2005 CAEP / ACMU Scientific Abstracts - Poster Presentations: 59-71

Development of an objective emergency physician practice profile.

MacLeod DB, Curry DG, McNaul M, Wang D. Department of Emergency Medicine, Calgary Health Region, Calgary, AB

INTRODUCTION: Emergency physicians find it difficult to obtain objective feedback on personal practice patterns. An iterative process is being undertaken to develop an objective profile by which emergency physicians can compare their own practice to that of peers in the same environment. METHODS: A profile was developed by combining a regional EDIS database, a regional diagnostic imaging database, a regional Health-Information database, and billing service databases tracking hours worked. Profiles are provided to individual physicians revealing personal data on each parameter, along with normative data of the entire group of 85 physicians, with personal and normative data being provided for practice at each of the four Calgary emergency departments. Profiles are provided in a non-threatening manner for personal information, but if individual physicians are outliers, results are discussed in biannual performance reviews.RESULTS: The third iteration of the profile included the following parameters: total number of patients seen, new patients seen per hour, CTAS breakdown of patients seen, admission rate for each CTAS level and overall admission rate, median length of time from physician sign-up for CTAS 3 patients until the patients leave the emergency department, percentage of patients seen with residents or clinical clerks, percentage of patients receiving Ultrasound, CT, VQ scan or IVP, and percentage of patients returning within 72 hours of discharge and requiring admission, with a list of these patients and diagnosis on both visits. Physicians receive information for each site worked, and normative data is separated by site. Physicians working on several sites demonstrate significant variability on different sites. The fourth iteration will include rates of consultation of various services on each site, and complaints received per 1000 patient care encounters. CONCLUSIONS: After initial apprehension prior to the first iteration release, the process has been very well accepted and requested by emergency physicians.Key words: medical administration; physician evaluation; CQI

Physician determinants of emergency care quality.

Innes G, Grafstein E, Christenson J, Stenstrom R. St. Paul's Hospital, Vancouver, BC

INTRODUCTION: High quality care is safe, timely, patient focused, effective and efficient. Care quality may be related to physician characteristics like age, training path and experience. Our objective was to determine whether shorter (CCFP-EM) residency training is associated with differences in emergency care quality.METHODS: We assessed safety by tracking bounce-back rate (% of patients hospitalized within 72 hrs of ED discharge) and we measured timeliness by waiting time to physician exam. We used random exit surveys to evaluate patient perception of MD communication and concern (patient focus) and to assess physician skill and care quality (effectiveness). Efficiency was based on admission rates for triage level 2-3 patients and on imaging rates (number of studies per 100 patients) in two cohorts: abdominal pain and extremity injury. For each outcome, physicians were stratified by quartile and the sum of their quartile ranks in 7 domains comprised an overall Q-score where low score = high quality. RESULTS: See Table 1, Abstract 60. 23 physicians were evaluated, including 12 CCFP-EM, 6 FRCP and 5 ABEM. Median Q scores (IQR) were 19 (17.8-20.2), 15.0 (14-16) and 15.5 (14.3-17.5) for CCFP-EM, ABEM and FRCP physicians, respectively (p = 0.16). Median Q scores (IQR) were 19 (17.8-20.2) for CCFP-EM vs. 15 (14-17) for other physicians (p = 0.06). This apparent difference was not apparent (p > .10) when ANCOVA was used to adjust for years in practice. CONCLUSION: Quality measures did not differ significantly based on training path.Key words: medical administration; CQI; emergency medicine, residency training

Emergency department triage: evaluating the reliability of a computerized triage tool and the effect of overcrowding.

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

INTRODUCTION: Emergency department (ED) triage prioritizes patients based on urgency of care. We describe agreement between duty triage nurses and study nurses using eTRIAGE, a web-based triage tool based on CTAS, and examine the effects of overcrowding on agreement.METHODS: This prospective study enrolled consecutive patients presenting to a tertiary care ED in a large urban centre. Patients were assessed by a duty triage nurse (TN) and a study nurse (SN). Both used eTRIAGE and were blinded to each other's assessment. SN collected data on the ED busyness every two hours. Agreement between TN and SN and the effects of ED busyness are reported. RESULTS: See Table 1, Abstract 61. Over a 9-week period, 577 patients were assessed and 569 (98.6%) data pairs were available. The mean age was 49.4 and 51.0% were male. Overall agreement was moderate (weighted κ = 0.518, 95% CI 0.464, 0.573). ED busyness questionnaires were completed for 353 (62.0%) data pairs. The most common reason for not completing the questionnaire was "too busy". There was no significant difference in agreement when busyness data was available (weighted κ = 0.558, 95% CI 0.491, 0.625) or not (weighted κ = 0.445, 95% CI 0.353,0.537). ED busyness had no significant effect on agreement:

CONCLUSIONS: Real-time, prospective evaluation of eTRIAGE use by two independent groups of nurses during a climate of ED overcrowding generated moderate agreement. Surges in overcrowding did not significantly influence agreement. More study is warranted to determine the impact of distractions on triage performance. Key words: triage; information systems; emergency medicine; overcrowding

The relationship between preventive health practices of emergency department patients and access to family physicians.

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

OBJECTIVE: Many Emergency Department (ED) patients report not having a family physician (FP). This study assesses FP contact and use of preventive services prior to the ED encounters.METHODS: Patients > 17 years of age were randomly selected from computerized ED records at 2 urban ED sites in Edmonton, AB (UAH; RAH). Following initial triage, stabilization, and informed consent, patients were asked to complete an on-line or paper survey. Survey data were cross-referenced to a minimal patient dataset. The questionnaire asked various demographic, presentation, primary care visit contact and preventive health practice questions. RESULTS: Of the 1425 patients approached, 904 (63%) surveys were completed; mean age was 44, 51% were female. Overall, 713 (78.9%) reported that they had a family physician (FP), while 191 (21.15%) reported that they did not have a FP (NFP). FP patients were more likely to report receiving a flu shot in the past year (37% vs. 19.4%; p < 0.001), less likely to smoke daily (24% vs. 44%; p < 0.001), less likely to smoke occasionally (7.2% vs. 15.7%; p < 0.001); and more likely to always wear a seat belt (71.7% vs. 62.5%; p = 0.077) than NFP patients. Female FP patients were more likely to have had a pap smear in the past 2 years (33.5% vs. 25.1%; p < 0.001) and male FP patients were more likely to have had a prostate exam in the past 2 years (13.5% vs. 2.1%; p < 0.001) than NFP patients. FP patients were less likely to receive a triage score of 4 or 5 (44.0% vs. 57.9%; p < 0.001) and less likely to be non-heterosexual (2.2% vs. 5.2%; p = 0.047).CONCLUSIONS: Nearly 1 in 5 patients presenting to urban Alberta EDs have no link to a FP and FP patients demonstrate better access and use of preventive health practices. This indicates the important role FPs can play on the health of their patients and why EDs should link patients with primary care whenever possible. The relationship between health practices and ED utilization is worthy of more detailed investigation. Key words: emergency medicine; public health

Contribution of various components of emergency department length of stay to emergency department overcrowding.

Curry DG, McLeod DB, Wang DM, McNaul MA. Department of Emergency Medicine, Calgary Health Region, Calgary, AB

INTRODUCTION: Creating and measuring impacts of successful initiatives to reduce Emergency Department (ED) overcrowding requires an understanding of the various factors that contribute to flow through EDs. Limited published data exists to quantify the contribution of various components to overall ED length of stay (LOS).METHODS: A retrospective database audit was conducted for all patients admitted to hospital from three urban ED sites over a 26 week period. Multiple time stamps were extracted from a computerized ED information system to measure the following intervals: triage to ED bed assignment (waiting room time), bed assignment to ED physician signup (bed to physician time), ED physician sign up to consultation (workup time), consultation to decision to admit (consultant time) and decision to admit to discharge to ward time (boarding time). Average and median LOS components were analyzed for admitted patients, by site and consultation service (including a category for patients with multiple consultations). LOS for each service was multiplied by number of patients to create a utilization measure of the relative contribution of each service to total ED LOS.RESULTS: Significant variability between different consultation services at the same site and significant consistency between the same services at different sites exists. Variable waiting room times confirmed the effect of triage decisions. Bed to ED physician times were usually longer than waiting room times, consultant times contributed more to ED LOS than boarding times and patients admitted to hospitalist services or those requiring multiple consults were the major contributors to ED LOS in our system. CONCLUSIONS: Multiple components of ED LOS can be identified, each requiring focused efforts to reduce. The majority of ED LOS for admitted patients is outside the direct control of the ED. Increasing access to consultants, inpatient beds and ED physicians are all keys to reducing ED overcrowding.Key words: overcrowding; CQI; emergency medicine

A quality improvement process enhances vital signs documentation.

Al Darrab A, Fernandes CMB, Zimmerman R, Smith R, Smith T, Worster A. Division of Emergency Medicine, McMaster University, Hamilton, ON

INTRODUCTION: Patient safety is a key concern in the Emergency Department. Documentation of full vital signs enhances screening for potential catastrophes in the Emergency Department. The objective of this study was to enhance complete documentation of vital signs upon initial triage. METHODS: A quality improvement process was undertaken to ensure full vital signs for all patients in the ED. Solutions included a departmental vital signs policy, availability of equipment, vital signs audits, a defined triage nurse role, and not to pull the triage nurse. A 48 h before-after intervention analysis was undertaken for Aug 2003 and 2004. Vital signs (temperature, blood pressure, respiratory and heart rates) were assessed. RESULTS: In the pre-intervention evaluation 578 charts were reviewed; of those 66% (383/578) had complete vital signs documented (95% CI: 62-70%). In the post intervention evaluation 492 charts were reviewed; of those 74% (365/492) had complete vital signs documented (95% CI: 70-78%) (p < 0.005). CONCLUSIONS: The application of a quality improvement process may improve complete vital signs documentation upon initial triage. Key words: quality improvement; triage

Correlation between the number of patients emergency physicians see on a shift and the percentage of patients who leave without being seen.

Abu-Laban RB, McKnight RD. Division of Emergency Medicine, University of British Columbia, Vancouver, BC

OBJECTIVES: It is known that a significant proportion of emergency department patients who leave without being seen (LWBS) have serious pathology. One factor shown to influence a decision to LWBS is the Emergency Department waiting time. We sought to determine the correlation between the number of patients Emergency Physicians (EPs) see on a shift and the percentage of patients who LWBS.METHODS: Data captured on a patient care information system from April 1, 2003 to April 30, 2004 for 19 board certified EPs (12 male, 7 female) practicing at a tertiary care teaching hospital were retrospectively reviewed using explicit criteria. The mean/median number of years since medical school graduation of the EP cohort was 21.7/21 (range 8-35).RESULTS: During the one year study period there were approximately 52,000 patient encounters. The mean/median number of patients seen per shift by EPs was 20.5/20.5 (range 17.7-23.6), and the mean/median percentage of patients who LWBS was 3.7%/3.9% (range 1.5% to 5.9%). Spearman's coefficient of correlation between the mean number of patients seen per shift and percentage of patients who LWBS for each EP was 0.62. A linear regression model, fit for hypothesis generating purposes, showed that the percentage of patients who LWBS was independent of EP gender or years of experience, but was significantly associated with the number of patients seen per shift (p = 0.027).CONCLUSIONS: For reasons that are unclear, there is a striking (almost four-fold) variability between different EPs in the percentage of patients under their care who LWBS. A moderate correlation exists between the percentage of patients who LWBS and the mean number of patients EPs see per shift. Whether this finding represents association or causation, whether confounding exists, and the relationship between individual EP efficiency and Emergency Department waiting time, merit further evaluation. Key words: quality improvement; emergency medicine; medical administration

A cross-sectional analysis of high frequency users of the emergency department and their reasons for visiting.

Brooks S, Brubacher JR, Sernik J. Vancouver General Hospital and University of British Columbia, Vancouver, BC

INTRODUCTION: Understanding why high frequency users (HFUs) visit the ED may allow us to improve their care and minimize their impact on ED resources. Our objective was to classify the reason for each HFU ED visit over a one year period. METHODS: An electronic database identified patients who visited the ED >10 times between 06/01/2001 and 05/31/2002. We randomly selected age and gender matched controls who visited the ED <2 times over the same period. One author retrospectively reviewed all ED visits and, using explicit criteria, abstracted data to classify patient visits. RESULTS: During the study period 33,646 patients made 44,954 ED visits. There were 1212 visits by 96 HFUs (10-33 visits/patient). The 56 male and 40 female HFUs ranged in age from 18 to 95. The 96 controls made 113 visits. Compared with controls, HFUs were more likely to have at least one visit with no listed GP (30/96 [31.2%; 95% CI = 22.2%-41.5%] vs 17/96 [17.7%; 95% CI = 10.7%-26.8%], p = 0.044}, no current employment (51/96 [53.1%; 95% CI = 42.7%-63.4%] vs 21/96 [21.9%; 95% CI = 14.1%-31.5%], p < 0.001), for substance problems (32/96 [33.3%; 95% CI = 24.0%-43.7%] vs 5/96 [5.2%; 95% CI = 1.7%-11.7%], p < 0.001), psychiatric problems (30/96 [3%; 95% CI = 24.0%-43.7%] vs 7/96 [7.3%; 95% CI = 3.0%-14.5%], p < 0.001), chronic pain (16/96 [16.7%; 95% CI = 9.8%-25.7%] vs 2/96 [2.1%; 95% CI = 0.3%-7.3%], p = 0.001), or drug seeking (8/96 [8.3%; 95% CI = 3.7%-15.8%] vs 0/96 [0%; 95% CI = 0-3.1%], p = 0.011). To reduce the impact of detection bias we also compared these categories as percentages of total HFU vs control visits. All findings except drug seeking and homelessness remained significant when analyzed in this way.CONCLUSIONS: Our data suggests that HFUs may have higher rates of unemployment, homelessness, substance problems, psychiatric problems and chronic pain issues and may be less likely to have a GP than controls. Appropriate care plans may be beneficial. Hypotheses formed from this data can be used to guide experimental ED interventions for HFUs. Key words: emergency medicine; medical administration

A quality improvement process enhances armband use in the emergency department.

Al Darrab A, Fernandes CMB, Zimmerman R, Smith R, Smith T, Worster A. Division of Emergency Medicine, McMaster University, Hamilton, ON

INTRODUCTION: Patient identification is an important step in error reduction in the Emergency Department. Armband use can enhance patient safety and reduce resource utilization. The objective of this study was to ensure patient identification with an armband upon registration to the Emergency Department. METHODS: A quality improvement process was undertaken to ensure all patients registered to the ED wear armbands. As a result of this process, the following solutions were put in place: business clerk accountability to apply armbands to ambulatory patients at time of registration, regular armband audits were completed and posted and followed up with education. A two-day pre-post intervention analysis was undertaken for August 2003 and 2004.RESULTS: In the pre-intervention evaluation 142 patient were checked; of those 83% (118/142) were identified with armbands (95% CI: 77-89%). In the post intervention evaluation 129 patients were checked; of those 98% (126/129) were identified with armbands (95% CI: 94-99%) (p < 0.0001).CONCLUSIONS: The formal application of a quality improvement process can improve patient identification upon registration to the ED. Key words: emergency medicine; medical administration; quality improvement

How does fast-track affect quality of care in the emergency department?

Al Darrab A, Fernandes CMB, Zimmerman R, Smith R, Smith T, O'Connor K, Worster A. Division of Emergency Medicine, McMaster University, Hamilton, ON

INTRODUCTION: The balanced scorecard includes business process measurement and customer satisfaction; in the Emergency Department, this includes length-of-stay and left-without-being-seen rate. The objective of this study was to determine if a dedicated fast-track for CTAS 4/5 patients affected: (1) the time to assessment for CTAS 3 patients; (2) the length-of-stay for CTAS 4/5 patients and; (3) the left-without-being-seen rate? METHODS: In June 2003, fast-track was opened in our Emergency Department from 1300-1900. A before-after intervention comparison analysis was completed for one week in Aug 2002 and the same week in Aug 2003. Data collected included: (1) time to assessment of CTAS 3 patients; (2) the length-of-stay for CTAS 4/5 patients and; (3) percentage of patients left-without-being-seen. RESULTS: 373 patients were reviewed for 2002, with 253 patients triaged as CTAS 4/5. 375 patients were reviewed for 2003 with 274 triaged as CTAS 4/5. Median time to assessment of CTAS 3 patients presenting 0930-1259 for 2002 was 0.92 h (95% CI: 0.56-1.27) compared to 1.8 h (95% CI: 1.38-2.22) in 2003 when fast-track was not open. Median time to assessment of CTAS 3 patients presenting 1300-1900 was reduced from 1.08h (95% CI: 0.87-1.29) to 1.00h (95% CI: 0.73-1.26) after fast-track was opened (p = 0.6). Median length-of-stay of CTAS 4/5 patients decreased from 2.83 h (95% CI: 2.55-3.11) to 1.83 h (95% CI: 1.63-2.03) (p < 0.0001). For those who did not require diagnostics length-of-stay decreased from 2.16h (95% CI: 1.74-2.58) to 1.33h (95% CI: 1.15-1.51) (p < 0.0005). Left-without-being-seen rate of CTAS 3 patients reduced from 4.2 % to 3.4%, while left-without-being-seen rate of CTAS 4/5 patients reduced from 4% to 1%. CONCLUSIONS: A dedicated fast-track for CTAS 4/5 patients reduced length-of-stay and left-without-being-seen rate with minimal impact on CTAS 3 patients to be seen in the main ED. Key words: fast track; emergency medicine; medical administration; quality improvement

Pareto analysis to identify delays of the admitted patients in leaving the emergency department.

Al Darrab A, Frenandes CMB , Zimmerman R, Smith R, Smith T, Worster A. Division of Emergency Medicine, McMaster University, Hamilton, ON

INTRODUCTION: Emergency department overcrowding reflects a crisis. One of the determinants of overcrowding is the inability to move admitted patients from the Emergency Department to an inpatient bed. The objective of this study was to identify the root causes for delays in the admitted patients leaving the Emergency Department.METHODS: A continuous quality improvement process was undertaken to examine the flow process of the admitted patients in an urban Emergency Department from the time of decision to admit to the time the patient leaves the Emergency Department.RESULTS: A flow chart of the admitted patients was developed for a two-week period in October 2003. The median time interval from the decision to admit until the time the patient left the ED was (20.65) hours (95% CI: 16.54-24.75). The longest delay was attributed to the time from "admitting has returned paperwork" to "Bed assigned" (15.55) hours (95% CI: 10.91-20.18). The causes of delay included bed management processes, discharge planning/communication, repeated phone calls, unclear overall picture of bed availability and physician workload.CONCLUSIONS: A quality improvement process can identify root causes of delays of admitted patients to leave the ED. The greatest cause for delay is waiting for a bed on an inpatient unit. Key words: overcrowding; emergency medicine; medical administration; quality improvement

Epidemiology of heart failure in a Canadian emergency department.

Marsden J, Grafstein E. Div. of Emergency Medicine, Department of Family Medicine, University of British Columbia, Vancouver, BC

INTRODUCTION: Most heart failure (HF) literature originates outside Canada and focuses on the epidemiology of the chronic heart failure patients. We describe the characteristics of patients with heart failure at an urban Canadian Emergency Department (ED) in Vancouver, BC.METHODS: A retrospective study utilizing the New Emergency Resource Database (NERD), the administrative ED database at Providence Health Care, from Jan. 2003-Jan. 2004. HF was defined as a discharge diagnosis using ICD-9CM codes (428.0). Only patients with the primary diagnosis of HF were analyzed.RESULTS: 242 of 49,632 ED patients (0.56% of overall visits) had a diagnosis of HF. These 242 patients had 282 encounters for HF of which 197 (70%) had a Canadian Triage Acuity Score Level of 1-3. Median age was 76 years and 60% were male. Disposition was: 65.6% (183) admitted; 32% (90) discharged; 1.4% (4) left against medical advice and 1% (3) discharged to a Long Term Care Facility. Of encounters without index admission to hospital, mean ED length of stay (LOS) was 5.2 hours (SD 4.4) and 15 (16.7%) revisited the ED within 30 days and were readmitted having a mean hospital LOS 19.6 days (SD 21.9). HF was the twenty-second most common admission diagnosis overall. The 183 encounters that were admitted accumulated a total hospital LOS of 2454 days with an mean hospital LOS of 13.8 days (SD 16.5). 11% (27) of patients died during their hospitalization. Median age of this cohort was 82 years. 49 patients (20%) had 98 visits to the other 4 regional hospitals during this period and were admitted on 33 of these encounters. CONCLUSION: The burden of CHF on acute care is significant and although one third of ED patients with CHF are discharged some revisit. Investigation of newer diagnostic techniques, clinical care pathways, and specialized follow up clinics may show enhanced the provision of care to this patient population. Key words: retrospective study; congestive heart failure; epidemiology; emergency medicine

Cardiac arrest care and emergency medical services in Canada.

Vaillancourt C, Stiell IG. Department of Emergency Medicine, University of Ottawa, Ottawa, ON

INTRODUCTION: Cardiac arrest is the primary cause of mortality in Canada and survival to hospital discharge from out-of-hospital cardiac arrest is low. The purpose of this study is to provide an overview of the outcomes for out-of-hospital cardiac arrest in Canada.METHODS: We conducted a national descriptive analysis of cardiac arrest care and emergency medical services, Utstein style. We approached a convenient sample of EMS directors and researchers in the field of cardiac arrest from across Canada. We compiled data from five separate sources: the City of Edmonton Emergency Response Department, the British Columbia Ambulance Service, the Nova Scotia Emergency Health Services, the "Urgences-santé" corporation of the Montreal Metropolitan region, and the Ontario Prehospital Advanced Life Support (OPALS) Study database. Data was analysed using descriptive statistics with 95% confidence intervals when available. RESULTS: There were 5,288 cardiac arrests from a range of small communities to large provincial cardiac arrest registries in 2002. They were men in their late sixties and early seventies, witnessed (35.2% to 55.0%), rarely receiving bystander CPR (14.7% to 46.0%), most often in asystole, most often at home (56.1%), and rarely surviving to hospital discharge (4.3% to 9.0%). Bystander CPR and early first responder defibrillation were significantly associated with increased survival. Cardiac arrest incidence rates per 100,000 varied between 53 and 59 among provinces and followed a downward trend. CONCLUSIONS: This paper is an important first step toward a national cardiac arrest registry comparing the impact of regional variations in patient and system characteristics. Most communities do not have accurate data on their performance with regards to the chain of survival, and need to significantly improve their capacity of providing citizen bystander CPR and rapid, first responder, defibrillation.Key words: EMS; cardiac arrest; epidemiology