ratios (95% CI) for the occurrence of specific risk criteria in the missed ACS / no ACS groups follow. Male, 1.03 (0.44, 2.39); >3 risk factors, 2.10 (0.81, 5.43); ECG not normal, 2.48 (1.05, 5.85); previous MI or angina, 3.06 (1.28, 7.35); age > 65, 5.97 (2.32, 15.35); any initial TnI elevation, 9.05 (3.10, 26.42). Of the 22 missed patients, 18 met 1 or more, 16 met 2 or more and 12 met 3 or more of the following criteria: >3 risk factors, ECG not normal, previous AMI/angina, age >65, or elevation of TnI. CONCLUSION: An abnormal ECG, previous MI or angina, age >65 and any elevation in TnI distinguish missed ACS patients from those without disease. Practical decision aids to guide individual diagnosis and disposition of patients with undifferentiated chest pain are likely possible and urgently needed.

042 Multicentre Controlled Clinical Trial to Evaluate the Impact of Advanced Life Support on Out-of-Hospital Respiratory Distress Patients.
Stiell IG, Wells GA, Spaite DW, Nichol G, Nesbitt L, De Maio VJ, Lyver MB, Brisson D, Martin MT, Doherty J, Beaudoin T, Cousineau D, for the OPALS Study Group. Division of Emergency Medicine, University of Ottawa. Ottawa, ON.

OBJECTIVES: There is little published evidence regarding the optimal EMS management of respiratory distress. Our study evaluated the impact of advanced life support (ALS) EMS programs on respiratory patient outcomes. METHODS: This multicentre before-after controlled clinical trial was conducted in 20 communities (population 15,000 to 750,000) as part of the Ontario Prehospital Advanced Life Support (OPALS) Study, which evaluates the impact of EMS programs for multiple conditions. During the before phase, care was provided at the BLS-D level. During the after phase, ALS providers performed endotracheal intubation and administered nebulized and IV drugs. Data were collected from ambulance reports, centralized dispatch data, ED records, and in-hospital records. Chi-square and Student's t-test analyses were performed. RESULTS: The 7,716 patients enrolled during the two 6-month BLS and ALS phases were well matched for clinical and demographic features and had these characteristics: mean age 74.0 (16-107), female 53.6%, EMS status 'severe/life threatening' 51.8%, mean RR 28, final hospital diagnoses: CHF 16.9%, COPD 16.3%, pneumonia 9.8%, asthma 5.7%, other cardiac 4.8%, CHF/COPD 3.4%, cancer 2.9%. During the ALS phase, patients received these EMS interventions: nebulized salbutomol 55.3%, IV furosemide 15.2%, SL NTG 9.6%, IV morphine 1.5%, intubation 1.2%. There was a 23.4% relative reduction in the primary outcome, overall mortality, from the BLS to the ALS phase (15.4% vs 11.8%; P < .001). This Table compares other outcomes.

CONCLUSIONS: This is the largest controlled trial of out-of-hospital respiratory distress patients and clearly shows important benefit from ALS programs for mortality and other outcomes.

043 The Use of Magnesium Sulfate in the Treatment of Acute Asthma.
Rowe BH, Roberts J, Camargo CA Jr, for the MARC Investigators. University of Alberta. Edmonton, AB.

OBJECTIVE: Recent systematic review evidence supports the use of magnesium sulfate (MgSO4) in the treatment of severe acute asthma. However, little is known about the clinical application of these results. We surveyed emergency department-based airway researchers to determine their use of MgSO4 in the treatment of asthma. METHODS: We contacted site investigators in the Multicentre Airway Research Collaboration (MARC) in February 2001 by e-mail. Reminders were sent 1 week later, and respondents were requested to complete an 11-question internet-based survey. The main outcomes were the percentage of sites using MgSO4, and the most common modifying patient factors in the use of this agent. RESULTS: Of 119 different sites approached, 58 sites (49%) responded to the first mailing and 32 sites (27%) responded to the second mailing, for a total of 90 (76%) site responses. There were a total of 93 different respondents from the 90 responding sites. Overall, 84 (90%) respondents stated their EDs have MgSO4 available and its use has occurred; 63 (68%) have personally used MgSO4 in the preceding 6 months. Fewer sites (36, 39%) reported their ICUs using this agent in acute asthma. More respondents listed severity (87%) and failure to respond to initial beta-agonists (80%) as factors that prompt their use of MgSO4. Other factors such as age, gender and duration of exacerbation were less commonly sited as factors in MgSO4 use. Using a 1 to 5 scale (mean ± SD), with 1 = strongly agree and 5 = strongly disagree, respondents disagree (3.9 ± 0.9) with the general use of MgSO4 in all patients presenting to the ED with asthma. Conversely, 72% of respondents agreed or strongly agreed with the statement that MgSO4 is beneficial in severe acute asthma (2.1 ± 0.8). Most (97%) respondents were interested in learning more about MgSO4 or being involved in future MgSO4 research. CONCLUSION: Most asthma researchers seem to accept the efficacy of MgSO4 in acute asthma, and appropriately restrict its use to patients with severe asthma. Interest in future educational and research activities involving MgSO4 appears of to be high within this research group.

044 Clinical Decision Rule for Emergency Department Patients with Acute Headache.
Perry JJ, Stiell IG, Wells GA, Mortensen M, Lesiuk H, Wallace G, Sivilotti M, Kapur A. Division of Emergency Medicine, University of Ottawa, Ottawa, ON.

OBJECTIVE: Currently it is unclear which patients with an acute headache require investigation with lumbar puncture (LP) and computed tomography (CT) for subarachnoid hemorrhage (SAH). This study derived a preliminary decision rule for the investigation of acute headache patients with normal neurological exam. METHODS: This prospective cohort study was conducted at 2 university tertiary care EDs. Patients >15 years of age, with normal neurological examination, GCS 15, and a complaint of a non-traumatic acute headache were enrolled over a 1-year period. Exclusion criteria included: history of recurrent headaches, referral of confirmed SAH, papilledema, previous SAH or known brain neoplasm. Emergency physicians completed data forms prior to investigation with LP and CT. The outcome criterion was SAH on CT, xanthochromia in the cerebral spinal fluid (CSF), or the presence of red blood cells in the final tube of CSF with positive cerebral angiography. Analysis included univariate association, odds ratios with logistic regression and recursive partitioning. RESULTS: The 222 enrolled patients had the following characteristics: mean age 43.4 years, 58.1% female, 9.0% SAH, onset to peak 13.2 minutes, worst headache of life 78.4%, 83.2% underwent CT, 41.9% underwent LP. This Table shows % of SAH and no-SAH patients with findings, p-value, adjusted odds ratio. The recursive partitioning model with 'age >50 years' and 'neck stiffness with flexion/extension' was found to have 100% (95% CI 83-100) sensitivity, 62% (95% CI 55-68) specificity, and would require only 41% patients to undergo investigation. CONCLUSIONS: We have derived a highly sensitive, preliminary decision rule for the investigation of acute headache. Future studies will evaluate the robustness of this rule in a larger patient sample.