Building Models and Algorithms for Enhanced Emergency Medical Services

Awarded the Winspear Chair in Business Analytics and Accounting in 2024, Armann Ingolfsson is recognized for his research and for developing customized tools for decision-makers that address important planning problems—all with the goal of saving lives and increasing our quality of life.

As this year’s Winter term wraps up, I had the pleasure of speaking with Armann to learn more about how his research positively impacts individuals, the healthcare profession, and communities, locally and internationally.

_{Image: Business Dean Vikas Mehrotra and Winspear Chair Armann Ingolfsson discuss how his work is applied to wait lines in emergencies, EMS tactical planning, fatigue and service slowdowns, and helicopter ambulance bases.} 

Your research focuses on service operations management. Can you explain this research further?

A: Service operations management focuses on ways of delivering services more effectively and efficiently. A service system is congested if clients must wait. We have all experienced these waits: waiting to see a health care provider, waiting for a package to be delivered, or even just waiting to buy a cup of coffee.

My work involves creating mathematical models of congested systems to predict how an output changes if an input changes. How does an average waiting time change, for example, if the order in which clients are served changes? There are two types of algorithms at work here: algorithms to compute output predictions for given input values; and algorithms to find optimal input values.

A large portion of our nation’s workforce—approximately four out of five Canadians—are employed in services-producing sectors such as retail, education, and healthcare. My primary focus is on the healthcare sector in general, and emergency medical services in particular, where I feel my research can have a strong positive impact.

You are the School’s newest Chair appointment, the Winspear Chair of Accounting and Business Analytics. How do you see this award enhancing your research program?

A: I am honoured and grateful to be appointed to a chair named after Dr. Winspear, a business leader and philanthropist who was instrumental in the founding of our business school, and whose support has strengthened the economic, cultural, and artistic life of Edmonton and beyond.

This award will provide me with more of the two main ingredients needed to advance research: time and good collaborators. It will free more of my time to devote to research and allow me to employ research and teaching assistants, including one of our PhD students, Masoum Talebi. This appointment will also allow me to strengthen connections with the accounting profession. I am not an accountant by training, but I am eager to explore ways in which business analytics can be relevant to accountants, such as through environmental, social, and governance (ESG) reporting, fraud detection, and professionalism in reporting that includes data visualization and dashboards.

What type of tools have you developed to improve ineffective systems?

A: Many years ago I developed what I called the Queueing ToolPak to make it easier for analysts to apply basic mathematical models of waiting lines, also known as queueing theory. This tool allows analysts to type formulas into an Excel spreadsheet, rather than writing a computer program, which then allows them to compute such outputs as average queue lengths or recommendations for optimal inputs, like the number of servers. This tool is freely available and I have learned a great deal about the practical application of queueing theory through discussions with people who have used this tool.

More recently, we developed a dashboard that allows users to experiment with the design of an emergency medical service system, including decisions about where to locate helicopter bases, which pre-hospital time standard to use, and which hospitals to transport patients to. The dashboard is customized to inform a current debate in Iceland about whether to transport patients by ground ambulance, helicopter, or fixed-wing aircraft. 

Similar dashboards could be created to help policy makers and others to explore trade-offs in the design of emergency medical systems in many other jurisdictions, including Alberta.

_{Image: Ingolfsson’s customized dashboard showing possible locations for emergency helicopters in Iceland, and the resulting coverage.}

Our Alberta Health Services has areas that could benefit from this work. How do you see your work impacting their key areas of concern and being translated into practical solutions for patients and healthcare providers?

A: The models and tools that my collaborators and I develop are not magic bullets. The tools can be a valuable part of the solution for improving healthcare systems, but only if people in the healthcare sector find them credible and have the expertise to successfully implement systems that incorporate these tools in their processes.

I hope to contribute in two ways: first, by collaborating with and learning from healthcare practitioners in order to develop more credible tools; and second, by training students who are pursuing careers in healthcare analytics.

Have you ever come across a point in your research that was an ‘aha’ moment? Something that presented itself in a context that you never considered before?

A: About 25 years ago, I was working with colleagues on developing a simulation model of the ambulance service in Edmonton. The model had many moving parts, and we were struggling to make the model valid, in the sense that, if fed with the same inputs as the real system, the model would replicate the outputs of the real system to a reasonable approximation.

My ‘aha’ moment came when I realized that the single most important ingredient needed to obtain a valid model was a reliable formula for the time it takes an ambulance to drive from one point to another and that it was not obvious which formula to use. We rediscovered a simple formula that had been used 25 years earlier for fire-engine drive times in New York City and found that it also worked well for ambulance drive times in Edmonton and Calgary.  We have recently used the same formula to estimate air ambulance flying times in Iceland.

How are you engaging with our U of A students? Are you incorporating your research in the classroom and if so, how?

A: I have taught a variety of courses on operations management, statistics, and business analytics.  I am currently teaching a new course on data visualization. On the one hand, data visualization is a topic that I have had less formal preparation to teach than any of the courses I’ve taught in the past. On the other hand, it is a topic that has always fascinated me. I believe a well-designed chart not only tells a clear and interesting story, but it can also be a thing of beauty.

For example, take the image below, reproduced from my recent paper in Health Care Management Science. It shows an interaction between spatial and temporal patterns in EMS call volumes in Calgary, Alberta. The graphs on the right show the weekly temporal call pattern for each of five regions (bold lines), superimposed on the weekly temporal call pattern for the whole city (gray lines). The story that this image tells is that call volumes are elevated late on Friday and Saturday nights, but only in the downtown core (Region 1)—the times and the place where risky behaviours are more likely.

Every paper I have ever published has included charts, maps, or other data visualizations. A well-designed chart can be worth a thousand words. But a poorly designed chart merely leads to confusion. I hope to convey to my students some of what I’ve learned through my struggles to create effective charts for my research papers, and I hope it helps them to tell more informative and engaging stories about the findings and implications of their analyses.

What grand challenges do you see your work addressing at an academic level and practical level?

A: I see the incorporation of human behaviour in mathematical models of service systems as a grand challenge. All models are simplifications of reality. It is not true that a model that captures more of reality is always better. The real world is too complex for us to understand fully. Models can be useful if they help us understand more of reality, and they can do that only if they are simpler than reality.

That said, I do believe there are opportunities to create new models that are more useful for planning purposes because they include more aspects of the way in which real human servers and clients react to their environment. Capturing how paramedics change the way they do their work if they are fatigued, or if they are nearing the end of their shift and want to avoid working overtime, could lead to more accurate workforce planning. Developing such models is a challenge because it requires the use of a multi-disciplinary approach, combining mathematical modeling, empirical studies, and psychology.

Lastly, many research projects require a blend of rigorous methodology and creative insight. Where do you find your most fruitful sources of inspiration and creativity? Are there environments, activities, or people that help you find new approaches in your work, or that help you get the distance you need to spark a new idea?

A: Edmonton’s river valley is where I go for rejuvenation. A walk at the end of the day from the University to downtown is often what I need to find a solution to a dilemma that I have been pondering. 

As for people, all of the healthcare providers that I have worked with who have taken the time to share stories and knowledge about their work and their challenges inspire me.

Read more about Armann Ingolfsson and his latest research.