Research Day 2026

Megan Harmon
Institute: University of Calgary
Department: Emergency Medicine
Title: Understanding Frequent Use of the Emergency Department: A Cluster Analysis
Authors: Jessalyn Holodinsky, Megan Harmon, Christine East, Kenneth Mont, Catherine Patocka, Tyler Williamson, Andrew McRae, Patrick McLane, Cody Dunne, Ni Lam, Shallen Moore, Domhnall O’Dochartaigh, Rhys Clark, Ryan Lee, Karen Moffat

Introduction: Frequent use of the emergency department (ED) is problematic on a patient, provider, and health system level. Although those who rely on ED frequently are a small proportion of patients, they account for a substantial number of ED visits. Past approaches to reducing frequent ED use have treated these individuals as homogeneous or characterized them using simple demographics, obscuring distinct needs of different patient groups. This project aimed to define homogeneous groups who use the ED frequently.

Methods: Using administrative data from Alberta (2022-24), we identified patients ≥16 years with ≥3 ED visits over 12-months. Using k-means clustering, we segmented this population using visit patterns (number of visits, visit regularity, time period of frequent use, hospital admission rate, ambulance use rate) and basic demographics as input variables. Clusters were discussed with a group of expert clinicians, patient partners, and health services researchers.

Results: 306,728 patients (1.8M visits) were included. Ten clusters (size 147 to 97,824 patients) were identified. Median visits/patient ranged from 4-141 across clusters. Three high-utilizing clusters (18-141 visits/patient) had elevated rates of substance use (43-77%), mental health conditions (71-82%), and housing instability (32-100%). The other 7 clusters showed more moderate use (3-14 visits/patient). One was characterized by repeated use in a short time frame (4-8 days). Two were comprised of older patients (median 73 and 69 years) with high burden of medical and mental health comorbidities, respectively. Both often presented by ambulance and were admitted to hospital, but this was more common in the medically complex group. One cluster was uniquely characterized by rural geography. Three clusters were sex specific; two all-female clusters which differed by age (median age 66 vs 32 years) and one all-male cluster. Variation for reason of visit was also noted which is to be explored further.

Conclusion: Our findings demonstrate that those who rely frequently on the ED represent a heterogeneous population with distinct subgroups characterized by differing visit patterns, comorbidities, and social circumstances. Future work includes qualitative exploration of each group to understand reasons for reliance on the ED. This will inform development and evaluation of cluster-specific interventions in partnership with clinical teams, patient partners, and individuals with lived experience.

Dr. Sierra Casey
Institute: University of Calgary
Department: Pediatric Emergency Medicine
Title: Mycoplasma pneumonia: A comparison of pediatric practice guidelines
Authors: Sierra Casey, Jade Chow, Stephen Freedman 
This abstract was submitted as a Work in Progress

Introduction: Mycoplasma pneumoniae (M. pneumoniae) is a common respiratory pathogen and a leading cause of community-acquired pneumonia in children and adolescents. Patients with M. pneumoniae classically experience non-severe and self-limiting lower respiratory tract infection.  In 2024, a worldwide resurgence of M. pneumoniae occurred. This resurgence led to front-line care providers to seek guidance on the testing and treatment of children with M. pneumoniae infection. Despite the need for evidence to inform practice, such evidence is lacking. For example, the optimal approach to treating M. pneumoniae infection is unclear. M. pneumoniae is inherently resistant to antibiotics traditionally used to treat community acquired pneumoniae. Macrolide antibiotics have predictable activity against M. pneumoniae, but there are increasing reports of resistance in Asia and in North America. A Cochrane review published in 2015 found that administration of targeted antibiotics to treat M. pneumoniae in hospitalized pediatric patients with M. pneumoniae infection did not alter their clinical course.  Given that inappropriate antibiotic use increases antimicrobial resistance and has potential for long-term side effects in pediatric patients21 it is essential to appropriately diagnose and treat M. pneumoniae infection. During the worldwide resurgence of M. pneumoniae infection, clinicians often sought guidance from clinical practice guidelines. This review seeks to determine if guidelines for the diagnosis and treatment of M. pneumoniae have been published by North American academic organizations (e.g., Infectious Diseases Society of America, Canadian Pediatric Society, etc.) and to evaluate the consensus in the recommendations provided.  We will also undertake a similar review of local institutional guidelines across Canada’s 15 pediatric tertiary care hospitals.

Methods: To identify published guidelines, we will follow the Joanna Briggs Institute Research Methodology for conducting scoping reviews. A comprehensive search of Ovid EMBASE and Ovid MEDLINE will be conducted with assistance from research librarians at the University of Calgary. The search will comprise of index terms “clinical practice guidelines”, “pneumonia”, and “pediatrics”. “Pneumonia” will be used as the index term because guidelines for testing and treatment of M. pneumoniae in pediatrics are often embedded into guidelines for community acquired pneumonia in pediatrics. To identify local institutional guidelines, we will contact the 15 Canadian pediatric tertiary care institutions. Inclusion and exclusion criteria will be specified prior to beginning the search.

Title and abstract screening will be completed by two independent reviewers to ensure that articles meet inclusion and exclusion criteria. Full text screening will then be completed by the same two reviewers. A data extraction form will be developed in a COVIDENCE database and data will be extracted in duplicate by two independent reviewers. Data to collect will include recommended indications for testing for M. pneumoniae and recommended diagnostic tests (serology or nucleic acid testing, location of swabs). We will also collect recommendations for treatment (no antibiotic/watchful waiting as first line, indications to start antibiotics, and first/second-line antibiotics). The appraisal of guidelines for research and evaluation (AGREE) II instrument will also be used to assess quality of all guidelines. 

Results: This review will follow reporting guidelines from the Joanna Brigg’s Institute guide for scoping reviews. The process of article selection for the scoping review will be presented in a PRISMA flow chart. Data analysis will be mainly descriptive. The frequency count of each specific diagnostic and treatment recommendation will be presented in data tables. Aggregate scores of the AGREE II instrument for each guideline will be presented in a table. 

Conclusions: We expect this review to identify key recommendations where there is consensus, identify knowledge gaps and areas of controversy and inform future research initiatives for the diagnosis and treatment of M. pneumoniae infection in children and adolescents.

Janna Newton
Institute: University of Calgary
Department: Emergency Medicine, Internal Medicine
Title: Intramuscular midazolam for methamphetamine-related agitation in the prehospital setting: a retrospective review of EMS encounters
Authors: Stephanie Vandenberg, Monty Ghosh 
This abstract was submitted as a Work in Progress

Introduction: Methamphetamine use is a growing public health concern, frequently resulting in acute behavioural disturbances that require emergency medical services (EMS) intervention. Midazolam, a benzodiazepine with rapid sedative effects, is commonly administered by EMS to manage methamphetamine-associated agitation. However, there is a limited understanding of ED outcomes after EMS midazolam administration. This study aimed to characterize the clinical outcomes of intramuscular (IM) midazolam administered by EMS for methamphetamine-related agitation.

Methods: We conducted a retrospective chart review of EMS and ED records from Calgary, Alberta, between January 1 and December 31, 2023. Adult (>18 yrs) patients who received IM midazolam in the setting of suspected methamphetamine-related agitation were identified. Demographic variables, vital signs, medication doses, and hospital disposition were extracted and analyzed descriptively. Secondary outcomes of interest included length of stay in ED/UCC, admission status, Canadian Triage and Acuity Scale. In addition to descriptive analysis, machine learning models are being developed to identify patterns in sedation response and predict ED resource utilization.

Results: A total of 104 patients met the inclusion criteria for initial analysis of data from 2019. Most patients were male (63%) with a median age of 35 years. The majority received an initial and total dose of 10 mg IM midazolam. Adequate sedation was achieved in 85% of cases, with 15% requiring a second dose, and 3% of patients requiring a third dose. All patients were transported to the hospital; 8% were admitted for further management, while 1% required intensive care unit involvement.  Ongoing AI-driven analysis aims to uncover predictors of repeat dosing and admission risk.

Conclusion: IM midazolam appears to be an effective and relatively safe option for prehospital management of methamphetamine-related agitation. Integrating AI into outcome analysis may enable predictive insights, inform EMS protocols, and optimize care for high-risk populations.

Dr. Mary Freymond
Institute: University of Calgary
Department: Emergency Medicine
Title: Clinical outcomes and characteristics of patients transported by helicopter emergency medical services to a comprehensive stroke centre.
Authors: Mary Freymond, Kira Genise, Domhnall O’Dochartaigh, Efrem Violato, Jessalyn Holodinsky, J.P. Berard, and Katie Lin

Introduction: Helicopter emergency medical services (HEMS) provide rapid access to comprehensive care for patients with ischemic stroke. Expansion of the endovascular thrombectomy (EVT) treatment time window has enabled more ischemic stroke patients to be considered for intervention, but also presents challenges for triage and transport decisions in strained healthcare systems. This study aimed to describe the clinical outcomes of HEMS transports for confirmed or suspected ischemic stroke patients, compare intervention rates between early (<6h from last seen normal to HEMS consult) and late (≥6h) presenters, and identify pre-transport factors associated with neuro-intervention.

Methods: This population-based retrospective chart review included adults with confirmed or suspected ischemic stroke transported by HEMS to a comprehensive stroke centre (CSC) in Alberta between Jan 1, 2019 and Dec 31, 2022. The primary outcome was neuro-intervention, including EVT and/or other neurosurgical procedure. Secondary outcomes included “low-yield” transfers (i.e., patients died/palliated or discharged/repatriated within 24h without neuro-intervention) and a comparison of pre-transport characteristics between patients who received intervention to those who did not, using Pearson’s Chi-square and Student’s T-test. An exploratory best-subset multivariable logistic regression identified pre-transport factors associated with neuro-intervention. Sample size was calculated a priori to achieve >80% power.

Results: Of 226 patients included, 35% (n=80) received neuro-intervention with 27% (n=62) receiving EVT, while 23% (n=53) met criteria for “low-yield” transfers. Patients who received neuro-intervention had similar times from last seen normal to CSC arrival (6.65.3 vs. 6.85.4 h, p=0.12) as those who did not, respectively. Neuro-intervention rates were similar for early presenters (<6h; 55/154, 36%) and late presenters (≥6h; 20/64, 31%; p=0.53). Pre-transport imaging (OR 5.26; 95% CI 1.79–10.67) and pre-transport thrombolysis (OR 2.82; 95% CI 1.05–12.94) emerged as the strongest predictors for intervention.

Conclusion: Approximately one-third of HEMS transfers for ischemic stroke resulted in neuro-intervention, and nearly one-quarter of transports were potentially low-yield. Early and late presenting cases had similar intervention rates. Pre-transport imaging and thrombolysis are strongly associated with neuro-intervention, underscoring the importance of access to neuroimaging to optimize care and resource utilization.

Dr. Cory Lefebvre
Institute: University of Calgary
Department: Pediatrics
Title: Development of a specific Connect Care measure for the number of visits to the Alberta Children’s Hospital (ACH) Emergency Department (ED) presenting with primary care appropriate condition (PCAC)
Authors: Cory Lefebvre, Jessalyn Holodinsky, and Jennifer Thull – Freedman

Introduction: Emergency departments (EDs) provide essential care for acute illness and injury; however, a substantial proportion of pediatric ED visits may be for conditions more appropriately managed in primary care settings. Limited access to timely primary care for children in Alberta may contribute to nonurgent ED use, impacting patient experience, continuity of care, and system efficiency. Existing administrative measures of primary care sensitive conditions have not been specifically validated in pediatric emergency settings, and their performance using Alberta administrative data remains uncertain. A locally relevant, validated measure of primary care appropriate conditions (PCAC) could support quality improvement and health system evaluation.

Methods: We conducted a retrospective administrative data analysis of all pediatric ED visits at Alberta Children’s Hospital from May 2022 to December 2025 using Connect Care data. An operational PCAC definition was developed based on published literature, expert consensus, and locally available data. A visit was classified as primary care appropriate if the patient was <18 years of age and had none of the following: CTAS 1-2; presenting concern consistent with injury; disposition of admission, transfer, or death; blood tests or imaging ordered; IV insertion ordered; blood product ordered; or time from MD to discharge >4 hours. Monthly PCAC proportions were examined using statistical process control (SPC) methods to assess stability and detect special cause variation. 

Results: Among 253,295 pediatric ED visits, 83,109 (32.8%) met operational PCAC criteria. SPC analysis demonstrated an initial system mean of approximately 34.4%, with isolated episodes of special cause variation, including summer 2022 and the STEC outbreak in fall 2023. Beginning in August 2024, a downward shift was observed, with stable performance throughout the period studied with a new mean of 31.0%.

Conclusions: An operational PCAC measure identified nearly one-third of pediatric ED visits as potentially suitable for primary care and demonstrated longitudinal stability following an identifiable downward shift. These findings support the construct validity and robustness of the measure and provide a strong foundation for prospective diagnostic accuracy validation. The project team is currently engaged in evaluation of diagnostic performance of the measure. A stratified random sample of ED visits will undergo independent chart review by two physician reviewers using a standardized PCAC definition, with a third reviewer adjudicating discordant cases and a 10% subset for agreement assessment. Sensitivity, specificity, and predictive values of the Connect Care PCAC measure will be calculated and compared with existing CIHI and HQCA measures, using physician adjudication as the reference standard. A validated PCAC measure will be valuable for monitoring health system performance for children and informing targeted quality improvement initiatives

Dr. Kathryn Crowder
Institute: University of Calgary
Department: Emergency Medicine
Title: Machine Learning-Driven Solutions in Emergency Department Settings: A Needs Assessment and Perception Exploration of Clinician Experience with Artificial Intelligence (AI)-based tools
Authors: Kathryn Crowder, Hoang Hong, Sana Sajjad, Jacob Idoko, Hossein Piri, Dominic Skitsko, William Betzner, Gouri Ginde

Introduction: Artificial intelligence–enabled clinical decision support systems (AI-CDSS), including AI-based documentation tools, have the potential to reduce documentation burden and support clinical decision-making in the emergency department (ED). However, clinician adoption and trust remain uncertain. This study aimed to assess current documentation practices, perceptions of AI tools, and perceived areas of utility for AI-CDSS among ED clinicians.

Methods: We conducted a cross-sectional, anonymous survey of ED clinicians working at multiple Calgary-area hospitals. Survey items included clinician role and experience, documentation workflows, time spent on documentation, use of AI and computerized clinical decision support tools, trust in AI-generated outputs, and perceived benefits and risks of AI scribes. Descriptive statistics were used to summarize responses.

Results: Ninety-two clinicians completed the survey, including registered and licensed practical nurses (63%), emergency medicine attending physicians (24%), and residents (4%). Most respondents had 6–10 years of practice (32%). Documentation was primarily manual, with 65% typing notes without templates and 56% using typed templates. Only 11% reported current use of an AI-enabled scribe. Most clinicians spent 5–10 minutes per patient encounter on documentation. 

Among clinicians with AI-scribe experience, 64% rated AI-assisted documentation as somewhat or very accurate, and 57% reported improved workflow efficiency. Provider-reported patient concerns regarding AI scribe use were rare. The most frequently cited concern was inaccurate or hallucinated outputs. Overall trust in AI-generated clinical outputs was moderate, with 56% reporting moderate trust. A majority of respondents preferred AI systems that both provide outputs and explain their reasoning (71%). AI support was most welcomed for alerting to contradictions (64%), risk prediction (49%), and patient triage (41%). AI-generated outputs for established clinical decision rules were viewed as useful by 46% of respondents, while 33% were uncertain. 

Conclusions: ED clinicians demonstrate cautious openness toward AI-enabled tools, particularly for documentation support and error detection, while expressing ongoing concerns regarding accuracy, transparency, and over-reliance. Explainability, validation, and integration into existing workflows will be critical for successful adoption of AI-CDSS in emergency care.

Dylan Viste
Institute: University of Calgary
Department: Emergency Medicine
Title: Mood trends and mental health interventions of a Canadian overdose response hotline.
Authors: Dylan Viste, BHSc; Zainab Abidi; S. Monty Ghosh, MD, MPH; Taylor Orr, MSc
This abstract was submitted as a Work in Progress

Introduction: Canada’s National Overdose Responses Service (NORS) was started in response to the opioid toxicity crisis but has unexpectedly received a large proportion of mental health calls, ranging from friendly peer support to emergency suicide  prevention. At the suggestion of other mental health hotlines, NORS now records the overall mood of callers at the beginning and end of sessions as a means of quantifying mental health impact. Our project seeks to investigate the mood change trends by demographics and in-call behaviours and then to qualitatively describe callers’ stressors as well as  hotline operator interventions to better understand how a substance-focused hotline provides mental health support.

Methods: Data from 18 months of NORS call logs was examined for mood change. NORS uses self-reported measures to score mood levels on a scale ranging from 0 (baseline - “I’m doing OK”) to -4 (Extremely Upset). Changes were measured by comparing the starting mood number from the ending mood number. A mixed-effect linear regression was performed to help control for the confounding effects of returning callers. Additionally, we qualitatively coded calls (which featured free-text notes and at least one negative mood) for occurrences of personal stressors (i.e. withdrawal symptoms) or operator interventions (i.e. redirecting conversations). Codes were then analyzed by their mood change scores, providing a descriptive summary of both average mood score per stressor and mood change per intervention.

Results: Of 6681 calls, 60% began the call “feeling OK”. Across all demographics, there was roughly an 83% reduction in negative mood states, and this effect was more pronounced amongst the more extreme starting moods. Odds ratios indicate that callers from communities with populations between 10-100k (OR 0.34, CI 95% 0.21-0.57), other call types (OR 0.12, CI 95% 0.21-0.81), texting (OR 0.70, CI 95% 0.55-0.90), and emergency situations of overdoses (OR 0.26, CI 95% 0.12-0.61) and mental health crisis or suicide prevention (OR 0.36, CI 95% 0.23-0.59) had a significantly lower likelihood of having their mood improve during a call. Qualitatively, we observed 41 different thematic client stressors, with notable average mood scores for challenges such as suicidality (-2.57 start, -1.20 end, STD 1.15), family conflicts (-1.85 start, -0.46 end, STD 0.96), and anxiety (-1.56 start, -0.59 end, STD 0.82). We observed 25 unique operator interventions, with notable items (with average mood change  scores) being redirected conversations (+1.74, STD 1.08), breath coaching (+1.56, STD 0.94) and peer support (+1.37,  STD 1.00)

Conclusions: NORS uses a variety of interventions to provide a net positive impact on the mental state of callers facing a myriad of life challenges. Other substance use hotlines or mental health crisis lines and in-person peer support worker programs may be able to improve their services based on the example of NORS

Robert Mayall
Institute: University of Calgary
Department: Undergraduate Medical Education, Obstetrics & Gynecology
Title: Wait times and resource use for Conditions Threatening Reproductive Health in Emergency Departments: Does Patient Sex Make a Difference?
Authors: Robert M. Mayall*, Sergiu Cociuba, Niloofar Taghizadeh, Tim Zhang, Fareen Zaver, Jessalyn K. Holodinsky, Amity E. Quinn
This abstract was submitted as a Work in Progress

Introduction: Threats to the reproductive system are an important presentation in healthcare due to the potentially catastrophic downstream effects on patient fertility. Prior research has demonstrated sex-based differences in Emergency Department (ED) treatment of patients with specific conditions, such as torsion or sexually transmitted infections. Understanding this variation across different patient and facility characteristics can inform resource allocation and equity in care.

Methods: Using data from the national ambulatory care reporting system (NACRS) from April 1st 2018 – March 31st 2024, we included adults (≥18) presenting to any ED in Alberta for ICD-10 codes associated with negative fertility outcomes (e.g., infection, trauma, vascular). Descriptive statistics summarized median wait times and procedure counts across sex, age, acuity (defined by CTAS score), diagnosis, facility type, and temporal factors. Adjusted negative log binomial and negative log binomial zero-inflated count models were used adjusting for age, diagnosis, acuity, temporal factors, and facility-type; effect modification was assessed using interaction terms.

Results: 12,059 patients were included in the study, with 93% assigned female at birth and 94% presenting for infections. For both wait times and procedure numbers the magnitude of sex differences depended on facility type. In academic facilities males had 15% shorter expected wait times (95% CI; 1.9% – 27.4%) compared to females, but no difference was seen in community facilities (IRR: 0.944; 95% CI; 0.851 – 1.160). In community facilities, males had a lower expected number of procedures compared to females, and the magnitude of this difference decreased with age (IRR: 0.33; 95% CI: 0.21–0.51 at age 20 to 0.49; 95% CI: 0.32–0.75 at age 46). This same age pattern was seen in academic facilities although the magnitude of difference was larger overall (IRR: 0.44; 95% CI: 0.29–0.66 at age 20 to 0.64; 95% CI: 0.44–0.94  at age 46). 

Conclusions: Males at academic centers experienced significantly shorter waits than females. Males also received fewer procedures than females both at academic and community centres; however the gap began to close with increasing age. These findings underscore the need for EDs to examine workflow and resource allocation to understand the reasons behind these differences to ensure equitable, consistent care across sexes.

Simran Dhillon
Institute: University of Calgary
Department: Emergency Medicine
Title: Chronic Pain, Addiction, and Emergency Department Care Among Survival Sex Workers in North America: A Scoping Review
Authors: Simran Dhillon, Stephanie VandenBerg
This abstract was submitted as a Work in Progress

Introduction: Survival sex workers experience intersecting health and social inequities shaped by chronic pain, substance use, trauma, and criminalization. Chronic pain in this population is frequently unmanaged and occurs alongside high rates of substance use, often in the context of violence, housing instability, and limited access to primary or specialty care. While survival sex work has been widely examined in relation to infectious disease risk and social vulnerability, the co-occurrence of chronic pain and addiction, and its implications for emergency department care, remains under explored. Emergency departments frequently serve as a primary or sole point of healthcare access for survival sex workers experiencing acute pain crises, substance-related complications, or untreated chronic conditions. This scoping review aims to map existing literature on chronic pain and addiction among survival sex workers in North America and to examine how structural factors influence emergency department utilization, care experiences, and health outcomes, with relevance to health equity-oriented emergency medicine practices in Canada.

Methods: A comprehensive search of MEDLINE, Embase, and relevant grey literature was conducted for studies published between 2012 and 2025. Eligible sources include qualitative, quantitative, and mixed-methods studies examining chronic pain, substance use, emergency department utilization, healthcare access, or related social and structural determinants among survival sex workers in North America. Data extraction and thematic synthesis are currently ongoing, with findings being analyzed using an iterative thematic approach.

Results: Preliminary synthesis suggests that chronic pain and addiction frequently co-occur among survival sex workers and are shaped by overlapping experiences of violence, economic precarity, criminalization, and structural exclusion. Emergency departments are commonly accessed for acute exacerbations of pain, substance-related presentations, and complications arising from untreated or inadequately managed chronic conditions. Across studies, emergency care encounters are often described as time-limited and episodic, with experiences of stigma and mistrust reported by patients. Limited access to longitudinal pain management, addiction treatment, and coordinated follow-up contributes to repeated emergency department use, delayed care-seeking, and fragmented care trajectories. Final results and thematic conclusions are pending completion of data extraction and analysis. 

Conclusions: This ongoing scoping review is expected to identify critical gaps in both emergency and longitudinal care related to the integrated management of chronic pain and addiction among survival sex workers. Anticipated findings will inform trauma-informed, structurally competent emergency department approaches that support appropriate pain management, reduce stigma, and strengthen pathways to ongoing addiction and pain care. Such approaches may improve care continuity, reduce recurrent emergency department utilization, and advance equity in emergency medicine for structurally marginalized populations.

Brian Steele
Institute: University of Calgary
Department: Cumming School of Medicine
Title: Transfusion Practices for Iron Deficiency Anemia in Calgary Emergency Departments
Authors: Nojan Mannani, Brian Steele, Kathryn Crowder, Melanie Sohn
This abstract was submitted as a Work in Progress

Introduction: Iron deficiency anemia (IDA) is the leading global cause of anemia, frequently resulting from blood loss, decreased iron absorption, or increased iron requirements. Symptoms of IDA range from weakness, fatigue, and light-headedness to more severe complications such as dyspnea, angina, and syncope. Emergency physicians (EPs) may transfuse patients presenting with IDA with red blood cells (RBCs), which temporarily elevates hemoglobin but does not address the underlying cause and carries transfusion-related risks. This study assesses whether Calgary EPs are over-transfusing patients who present with IDA as compared to the Alberta Medical Association Toward Optimized Practice (AMA TOP) Guidelines. 

Methods: A retrospective cohort study was conducted of adults presenting to four Calgary emergency departments between January 2021 and December 2024 who were discharged with a diagnosis description including “anemia.” Electronic medical record data (Sunrise Clinical Manager and Connect Care) were used to identify eligible visits and extract demographics, triage vitals, hemoglobin (Hgb), mean corpuscular volume (MCV), RBC transfusion, iron supplementation, and disposition variables. Patients who met the AMA definition for IDA (Hgb <135 g/L for males or Hgb <120 g/L for females and MCV < 75fL) were eligible for inclusion. Hemodynamically unstable patients were excluded. The primary outcome was adherence to appropriate RBC transfusion based on the AMA TOP guidelines. The secondary outcome was iron supplementation adherence in asymptomatic patients as compared to the AMA TOP guidelines. 

Results: A total of 3,906 ED visits met the inclusion criteria. RBC transfusions were administered in 53.9% of visits (n = 2,106). Overall adherence to AMA TOP transfusion guidance was 55.4% (n = 2,164), with non-adherent transfusions occurring in 44.6% of visits (n = 1,742). Among asymptomatic cases, transfusion was ordered in 56.4% of visits despite guideline recommendations favoring oral iron. Among symptomatic cases, over-transfusion occurred in 26.6% of visits where Hgb was <70 g/L and 16.7% of visits where Hgb was 70–100 g/L. Under-transfusion was less frequent, occurring in 4.7% of symptomatic patients with Hgb <70 g/L. Based on preliminary data available from Connect Care prescription information, iron supplementation adherence was low; among asymptomatic visits (n = 1,437), 71.3% demonstrated non-adherence, including no iron provision (36.4%) or IV iron instead of oral (34.9%), however, more fulsome data, including PIN data for oral iron prescriptions is forthcoming. 

Conclusions: This study demonstrates there is a trend towards over-transfusion of IDA patients in Calgary EDs when compared to the AMA TOP guidelines, particularly in asymptomatic patients. These findings highlight potential opportunities for quality improvement and continuing medical education initiatives to support guideline-concordant IDA management, optimize resource utilization, and improve patient outcomes.

Jiajun Wu
Institute: University of Calgary
Department: Electrical and Computer Engineering
Title: EDSim: An Agentic Simulator for Emergency Department Operations 
Authors: Jiajun Wu, Hutton Ledingham, Zirui Wang, Alex Burn, Oussama Ouadihi, Darin Vicaldo, Mohamad Ghattas, Gisa Tchaka Gashagaza, Sergiu Cociuba, Megan Harmon, Braden Teitge, Jessalyn Holodinsky, Steve Drew 
This abstract was submitted as a Work in Progress

Introduction: Emergency departments (EDs) face chronic crowding and complex patient-flow challenges that are difficult to study safely and quickly in the real world. Conventional discrete-event simulation (DES) and traditional agent-based simulation (ABS) can often match high-level metrics (e.g., throughput or wait-time distributions), but they struggle to reproduce the fine-grained behaviors, communication, and dynamic decision-making of clinicians and patients that can meaningfully shape ED performance. Recent progress in large language models (LLMs) enables agents that can perceive context, interact naturally, and adapt to evolving conditions, creating an opportunity for more realistic, workflow-sensitive ED simulation.

Methods: We developed EDSim, an agentic ED flow simulator powered by LLM agents that supports a modular patient journey from triage through treatment and discharge. Virtual patients and healthcare providers converse and make decisions in natural language, while being constrained by clinical rules and global ED states to preserve operational realism. EDSim enforces spatial awareness and movement constraints grounded in an editable ED floor plan, enabling environmentaware workflows. EDSim couples natural-language interaction with operational state tracking. Agents communicate through logged conversations for tasks such as triage questioning, care coordination, and clinical reasoning, while the simulation engine simultaneously records structured events and timestamps for key operational milestones. Decisionmaking is constrained by global ED state so that actions depend on real-time resource availability, queue status, and competing demands.We integrate real-world ED statistics (Foothills Medical Centre, Calgary) to parameterize arrival patterns, acuity mix, and diagnosis distributions, and assign triage acuity using standard mappings to support realism without overcomplicating triage RESULTS: Given the limited utilization of RCNs in Canada, with Ontario being the only province currently employing RCNs in its EDs, the survey findings are expected to reflect limited familiarity with recovery coaches among healthcare providers (Ontario Health, 2023). 

Results: EDSim produces baseline outputs that align with acuity-stratified waiting-time patterns, demonstrating macro-level fidelity in how delays differ across triage levels. Beyond aggregate metrics, the simulator generates granular, interpretable traces of ED operations: conversational logs that capture handoffs and coordination, movement trajectories across the floor plan, and event-level timelines linking resource contention to downstream delays. These traces make it possible to explain not only what changed in performance but why it changed, by connecting bottlenecks to specific constraints such as bed saturation, staff availability, or diagnostic-process delays. We also demonstrate rapid what-if experimentation capabilities. EDSim can be used to stress the system with demand surges and to test operational interventions such as staffing or capacity adjustments, while preserving consistent logging of patient-level and system-level outcomes. In these scenarios, the simulator reveals non-linear congestion effects, where delays compound across stages and shift bottlenecks  downstream, and it provides patient-level distributions that help characterize which acuity groups are most affected. 

Overall, EDSim supports both outcome comparison across scenarios and mechanistic insight into how workflow interactions and resource constraints drive observed wait times. 

Conclusion: EDSim introduces a new paradigm for healthcare operations research by combining data-driven ED flow modeling with LLM-generated, environment-aware agent behavior. It provides a realistic and configurable testbed for exploring operational changes with both quantitative agreement to historical patterns and qualitative insight into communication and workflow dynamics, supporting use by researchers and practitioners to improve emergency care delivery.

Jaden Frizell
Institute: University of Calgary
Department: Bachelor Health Science
Title: Emergency Department Wait Times by Housing Status: Did the COVID-19 Pandemic Exacerbate Disparities?
Authors: Jaden Frizzell*, Megan Harmon, Niloofar Taghizadeh, Jessalyn Holodinsky

Introduction: Emergency Department (ED) overcrowding is a prominent public health issue, with the COVID-19 pandemic substantially affecting ED utilization and wait times. Disparities in ED wait times between groups have become a critical area of research, with housing status emerging as a factor of growing interest. The unhoused population have an increased reliance on the ED and are disproportionately affected by the factors that contribute to longer ED wait times. This study aimed to describe differences in Alberta ED wait time between the housed and unhoused patient populations and to compare wait time differences before, during, and after the COVID-19 pandemic.

Methods: Using data from the National Ambulatory Care Reporting System, we included all adult (>18) visits to an ED or Urgent Care Clinic in Alberta from April 1st, 2018, to September 30th, 2023. Housing status was defined using ICD-10 code Z59.0. The pandemic period was defined as starting the week of March 15, 2020, and ending the week of June 14, 2022. Piecewise negative binomial regression models were fit with robust standard errors to assess changes in wait times pre-, during-, and post-COVID-19 pandemic among the housed and unhoused. All models were adjusted for age, sex, and time of day. Separate analyses were performed for each Canadian Triage Acuity Score (CTAS). Analyses were conducted using Stata19.

Results: 8,700,411 observations were included in this study, where 1.27% of these visits were unhoused. Over the course of the study, mean wait times increased from 73.5 to 135.5 minutes among the housed and from 142.2 to 216.8 minutes among the unhoused. During the pre-pandemic period (April 2018 to March 2020), changes in wait times did not differ meaningfully between the housed and unhoused across all CTAS levels. At pandemic onset, wait times dropped significantly for both groups, with the unhoused experiencing a more drastic decrease in wait times among CTAS 3-5. Throughout the COVID-19 Pandemic, wait times increased at the same rate for both groups. At the end of the COVID-19 period (June 2022), wait times increased among both groups, with the unhoused experiencing a significantly larger increase in wait times compared to the housed across all CTAS levels. Following the COVID-19 pandemic, wait time trends showed stabilization or gradual declines in wait times for both groups. 

Conclusion: Throughout the study period, emergency department wait times were consistently longer for unhoused patients than for housed patients, with the magnitude of this disparity increasing as patient acuity decreased. Additionally, unhoused patients exhibited greater sensitivity to pandemic-related disruptions, experiencing more pronounced changes in wait times compared with housed patients. Future work will continue to explore reasons for these differences, such as time of presentation and specific reasons for presentation. These results can inform future interventions to improve access to care for the unhoused population, reduce ED wait times and better support unhoused patients.

Jaden Frizell
Institute: University of Calgary
Department: Bachelor Health Science
Title Housing Status and Preventable Emergency Department Visits: An Analysis of Ambulatory Care–Sensitive Conditions in Alberta (A Work in Progress)
Authors: Jaden Frizzell*, Megan Harmon, Niloofar Taghizadeh, Jessalyn Holodinsky
This abstract was submitted as a Work in Progress

Introduction: With Canadian Emergency Departments (ED) continuing to be overwhelmed by increasing patient volumes, determining what ED visits may be preventable has become an increasingly critical area of research. Ambulatory care sensitive conditions (ACSC) are conditions which can be managed in a primary care setting but can lead to hospitalization when unmanaged. As such, ED visits for ACSCs can be considered preventable with access to timely primary care. However, populations that have limited access to primary care may be disproportionately burdened by ED visits for ACSCs. The unhoused population often rely on the ED as their main source of healthcare, due to unique barriers to primary care. The purpose of this project is to understand if ACSC-related presentations to the ED differ between the housed and unhoused populations in Alberta.

Methods: This study will use data from the National Ambulatory Care Reporting System (NACRS) of all ED and urgent care clinic visits in Alberta, between April 1st, 2018 and September 30th, 2023. The primary outcome of interest is a binary indicator for the presence of an ACSC diagnosis, according to the definition of ACSC developed by Purdy et al (2009). ED visits are to be considered ACSC-related visits if any of the first three discharge diagnostic positions contain an ACSC diagnosis. The exposure variable of interest is the patient's housing status at a given visit. A visit will be categorized as unhoused if any of the ten discharge diagnostic positions contain the International Classification of Diseases 10 (ICD-10) code Z59.0, representing ‘homelessness’.

Descriptive statistics will be used to characterize baseline characteristics for both the housed and unhoused groups, including age, sex, and discharge disposition. These baseline characteristics will be compared between groups using independent sample t-tests and chi-squared tests of independence, with an alpha value of 0.05 as the significance cut-off. Multivariable logistic regression models will be used to estimate the odds of an ACSC visit among unhoused visits compared to housed visits (reference group). Odds ratios, 95% confidence intervals, and p-values (alpha = 0.05) will be reported. All models will be adjusted for age and sex among other covariates. Robust standard errors will be used to account for multiple visits per individual.

Discussion: Current literature lacks a comprehensive investigation of ACSC-related ED utilization among the unhoused population in Canada. This analysis will attempt to fill this gap and help determine if housing status affects reliance on the ED for ACSCs. Through measuring ACSC-related ED utilization, as an indicator of primary care access, we can better assess the unhoused population’s reliance on the ED. This knowledge can inform future interventions to address the healthcare inequities among the unhoused population.

Mackenzie Hancock
Institute: University of Calgary
Department: Cumming School of Medicine
Title: Urgent emergency department follow up clinics for recently discharged patients - a scoping review
Authors: Kathleen Fougere, Samantha McIntosh, Mackenzie Hancock, Kathryn Crowder, Ayesha Khory, and Stuart Turner
This abstract was submitted as a Work in Progress

Introduction: Urgent emergency department (ED) follow up clinics provide an opportunity to reassess patients who are deemed safe to go home with plans for short term reassessment after ED discharge. These clinics have proven to avoid unnecessary inpatient admissions and reduce health system costs by reducing acute care utilization. Despite the growing evidence in favour of urgent follow up clinics, there are relatively few clinics in practice. Further, the characteristics of existing clinics and their impact on repeat ED presentations and subsequent admissions are not well understood. The objective of this scoping review is to outline the characteristics of, and reported outcomes associated with urgent follow up clinics for patients recently discharged from the ED. We specifically aim to understand the reported effects of these clinics on rates of return to the ED and hospital admissions. Further, we aim to compare outcomes in clinics led by emergency physicians and primary care physicians. A secondary aim of this review is to characterize the structure of these clinics, including strategies used to predict clinic volumes. 

Methods: Literature involving adult patients, aged 16 and older who were seen in an urgent follow up clinic after their initial presentation to the ED (index visit) will be included. All study types discussing urgent follow up clinics will be included. Urgent follow up clinics will be defined as ambulatory clinics led by emergency or primary care physicians, including those independent from the ED, where patients are seen within 30 days (or 1 month) of their index visit for reassessment of their initial presenting complaint. A comprehensive search will be conducted in Medline, Embase, Scopus, Web of Science, CINAHL, and Psycinfo for academic literature. Grey literature will be identified through Google search and grey literature databases. 

Results: Across databases, the search strategy returned 9719 articles, which were then pulled into Covidence systematic review software for title and abstract screening. 3841 articles were identified as duplicates and removed. Title and abstract screening was then completed for the remaining 5878 articles. Full text screening of 286 articles is currently underway. Studies meeting inclusion criteria will undergo data extraction using a standardized data extraction instrument. 

Conclusions: This review will synthesize existing evidence on urgent follow up clinics for patients recently discharged from the ED. By identifying and reporting the characteristics of existing clinics, we can begin to better understand which patients are appropriate for urgent follow up, and how we can best structure these follow up clinics to reduce ED revisits and hospital admissions. Future research will be able to use these findings to inform the creation and/or replication of evidence-based, urgent follow up clinic models, with the potential to mitigate ED crowding and reduce physician burn out while improving patient outcomes and resource allocation.

Simran Dhillon
Institute: University of Calgary
Department: Emergency Medicine
Title: Association Between Supervised Consumption Site Closures and Opioid-Related Mortality and Emergency Department Visits in Alberta 
Authors: Simran Dhillon, Eddy Lang
This abstract was submitted as a Work in Progress

Introduction: Supervised consumption sites (SCS) are evidence-based harm reduction interventions that reduce overdoserelated morbidity and mortality and improve access to healthcare for people who use drugs. These services play a critical role in overdose prevention, early intervention, and linkage to health and social supports. In recent years, multiple SCS across Alberta have closed or experienced reductions in service availability, prompting concern regarding the potential population-level consequences of decreased access to supervised consumption services during an ongoing overdose crisis. Evidence examining the downstream health impacts of SCS closures remains limited in the Canadian context. This study examines the association between SCS closures and changes in opioid-related mortality and emergency department (ED) visits in Alberta. 

Methods: We are conducting a retrospective observational study using published administrative health data. Opioid-related deaths and ED visits will be examined before and after SCS closures using interrupted time-series and comparative trend analyses. Primary outcomes include monthly opioid-related mortality rates and associated trends. Secondary analyses will explore regional variation in outcomes based on the timing and extent of service reductions. Analyses will assess temporal trends and changes associated with supervised consumption service availability across affected regions of Alberta.

Results: This study is currently in progress. Planned analyses will evaluate changes in opioid-related mortality and ED visit trends following SCS closures and compare patterns across regions with differing levels of supervised consumption service availability. Results will describe temporal associations between SCS closures and opioid-related health outcomes at the population level. 

Conclusion: This study will generate timely Canadian evidence on the population-level and health system impacts of supervised consumption site closures within the context of the ongoing overdose crisis. Findings will contribute to public health and policy discussions regarding harm reduction service accessibility, overdose prevention strategies, acute care utilization, and health equity for people who use drugs in Alberta and across Canada.

Taylor Krawec
Institute: University of Calgary
Department: Emergency Medicine
Title: Short-term cardiovascular outcomes among emergency department patients ruled out using high-sensitivity cardiac troponin assays: a stratified meta-analysis of assays and algorithms.
Authors: Taylor Krawec, Niklas Bobrovitz, Cody Dunne, Jinan Daqqa, Anabel Selemon, Christian Cao, Caitlin McClurg, Jessalyn Holodinsky, Paul Ronksley, Andrew McRae
This abstract was submitted as a Work in Progress

Introduction: Chest pain is among the most common emergency department (ED) complaints in Canada and high-sensitivity cardiac troponin (hs-cTn) assays are central to ruling out myocardial infarction (MI). Substantial variation exists in assays and rule-out algorithms, but it remains unclear if these differences affect patient outcomes. We conducted a systematic review and stratified meta-analysis to compare 30-day incidence of major adverse cardiac events (MACE) among patients with suspected acute coronary syndrome (ACS) ruled out for MI using various hs-cTn assay types and algorithms.

Methods: We searched MEDLINE (Ovid), Embase (Ovid), Web of Science (Core Collection), ClinicalTrials.gov, Cochrane CENTRAL (Ovid), and Europe PMC (preprints) from January 2009 to October 2025 for observational, quasi-randomized, and randomized studies of adults (≥18 years) with suspected ACS in whom MI was ruled out by ECG and hs-cTn. The primary outcome was cumulative incidence of MACE within 30 days of ED discharge. Random-effects meta-analyses were stratified by hs-cTn assay type (manufacturer; troponin I vs T) and rule-out algorithm (single-sample vs serial; 0/1h, 0/2h, 0/3h).

Results: We screened 7,683 abstracts and 930 full-text articles, resulting in 28 studies included (127,841 patients). Nine assays were used in 39 distinct rule-out algorithms, measuring troponin I (51.3%, n=20 studies) or T (49.7%, n=19 studies). Single-sample strategies were used in 59% of studies (n=23) while serial testing was used in 41% of studies (n=16), including intervals at 0/1h (56.3%, n=9 studies), 0/2h (18.8%, n=3 studies), and 0/3h (25.0%, n=4 studies). 

The pooled 30-day MACE incidence was similar between single and serial rule-out algorithm strategies (0.80%, 95% CI 0.20-1.60%, I²=92%, n=13 vs 0.80%, 95% CI 0.10-2.00%, I²=98%, n=8; p=0.75). Incidence of individual MACE components was also similar between single and serial strategies: MI 0.20% (95% CI 0.00-0.30%, I²=56%, n=9) vs 0.40% (95% CI 0.10-0.80%, I²=73%, n=6; p=0.08); cardiac death 0.00% (95% CI 0.00-0.10%, I²=0%, n=5) vs 0.00% (95% CI 0.00-0.10%, I²=14%, n=4; p=0.92); and all-cause death 0.00% (95% CI 0.00-0.00%, I²=13%, n=10) vs 0.00% (95% CI 0.00-0.00%, I²=19%, n=10; p=0.64). There were no differences in pooled 30-day MACE between serial testing algorithms using different time points.

Conclusions: This preliminary analysis demonstrated a consistently low incidence of 30-day MACE across assay types and rule-out algorithms for ED patients ruled out for MI using hs-cTn.

Dr. Graham Thompson
Institute: University of Calgary
Department: Pediatrics and Emergency Medicine
Title: Pragmatic Trial of Balanced Versus Normal Saline Fluids in Sepsis (PRoMPT BOLUS)
Authors: Graham Thompson, Beata Mickiewicz, Waleed Alqurashi, Simon Berthelot, Sarah Curtis, Adrienne Davis, Jason Emsley, Stephen Freedman, Karen Gripp, Pavan Judge, April Kam, Ahmed Mater, Gary Joubert, Anupam Sehgal, Yasaman Shayan, Fran Balamuth, Nate Kuppermann, Scott Weiss, Franz Babl, Stuart Dalziel, Elliot Long
This abstract was submitted as a Work in Progress

Introduction: - International guidelines recommend crystalloid resuscitation in children presenting with septic shock. However, conflicting evidence as to whether normal saline (NS) or balanced fluids (BF) results in superior outcomes has led to continued equipoise and significant practice variation. The objective of this study was to determine if fluid resuscitation with BF will improve clinical outcomes compared to fluid resuscitation with NS in pediatric septic shock. Our secondary objectives were to evaluate the safety of BF versus NS in pediatric septic shock, and to conduct a cost-effectiveness analysis. 

Methods: We conducted a multi-national, pragmatic, open-label, randomized controlled trial (RCT) of children aged 2 months through 17 years with ED presentations concerning for septic shock. Clinicians identified eligible children expected to receive at least one bolus of resuscitation fluid to treat poor perfusion associated with infection. Participants were enrolled using a deferred consent model, and randomized to either NS or BF for all further bolus and maintenance fluid administration. The intervention window continued until midnight the day following randomization. Fluid dosing, diagnostics and all other management were at the discretion of the managing health care team. Follow up continued through the first of either hospital discharge or 90 days post enrolment.

Our primary outcome measure was Major Adverse Kidney Events (MAKE30), measured as a composite of death, initiation of new renal replacement therapy (RRT) or persistent kidney dysfunction (KD). Secondary efficacy measures included persistent KD, new RRT, hospital-free days and all-cause mortality (discharge and 90 days). Secondary safety measures include clinically significant laboratory abnormalities, therapy for thromboembolism, and brain herniation. Societal outcomes are measured by incremental cost-effectiveness ratio. A sample size of 8800 evaluable participants across 5 participating countries was required to detect a 1.7% absolute risk reduction in MAKE30 with a power of 95%. Data was analyzed using intention-to-treat and perprotocol approaches. Subgroup analyses evaluated the impact of age, sex, known underlying cancer and abnormal kidney function at presentation 

Results: This study remains in progress. Enrolment was completed on Oct 31, 2025. Globally, over 9000 participants were randomized, including over 900 from across 12 Canadian pediatric institutions. Data cleaning has been completed. Statistical analyses are ongoing. Over 20 secondary analyses have been planned.

Conclusion: This multinational RCT will provide definitive evidence identifying the optimal fluid for resuscitating children with septic shock. As both fluid types are routinely available world-wide, results can be immediately transferable to clinical practice. Furthermore, this study demonstrates the power of pediatric research network collaborations (PERC, PECARN and PREDICT), as well as interdisciplinary teamwork (emergency medicine, intensive care and hospital medicine)