SUMMER TRAINING AND RESEARCH IN EMERGENCY MEDICINE (STAR-EM)

Sickle Cell Disease Working Group & Behavioral Safety Alerts 

Summary: The University Health Network emergency departments (UHN EDs) care for 250-300 patients with Sickle Cell Disease (SCD) per year. Patients with SCD experience acute pain episodes requiring timely medical intervention in the ED. The SCD working group has identified that patients living with SCD frequently have behavioral safety alerts (BSAs) in their medical record.  Electronic BSAs notify staff of prior behavioural and/or violent events that have occurred during hospitalization. This project will improve our understanding of the antecedent factors that contribute to behavioural escalations, what behaviours are documented, and how the presence of a BSA in the chart may impact future care for patients with SCD. Additionally, we will quantify the frequency of new BSA generation and the overall prevalence of BSAs for UHN ED patients with SCD. Through improving our understanding of BSAs, this study will help to improve both workplace safety and care of patients with SCD.  
 

Large Language Model (LLMs) in automated patient self-triage

This project aims to address the critical global health issue of Emergency Department (ED) overcrowding by investigating the performance and safety of Large Language Models (LLMs) in automated patient self-triage. The project will involve conducting a comprehensive scoping review following PRISMA-ScR guidelines. This work is specifically designed to investigate the "Safety Paradox," where models exhibit high aggregate accuracy yet fail predictably on life-threatening presentations. I will also evaluate how technical refinements, such as Retrieval-Augmented Generation (RAG), might mitigate these risks. This project will provide an evidence-based framework for evaluating generative AI tools, ensuring that "human-in-the-loop" systems remain the standard for safe emergency care. This structured approach ensures the project is largely achievable within the summer timeframe while producing a high-impact, submission-ready contribution to the field of emergency medicine and clinical informatics.

​​

PREVIOUS PROJECTS - 2025: 

​

Evaluating the Acute Care Echocardiography (ACE) Training Program
Summary: The Toronto Addis Ababa Academic Collaboration in Emergency Medicine (TAAAC-EM) has developed an Acute Care Echocardiography (ACE) training program to train emergency and critical care medicine residents at Addis Ababa University (AAU) on focused transthoracic echocardiography. During this project, the student will look into the perceived effect and impact on knowledge of residents participating in this pilot ACE workshop. As a pre and post cohort study, this will be done by analyzing pre-reading surveys, as well as pre- and post-workshop evaluations to inform future iterations of the workshop.
 

Formative evaluation of a Stabilization and Connection Centre (SCC) for stable patients using drugs and alcohol
Summary: The dual crises of ED overcrowding and highly prevalent alcohol and substance use disorders require 
novel models of care provision. This is a retrospective, formative evaluation of Stabilization and Connection Centre (SCC) operations and outcomes from Dec 2022 to Dec 2024. Comparative data for intoxicated patients attending UHN Emergency Departments during the same period will be extracted from EPIC. Comparisons between the SCC visits and ED visits will be made for demographics, substances used, EMS offload time, time spent on site, and disposition.
 

Evaluating impact and efficacy of iterative changes in FAST-HIV testing in the Emergency Department
Summary: The student will conduct a quantitative evaluation of the FAST-HIV testing program's performance at UHN, focusing on the impact of three iterative changes implemented over the past 12 months. They will analyze de-identified patient chart data from three distinct phases