Background: Pulmonary embolism (PE) is a common illness with significant mortality without appropriate treatment. Its disease severity is variable, difficult to prognosticate and triage of severe PE remains a patient safety concern. Some PE may benefit from invasive and advanced medical therapy, but these decisions require complex multi-disciplinary coordinated care. We have launched a multi-disciplinary rapid response team at the Foothills Medical Center Hospital (FMC) to assist prognostication, treatment, disposition planning, and followup for high-risk PE: The Pulmonary Embolism Response Team (PERT). Aim Statement: PERT has been implemented to improve patient-oriented outcomes however, as severe PE is infrequent, we initially target process measures. In the first year of PERT rollout, we aim for: 1) 100% of high risk PE be detected by emergency for PERT consult 2) PERT response be within 45 minutes of activation 3) PERT treatment and disposition be made within 1 hour of consult. 4) > 80% of patient dispositions match those informed by evidence-based risk stratification tools. Measures & Design: Through collaboration between emergency medicine, radiology, cardiac sciences, medical specialties and critical care, a collective evidence-based PE risk stratification/treatment pathway was developed. This has been disseminated to providers and embedding into electronic medical records (EMR) for computer assisted decision-making support. EMR data has been harmonized with standardized radiographic reporting for PE to cue reporting of high risk imaging findings. Standardized imaging and EMR prognostic factors flag high risk PE suggesting PERT activation. PERT standard operating procedures have been developed, including evidenced-based pathways for further therapy, advanced imaging, and subspecialized disposition planning. Clinical services meet quarterly, and review dashboard summary data on clinical adverse events, resource utilization, and time data of patient flow to revise PE care pathways. Evaluation/Results: PERT activations occur approximately 2 times weekly. Adherence to operating procedures is high. Feedback post implementation cites improved adherence to evidence-based practice, clearer communication, and faster patient disposition. Quantitative analysis of performance is limited by infrequency of cases. Discussion/Impact: Our project shows feasibility of a PERT service. Pre-implementation data is collected, and we are currently measuring these post. We suspect signal of improved patient-oriented outcomes will be detected with more cases.

Background: Patients presenting to the Emergency Department (ED) may be subjected to unnecessary bloodwork. This leads to excessive work for front-line nurses, physicians and laboratory staff, contributing to increased ED length of stay (LOS), patient discomfort, and health care costs. Aim Statement: By January 1, 2020, we will reduce the number of targeted blood tests (AST, GGT, aPTT and CK) by 40% in the Mount Sinai ED, as measured by the percent per 1000 ED visits of AST to ALT, GGT to ALT, aPTT to INR and CK to troponin. Measures & Design: This was a prospective time series quality improvement study. Using the Model for Improvement, we engaged front-line ED staff, as well as stakeholders from Consultant, Laboratory and Information Services. Data was analyzed using run chart rules. Intervention: a) Removed rarely used tests from electronic nursing order sets b) Uncoupled order panels c) Developed six presentation-based medical directives with appropriate blood testing. d) Staff education Family of measures Outcomes: percent of targeted uncoupled test per 1000 ED visits for each of AST to ALT, GGT to ALT, aPTT to INR, and CK to troponin; Total number of blood tests ordered per 1000 ED visits Process: number of “separate and hold” tubes; number of blood tubes used in the ED; proportion of staff attending education Balancing: volume of blood drawn; LOS Evaluation/Results: Outcome: Estimated relative reduction in proportion of all uncoupled tests per 1000 ED visits by: • 33% AST/ALT • 52% GGT/ALT • 50% CK/troponin •18% aPTT/INR Total number of lab tests per 1000 ED visits decreased by 7.7% (5742 to 5331). Evidence of special cause variation on all outcomes. Process measures: 1. 100% reduction in weekly “Separate and Hold” tubes (56 to 0). 2. Monthly total of blood tubes used in the ED decreased by 2.8% (11620 to 11300) 3. Attendance pending. Balancing measures: Monthly average volume of blood drawn decreased by 1.4L(2%) from 50.4L to 49.0L; LOS pending Discussion/Impact: A multi-pronged intervention resulted in a decrease in blood testing in the ED. We achieved the sub-aim of reducing targeted blood tests and are on track to achieve the overall aim of total lab reduction in the ED by April 2020. Final interventions to be implemented in the coming months include changes to the ED paper record and replacement of the paper add-on order process with an electronic ordering tool. Complete data will be available by April 2020. This intervention is scalable and has the potential to reduce costs and preventable harm to patients.

Background: Homelessness is a growing Canada-wide concern. Those with no fixed address have increased rates of emergency department (ED) utilization and increased healthcare spending compared to the general population, with higher rates of acute and chronic illnesses, as well as all-cause mortality. EDs are uniquely situated to act as an access point to the network of available community services, however referral rates from the ED is uncertain. To date, there has been no data collected on London, Ontario's homeless population, their health burden, and their utilization patterns of the ED. Aim Statement: The primary objective of this study is to describe ED visits for adult patients with no fixed address in London, Ontario to assess for potential areas to improve care. Measures & Design: This is a retrospective chart review, of patients with no fixed address visiting London, Ontario Emergency Departments in 2018. ED visits were identified and pulled using either a diagnosis of “homeless”, a lack of postal code, or a postal code for a known shelter. Cases included based on postal code were manually reviewed to determine whether the patient had a resident address with the same postal code. Evaluation/Results: From this search, 4,294 visits were identified for 1237 unique patients. The median visits per person was 1 (IQR 1-2), with 388 patients having 3 or more visits, and the max being 138 visits. The median age was 38 (IQR 28-52), with 73% male. Ground ambulance was used for 46% of visits. 28% of visits were CTAS 1&2 and 5% were CTAS 5. Police facilitated visits in 401 cases. Top 3 discharge diagnosis categories were mental health (19%), infection (18%), drug misuse (17%). Discussion/Impact: Several errors were identified with our search strategy suggesting the current system of capturing homelessness in the EPR is not accurate, leading to an underestimation of the problem and limiting our ability to describe this population. The Ministry of Health mandates homelessness be applied as a tertiary discharge diagnosis during coding of the patient visit if possible. However, use of this code is inconsistent leading to large-scale omission of visits and an underrepresentation of pediatric cases. Systemic steps should be taken to improve identification of these patients moving forward.

Background: Preliminary reports and subsequent immediate management decisions of radiological scans are often performed by emergency physicians and on-call radiology residents. Many academic hospitals have resident-only coverage for after-hour shifts. Generally, these preliminary reports are eventually reviewed by a staff radiologist, during which discrepancies may be identified. Depending on the severity of the discrepancy and the time taken to notify the treating physician, there is potential for significant impact on the patient's care. Aim Statement: In an attempt to identify and minimize errors in radiological readings, and to improve the communication of discrepancies, our project aims to retrospectively audit all radiological discrepancies that have occurred at The Ottawa Hospital's emergency departments from April 2018 to May 2019. Measures & Designs: A systematic review of all cases with noted radiological discrepancies was obtained from the Picture Archive and Communication System software and EPIC platform. Analysis of these cases will allow us to define when errors occur, what is the type and severity of the error, how long it took to relay the discrepancy to a treating physician, and what was the subsequent management impact. Evaluation/Results: We discovered 712 cases with radiological reading discrepancies, 168 major, 527 minor, and 17 incidentals. Interestingly, a significant portion of major (severely affecting care/life-threatning) discrepancies were reported from radiology residents, especially on CT images, although emergency physicians had the most discrepancies (mostly minor). Radiology residents were seen to have more discrepant reports during after-hour services while emergency physicians did not show any specific pattern of discrepant reporting. The average time to report a major discrepancy to a treating physician is 8.8 hours, where the maximum time taken was 104 hours and the minimum was 0.2 hours. 56% of reports with major discrepancies made no mention of who was notified. Discussion/Impact: By identifying weak points in radiological reporting, our results will allow us to provide suggestions at an administration and teaching level to minimize discrepancies. It is critical to create a workflow where mistakes are mitigated, and communication is efficient and standardized to prevent patient harm from delayed or incorrect diagnosis.

Background: In patients with acute hip fracture, a fascia iliaca compartment block (FICB) has been shown to provide effective non-opioid analgesia, reduce the incidence of pneumonia, and potentially decrease the rate of delirium [1]. However, this procedure was infrequently used in the St. Michael's Hospital (SMH) emergency department (ED). Aim Statement: Our aim was to increase the proportion of patients with hip fracture receiving FICB in the ED to 50% in six months. Measures & Design: We completed two Plan-Do-Study-Act (PDSA) cycles, measuring rates of FICB before and after each cycle. The first was a departmental rounds presentation with information about the process and benefits of FICB, addressing barriers identified by surveying the group. The second cycle included a bundle of interventions comprising of an “instruction card” with the steps required to do the procedure, access to a video tutorial, and a list of experienced physicians willing to help less experienced providers perform FICB. Evaluation/Results: In the three months prior to the project, the rate of FICB in the ED was 12.5% (3/24). For the three months after the first PDSA cycle, the rate increased to 22.2% (8/36). Then, the second cycle was performed. In the following two months the rate further increased to 36.8% (7/19). Discussion/Impact: Despite the clear increase in FICB rate, these changes were not statistically significant (p = 0.063). Our methodology was shown to be safe and effective, and our model can be applied to other ED groups looking to increase their rates of FICB.

Background: Clinical handover between emergency medical services (EMS) and the hospital trauma team can be subject to errors that may negatively affect patient care. Thus far, there has been limited evaluation of the quality of EMS handover. As such, we sought to characterize handover practices from EMS to the trauma team, identify areas for improvement, and determine if there is a need for standardization of current handover practices. Aim Statement: Identify areas for improvement in handover from EMS to the trauma team, specifically examining handover content, structure, and discordances between different team members regarding handover expectations. Measures & Design: Data were prospectively collected over a nine week period by a trained observer at Canada's largest level one trauma centre. A randomized scheduled was used to capture a representative breadth of handovers. Data collected included outcome measures such as duration of handover, structure of the handover, and information shared, process measures such as questions and interruptions from the trauma team, and perceptions of the handover from nurses, trauma team leaders (TTLs) and EMS according to a bidirectional Likert scale. Evaluation/Results: Of 410 trauma team activations, 79 verbal handovers were observed. Information was often missing regarding airway (present 22%), breathing (54%), medications (59%), and allergies (54%). Handover structure lacked consistency beyond the order of identification and mechanism of injury. Only 28% of handovers had a dedicated question and answer period. Of all questions asked, 35% were questioning previously given information. EMS returned to categories of information unprompted in 84% of handovers. The majority of handovers (61%) involved parallel conversations between team members while EMS was speaking, which was associated with a greater number of interrupting questions from the trauma team (3.15 vs. 1.82, p =.001). There was a statistically significant disparity between the self-evaluation of EMS handovers and the perceived quality determined by nurses and trauma team leaders. Discussion/Impact: At our trauma centre, we have identified the need for handover standardization due to poor information content, a lack of structure and active listening, significant information repetition, and discordant expectations between EMS, nurses, and TTLs. We intend to use our results to guide the development of a co-constructed framework integrating the perspectives of all team members on the trauma team.

Background: Curbing unnecessary laboratory testing represents a significant opportunity for cost reduction in the Canadian health care system. A Choosing Wisely report cited a 31% decline in the number of tests ordered in a Canadian emergency department (ED) after implementation of recommendations. The international normalized ratio (INR) remains frequently ordered in emergency departments without an appropriate indication. Aim Statement: We aimed to reduce the number of INR tests completed in the St. Joseph's Healthcare Hamilton Emergency Department by 50% by April 30, 2019. Measures & Design: We conducted the study in an urban, academic ED employing the Epic electronic health record (EHR). We tailored interventions according to the Hierarchy of Effectiveness to address root causes revealed by analysis of our baseline ordering behaviour. Interventions included provider education around evidence-based ordering indications and removal of the INR from our “chest pain” bloodwork panel. Our outcome measure was the weekly number of INR tests completed per ED visit. Process measures included the proportion of INR tests ordered for inappropriate indications on monthly audits of 20 charts where an INR was completed. Balancing measures included average ED length of stay for patients receiving INR testing. Evaluation/Results: We collected outcome, process, and balancing measures through the EHR and analyzed this data using statistical process control charts. Over the nine-month study period, we decreased weekly INR tests from 248.4 to 115.0, a reduction of 56% which met criteria for special cause variation. This amounts to a cost savings of $43,008 per year. ED length of stay for patients receiving INR testing did not change significantly. Discussion/Impact: Our interventions were successful in realising our 50% target reduction in INR tests without an increase in ED length of stay from repeat venipuncture. This result is in keeping with similar efforts in other Canadian EDs. Our interventions could likely be spread to other settings where an INR is included as part of a “chest pain” panel. This may represent a substantial cost reduction opportunity on a national scale. Further work is needed in order to assess long term sustainability, which can be supported by employing high effectiveness mechanisms such as automation of optimal behaviour.

Background: Hemolysis of blood samples is the leading cause of specimen rejection from hospital laboratories. It contributes to delays in patient care and disposition decisions. Coagulation tests (prothrombin time/international normalized ratio [PT/INR] and activated partial thromboplastin time [aPTT]) are especially problematic for hemolysis in our academic hospital, with at least one sample rejected daily from the emergency department (ED). Aim Statement: We aimed to decrease the monthly rate of hemolyzed coagulation blood samples sent from the ED from a rate of 2.9% (53/1,857) to the best practice benchmark of less than 2% by September 1st, 2019. Measures & Design: Our outcome measure was the rate of hemolyzed coagulation blood samples. Our process measure was the rate of coagulation blood tests sent per 100 ED visits. Our balancing measure was the number of incident reports by clinicians when expected coagulation testing did not occur. We used monthly data for our Statistical Process Control (SPC) charts, as well as Chi square and Mann-Whitney U tests for our before-and-after evaluation. Using the Model for Improvement to develop our project's framework, we used direct observation, broad stakeholder engagement, and process mapping to identify root causes. We enlisted nursing champions to develop our Plan-Do-Study-Act (PDSA) cycles/interventions: 1) educating nurses on hemolysis and coagulation testing; 2) redesigning the peripheral intravenous and blood work supply carts to encourage best practice; and 3) removing PT/INR and aPTT from automatic inclusion in our electronic chest pain bloodwork panel. Evaluation/Results: The average rate of hemolysis remained unchanged from baseline (2.9%, p = 0.83). The average rate of coagulation testing sent per 100 ED visits decreased from 41.5 to 28.8 (absolute decrease 12.7 per 100, p < 0.05), avoiding $4,277 in monthly laboratory costs. The SPC chart of our process measure showed special cause variation with greater than eight points below the centerline. Discussion/Impact: Our project reduced coagulation testing, without changing hemolysis rates. Buy-in from frontline nurses was integral to the project's early success, prior to implementing our electronic approach – a solution ranked higher on the hierarchy of intervention effectiveness – to help sustainability. This resource stewardship project will now be spread to a nearby institution by utilizing similar approaches.

Background: Over 35,000 Canadians lose their lives to cardiac arrest each year. CPR and automated external defibrillator (AED) use are modifiable factors. Survival rates drop by 7-10% each minute that defibrillation is delayed, and survival rates are less than 5% after 12 minutes of ventricular fibrillation which stresses the need for bystander AED use in out-of-hospital arrests. Niagara Region lacks a publicly accessible registry of AEDs. AED access is a major focus in King County, Washington which has higher survival rates and has all AEDs registered with Emergency Medical Services. Aim Statement: This project aims to log 100 or more AEDs within a year into a publicly accessible registry and to connect the registry information to medical trainees in the Niagara region and all employees of the Niagara Health System involved in patient care. Measures & Design: PulsePoint is an application used to register AEDs within the Niagara region. PulsePoint allows users to geotag AEDs while tracking data entries. Over 16 weeks, 4 PDSA cycles tested the effectiveness of logging methods for AEDs including opportunistic logging, daily emailed reminders, and contacting organizations with high likelihood of having an AED. Information about the project and registry was shared with residents and medical students in Niagara. A second phase of cycles involves relaying information to Niagara Health system employees and the medical community. A final cycle will target a broader group of local organizations with intermediate probability of having AEDs. Primary outcome measures include the numbers of regional AEDs logged and members reached by knowledge sharing cycles. Evaluation/Results: PulsePoint was found to be an effective, free, publicly accessible resource to log AEDs within the Niagara region. The initial round of 4 PDSA cycles added a total of 56 new AEDs within the region, which were logged into PulsePoint app and the Excel spreadsheet. Through the fourth PDSA cycle, 136 businesses were contacted and made aware of the project and the AED application. In addition,138 health-related colleagues and medical students were contacted to raise awareness. PDSA cycles five through eight are currently ongoing or in the planning stages. Discussion/Impact: Raising awareness among emergency services and sharing information about the registry to local CPR training providers will be paramount. Creating awareness of PulsePoint and installing AEDs in locations that currently lack such devices could ultimately improve cardiac arrest survival rates within Niagara Region.

Background: In Canada, injuries represent 21% of Emergency Department (ED) visits. Faced with occupational injuries, physicians may feel pressured to provide urgent imaging to facilitate expedited return to work. There is not a body of literature to support this practice. Twenty percent of adult ED injuries involve workers compensation. Aim Statement: Tacit pressures were felt to impact imaging rates for patients with workplace injuries, and our aim was to determine if this hypothesis was accurate. We conducted a quality review to assess imaging rates among injuries suffered at work and outside work. A secondary aim was to reduce the harm resulting from non-value-added testing. Measures & Design: Information was collected from the Emergency Department Information System on patients with acute injuries over the age of 16-years including upper limb, lower limb, neck, back and head injuries. Data included both workplace and non-work-related presentations, Canadian Triage and Acuity Scale (CTAS) levels and age at presentation. Imaging included any of X-ray, CT, MRI, or Ultrasound ordered in EDs across the central zone of Nova Scotia from July 1, 2009 to June 30, 2019. A total of 282,860 patient-encounters were included for analysis. Comparison was made between patients presenting under the Workers’ Compensation Board of Nova Scotia (WCB) and those covered by the Department of Health and Wellness (DOHW). Imaging rates for all injuries were also trended over this ten-year period. Evaluation/Results: In patients between 16 and 65-years, the WCB group underwent more imaging (55.3% of visits) than did the DOHW group (43.1% of visits). In the same cohort, there was an overall decrease of over 10% in mean imaging rates for both WBC and DOHW between the first five-year period (2009-2013) and the second five-year study period (2013-2018). Imaging rates for WCB and DOHW converged with each decade beyond 35 years of age. No comparison was possible beyond 85-years, due to the absence of WCB presentations. Discussion/Impact: Patients presenting to the ED with workplace injuries are imaged at a higher rate than those covered by the DOHW. Campaigns promoting value-added care may have impacted imaging rates during the ten-year study period, explaining the decline in ED imaging for all injuries. While this 10% decrease in overall imaging is encouraging, these preliminary data indicate the need for further education on resource stewardship, especially for patients presenting to the ED with workplace injuries.

Background: Pediatric pain is often under-treated in emergency departments (EDs), causing short and long-term harm. In Alberta EDs, children's pain outcomes were unknown. A recent quality improvement collaborative (QIC) led by our team improved children's pain care in 4 urban EDs. We then spread to all EDs in Alberta using the Institute for Healthcare Improvement Framework for Going to Full Scale. Aim Statement: To increase the proportion of children <12 years who receive topical anesthetic before needle procedures from 11% to 50%; and for children <17 years with fractures: to 1) increase the proportion receiving analgesia from 31% to 50%; 2) increase the proportion with pain score documentation from 24% to 50%, and 3) reduce time to analgesia from 60 to 30 minutes, within 1 year. Measures & Design: All 97 EDs in Alberta that treat children were invited. Each was asked to form a project team, attend webinars, develop key driver diagrams and perform PDSA tests of change. Sites were given a monthly list of randomly selected charts for audit and entered data in REDCap for upload to a provincial run chart dashboard. Baseline performance measurement informed aims. Measures included proportion of children <12 years undergoing a lab test who received topical anesthetic, and for children <17 years with fracture, the proportion with a pain score, proportion receiving analgesia and median minutes to analgesia. Length of stay and use of opioids were balancing measures. Control charts were used to detect special cause. Interrupted time series (ITS) was performed to assess significance and trends. Evaluation/Results: 36 sites (37%) participated, including rural and urban sites from all regions. 8417 visits were audited. 23/36 sites completed audits before and after tests of change and were analyzed. Special cause occurred for all aims. The proportion receiving topical anesthetic increased from 11% to 30% (ITS p < 0. 001). For children with fractures, the proportion with pain scores increased from 24% to 34% (ITS p = 0.21, underlying trend present), proportion receiving analgesic medication increased from 31% to 39% (ITS p = 0.41, underlying trend present) and minutes to analgesia decreased from 60 to 28 (ITS p < 0. 01). There was no increase in length of stay or use of opioid medications. Discussion/Impact: A pragmatic approach encouraging locally led change was well-received and key to success. The QIC method shows promise for improving outcomes in diverse EDs across large geographic areas. Next steps include further spread and sustainability measurement.

Background: Unintentional opioid overdoses in and around acute care hospitals, including in the ED, are of increasing concern. In April 2018, the Addiction Recovery and Community Health (ARCH) Team at the Royal Alexandra Hospital opened the first acute care Supervised Consumption Service (SCS) in North America available to inpatients. In the SCS, patients can consume substances by injection, oral or intranasal routes under nursing supervision; immediate assistance is provided if an overdose occurs. After a quality assurance review, work began to expand SCS access to ED patients as well. Aim Statement: By expanding SCS access to ED patients, we aim to reduce unintentional and unwitnessed opioid overdoses in registered ED patients to 0 per month by the end of 2020. Measures & Design: Between June 13-July 15, 2019, ARCH ED Registered Nurses were asked to identify ED patients with a history of active substance use who may potentially require SCS access. Nurses identified 69 patients over 43 8-hour shifts (range 0-4 patients per shift); thus, we anticipated an average of 5 ED patients per 24-hour period to potentially require SCS access. Based on this evidence of need, ARCH leadership worked with a) hospital legal team and Health Canada to expand SCS access to ED patients; b) ED leadership to develop a procedure and flowchart for ED SCS access. ED patients were able to access the SCS effective October 1, 2019. Evaluation/Results: From October 1 to December 1, 2019, the SCS had 35 visits by 23 unique ED patients. The median time spent in the SCS was 42.5 minutes (range 14.0-140.0 minutes). Methamphetamine was the most commonly used substance (19, 45.2%), followed by fentanyl (10, 23.8%); substances were all injected (91.4% into a vein and 8.6% into an existing IV). In this time period, there were zero unintentional, unwitnessed opioid poisonings in registered ED patients. Data collection is ongoing and will expand to include chief complaint, ED length of stay and discharge status. Discussion/Impact: Being able to reduce unintentional overdoses and unwitnessed injection drug use in the ED has the potential to improve both patient and staff safety. Next steps include a case series designed to examine the impact of SCS access on emergency care, retention in treatment and uptake into addiction treatment.

Background: Increasing Emergency Department (ED) stretcher occupancy with admitted patients at our tertiary care hospital has contributed to long Physician Initial Assessment (PIA) times. As of Oct 2019, median PIA was 2.3 hours and 90th percentile PIA was 5.3 hours, with a consequent 71/74 PIA ranking compared to all Ontario EDs. Ambulatory zone (AZ) models are more commonly used in community EDs compared to tertiary level EDs. An interdisciplinary team trialled an AZ model for five days in our ED to improve PIA times. Aim Statement: We sought to decrease the median PIA for patients in our ED during the AZ trial period as compared to days with similar occupancy and volume. Measures & Design: The AZ was reserved for patients who could walk from a chair to stretcher. In this zone, ED rooms with stretchers were for patient assessment only; when waiting for results or receiving treatment, patients were moved into chairs. We removed nursing assignment ratios to increase patient flow. Our outcome measure was the median PIA for all patients in our ED. Our balancing measure was the 90th percentile PIA, which could increase if we negatively impacted patients who require stretchers. The median and 90th percentile PIA during the AZ trial were compared to similar occupancy and volume days without the AZ. Additional measures included ED Length of Stay (LOS) for non-admitted patients, and patients who leave without being seen (LWBS). Clinicians and patients provided qualitative feedback through surveys. Evaluation/Results: The median PIA during the AZ trial was 1.5 hours, compared to 2.1 hours during control days. Our balancing measure, the 90th percentile PIA was 3.7 hours, compared to 5.0 during control days. A run chart revealed both median and 90th percentile PIA during the trial were at their lowest points over the past 18 months. The number of LWBS patients decreased during the trial; EDLOS did not change. The majority of patients, nurses, and physicians felt the trial could be implemented permanently. Discussion/Impact: Although our highly specialized tertiary care hospital faces unique challenges and high occupancy pressures, a community-hospital style AZ model was successful in improving PIA. Shorter PIA times can improve other quality metrics, such as timeliness of analgesia and antibiotics. We are working to optimize the model based on feedback before we cycle another trial. Our findings suggest that other tertiary care EDs should consider similar AZ models.

Background: Trauma care represents a complex patient journey, requiring multi-disciplinary coordinated care. Team members are human, and as such, how they feel about their colleagues and their work affects performance. The challenge for health service leaders is enabling culture that supports high levels of collaboration, cooperation and coordination across diverse groups. Aim Statement: We aimed to define and set the agenda for improvement of the relational aspects of trauma care at a large tertiary care hospital. Measures & Design: We conducted a mixed-methods collaborative ethnography using the Relational Coordination survey – an established tool to analyze the relational dimensions of multidisciplinary teamwork – participant observation, interviews, and narrative surveys. Findings were presented to clinicians in working groups for further interpretation and to facilitate co-creation of targeted interventions designed to improve team relationships and performance. Evaluation/Results: We engaged a complex multidisciplinary network of ~500 care providers dispersed across seven core interdependent clinical disciplines. Initial findings highlighted the importance of relationships in trauma care and opportunities to improve. Narrative survey and ethnographic findings further highlighted the centrality of a translational simulation program in contributing positively to team culture and relational ties. A range of 16 interventions – focusing on structural, process and relational dimensions – were co-created with participants and are now being implemented and evaluated by various trauma care providers. Discussion/Impact: Through engagement of clinicians spanning organizational boundaries, relational aspects of care can be measured and directly targeted in a collaborative quality improvement process. We encourage health care leaders to consider relationship-based quality improvement strategies, including translational simulation and relational coordination processes, in their efforts to improve care for patients with complex, interdependent journeys.

Background: The diagnosis of urinary tract infection (UTI) is made based on symptoms, urinalysis and urine culture. While simple urinary tract infections do not require routine culture, the Infectious Disease Society of America (IDSA) Guidelines state that complicated urinary tract infections should have urine cultures performed to determine which antibiotics are effective, as there is a higher risk of infection with resistant organisms. We hypothesized that the rate of urine cultures sent for complicated UTI is less than is recommended by the literature. Aim Statement: We aimed to implement a follow-up reporting system for Urinary Culture in patients diagnosed with complicated UTIs and raise our Urinary Culture rates in this population to 80% by June 2019. Measures & Design: We performed a single-center chart review using Emergency Department (ED) charts of non-admitted patients. They were audited daily for two weeks to obtain a sample of patients who had a discharge diagnosis of urinary tract infection, pyelonephritis or cystitis. Charts capturing these diagnoses were assessed to see if a culture was clinically indicated and if it was ordered. Charts were screened for the presence of any of the following criteria indicating complicated UTI: known structural or functional abnormality of the urinary tract, genitourinary obstruction, pregnancy, immunosuppression, diabetes, indwelling or intermittent catheter use, fever, male patient, clinical pyelonephritis, antimicrobial failure, or transfer from a nursing home. Data was then compiled to determine culture rates in complicated and uncomplicated UTIs. This prevalence rate established the baseline performance in the ED which was used to inform the quality improvement project. Evaluation/Results: Over a two week period, 26 patients were discharged from the ED with a diagnosis of UTI, with 17 of these patients meeting criteria for complicated UTI. Only 6 of 17 complicated UTIs were sent for urine culture, therefore our pre-implementation culture rate was 35%. After initial data collection, a follow-up system was designed ensuring that urine culture and sensitivities results would be compiled and reviewed daily at Hamilton Health Sciences. This system was created with input from key stakeholders including department chiefs, core lab services, ED physicians and business clerks. A discrepancy form was created for documentation of culture result recognition and any required patient follow up ie. antibiotic change. In October 2019, the system had been implemented for a month, after which another chart review was completed. 27 cases were captured, 18 of which were complicated. The complicated culture rate had increased significantly from 35% to 72%. Discussion/Impact: In the ED, ordering of cultures for patients being discharged, regardless of type, is commonly associated with concern of result follow up, which may take up to 72 hours. This discrepancy system was implemented to ensure that all urine cultures ordered had appropriate follow up, thus supporting physicians in ordering cultures when indicated. The significant improvement in culture rate from 35% to 72% is balanced by one single culture of all 9 simple UTIs (11%). In PDSA cycle 2, we hope to increase rates to 90% by improving current challenges with the system.

Background: Emergency department (ED) overcrowding is associated with a broad spectrum of poor medical outcomes, including medical errors, mortality, higher rates of leaving without being seen, and reduced patient and physician satisfaction. The largest contributor to overcrowding is access block – the inability of admitted patients to access in-patient beds from the ED. One component to addressing access block involves streamlining the decision process to rapidly determine which hospital service will admit the patient. Aim Statement: As of Sep 2011, admission algorithms at our institution were supported and formalised. The pancreatitis algorithm clarified whether general surgery or internal medicine would admit ED patients with pancreatitis. We hypothesize that this prior uncertainty delayed the admission decision and prolonged ED length of stay (LOS) for patients with pancreatitis. Our project evaluates whether implementing a pancreatitis admission algorithm at our institution reduced ED time to disposition (TTD) and LOS. Measures & Design: A retrospective review was conducted in a tertiary care academic hospital in Montreal for all adult ED patients diagnosed with pancreatitis from Apr 2010 to Mar 2014. The data was used to plot separate run charts for ED TTD and LOS. Serial measurements of each outcome were used to monitor change and evaluate for special cause variation. The mean ED LOS and TTD before and after algorithm implementation were also compared using the Student's t test. Evaluation/Results: Over four years, a total of 365 ED patients were diagnosed with pancreatitis and 287 (79%) were admitted. The mean ED LOS for patients with pancreatitis decreased following the implementation of an admission algorithm (1616 vs. 1418 mins, p = 0.05). The mean ED TTD was also reduced (1171 vs. 899 mins, p = 0.0006). A non-random signal of change was suggested by a shift above the median prior to algorithm implementation and one below the median following. Discussion/Impact: This project demonstrates that in a busy tertiary care academic hospital, an admission algorithm helped reduce ED TTD and LOS for patients with pancreatitis. This proves especially valuable when considering the potential applicability of such algorithms to other disease processes, such as gastrointestinal bleeding and congestive heart failure, among others. Future studies demonstrating this external applicability, and the impact of such decision algorithms on physician decision fatigue and within non-academic institutions, proves warranted.

Background: Emergency Department overcrowding remains a significant problem. Interventions have often focused on areas external to the ED, with patient flow in the ED receiving less attention. Efforts to address ED flow are complicated by daily fluctuations in patient volume and acuity. Our local protocol brings in additional physicians when internal metrics indicate patient demand can't be met by current physician resources (a ‘surge’ period). However, anecdotal evidence suggests a lack of satisfaction and efficacy. We therefore undertook a project to improve our local management of these surge periods. Aim Statement: To improve the effectiveness of an ED Physician Surge Protocol to allow for a physician scheduling strategy that is reflective of the needs of the ED. Measures & Design: This project consists of 3 phases. Phase 1 was an analysis of current surge metrics (including frequency, temporal patterns and physician response), with concurrent literature search to identify any best practices or easily addressable protocol changes, with first planned PDSA cycle. Phase 2 is a mixed methods survey of local staff to identify barriers and enablers of our current protocol, concurrent with a national survey of current practices. Phase 3 will be the implementation of a revised protocol, followed by a second mixed methods survey and analysis of metrics of interest. Evaluation/Results: Analysis of surge data (Oct 2018-Oct 2019) demonstrated a high volume of surges per month (78.7 +/- 10.9), highest at Foothills Medical Centre (94.3). Across all sites, afternoon periods had highest frequency of surges (absolute peak 1400 - 1500) with a secondary peak 2200–2300, both peaks occurring most frequently on weekends (Fri-Sun) However, physician response to surge calls was < 10% (5.8-9.1%), with no discernable temporal pattern, even accounting for the significant number of automatic surge calls cancelled by clinicians. Analysis of data, in addition to literature review and engagement with senior administration suggested no immediate protocol changes, therefore project moved to 2nd phase. This phase is currently in progress, with planned analysis using Pareto Chart methodology. Discussion/Impact: Our initial data clearly demonstrates that current procedures are inadequate to address this ongoing issue, with no readily apparent solutions. Analysis of local barriers and enablers is currently underway, in addition to a national survey, with the results expected to inform an extensive redesign of current procedures.

Background: Urinary tract infections (UTI) are a common emergency department (ED) presentation. Urine cultures (UC) are frequently ordered to confirm the diagnosis, however, it can be challenging to differentiate between a true infection and asymptomatic bacteriuria (ASB) which does not generally benefit from antibiotics. This over-treatment of ASB leads to serious adverse side effects, growing antimicrobial resistance and increased healthcare costs. By reducing inappropriate ED urine culture testing, we can concomitantly avoid the false positives that contribute to this large-scale problem. Aim Statement: We aimed to reduce ED urine culture testing at Credit Valley Hospital, a large community hospital based in Mississauga, Ontario by 30%, from a baseline average of 97 cultures per 1000 ED visits in 2017, to 68 cultures per 1000 ED visits by year end 2019. Measures & Design: Multiple PDSA cycles were ran with our multi-disciplinary ED team. Our interventions to encourage rational urine culture testing are three-fold, including (1) medical directive optimization (removal of routine sending of UC), (2) individualized physician feedback and (3) physician education with introduction of a clinical decision aid. Our outcome measure is rate of UC per 1000 ED patient visits with a balance measure of rate of 30-day ED return visit of hospital admission for patients with a UTI. Evaluation/Results: Despite a parallel surge in ED volumes, we observed a significant decrease in urine culture testing, from an annual average of 97 cultures per 1000 ED visits to 60 cultures per 1000 ED visits in 2019 year-to-date. There was no increase in the rate of ED 30-day return visit or admission for UTI or a diagnostic equivalent. Discussion/Impact: Our multipronged approach effectively decreased the rate of UC testing during the study period. ED physicians provide higher quality care with judicious use of resources to guide diagnosis and management. Active ongoing interventions include our transition to a 2-step UC order protocol (uncoupling urinalysis with culture) using BD vacutainer urine collection products, which will allow for 48 hour storage of uncompromised urine. Further work will leverage our knowledge and experience with optimizing urine culture testing to other culture specimens.

Background: Telemedicine has been defined as the use of technology to provide healthcare when the provider and patient are geographically separated. Use of telemedicine to meet the needs of specific populations has become increasingly common across Canada. The current study employs the Ontario Telemedicine Network (OTN) to connect the emergency departments of a community hospital system and a pediatric tertiary care hospital. OTN functions through a two-way video conferencing system, allowing physicians at the tertiary site to see and hear the patient being treated in the community hospitals. Aim Statement: The aim of this project is to ensure essential care is provided to CTAS 1 and 2 pediatric patients who present to Niagara Health emergency departments, to increase the number of appropriate patient transfers. Measures & Design: Data for this project include a) description of common diagnoses, b) time of call, c) occurrence of transfers, and d) professional perceptions of the technology. A descriptive design was used together with the implementation of quality improvement cycles as the intervention occurred. Quality improvement methodologies including plan-do-study-act (PDSA) cycles ensured continuous improvement to the process of OTN use and therefore patient safety throughout the study. Evaluation/Results: Since the intervention was employed on December 17, 2018 there have been a total of 19 cases for which 4 transfers were requested. Changes to the process were made including the addition of weekly technology tests and feedback to health professionals involved to garner further support for the use. Results have indicated that seizure was the most common diagnosis, accounting for 37% of cases. The majority of calls were placed after 19:00 hours with no calls being placed between 24:00 and 10:00. Discussion/Impact: Healthcare providers had positive perceptions of the technology agreeing that decision making between on-site and remote teams was timely and collaborative, as well as that patient care and outcomes were improved with its use. The results of this study will be used to determine the benefits of employing telemedicine in the emergency departments of other hospital systems.

Background: Liberal prescribing of opioids is a major contributing factor to the opioid crisis. Patients who take opioids for >5 consecutive days are at greater risk of long-term use. Evidence shows that significantly more opioids are prescribed for emergency department (ED) patients with acute pain compared to amounts consumed. Guidelines recommend prescribing a 3-day supply or 10-15 tablets of opioids for patients with acute pain Aim Statement: By January 2020, >70% of opioid prescriptions from our ED will be for <15 tablets of morphine 5 mg equivalents. Measures & Design: Emergency physicians were educated on best practice of prescribing opioids for discharged patients. An electronic prescription writer was built for discharged ED patients with a pop-up reminder for quantities >15 tablets (indicating a recommended quantity of 10-15 tablets) and a pop-up reminder for quantities >30 tablets (indicating a maximum quantity of 30 tablets and recommended quantity). A feature was built to auto-populate a prescription for morphine 5 mg po q4h prn x 10 tablets to facilitate adherence to guidelines. Outcome Measure % opioid prescriptions for <15 tablets of morphine 5 mg equivalents Process Measure Amount of opioids prescribed for discharged ED patients, measured as morphine 5 mg equivalents Number of opioid prescriptions for >30 tablets of morphine 5 mg equivalents Balancing Measure Number of patients that return to ED within 7 days and receive a repeat opioid prescription. Evaluation/Results: Prior to implementation of the electronic prescription writer a sample audit revealed that 50% of opioid prescriptions were written for <15 tablets of morphine 5 mg equivalents. For the first three quarters of 2019, 62%, 61% and 69% of opioid prescriptions were written for <15 tablets of morphine 5 mg equivalents. Only two prescriptions during the study period were for >30 tablets of morphine 5 mg equivalents. An average number of 7 patients per quarter were given a repeat opioid prescription during a return ED visit. Discussion/Impact: We were successful in influencing emergency physicians to prescribe fewer opioids to discharged patients. This has the potential to avoid converting ED patients with acute pain into long-term opioid users and to avoid the diversion of unused opioid tablets.