CAEP/ACMU 2002 Scientific Abstracts: 60-74
2003 Scientific Abstracts
CJEM 2002;4(2):124-154
| Disclaimer: The large number of abstracts submitted and the short time interval between submission and publication did not permit communication with authors, abstract revision or CJEM editorial review. The following abstracts are presented, unedited, as they were submitted to the CAEP Research Committee. Abstract authors are from the department or division of emergency medicine of their respective universities unless otherwise specified. |
060 The Influence of the Residents' Training Level on their Evaluation of Emergency Medicine Clinical Teaching Faculty.
Steiner IP, Yoon PW, Kelly KD, Donoff MG, Diner BM, Rowe BH. University of Alberta. Edmonton, AB.
INTRODUCTION: The evaluation of clinical teaching faculty is necessary, and using accepted tools it can be valid and reliable. The medical education literature has a paucity of studies that examine resident-related factors that influence the evaluation of teaching faculty members. This study explores the influence of the training level of Family Medicine residents (postgraduate years {PGY} 1, 2 and 3) on the evaluation of Emergency Medicine (EM) clinical teaching faculty. METHODS: A retrospective analysis of data accumulated between 07/1994-07/2000 from 1st, 2nd and 3rd year Family Medicine residents' evaluations of EM clinical teachers at the University of Alberta was conducted. Resident and teaching faculty-related variables were entered anonymously using the ED Scale (Acad Emerg Med 2000;7[9]:1015-21). Descriptive statistics are presented. Univariate analysis of mean scores, standard deviation values, and significance levels were calculated, and groups were compared using ANOVA. RESULTS: 562 Family Medicine residents completed EM clinical rotations over the study period, for a total of 831 possible faculty evaluation forms. 777 (94%) had voluntarily returned anonymous completed evaluation forms. Of these, 705 (85%) had valid data. 115 clinical teaching faculty members were evaluated in 4 dispositional domains: Didactic teaching, Clinical teaching, Approachability and Helpfulness for a total of 12,816 individual evaluations. Scores by year of starting EM training, seasonal variations in the academic year, and trends over time were also analyzed. In all 4 categories the mean score differences (MSD) between resident evaluations were statistically and practically insignificant (MSD ¾ 0.1, [p > 0.05]). Seasonal variations were also not significant (MSD in each category ¾0.1 [p > 0.05]). Scores demonstrated a linear upward trend (MSD) in each category >=0.36 (p < 0.05). CONCLUSION: The training level of Family Medicine residents does not appear to be a significant moderator of teaching performance scores of EM clinical faculty.
061 Training BLS Paramedics in the use of the Laryngeal Mask Airway in Cardiac Arrest using a Classroom Manikin Model.
Murray MJ, Morrison L, Vermeulen M. Royal Victoria Hospital. Barrie, ON.
INTRODUCTION: The purpose of this study was to determine if a classroom mannequin training method for teaching Laryngeal Mask Airway insertion translated into successful use in adult out of hospital cardiac arrest patients by basic level paramedics. Methods. All pre hospital adult non traumatic cardiac arrest patients greater than twelve years of age or 40 kg, attended to by the Paramedics and transferred to hospital were included in the study. Eight ambulance services, 209 paramedics, and 8 hospital emergency departments were involved. All paramedics were trained during a 4 hour classroom and manikin course and certified to use the LMA in the field. RESULTS: Two hundred and eight paramedics completed the training and certification. During the training, of the 193 paramedics in which the number of attempts was recorded, the mean number of attempts was 1 and only 4 (2.1%) paramedics required a second attempt. During the subsequent study period a total of 291 cardiac arrests were observed. Insertion of the LMA was attempted in 283 (97.3%) of these arrests. Insertion was successful after 1 or more attempts in 199 (70%) patients. However, the LMA became dislodged in 5 (2.5%) of these patients and 12 patients (6.0%) required its removal due to inability to clear vomit in the airway. Thus, the number of successful insertions as defined by the study criteria was 182 (64%). CONCLUSIONS: This study shows that paramedics can be trained in the use of the LMA on manikins in the classroom and this translates into successful insertion and ventilation in the field in about 64% of adult cardiac arrest patients. Therefore it is our conclusion that the LMA would be a useful BLS adjunct and a useful alternative to the oral airway and BVM for ventilation in adult out of hospital cardiac arrest.
062 Evaluation of Medic Electronic Data and Information Capture (e-MEDIC) Phase I: A Reliability and Validity Assessment Comparing Electronic Data Entry by Data Clerk with Paramedic.
Morrison LJ, Vermeulen MJ, on behalf of e-MEDIC Study Group. Sunnybrook & Women's College Health Sciences Centre, University of Toronto. Toronto, ON.
OBJECTIVE: To evaluate the reliability and validity of data entry by data clerk (DC) compared with paramedic (P) on a representative retrospective cohort of ambulance call reports (ACR). METHODS: 28 fields on the ACR were defined by a focus group as critical to the data needs of Emergency Medical Services (EMS). 250 ACRs (EMT-D and EMT-P) representing 5 sentinel cases (hypoglycemia treated with glucagon [EMT-D] and dextrose [EMT-P], supraventricular tachycardia, chest pain and seizure) were randomly selected from 1999 records of an EMS system, of which 242 were evaluable. All ACRs were entered twice by DC and P. Inter- and intra-rater reliability were evaluated using intraclass correlation (ICC) and kappa (K), and overall accuracy by total number of correctly entered fields. Selected fields were grouped for evaluation as clerical (e.g., call date, call times), patient history, and vital signs. RESULTS: Mean (95% CI) number of correct entries: DC = 25.9 (25.7, 26.0); P = 27.5 (27.4, 27.6). Inter-rater reliability (ICC or K, 95% CI) was 0.23 overall; patient history = 0.19; vital signs = 0.14; clerical = 0.22; primary problem code = 0.30 (0.14, 0.46); procedure codes = 0.19 (0.06, 0.32). Intra-rater reliability (ICC or K, 95% CI) for DC was 0.57 overall; patient history = 0.87; vital signs = 0.26; clerical = 0.21; primary problem code = 0.60 (0.48, 0.73); procedure codes = 0.48 (0.38, 0.59). Intra-rater reliability (ICC or K, 95% CI) for P was 0.38 overall; patient history = 0.39; vital signs = 0.46; clerical = 0.36; primary problem code = 0.76 (0.60, 0.92); procedure codes = 0.29 (0.14, 0.44). CONCLUSIONS: Paramedic data entry is more accurate than data clerk entry. Agreement between data clerk and paramedic is poor to moderate overall and in sub-categories. Data clerks are more reliable on re-entry overall and in history and procedure codes. Paramedics are more reliable in vital signs, clerical and primary problem code fields.
063 Evaluation of Medic Electronic Data and Informaltion Capture (e-MEDIC): An Assessment of Accuracy of the Data Recorded on the Ambulance Call Report Compared with a Gold Standard.
Morrison LJ, Vermeulen MJ, on behalf of e-MEDIC Study Group. Sunnybrook & Women's College Health Sciences Centre, University of Toronto. Toronto, ON.
OBJECTIVE: To evaluate the accuracy of data recorded on the ambulance call report (ACR) by the paramedic as compared with a gold standard on a representative retrospective cohort of ambulance calls. METHODS: A focus group of EMS stakeholders defined 28 fields in the ACR as critical to the data needs of Emergency Medical Services (EMS). 250 ACRs (EMT-D and EMT-P) representing 5 sentinel cases (hypoglycemia treated with glucagon (EMT-D) and dextrose (EMT-P), supraventricular tachycardia, chest pain and seizure) were randomly selected from 1999 records of an EMS system. Accuracy of data recorded by the paramedic on each ACR was evaluated by 2 independent reviewers (gold standards) and evaluated as % (95% CI) correctly recorded or coded in each critical field. RESULTS: See Table.
|
||||||||||||||||||||||||||||||||||
![[description missing]](../../../000.common/000-spc/spc-hr.570gy.gif){kind=small}
CONCLUSIONS: Overall accuracy of data recorded on ACRs is 84%. Most fields are correctly recorded. Procedure and primary problem codes and call received times were the fields most commonly recorded incorrectly. Recording of health insurance number, an important unique identifier for outcome data linkage, was poor.
064 Longitudinal Analysis of Effect of Resuscitation on Health-Related Quality of Life After Sudden Cardiac Arrest.
Huszti E, Nichol G, Wells GA, Stiell IG, Nesbitt L, Blackburn J. Division of Emergency Medicine, University of Ottawa, Ottawa, ON.
OBJECTIVES: The Ontario Prehospital Advanced Life Support (OPALS) Study is a large EMS trial that evaluates ALS interventions for out-of-hospital patients and conducts detailed measurements of cardiac arrest survivor outcomes. This study assessed novel methods of analyzing incomplete or missing longitudinal data related to health-related quality of life (HRQL). METHODS: This prospective cohort study included all adult out-of-hospital cardiac arrest patients treated in 20 cities with a mixed BLS-D/ALS EMS system. Patients were evaluated for the Health Utilities Index (HUI) Mark 3, which describes health on a scale from 0 (dead) to 1.0 (perfect health). Subjects were interviewed at 3, 6, 9 and 12 months after discharge and were subject to dropout and truncation. The Propensity Score Multiple Imputation (MI) method was used to estimate the propensity that a data element is missing. Imputed values are then generated from those observed values that haves similar propensity scores. Finally, generalized linear models are applied to the complete data set to test for independent associations between response intervals and longitudinal HRQL. Secondary analysis considered the Predictive Mean MI method. RESULTS: Of 8105 cardiac arrest patients (1995-2000), 418 (5.2%) survived to discharge and 342 (81%) completed at least 1 HRQL interview. The median HUI Score was 0.87 (IQR 0.71-0.95) and the majority of cases had a score exceeding 0.8. Bystander CPR (odds ratio 1.9; 95% CI 1.1-3.3) and age >80 (OR 0.3; 0.1-0.97) were associated with very good HRQL. Generalized linear regression found time to defibrillation was associated with better HUI scores (p value = 0.0465). CONCLUSIONS: This is the first study to apply MI methods to the analysis of longitudinal HRQL data and to demonstrate that shorter response intervals are associated with better HRQL in cardiac arrest. This increase in sample size due to imputation significantly improves the accuracy of the results in EMS research.
065 The Effect of ALS interventions on Scene Times in Patients with Non-Traumatic Chest Pain : A Retrospective Review.
Jones AE, Lewell M. Division of Emergency Medicine, University of Western Ontario. London, ON.
INTRODUCTION: Advanced life support (ALS) interventions are a recent addition to the skill set of basic paramedics. We wanted to determine what effect interventions by ALS paramedics had on scene times in patients with non-traumatic chest pain. METHODS: A retrospective case control study was carried out using ambulance call reports (ACRs) from the base hospital program, London, Ontario. Scene time and treatment data from 300 consecutive ACRs with the final diagnosis of "chest pain" was collected for both ALS and BLS calls. Charts were excluded if essential data was missing or if the etiology of the chest pain was traumatic. Mean scene times were then calculated and compared using a t-test. RESULTS: Mean scene time for BLS and ALS calls were 13:28 (95% CI, 13:04-13:52) and 14:07 (95% CI, 13:31-14:42) respectively. These did not differ significantly (t = 3.32, p = 0.196). In addition, more patients received nitroglycerin as treatment in the ALS group (81.6%) compared with the BLS group (69.3%). CONCLUSION: ALS interventions do not significantly increase scene time in patients with non-traumatic chest pain.
066 Epidemiology of Prehospital Care in a Large Urban EMS System.
Morrison LJ, Vermeulen MJ, Burgess R. Sunnybrook & Women's College Health Sciences Centre, University of Toronto. Toronto, ON.
OBJECTIVE: To describe the population served by a large urban EMS system according to prehospital conditions (Maio, 1999). METHODS: An EMS system providing service to a population of 2.5 million people in the year 2000 entered 106,552 emergency calls (74%) in its database. Paramedic data recording accuracy was 84% and data entry reliability was 0.57. RESULTS: See Table. Priority Conditions (PC) were: Trauma, RD = Respiratory Distress, AO = Airway Obstruction, CA = Cardiac Arrest, Seizure, CP = Chest Pain. Abdominal pain (61%), allergic reaction (60%) and fractures (61%) occurred more frequently in women. Cardiac arrest (61%), environmental emergency (63%), near drowning (88%), penetrating trauma (80%) were more common in men. Allergic reaction (22%), burns (25%), head injury (20%), near drowning (25%), seizure (25%) were most frequent in children. In the homeless (n = 1274), the most common conditions were trauma (13%), seizure (10%), overdose (8%), behavior disorder (8%). 90th percentile scene interval was 27.3 minutes, 28.5 for EMT-P (n = 54,321) and 25.5 for EMT-D (n = 49,790) calls. Scene intervals were longest in cardiac arrest (41.8), diabetic emergencies (35.3), major trauma (37.0) and shock (40.8). CONCLUSION: Frequencies of EMS conditions in an urban EMS system are described according to gender, age, homelessness and scene and transport intervals.
|
||||||||||||||||||||||||||||||||||||||||||||||||
067 Can the Ontario Universal Influenza Immunization Program Reduce Emergency Department Overcrowding?
Groll D. Kingston General Hospital. Kingston, ON.
OBJECTIVES: In 2000 the Ontario Minister of Health and Long-Term Care announced a universal influenza immunization program for Ontario, Canada. The 2 stated objectives of this $38 million program were to decrease seasonal impact of influenza on emergency department (ED) visits and to decrease the number of cases and severity of influenza. This paper examines the impact of influenza and influenza-like illness on ED volume, and describes changes in ED volume over a 5 year period. METHODS: This is a retrospective, observational population study reporting on ED volume and influenza rates in over a five-year time period (1996/97-2000/01). A count of all adult ED admissions were gathered for 5 tertiary care hospitals in 3 Ontario cities from November 1, 1996-March 31, 2001. The number of people admitted with diagnoses of influenza and upper respiratory illness (influenza-like illness) as a proportion of all admissions were examined in 2 hospitals. RESULTS: ED volume has increased annually, peaks in the summer months (May-August), and is lowest in the influenza season (November-April). Acute upper respiratory (AUR) diagnosis accounted for less than 10% of all ED admissions, and a diagnosis of influenza accounted less than 10% of AUR. There was no significant correlation between population influenza rates, AUR, and ED volume (p = 0.901, and p = 0.449). CONCLUSION: Based on this study, a Universal Influenza Immunization Campaign is not likely to have a significant impact on emergency department volume.
068 Evaluation of a New 3-Valve Non-Rebreathing Mask (Hi-Ox) for the Delivery of High FIO2 at Low Flows -- Comparison with the Hudson Type Non-Rebreathing Mask.
Somogyi R, Vesely A, Sasano H, Prisman E, Fisher JA. Department of Anesthesia, University Health Network. Toronto, ON.
INTRODUCTION: The non-rebreathing mask (NRM) (Hudson RCI, Temecula, CA) is the familiar mask used to deliver high concentrations of O2. The Hi-Ox (Hi-Ox, Bird Medical Products, Palm Springs, CA) is a new compact 3-valve mask designed to function like the 5-valve self-inflating bag in providing FiO2 approaching 1.0 when O2 flow is greater than or equal to minute ventilation. We compared the performance of the Hi-Ox to a standard NRM. METHODS: A single trained subject was used to evaluate the performance of the masks. The seal of each mask to the face was assured by tape. O2 flows were set at 8 and 15 L/min. The subject breathed at his resting level and at levels resulting in end-tidal PCO2s of 35 and 30 mm Hg in response to visual feedback from the capnograph tracing. Gas was sampled from a catheter with the opening in the nasopharynx. Net FiO2 was calculated from expired FO2 using the alveolar gas equation.
RESULTS: See Table. FIO2 (mean ± SD) for all conditions.
|
||||||||||||||||||||||||||||||||
CONCLUSION: We conclude the Hi-Ox provides FiO2 near 1.0 when O2 flow is equal to or greater than minute ventilation in a method analogous to a self-inflating bag. Like the self-inflating bag, when minute ventilation exceeded O2 flow, outside air was entrained, resulting in a fall in FiO2. The lower FiO2 observed with the NRM resulted from obligatory air entrainment throughout inspiration at both O2 flows.
069 Predicted Impact of Citizen CPR Training on Cardiac Arrest Survival, Based on Location of Cardiac Arrest.
Vaillancourt C, Stiell IG, Wells GA, De Maio VJ, Nesbitt L, Martin MT, Cousineau D, for the OPALS Study Group. Division of Emergency Medicine, University of Ottawa. Ottawa, ON.
INTRODUCTION: Bystander CPR rates are generally better in public places than residential locations. We sought to determine the potential impact of citizen CPR training on cardiac arrest survival, based on specific location of cardiac arrest. METHODS: We reviewed data prospectively collected within the Ontario Prehospital Advanced Life Support (OPALS) Study. The OPALS database includes a population-based cohort of adult out-of-hospital cardiac arrest cases in 20 communities with BLS-D and ALS paramedics. We merged the OPALS and Municipal Property Assessment Corporation databases to obtain precise description of cardiac arrest location. Data was analysed using descriptive statistics with 95% CI and sensitivity analysis. RESULTS: From 1995 to 2000, there were 6,816 consecutive cardiac arrest cases. Cardiac arrest occurred most often in residential locations 85.8% (House 57.4%, Apartment 22.5%, Nursing Home 6.0%) as opposed to public places 14.2% (Store 3.4%, Recreation Facility 2.1%, Street 1.4%, Office Building 1.3%, Shopping Mall 1.1%). Overall survival rate was 3.8% (95% CI 3.4-4.3) and varied between 0.68% (0.38-0.98) and 18.5% (14.0-23.0) depending on location, witnessed status and bystander CPR rate. Assuming that better CPR technique could increase survival by 3%, increasing bystander CPR rates to 35% and 50% would result in overall survival rates and additional number of lives saved per year in residential locations and public places in the OPALS communities, respectively, of 5.38% (18, 3) and 6.19% (28, 24). Similarly, relative increase and absolute number of lives saved in specific residential locations would be, respectively: House 63.4% (12) and 63.3% (18), Apartment 30.1% (5) and 29.5% (8), Nursing Home 6.5% (1) and 7.2% (2). CONCLUSION: Citizen CPR training could significantly increase survival to cardiac arrest, especially in private houses and apartment buildings. An intervention focussed at improving bystander CPR rates and quality in those locations should be developed.
070 A New Highly Efficient Breathing Circuit for Oxygen Therapy.
Vesely A, Somogyi R, Sasano H, Preiss D, Prisman E, Adams T, Volgyesi G, Stenzler A, Azami T, Iscoe S, Fisher JA. University of Toronto Faculty of Medicine. Toronto, ON.
INTRODUCTION: In treating mass casualties in the field, 1 of the first medications to be depleted, and among the most difficult to replenish, is O2. Even so, at 8 L/min O2 Venturi masks provide only 32 L/min of 40% O2, a flow insufficient to maintain the inhaled fraction of inspired O2 (FiO2) in hyperventilating patients. Our goal was to design a breathing circuit that both increases the efficiency of oxygen use and maintains a high FiO2 in the face of hyperventilation. METHODS: We designed and constructed from tubing, plastic valves, and a rubber ventilator bellows, a breathing circuit that, at an O2 flow of 2 L/min, provides an FiO2 of 0.4 independent of minute ventilation. It does this by, first, ensuring that all of the O2 reaches the lungs (i.e., none is lost during exhalation or in ventilating the anatomical dead space) and, second, by allowing rebreathing only when minute ventilation (and thus expired PO2) increases. RESULTS: In 6 volunteers, FiO2 (calculated from PO2 of expired gas) averaged 0.41 ± 0.03 (SD) at rest. At a minute ventilation of 20 L/min, FiO2 was 0.37 ± 0.04 (N.S.) and PCO2 was 37 ± 3 mm Hg (p < 0.05). CONCLUSIONS: This new breathing circuit provides an FiO2 of 0.4 with 1 quarter of the O2 flow required for the Venturi mask and maintains the FiO2 despite marked increases in minute ventilation. The efficiency of the circuit may make it useful for field, military and third world use while the independence of FiO2 from minute ventilation should permit precise O2 dosing in the emergency department.
071 Computerized Physician Order Entry as a Cost-Effective Method of Capturing Emergency Department Quality and Utilization Data.
Innes G, Grafstein E, Christenson J, Epstein J. St. Paul's Hospital, UBC, Providence Health Care. Vancouver, BC.
OBJECTIVE: To compare the accuracy and cost of acquiring important quality and utilization data from a computerized physician order entry (POE) system vs. a standard chart review. METHODS: We developed a POE system by modifying order transmission processes and adding ED-specific order sets to an existing clinical information system. POE links physicians to their patients, populates a utilization database and captures time of MD exam, order entry and consult request. Linking to the POE and other existing databases provides electronic access to patient demographics, case-mix, waiting time, LOS, disposition and diagnostic utilization, stratified by physician. To assess data accuracy, we generated a report looking at all patients with CTAS level 3 abdominal pain (AP3) who presented between June 1-Oct 30, 2001. POE data were compared to manual chart review data from a random sample of 200 AP3 patients studied between June 1-Oct 30, 2000 (control data). Personnel time for data acquisition was also compared. RESULTS: Data from 864 POE and 200 control patients were analyzed. Age (40.8 vs 40.3 years) and gender (49.3% vs 49.5% male) were similar in the 2 cohorts, and diagnostic spectrum was identical. Other quality and utilization measures, including physician variability (in the POE set), are reported below. Chart review and data entry for the control cohort consumed 120 hours of staff time and 6 hours of physician time; POE report generation consumed 16 hours of physician time. Physician satisfaction with POE was high. CONCLUSION: POE can provide accurate utilization and quality data with more detail and lower cost than manual chart review.
|
||||||||||||||||||||||||||||||||||||
072 Does Computerized Physician Order Entry Reduce Emergency Department Length of Stay?
Innes G, Grafstein E, Christenson J, Epstein J. St. Paul's Hospital, UBC, Providence Health Care. Vancouver, BC.
OBJECTIVE: In November 2000, we introduced ED computerized physician order entry (POE). Before this, unit clerks transcribed writ10 orders into the hospital's clinical information system. Our objective was to determine the impact of POE on ED length of stay (LOS) in a group of patients with abdominal pain. METHODS: The POE cohort included all patients triaged with CTAS level 3 abdominal pain (AP3) between June 1 and Oct 31, 2001. The control cohort was a random sample of AP3 patients seen between June 1 and Oct 31, 2000. POE data were gathered electronically from the POE and linked databases; control data were gathered by manual chart review. Confounding variables were analyzed. The primary outcome was ED LOS (registration to discharge) for discharged patients. RESULTS: See Table. Data from 864 POE and 200 control patients were studied. Age (40.8 vs. 40.3 years), gender (49.3 vs. 49.5% male) and disease spectrum were similar in the POE and control groups. Admission rates were 22.8% and 19.5% respectively. ED overcrowding increased during the POE period: More patients were treated in hallways, the mean number of admitted patients held (at 24:00) rose from 16 to 18.4, and median ED LOS for admitted AP3 patients (waiting for inpatient beds) rose from 600 to 693 minutes. Critical time outcomes (median/IQR) are summarized below. CONCLUSIONS: Despite confounding factors that would tend to slow ED processes, LOS fell by 30 minutes (14%) after POE implementation -- a potential saving of 432 hours of ED stretcher time in this cohort. If similar benefits are apparent in other case-mix groups, POE may be a valuable tool to gather important data and improve ED efficiency.
|
||||||||||||||||||
073 The Ethics of Emergency Research: Feasibility and Application of Canadian Guidelines.
McRae AD, Weijer C. Dalhousie University Faculty of Medicine; Department of Bioethics, Dalhousie University. Halifax, NS.
INTRODUCTION: The past 2 decades have seen extensive ethical debate over clinical research on critically ill patients in emergency settings. Though Canadian guidelines and US regulations have changed in recent years, it remains unclear how these are best applied, or whether they sufficiently protect vulnerable patients. One problem with these rules is that they were devised using an ethical framework that analyses the risk of whole research protocols. This approach fails to recognize the morally relevant differences between the therapeutic interventions in a clinical trial and the non-therapeutic procedures used solely to answer the scientific question of the same trial. METHODS: We reviewed 70 studies that used a waiver of consent published between 1996-2000. We employed an ethical framework that analyzes the risk of therapeutic procedures and non-therapeutic procedures separately, and is essential for the proper ethical review of protocols. Studies were classified as posing minimal risk, probably minimal risk, or probably more than minimal risk. This data was combined with a conceptual analysis of current Canadian emergency research guidelines to suggest how these could be better applied, or altered, in order to adequately protect patients while permitting important research to proceed. RESULTS: 98.6% of published studies using a waiver of informed consent were either minimal risk or probably minimal risk. CONCLUSIONS: The vast majority of emergency research poses only minimal risk to patients, even in dire clinical situations. Current Canadian guidelines rely on an outdated approach to the ethical review of research protocols. Canadian guidelines do not sufficiently restrict the risks of non-therapeutic interventions in emergency clinical trials. However, we present empirical data showing a more stringent risk threshold still permits important emergency research to be completed. We offer suggestions for facilitating ethics review of emergency research in a way that both protects patients and advances emergency research.
074 Safety and Efficiency of an Individualized Approach to Patients with Chest Pain.
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, BC.
INTRODUCTION: Many US emergency departments (ED) have developed diagnostic pathways for chest pain evaluation. Canadian EDs tend to use an unstructured/individualized approach to each patient. Data are unavailable regarding the safety and efficiency of this practice. OBJECTIVE: To determine the proportion of patients with acute coronary syndrome (ACS) inappropriately discharged from the ED and the proportion of those without ACS (or an adverse event) held for prolonged investigation. METHODS: Consecutive, consenting patients >24 years old who presented with chest discomfort to 2 urban, tertiary care EDs between May 1, 2000 and April 25, 2001 were prospectively enrolled. Exclusion criteria were inability to communicate or contact, terminal illness, or obvious traumatic or radiographic cause. Disposition and ED length of stay (LOS) were documented, and LOS >3 hours was considered prolonged investigation. At 30 days, investigators used pre-defined explicit criteria to assign an outcome diagnoses: definite ACS (AMI or objective unstable angina) or non ACS. Patients who were discharged from the ED without planned urgent outpatient investigations and with a non-ACS diagnosis were categorized as ACS Not Suspected. RESULTS: Of 1831 patients, 244 (13.3%) had AMI, 161 (8.8%) definite unstable angina, and 1426 (77.9%) no ACS. 22/405 patients with ACS were discharged with ACS unsuspected. 373/1426 patients without ACS were admitted or had ED diagnoses suspecting ACS. Current clinical sensitivity is 94.6% (95% CI: 91.8, 96.5) with specificity of 73.8% (95% CI: 71.5, 76.1). 69.9% of patients without ACS or adverse event were admitted or held in the ED for >3 hours. CONCLUSION: The current individualized ED evaluation and disposition of patients with chest discomfort in 2 Canadian hospitals "misses" 5.4% of definite ACS. Only 50% of those suspected or admitted prove to have ACS. Opportunities exist to reduce numbers of missed ACS and to reduce the cost of investigation of patients without ACS.
Search
Downloads
-
217.15 KB
