Introduction: It is typical to look for UTI in delirious elderly patients, despite a high prevalence of asymptomatic bacteriuria (ASB) in this population. A common presentation of infection is delirium, which often has a non-specific and multifactorial etiology. Therefore, when bacteriuria is present with delirium in the absence of urinary symptoms, physicians prescribe antibiotics for the suspected UTI-induced delirium. We set to determine whether antibiotic treatment in the elderly presenting with delirium in the presence of ASB resulted in resolution of delirium. Methods: Literature searches were performed in MEDLINE, EMBASE, CINAHL and Cochrane Library. Abstracts were independently reviewed by two authors for decision to include for full-text review. Inclusion criteria included female gender, >65 years of age, presenting in an acute care setting with delirium and ASB. The primary outcome was resolution of delirium. The secondary outcomes were mortality, frequency of side effects from antibiotics, length of hospital stay and readmission for delirium. Results: 930 abstracts published from 1946-2017 were screened, and 42 were included for full text review. No studies were eligible for inclusion in the systematic review, as none addressed the primary outcome. One study addressed the outcomes of poor functional recovery after delirium and the rate of improvement of delirium symptoms after presentation of delirium with ASB. Conclusion: Even though current guidelines recommend against treatment of ASB, no guideline states whether ASB should be treated in elderly patients with delirium. Little evidence exists to elucidate whether treating delirious patients with ASB results in improvement in outcomes. Future studies should focus on demonstrating the relationship between resolution of delirium with antibiotic treatment. This will clarify whether delirium is a true symptom of ASB and whether treatment results in faster resolution of delirium.

Introduction: Cognitive processing theories postulate that decision making depends on both fast and slow thinking. Experienced physicians (EPs) make diagnoses quickly and with less effort by using fast, intuitive thinking, whereas inexperienced medical students rely on slow, analytical thinking. This study used a cognitive task analysis to examine EPs cognitive processes and ability to provide knowledge translation to learners. Methods: A novel mind mapping approach was used to examine how EPs translate their clinical reasoning to learners, when evaluating a patient for a possible venous thromboembolism (VTE). Nine EPs were interviewed and shown two different videos of a medical student patient interview (randomized from six possible videos). Results: EPs were asked to demonstrate their clinical approach to the scenario using a mind map, assuming they were teaching a learner in the Emergency Department. EPs were later re-interviewed to examine response stability, and given the opportunity to make clarifying or substantive mind map modifications. Maps were broken into component pieces and analyzed using mixed-methods techniques. A mean of 15.7 component pieces were identified within each mind map (standard deviation (SD) 7.8). Maps were qualitatively coded, with a mean of 2.8 clarifying amendments (e.g. adding a time course caveat) (SD 1.5-5.75) and 4.4 substantive modifications (e.g. changing the flow of the map) (SD 2-5). Conclusion: Resulting mind maps displayed significant heterogeneity in teaching points and the degree to which EPs used slow thinking. EPs frequently made fast thinking jumps, although learners could prompt slow thinking by questioning unclear points. This is particularly important as learners engage in cognitive apprenticeship throughout their training. An improved understanding of EPs cognitive processes through mind mapping will allow learners to improve their own clinical reasoning (Merrit et al., 2017). Educating EPs on these processes will allow modification of their teaching styles to better suit learners.

Introduction: Non-publication of trial findings results in research waste and compromises medical evidence and the safety of interventions in child health. The objectives of this study were to replicate, compare and contrast findings of a previous study (Klassen et al., 2002) to determine the impact of ethical and editorial mandates to register and publish findings. Methods: Abstracts accepted to the Pediatric Academic Societies meetings (2008-2011) were screened in duplicate to identify Phase-III RCTs enrolling pediatric populations. Subsequent publication was ascertained through a search of electronic databases. Study internal validity was measured using Cochrane Risk of Bias and Jadad Scale, and key variables (e.g., trial design, study stage) were extracted. Pearson X2, t-tests and Wilcoxon rank sum tests were used to examine association between variables and publication status. Logistic regression, log-rank tests, rank correlation and Egger regression were used to assess predictors of publication, time to publication and publication bias, respectively. Results: Compared to our previous study, fewer studies remained unpublished (27.9% vs 40.9%, p=.007). Abstracts with higher sample sizes (p=0.01) and those registered in ClinicalTrials.gov were more likely to be published (p<.0001). There were no differences in quality measures/risk of bias or in preference for positive results (p=0.36) between published and unpublished studies. Mean time to publication was 26.5 months and published manuscripts appeared most frequently in Pediatrics, the Journal of Pediatrics, and Pediatric Emergency Care. The funnel plot (p=0.04) suggests a reduced but ongoing existence of publication bias among published studies. Overall, we observed a reduction in publication bias and in preference for positive findings, and an increase in study size and publication rates over time. Conclusion: Despite heightened safeguards and editorial policy changes in recent decades, publication bias remains commonplace and presents a threat to assessing the efficacy and effectiveness of interventions in child health. Our results suggest a promising trend towards a reduction in publication bias over time and positive impacts of trial registration. Further efforts are needed to ensure the entirety of evidence can be accessed when assessing treatment effectiveness.

Introduction: Take Home Naloxone (THN) programs prevent death from opioid poisoning by training laypersons to recognize an overdose and administer naloxone. Dispensing THN through the emergency department (ED) is particularly critical because an ED visit for opioid poisoning strongly predicts future mortality. Many EDs have implemented THN programs, yet almost no literature examines the reach of such initiatives. To address this gap, we conducted a chart review of all patients presenting for opioid poisoning to an urban tertiary hospital, with a large ED-based THN program. This exploratory study hypothesized that more than 50% of ED patients presenting for opioid poisoning would be offered a THN kit. Methods: Data on demographics, clinical characteristics, and THN kit dispensing were extracted and analyzed from the charts of all ED patients presenting with a primary diagnosis of opioid poisoning between April 1 2016 and April 30 2017. Logistic regression analyzed predictors of being offered a THN kit. Results: A total of 347 ED visits for 301 unique patients occurred during the study period. The mean age ± SD of patients was 38 ± 14 years, and 69% were male. In 49% of ED visits, a THN kit was offered; 73% of these episodes had a THN kit dispensation. Patients who were male (AOR=1.94; 95% CI 1.11 - 3.40), and reported that their overdose was unintentional (AOR=2.95; 95% CI 1.04 8.35) and caused by illegal opioids (AOR=4.73; 95% CI 2.63 8.52) were significantly more likely to be offered a THN kit. Conclusion: ED-based THN programs have the potential to reach significant proportions of patients at high risk of mortality. However, these programs may have differential reach within the target population. Further research is needed to examine barriers and facilitators to offering all eligible ED patients a THN kit.

Introduction: Optimal discharge communication between healthcare providers and parents who present to the emergency department (ED) with their children is not well understood. Current research regarding discharge communication is equivocal and predominantly focused on evaluating different delivery formats or strategies with little attention given to communication behaviours or the context in which the communication occurs. The objective of this study was to characterize the process and structure of discharge communication in a pediatric ED context. Methods: Real-time video observation and follow-up surveys were used in two academic pediatric EDs in Canada. Parents who presented with their child to the ED with one of six illness presentations, a Canadian Triage Acuity Score of 3-5 were eligible to participate. All ED physicians, learners, and staff members were also eligible. Provider-parent communication was analyzed using the Roter Interaction Analysis System (RIAS) to code each utterance. Parent health literacy and anxiety were measured upon admission to the ED. Parent recall of important discharge information and satisfaction with communication was assessed within 72 hours of discharge. Results: A total of 107 ED patient visits were video recorded and a total of 70,000 utterances were coded across six illness presentations: abdominal pain (n=23), asthma (n=7), bronchiolitis (n=4), diarrhea/vomiting (n=20), fever (n=27), and minor head injury (n=26). The average length of stay for participants was 3 hours, with an average of three provider interactions per visit. Interactions ranged in time from less than one minute up to 29 minutes, with an average of six minutes per interaction. The majority of visits were first episodes for the presenting illness (63.2%). Physician utterances coded most commonly involved giving medical information (22.9%), whereas nurses most commonly gave orientation instructions (20.9%). Learners were most likely to employ active listening techniques (14.2%). Communication that provided post-discharge instructions for parents comprised 8.5% of all utterances. Overall, providers infrequently assessed parental understanding of information (2.0%). Only 26% of parents recalled receiving important discharge information deemed relevant to their childs disposition. Yet, parent satisfaction with the amount of information communicated during the ED visit was generally high (89.6% agreed or strongly agreed). Conclusion: This is the first study of ED discharge communication to be conducted in a pediatric setting using video observation methods. Provider-parent communication was predominantly characterized by giving medical information, with little time devoted to preparing families to care for their child at home. Greater assessment of parent comprehension of discharge communication is needed to ensure that parents understand important instructions and know when to seek further care.

Introduction: NSAIDS offer more effective analgesia than opioids, require less rescue medication, and decrease the incidence of nausea and vomiting in renal colic patients. Alpha blockers and Opioids are also prescribed frequently, but doses used and treatment durations are not well described. Our objective was to investigate ED prescribing decisions and medication compliance by patients with acute renal colic. Methods: In this prospective two-city cohort study, we invited patients with a first ED visit for image-confirmed 2-10 mm ureteric stones to consent to a telephone survey 10 days after their ED visit. During follow-up interviews, patients were asked what drugs they were prescribed and how many doses they required. This study was REB approved. Results: A convenience sample of 224 patients, including 152 males (67.9%) and 72 females (median age= 52.4 years) completed 10-day surveys. NSAIDS were prescribed for 48.7%, tamsulosin for 65.2% and opioids for 81.7%. One-third received a tamsulosin-NSAID combination, 40% an opioid-NSAID combination and 28% a tamsulosin-NSAID-opioid combination. Of 109 patients prescribed an NSAID, only 70 (64.2%) took 1 dose/day; however an additional 28 who were not prescribed NSAIDs took 1 NSAID dose/day. Mean (sd) NSAID intake in the overall study group was 1.1 (1.5) doses/day from day 1-5 and 0.6 (1.1) doses/day on days 6-10, with 90%ile values of 3.0 and 2.0 doses/day. NSAID compliance was more common in patients who stated they received high quality discharge instructions (63.8% vs. 32.6%; RR=1.95; 95% CI 1.47-2.60). Mean opioid intake in the overall study group was 1.2 (1.7) doses/day from day 1-5 and 0.5 (1.3) doses/day on days 6-10, with 90%ile values of 4.0 and 2.0 doses/day. Among patients prescribed tamsulosin, the average was 4.0 days of compliance (sd=4.3), with a 90%ile value of 10 days. Conclusion: This study provides estimates for the amount of drug actually used by renal colic patients during the 10-days after their ED visit. Patients used fewer opioid doses than expected, and NSAID and tamsulosin compliance appears relatively poor. NSAID compliance was better in patients who perceived high quality discharge instructions. This study suggests there is room for improvement in medication prescribing and discharge instructions for ED patients with an acute episode of ureteral colic.

Introduction: The accuracy of ultrasound (US) for diagnosing lower extremity deep vein thrombosis (DVT) in non-pregnant patients has been well validated. However, in pregnant women with suspected DVT and an initial negative US (with imaging of the iliac veins), serial US is recommended. We aimed to determine the ability of single negative US to exclude DVT in symptomatic pregnant women. Methods: Two authors independently reviewed the following databases: MEDLINE, PubMed and EMBase from inception until May 2017. Three authors reviewed all full text papers and data were extracted from included studies by four authors. An overlap among study populations was identified in 4 of the manuscripts, all from one multicentre Canadian study. Two authors performed data re-extraction from the hard copy research charts from this study. We assessed the risk of bias using the CLARITY group tool for prognostic studies. Results: Of 109 potentially relevant articles, 8 studies (7 prospective studies and 1 retrospective) were included. Risk of bias was low for the included populations, and low or moderate for method of measurement and for completeness of follow up. A total of 635 pregnant patients with symptoms of DVT had an initial negative US examination. Of those, 6 had positive DVT during serial US (0.94%) and 3 developed DVT during 3-month follow-up after serial ultrasound (0.47%). Using random-effects model, the pooled false negative rate of a single ultrasound was 1.27% (95% confidence interval, 0.42 to 2.56), I2= 27%. Conclusion: The false negative rate of a single ultrasound with iliac vein imaging for DVT in pregnancy is low. Our results will help inform shared decision making around planning repeat ultrasound scans in these patients.

Introduction: Undertreated pain is known to cause short and long-term harm in children. Limb injuries are a common painful condition in emergency department (ED) patients, accounting for 12% of ED visits by children. Our city has one pediatric ED in a freestanding children’s hospital and 3 general ED’s that treat both adults and children. 68% of pediatric limb injuries in our city are treated in the pediatric ED and 32% are treated in a general ED. A quality improvement (QI) initiative was developed at the children’s hospital ED in April 2015 focusing on “Commitment to Comfort.” After achieving aims at the childrens hospital, a QI collaborative was formed among the pediatric ED and the 3 general ED’s to 1) improve the proportion of children citywide receiving analgesia for limb injuries from 27% to 40% and 2) reduce the median time to analgesia from 37 minutes to 15 minutes, during the time period of April-September, 2016. Methods: Data were obtained from computerized order entry records for children 0-17.99 years visiting any participating ED with a chief complaint of limb injury. Project teams from each site met monthly to discuss aims, develop key driver diagrams, plan tests of change, and share learnings. Implementation strategies were based on the Model for Improvement with PDSA cycles. Patient and family consultation was obtained. Process measures included the proportion of children treated with analgesic medication and time to analgesia; balancing measures were duration of triage and length of stay for limb injury and all patients. Site-specific run charts were used to detect special cause variation. Data from all sites were combined at study end to measure city-wide impact using 2 and interrupted time series analysis. Results: During the 3.5-year time period studied (April 1, 2014-September 30, 2017), there were 45,567 visits to the participating ED’s by children 0-17.99 years with limb injury. All visits were included in analysis. Special cause was detected in run charts of all process measures. Interrupted time series analysis comparing the year prior to implementation at the childrens hospital in April 2015 to the year following completion of implementation at the 3 general hospitals in October 2016 demonstrated that the proportion of patients with limb injury receiving analgesia increased from 27% to 40% (p<0.01), and the median time from arrival to analgesia decreased from 37 to 11 minutes (p<0.01). Balancing measure analysis is in progress. Conclusion: This multisite initiative emphasizing “Commitment to Comfort” was successful in improving pain outcomes for all children with limb injuries seen in city-wide ED’s, and was sustained for one year following implementation. A QI collaborative can be an effective method for spreading improvement. The project team is now spreading the Commitment to Comfort initiative to over 30 rural and regional EDs throughout the province through establishment of a provincial QI collaborative.

Introduction: Treat and Release (T&R) patients are seen and discharged home from the emergency department (ED), and asked to return within 12-72 hours for follow-up care (e.g., ultrasound, repeat blood work). Our two academic teaching hospitals see approximately 2,000 T&R patients per year. Handover of care for T&R patientsdone through charting only and therefore dependent on the charts adequacy and completenessis crucial to the safety and quality of care they receive. An 18-month retrospective chart audit at our sites identified quality gaps, including suboptimal documentation that ultimately impedes patient disposition. Our projects aim was to reduce the time-to-disposition (TTD; time spent by patients between provider initial assessment and discharge from the ED) by a third (from 70min) in 6-months time (March 2017), a target felt to be both meaningful and realistic by our stakeholder team. Methods: Our primary outcome measure was the TTD (in minutes). Our process measure was the quality of documentation, using a modified version of QNOTE, a validated tool used to assess the quality of health-care documentation. PDSA cycles included: 1) Involvement of stakeholders for the creation and refinement of an improved T&R handover tool to cue more specific documentation; 2) Education of health-care providers (HCPs) about T&R patients; 3) Replacement of the previous T&R handover tool with a newly designed and mandatory tool (i.e. a forcing function); 4) Refinement of the process for T&R patients and chart hold-over. Results: Run charts for both the median TTD and median modified QNOTE scores over time demonstrate a shift (i.e., run chart rule) associated with the second and third clustered PDSA cycles. After the first three clusters of PDSA cycles (i.e., before-and-after), mean TTD was reduced by 40% (70min to 42min, p=0.005). The quality of documentation (mean modified QNOTE scores) was also significantly improved (all results p<0.0001): patient assessment from 81% to 92%, plan of care from 58% to 85% and follow-up plan from 67% to 90%. Conclusion: We reduced the time-to-disposition for T&R patients by identifying gaps in the quality of documentation of their chart. Using iterative PDSA cycles, we improved their time-to-disposition through improved communication between health-care providers and a new T&R handover tool working as a forcing function. Other centers could use similar assessment methods and interventions to improve the care of T&R patients.

Introduction: Over 1 million patients with head injuries (HIs) are seen every year at emergency departments (EDs) in North America, with over 90% being minor HIs. Over-utilization of computed tomography (CT) scans in these patients results in unnecessary exposure to radiation and increases health-care resource utilization. Using recommendations from the Choosing Wisely Campaign (CWC) and quality improvement (QI) methodology, we developed a local initiative targeting this issue. Our aim was to reduce the CT scan rate for patients presenting with HIs by 10% over a 6-month period at two academic EDs. This was considered both achievable and meaningful by our stakeholders. Methods: Baseline CT scan rates for patients with HIs were determined through a 10-month retrospective cohort review. We used stakeholder engagement and provider surveys to develop our driver diagram and PDSA cycles, which included: 1) Assessing and improving provider knowledge about the CWC recommendations; 2) Testing, refining and implementing a modified Canadian CT Head Rule checklist in the ED; 3) Developing and giving patients CWC-themed handouts pertaining to HI best practice; 4) Bimonthly reporting of CT scan rates to providers. Our primary outcome measure was the number of CT scans performed for patients with HIs. Process measures included the number of checklists completed and ED length of stay (LOS). Our balance measure was return ED visits within 72 hours. Results: Baseline rate of CT scans prior to our interventions was 47.9%. Our QI initiative resulted in a significant shift in the run chart of the weekly CT scan rates, associated with the second PDSA cycle cluster. We observed a 16% relative decrease in CT scans at 3 months (47.9% to 40.5%, P=0.005) and 10.4% at 8 months (47.9% to 43.1%, P=0.02). Non-sustained trends and shifts were seen in the run chart of median ED LOS for HI patients, but overall before-and-after median times were not significantly different (237min to 225min, P=0.18). 33% of total checklists were completed. 72-hr return visits did not change during the 8-month study period (4.0% to 4.16%, P=0.85). Conclusion: Our local QI initiative was successful in decreasing CT rates for patients presenting with a HI. The decrease in effect at 8 months suggests the need for continued feedback and reminders to ensure long-term sustainability. Other centres could use similar QI methods, as well as the materials we developed, to achieve similar results of improved evidence-based utilization of diagnostic tests.

Introduction: There is a high risk for communication breakdown, discontinuity of clinical care, and medical errors during ED physician handover. Locally, there is no standardized handover process to ensure adequate communication of critical information. Our aim was to use a locally developed handover tool to increase frequency of adequate physician handover during overnight shift change by 50% in 4 months. Methods: Using published best practices, local observational data, and stakeholder input, we determined critical components of ED handovers. We developed a structured communication tool for two unique populations in our ED: ED-VITAL for patients receiving active ED care; ED-VSA for patients who are admitted/referred. Strategies used to implement the tool included: engagement of staff physicians to introduce & modify the tool; formal education and training to ED residents; and provision of cognitive aids. A QI coordinator conducted direct observations of handovers using convenience sampling. We provided feedback to staff and resident physicians, and used their input to continuously modify the tool. The main outcome measure was adequate patient handover, defined as verbal communication of 50% of critical handover components, or documentation of key information on an electronic note. Process measures included tool utilization characteristics. Balance measures included time metrics such as handover duration. We present run charts and qualitative statistics. Results: We assessed 368 individual patient handovers (93 pre- & 275 post-implementation). The median proportion of patients in active ED care who were verbally handed over increased from 75% to 100%. The median proportion of adequate handovers improved from 50% to 72%. The time to deliver handover increased by 13 seconds per patient. Qualitative feedback from end users was positive overall, particularly for communication quality and resident educational value. Conclusion: Use of a standardized handover tool improved both verbal and documented communication during shift change. A customized approach, sensitive to local context, was important to successful implementation. Residents play a large role in handovers; strategies to improve handover processes that emphasize medical education appear to enhance success. Future PDSA cycles will focus on interventions to further enhance the utilization of the tool, and to measure direct impact on clinical outcomes.

Introduction: CT scan is the most common imaging modality for suspected renal colic and is used for about 80% of presentations. Cumulative ionizing radiation exposure from repeat CT scans increases long-term cancer risk. Despite a 10-fold increase in CT use to detect kidney stones in the ED in just over a decade, there has been no increase in the proportion of kidney stones diagnosed, number of significant alternate diagnoses or admissions to hospital. Choosing Wisely recommends to avoid ordering CT of the abdomen/pelvis in otherwise healthy patients<age 50 presenting to the ED with known history of kidney stones and with symptoms consistent with uncomplicated renal colic. The aim is that >90% of patients < age 50 with a history of renal stones arriving in Sunnybrook ED with symptoms consistent with renal colic will be managed without a CT abdomen/pelvis. Methods: Emergency physicians were engaged in the process at various stages, including a brainstorming session to perform a root cause analysis. A Driver diagram was created to generate change ideas. Outcome Measure Number of CT scans ordered for target population (Results: Results to date indicate that there is a non-sustained decrease in the number of CT scans performed on ED patients < age 50 with recurrent renal colic. The STONE score was infrequently used, thus making it difficult to standardize CT ordering for presumed renal colic. Conclusion: As a result of this QI initiative, there is awareness amongst emergency physicians of a patient population that is over imaged with CT scan, often with no change in management. Introduction of a low dose CT scan order was the greatest gain from this QI initiative. In order to decrease CT utilization, physicians need to be shown the lack of benefit of CT use and a safe alternative diagnostic approach.

Introduction: With the current opioid crisis in Canada, presentations of acute opioid withdrawal (AOW) to emergency departments (ED) are increasing. Undertreated symptoms may result in relapse, overdose and death. Buprenorphine/naloxone (bup/nal) is a partial opioid agonist/antagonist used to mitigate symptoms of AOW, approved by Health Canad in 2007 for opioid use disorder. It is superior to clonidine, and increases follow up with addiction treatment programs when initiated in the ED. Nevertheless, in our inner-city ED in 2014, bup/nal was rarely prescribed. We aimed to increase ED physician prescribing of bup/nal for AOW by 50% over a 26-month period. Methods: Commencing in 2014, an interprofessional team of ED physicians, nurses (RN), pharmacists and QI specialists collaborated to improve the care of patients with AOW. PDSA cycles included: (1) needs assessment of emergency physicians knowledge and practices in 2014; (2) Grand Rounds and a web based information sheet in 2015; (3) ED stocking of bup/nal; (4) convenience order set to standardize AOW management; (5) Grand Rounds in 2016 and (6) peer-coaching for RNs, including case-based discussions and pocket card cognitive aids. The outcome was the number of times bup/nal was prescribed per month by ED physicians between Sept, 2015 and Oct, 2017. Data included the prescriber and use of order set as the process measure. The balancing measure was the number of patients referred to the Addiction Medicine Team who subsequently received bup/nal. Results: Bup/nal was prescribed by ED physicians 70 times, and 14 times by the Addiction Medicine Team. With each PDSA cycle, there was an increase in prescribing, with no significant shifts or trends. By all physicians, the median number of prescriptions per month was 3, and increased from 2 to 4 prescriptions/month after nursing education. There was a smaller increase in the median from 2 to 3 prescriptions/month by ED physicians alone. The order set was used 97% of the time. Conclusion: Bup/nal is safe, effective, and increases follow up with addiction programs for comprehensive assessment and treatment planning. We met our goal of increasing bup/nal prescribing in the ED for AOW by 50%. Moreover, prescribing increased by 100% with the addition of patients who received bup/nal after a referral to the Addiction Medicine Team. The intervention with the greatest impact was RN education, demonstrating that peer-coaching and teaching by an interprofessional team is key to changing practice. Unfortunately, overall prescribing remains low, and ED physicians may still be hesitant to prescribe bup/nal and defer to the specialists. It is unclear if this is due to a low number of patients presenting with AOW, patients with contraindications to bup/nal, or ED physician factors. The next step is an audit of all patients with AOW to see what percentage of those eligible are treated with bup/nal. A follow up survey to determine ongoing barriers will inform further PDSA cycles.

Introduction: Le syndrome aortique aigu (SAA) est une condition potentiellement létale et nécessitant une prise en charge immédiate et fréquemment chirurgicale. Lutilisation dun test de D-dimère a été proposé afin dexclure ce diagnostic suite à de nombreuses études récentes sur le sujet. Dans ce contexte, lobjectif de la présente revue systématique et méta-analyse était dévaluer et de synthétiser lévidence disponible quant à la valeur diagnostique dun dosage de D-dimères pour le SAA. Methods: Les bases de données Medline, Cochrane, ACP Journal Club, DARE, Health technology assessment, NHS Economic evaluation et Embase ont été fouillées en utilisant les mots clés «D-Dimers» et «Acute aortic». Les bibliographies des articles retenus ont également été consultées. Les résultats des études incluses ont été regroupés pour calculer la sensibilité, la spécificité, le rapport de vraisemblance positif (RV+) et négatif (LR-) du test de D-dimères aux seuils choisis par les auteurs. Les sensibilités et spécificités sont présentées avec leurs intervalles de confiance (IC) à 95% et les RV+ et RV- sont donnés à titre indicatif. Results: À partir des 6942 articles initialement identifiés, 34 études portant sur lutilité diagnostique des D-dimères en SAA, incluant un total de 7938 patients, ont été retenues. La prévalence globale de SAA était de 27,1%, mais variait considérablement dune étude à lautre (médiane=38,5%; intervalle interquartile=29% ; prévalence minimale=0.7% ; prévalence maximale=100%). Les seuils de positivité des tests de D-dimères variaient également selon lentreprise les fabriquant. La sensibilité dun test de D-dimères selon le seuil proposé par leur fabriquant était de 91% (IC 95% 86-96) et la spécificité 46% (IC 95% 39-54). Le RV+ était de 1,69 et le VR- était de 0,2. Conclusion: La prévalence très variable du SAA dans la population regroupée par ces études laisse présager une différence entre ces populations et celles rencontrées en médecine durgence. Un test de D-dimères a sensibilité acceptable, mais imparfaite, qui ne semble pas permettre dexclure une pathologie aussi grave que le SAA.

Introduction: Les outils de prédiction disponibles (score clinique, ratio des neutrophiles sur lymphocytes et dosage des D-dimères) dans le diagnostic du syndrome aortique aigu (SAA) demeurent imparfaits. Dans ce contexte, avec lobjectif de développer une règle de décision clinique pour le diagnostic du SAA, la présente étude visait à définir un ensemble de variables discriminantes chez les patients souffrant ou non dun SAA en utilisant les outils dintelligence artificielle. Methods: À partir de lensemble des données cliniques disponibles chez les patients investigués pour douleur thoracique au département durgence avec une angiographie par tomodensitométrie (angioCT) visant à éliminer un SAA entre 2008 et 2014, un programme dapprentissage a été chargé de construire un arbre de décision (Clustering And Regression Tree) identifiant les patients ne souffrant pas dun SAA. La variable dintérêt était labsence de SAA et 23 attributs ont été testés. Le diagnostic de SAA était établi avec les résultats de langioCT. Des échantillons aléatoires de 70% de la population étudiée ont été testés de façon récursive (maximum de 100 itérations) pour construire larbre de décision. Six algorithmes dapprentissage (Reg Tree, LR, KNN, Naive B, Random Forest et CN2) ont été comparés et loptimisation du gain dinformation a été mesurée par les techniques de Gain Ratio et de Gini. Results: Un total de 198 patients (99 hommes et 99 femmes) dun âge moyen de 63 ans (±16) ont été inclus dans létude, parmi lesquels 26 (13%) souffraient dun SAA. Trois attributs ou regroupements dattributs ont permis de construire un arbre de décision permettant didentifier 114 patients sur 198 (57,6%) ayant une très faible probabilité de SAA (sensibilité visée de 100%). La sensibilité et spécificité de larbre de décision clinique était respectivement de 100% (intervalle de confiance [IC] 95% 86,7-100,0) et 70,4% (IC 95% 62,7-77,3). Les attributs en question étaient labsence de tout facteur de risque (e.g. syndrome de Marfan, chirurgie aortique ou valvulaire, histoire familiale), les signes vitaux (tension artérielle systolique, pouls et choc index) et les D-dimères. Le seuil de D-dimères utilisé pouvait varier entre 1114 et 1211 mcg/L selon lhémodynamie et la présence de facteur de risque. Les attributs suivants nétaient pas discriminants : le sexe, un antécédent de diabète, dhypertension artérielle ou de dyslipidémie, le tabagisme, avoir un déficit de perfusion, une différence de tension artérielle entre les deux bras ou un souffle dinsuffisance aortique et la formule sanguine. Conclusion: Les attributs les plus discriminants pour le SAA sont les facteurs à risque, lhémodynamie et les D-dimères. Une étude prospective multicentrique devrait être réalisé afin de développer une règle de décision clinique afin didentifier les patients à très bas risque de SAA à partir de ces attributs.

Introduction: The prevalence of pulmonary embolism (PE) among patients with syncope is understudied. Based on a recent study with an exceptionally high PE prevalence, some advocate investigating all syncope patients for PE, including those with another clear cause for their syncope. We sought to evaluate the PE prevalence among emergency department (ED) patients with syncope. Methods: We combined data from two large prospective studies enrolling adults with syncope from 17 EDs in Canada and the United States. Each study collected the results of investigations related to PE (i.e. D-dimer or ventilation-perfusion (VQ) scan, or computed tomography pulmonary angiogram (CTPA)), and 30-day adjudicated outcomes including diagnosis of PE, arrhythmia, myocardial infarction, serious hemorrhage and/or death. Results: Of the 9,091 patients (median age 66 years, 51.9% females, 39.1% hospitalized) with 30-day follow-up, 546 (6.0%) were investigated for PE: 278 (3.1%) had D-dimer, 39 (0.4%) had VQ and 347 (3.8%) patients had CTPA performed. 30-day outcomes included: 874 (9.6%) patients with any serious outcome; 0.9% deaths; and 818 (9.0%) patients with non-PE serious outcomes. Overall, 56 patients (prevalence 0.6%; 95% CI 0.5% 0.8%) were diagnosed with PE, including 8 (0.1%) of those admitted to hospital at the index presentation. Only 11 patients (0.1%) with a non-PE serious condition had a concomitant underlying PE identified. Conclusion: The prevalence of PE is very low among ED patients with syncope, including those hospitalized following syncope. While acknowledging syncope may be caused by an underlying PE, clinicians should be cautious against indiscriminate over-investigations for PE.

Introduction: Le syndrome aortique aigu (SAA) comprend les dissections aortiques, les hématomes intramuraux et les ulcères de lintima, trois conditions difficiles à diagnostiquer, potentiellement létales et nécessitant une prise en charge immédiate et fréquemment chirurgicale. Lutilisation dun test de D-dimère a été proposé afin dexclure ces diagnostics et éviter une investigation plus poussée par angiographie par tomodensitométrie (angioCT). Cependant, il est peu plausible que les patients souffrant dhématomes intramuraux aient une valeur de D-dimères très élevée. Dans ce contexte, lobjectif primaire de la présente étude est de déterminer la valeur diagnostique (sensibilité et spécificité et rapport de vraisemblance négatif [RV-]) dun test de D-dimères chez les patients suspectés de SAA au département durgence. Methods: Les patients ayant subi une angiographie par tomodensitométrie (angioCT) à la recherche dune dissection aortique entre 2008 et 2014 à lurgence dun hôpital tertiaire montréalais ont été inclus dans cette étude de cohorte rétrospective. Les patients nayant pas eu de dosage de D-dimères en ont par la suite été exclus. La valeur diagnostique dun test de D-dimères de plus de 500 mcg/L a été comparée à celle du test de référence (angioCT) afin de calculer la sensibilité, la spécificité et le rapport de vraisemblance négatif et leurs intervalles de confiance (IC). Results: Un total de 139 patients ont été inclus dans létude, parmi lesquels 12 (8,6%) souffraient dun SAA. La sensibilité dun test de D-dimères avec un seuil de positivité de 500 mcg/L était de 83,3% (IC 95% 51,6-97,9), la spécificité de 52,8% (IC 95% 47,8-66,4) et le VR- de 0,32 (IC 95% 0,09-1,13). Les deux patients pour qui le résultat du test de D-dimère était un faux négatif souffraient dun hématome intramural. Les sept patients avec un D-Dimères de plus de 4000 mcg/L semblaient souffrir dun diagnostic grave (dissection aortique : n=5, liquide libre intra-abdominal avec état de choc : n=1 et tamponnade cardiaque : n=1). Conclusion: Avoir un test de D-dimères inférieur à 500 mcg/L ne permet pas déliminer un SAA, particulièrement un hématome intramural.

Introduction: Acute aortic dissection (AAD) is a rare but fatal condition where over-investigation and missed diagnosis are common. Our objectives were to derive a highly sensitive clinical risk score for AAD and perform pilot validation. Methods: We started with two independent systemic reviews to firstly identify clinical variables associated with AAD and secondly to determine reasons for missed diagnosis. We searched Medline, Embase and the Cochrane database (1968-July 2016). Two reviewers screened articles and extracted data. Agreement was measured by Kappa and study quality by the QUADAS-2 tool. Bivariate random-effects meta-analyses (Revman 5 and SAS 9.3) were performed. Due to sampling bias found in the systematic reviews a matched case control study confirming the strength and direction of predictor variables was performed. The cases (2002-2014) included new emergency department (ED) or in-hospital diagnosis of non-traumatic AAD confirmed by computed tomography (CT). The controls (2010-2011) were a random age/sex matched sample of patients triaged with undifferentiated acute truncal pain (< 14 days). Finally, we used the beta coefficients derived from multivariate logistic regression of our case control study to assign a numerical strength of association to predictor variables. To mitigate the bias inherent in case control studies we adjusted the beta coefficient for each variable by the diagnostic odds ratio calculated from each systematic review. Pilot validation was performed on a retrospective sample of all those undergoing CTA to rule out AAD at two tertiary care ED over 12 months. Two abstractors were blinded to the final diagnosis. Results: We derived a two-step risk score based on the derivation sample which included 4960 patients(Clinical variables systematic review -9 studies, N=2400, low risk of bias, Kappa 0.9 & Reasons for missed diagnosis systematic review - 11 studies, N=800, low-moderate risk of bias, Kappa 0.89 & Case control study -194 AAD, 776 Controls). Step one is a RAPID assessment for AAD 1) Risk factors 2) Alternative diagnosis in the differential that mimics AAD- ACS, PE, Stroke 3) Physical exam- hypotension, pulse deficit 4) Impression- clinical suspicion of AAD and 5) Discomfort- migrating, tearing, pleuritic, thunderclap, severe pain. If any of the above factors are present proceed to step two. Step two stratifies patients based on history (low, moderate, high suspicion), physical exam (hypotension/pulse deficit) and risk factors. In the pilot validation (N=375,AAD=16) sensitivity was 100% (95%CI 79.4-100) and specificity 36.5% (95%CI 31.5-41.7%). Patients were successfully stratified into low (<2, 0% AAD), moderate (2, 2.2% AAD), high (>2, 19.6% AAD) and critical probability (>3, 62.5% AAD), with up to 36% reduction in imaging. Conclusion: We derived a highly sensitive new clinical risk score with the potential to reduce missed cases of AAD, reduce unnecessary imaging and expedite care.

Introduction: Electrocardiographic (ECG) rhythms are used during resuscitation (ACLS) to guide resuscitation, and often to determine futility. Survival rates to hospital discharge have been reported to be higher for patients with PEA than asystole in out-of-hospital cardiac arrest. This study examines how well the initial ECG cardiac rhythm represents actual cardiac activity as determined by point of care ultrasound (PoCUS). Methods: A database review was completed for patients arriving to a tertiary ED in asystole or PEA arrest, from 2010 to 2014. Patients under 19y or with a previous DNR were excluded. Patients were grouped into those with cardiac activity (PEA) and asystole on ECG; as well as whether cardiac activity was seen on PoCUS during the arrest. Data was analyzed for visualized cardiac activity on PoCUS. Results: 186 patients met the study criteria. Those with asystole on ECG were more likely to have no cardiac activity than those with PEA (Odds 7.21 for initial PoCUS; 5.45 for any PoCUS). The sensitivity of ECG rhythm was 80.49% and 82.12%, specificity was 77.91% and 54.28%, positive predictive value was 94.28% and 88.57%, and negative predictive value was 30.43% and 41.30% for cardiac activity on initial PoCUS and on any PoCUS respectively. The positive and negative likelihood ratios for ECG were 3.47 and 0.25 for activity on initial PoCUS. The positive and negative likelihood ratios for activity on any PoCUS were 1.78 and 0.33. Conclusion: Our results suggest that although most patients with asystole on ECG demonstrate no cardiac activity, a small number actually had activity on PoCUS. This supports the use of PoCUS during cardiac arrest, in addition to ECG, to identify patients with ongoing mechanical cardiac activity.

Introduction: One of the most high-risk tasks regularly performed by emergency medicine (EM) physicians is airway management. Many studies identify an increase in adverse events associated with airway management outside of the operating theatre. Errors of omission are the single most common human error type. To address this risk, the checklist is becoming a common pre-intubation tool. Simulation is a safe setting in which to study the implementation of a new airway checklist. The purpose of this study was to determine if a novel airway checklist decreases practitioners rates of omission of important tasks during simulated resuscitation scenarios. Methods: This was a dual-centre, randomized controlled trial of a novel airway checklist utilized by EM practitioners in a simulated environment. The 29-item peri-intubation checklist was derived by experienced EM practitioners following a review of airway checklists in published and gray literature. Participants were EM residents or EM physicians who work more than 20 hours/month in an emergency department. Volunteers were recruited from two academic health centres to complete three simulated scenarios (two requiring intubation, one cricothyroidotomy), and were randomized to either regular care or checklist use. A minimum of two assessors documented the number of omitted tasks deemed important in airway management and the time until definitive airway management. Discrepancies between assessors were resolved by single-assessor video review. Results: Fifty-four EM practitioners participated. There was no significant difference in baseline characteristics between the two study groups. The average percentage of omitted tasks over the three scenarios was 45.7% in the control group (n=25) and 13.5% in the checklist group (n=29) an absolute difference of 32.2% (95% CI: 27.8%, 36.6%). Time to intubation (normally distributed) was significantly longer in the checklist group for the first two scenarios (mean difference 114.10s, 95% CI: 48.21s, 179.98s and 76.34s, 95% CI:31.35s ,121.33s), but there was no statistical difference in the third scenario where cricothyroidotomy was required (mean difference 33.75s, 95% CI: -28.14s, 95.65s). Conclusion: In a simulated setting, use of an airway checklist significantly decreased the omission rate of important airway management tasks, however it increased the time to definitive airway management. Further study is required to determine if these findings are consistent in a clinical setting and how they impact the rate of adverse events.