Innovation where it matters most: How clinician-led research is redefining the future of emergency care

HN Summary

• Clinician-led innovation is transforming emergency care at Humber River Health, where frontline teams and data scientists are using AI, predictive modelling, and simulation to address ED overcrowding and improve patient flow.

• A dedicated PIA Working Group and a high-fidelity digital model of the ED allow leaders to test staffing and workflow changes virtually, optimize physician schedules, and evaluate interventions before implementing them in real-world settings.

• New AI tools, including a surge event prediction model, will enable managers to anticipate overcrowding, deploy resources proactively, and enhance both patient and staff experiences, positioning Humber as a leader in applied digital health innovation.

Emergency department (ED) overcrowding remains a global concern, as it can compromise patient outcomes, increase wait times, and strain healthcare workers. As healthcare systems worldwide face unprecedented pressures from growing patient volumes and constrained resources, hospitals are increasingly turning to advanced analytics and artificial intelligence (AI) to find smarter, more sustainable solutions. 

At Humber River Health (Humber), known as North America’s first digital hospital and home to one of the busiest EDs in the country, innovation is happening where it matters most: on the front lines of patient care. Driven by clinicians and supported by a robust data science team, Humber is launching pioneering AI initiatives that are setting the stage for clinician-led, data-informed research with real-world impact.

Recognizing that physician initial assessment (PIA) time is a key driver of department flow and patient experience, Humber’s data science and clinical teams are applying AI-driven analytics, predictive modelling, and simulation techniques to manage overcrowding and test interventions virtually before implementation in practice. This approach exemplifies Humber’s strategic priorities to advance digital and community health research, adopt technology to promote exceptional care, as well as develop innovative strategies to deliver integrated care in the community.

Transforming care through collaboration: The creation of the PIA Working Group

Led by the ED’s Program Director, Linda Jorgoni, and former Chief of Emergency Services, Dr. Leon Rivlin, the PIA Working Group was formed to address increasing patient volumes and rising acuity levels contributing to longer wait times. The creation of this group was inspired by the insights generated from their ongoing work with Humber’s data science team to develop an online queuing tool in the ED, an initiative that had been awarded $1.5 million in funding from Scale AI in September of 2023. “Our teams experience firsthand the challenges that come with increasing patient volumes and acuity, and they’re often the ones with the most practical ideas for improvement,” says Linda Jorgoni. “By bringing together diverse perspectives and expertise, we can transform those ideas into measurable solutions that enhance both the patients and staff experience.”

This clinician-led initiative brings together ED leadership, management, physicians, program directors, clinical analysts, and the Humber Research Institute’s data science team, composed of Dr. Krutika Joshi, Christina Seo, Troy Gloyn, and Dr. Pete Wegier. From their first discussions, a multi-stage program emerged to enhance ED efficiency by combining clinical expertise with data intelligence to improve patient flow, optimize resources, and reduce staff burden.

Where insight meets data: Building a digital model of the ED

Drawing on anonymized data from our electronic medical records (Meditech), diagnostic imaging, laboratory services, and staff scheduling systems between January 2022 and August 2024, the first stage of this program involved a holistic understanding of how patients move through the ED and how workflows influence wait times and outcomes. From a subset of this data, the team developed a discrete-event simulation (DES) model—or “digital model”—of the ED, simulating variables such as patient arrival rates, service times, staffing levels, and care zones to replicate ED flow with high fidelity. Comparable to a SimCity-style model, this digital model allows our data scientists to run “what-if” scenarios provided by the ED team to evaluate potential interventions in a controlled, virtual environment. The model has been validated against historical data and can be used to predict performance metrics, including 90th percentile time to PIA and length of stay.

This tool provides a safe platform for decision-makers to test different staffing and resource allocation scenarios, such as adding triage nurses or modifying on-call physician schedules and observe their impact on key indicators before implementing them in real life. Our new Chief of Emergency Services, Dr. Tajinder Kaura, has worked with our data science team to explore potential staffing scenarios that could optimize physician scheduling. As Dr. Kaura states, “being able to use the digital model to build an optimal physician schedule that positively impacts patient flow is just the beginning.”  

Innovation at work: Building the Surge Event Prediction Model

The next stage of work is the development of a prediction model to forecast surge events, giving ED managers the ability to proactively deploy staff and resources before congestion occurs. Using anonymized historical data from Meditech and surge labels provided by the ED team, our data scientists are training multiple AI models to predict short-term spikes in patient volume and wait times. The outputs will then be combined into a real-time surge alert system, which flags potential overcrowding scenarios based on forecast thresholds. These forecast thresholds are being revised iteratively with the ED team to ensure surge alert accuracy and clinical relevance. The next step will be to further fine tune these alert thresholds by leveraging real-time data. 

This proactive solution will allow managers to mitigate bottlenecks before they arise, bringing in on-call staff earlier, adjusting resource levels dynamically, and ultimately improving both patient and provider experiences.

Looking ahead: Healthcare innovation at the front lines

Together, these initiatives represent a technology-enabled approach to health system design that leverages data science, simulation, and AI to achieve measurable improvements in care delivery. By transforming raw operational data into applied digital tools that guide front-line decision-making, this model of grassroots teamwork demonstrates its potential as a catalyst for innovation, a bridge between data insight and clinical impact. “We are building the infrastructure, skills, and partnerships needed to make Humber a hub for applied digital health innovation, where every project has a clear path from ideation to impact,” says Dr. Justin Grant, Vice President of Research and Innovation. “As we continue to expand our research capacity, this approach will position Humber as a leader in converting AI and data science into meaningful, patient-centred improvements in care.” 

The journey is just beginning, but the foundation being built today reflects the organization’s ongoing commitment to lighting new ways in healthcare through innovation and excellence in patient care. 

Related reading: Living classroom brings training into long-term care

Related reading: Interior Health expands use of virtual care in EDs

Related reading: Sunnybrook launches innovation and impact award

Related reading: Ava adds to scribe, automatically organizes chart

Related reading: UHN surgeons curing rare lung condition CTEPH with world-class procedure

Related reading: RapidAI and AWS strengthen relationship

Related reading: UniDoc sells three telehealth Cubes to Verizon

Related reading: Paratus launches AI-driven support for EDs