Combining artificial intelligence and health analytics to transform health systems
Christopher Sun, assistant professor at the Telfer School of Management and scientist at the University of Ottawa Heart Institute, has been awarded the Tier 2 Canada Research Chair (CRC) in Data Analytics for Health Systems Transformation. The prestigious Canada Research Chairs program invests in world-class researchers to promote research excellence in Canadian postsecondary institutions.
Innovation: the key to transforming our health systems
As they evolve, technology and innovation continually enhance our lives and contribute to our well-being. They have the potential to transform many industries and improve the world as we know it. We use new tools daily, sometimes without even realizing it – why then are we so resistant to implementing new tools and technologies in our health-care systems?
According to Professor Christopher Sun, who was recently awarded the Canada Research Chair in Data Analytics for Health Systems Transformation, progress and innovation in healthcare stall partly due to barriers, such as siloed data, mistrust of algorithms, bias and discrimination in artificial intelligence, and lack of engagement with stakeholders. Although we like to think of Canadian healthcare as world-class, a recent report by The Commonwealth Fund ranked health services in Canada 10th out of 11 high-income countries, with poor ratings in access to care, efficiency, equity, and health outcomes. Professor Sun’s work aims to reduce and remove obstacles to improving the performance of healthcare systems, in Canada and globally.
A successful collaborative approach
Professor Sun came to Telfer armed with a Bachelor of Applied Science (BASc) in biomedical engineering and a PhD in industrial engineering from the University of Toronto, having also completed postdoctoral training in operations management at the Massachusetts Institute of Technology (MIT). His journey through different disciplines at renowned institutions enabled him to build an impressive network of specialists and experts in the field of healthcare, leading to collaborations with medical institutes across the globe.
One such collaboration is with two doctors at the Montefiore Medical Center: Dr. Michelle Gong, Chief, Division of Critical Care Medicine, Jay B. Langner Critical Care Service, Chief, Division of Pulmonary Medicine, Director of Critical Care Research, Department of Medicine and Dr. Ari Moskowitz, Medical Intensive Care Unit. The team, which is supported by the University of Ottawa Heart Institute (UOHI), was recently awarded a $1.45 million Project Grant from the Canadian Institutes of Health Research (CIHR) for a new research project on predicting and preventing in-hospital cardiac arrest.
Dr. Gong highlights Professor Sun’s rigorous approach and his desire to integrate the reality of healthcare settings into his research as very valuable: “This is an arena that is ripe for medicine and data science to come together. There have been studies published on this type of tool, but one of the challenges is that there isn’t a link between clinicians and data scientists. Professor Sun works closely with physicians; he actively seeks information and feedback from clinicians, which impressed me and my colleagues.”
Though still in the early stage of his career, Professor Sun has established an international reputation as an innovative and impactful researcher. Some highlights and recognition of his work include:
- Numerous articles introducing creative solutions published to international acclaim in top-tier journals such as Circulation, the Journal of the American College of Cardiology and Health Affairs.
- Recognition from the American Heart Association (AHA) for his work on optimizing the placement of public access defibrillators (PAD) to improve cardiac arrest outcomes.
- Prestigious awards, namely the CIHR Vanier Scholarship, the AHA Young Investigator Award, and the Citizen CPR Foundation (CCPRF) Excellence in Education Award.
A diversity of perspectives
As a student, Professor Sun moved through different disciplines, starting in biomedical studies then shifting to systems optimization work in healthcare analytics. In combination with his experience leading interdisciplinary research in medical practice and health policy, the diversity of his experiences provides a fresh perspective on how to successfully implement healthcare transformation.
In addition to his focus on cardiac health, Professor Sun has woven the principles of equity, diversity, and inclusion (EDI) throughout his studies and research projects. His commitment to EDI is essential in healthcare analytics, a field deeply intertwined with public well-being. It’s also evident in his professional development and community service activities. One of many examples is his work supporting underprivileged high school students through outreach programs; another is his collaboration with Girls Leadership in Engineering Experience at the University of Toronto, an event that supports women in engineering. He is also currently a member of Telfer’s EDI committee, providing the committee with valuable insight gained from his vast experience.
Heart-healthy collaboration
Cardiac health holds a special place in Professor Sun’s heart; it’s not surprising then that one of his closest partnerships is with the world-renowned University of Ottawa Heart Institute (UOHI), where he holds the position of scientist. The UOHI is one of the largest heart health centres in Canada, treating over 2.5 million cardiovascular patients in Eastern Ontario and Western Québec.
One of the challenges of such a large medical institution is trying to balance efficiency and outcomes: how do you allocate resources fairly and adequately while simultaneously ensuring that the needs of individual patients are prioritized?
According to Dr. Peter Liu, Chief Scientific Officer and Vice President of Research at the UOHI, this requires knowledge of business, artificial intelligence, and data science, which is precisely what Telfer brings to the table. Professor Sun and Dr. Liu agree that the collaboration between Telfer and the UOHI will make smart use of the power of data to solve real-world problems. The UOHI has been gathering data on its large patient population for over 10 years, which Professor Sun will be able to use in his research.
Professor Sun explains: “This unique collaboration between Telfer and UOHI, where innovative research at Telfer can be directly applied at UOHI to enhance patient care and healthcare services, is very exciting. Working with forward-thinking clinicians and agile health institutes that support data-driven decision-making will help close the gap between theory and practice in healthcare analytics.”
Dr. Liu believes that the collaboration with Telfer is very advantageous for the institute and for healthcare more broadly, noting Telfer’s specialization in health systems management. “Health institutions are often run like businesses, but as a physician, in medical school, no one teaches the fundamentals of business administration,” he says. “Physicians are busy providing medical care for patients, that’s our knowledge, our specialty; we think of it one patient at a time. A business school is a fantastic partner, since they bring knowledge of systems organizations, performance management, optimization – all things that can benefit healthcare worldwide.”
Learn more about practice-changing research at the Heart Institute.
New research will lead to new tools
The CRC in Data Analytics for Health Systems Transformation will allow Professor Sun to continue his efforts to reduce inequity and transform our healthcare systems. His goal is to develop and implement prescriptive and predictive models and tools supported by artificial intelligence, involving many stakeholders in the process, and using their lived experience and knowledge to build effective tools.
When we think of using artificial intelligence to improve our healthcare systems, we might think of ChatGPT or robots that can perform surgery. As it turns out, artificial intelligence has many different applications; where ChatGPT is a generative tool, Professor Sun’s work tends toward predictive AI and optimization models that will help support decision-making.
In the field of data science and health analytics, as in many others, there can be resistance to implementing and using innovations borne of machine learning. To help lessen that resistance, Professor Sun is prioritizing co-development, a research method that focuses on the needs of “knowledge users” when creating a research project.
Knowledge users, or end users, are the people who will be most affected by the research. In healthcare, knowledge users could be patients, physicians, nurses, technicians, or administrators. Co-development requires engagement with all the relevant stakeholders to ensure their needs are considered and addressed. Involving them in multiple stages of the research helps build confidence that these new tools will benefit them, as well as healthcare in general. Professor Sun explains: “In high-stake domains like healthcare, it’s important for human experts to be actively involved in decision-making processes. So we’re really trying to build analytical models to help them be more efficient and more effective so they can reach more patients, which requires their valuable input during model development.”
Tools to improve cardiac event outcomes
Professor Sun’s newly awarded CRC outlines three main research objectives, all of which involve some measure of human-algorithm collaboration:
- Prediction and prevention of emergency cardiac events
- Cardiac imaging diagnostic support
- Automated operating room scheduling
In line with his interest in cardiac health and his role at the University of Ottawa Heart Institute, one of Professor Sun’s research objectives is to develop models to help predict and prevent emergency cardiac events. For example, given that Canadians who experience a out-of-hospital cardiac arrest face survival rates of less than 10%, our healthcare system is in dire need of such tools.
His second research objective is to develop and assess the value of cardiac imaging diagnostic support tools that will be specifically tailored to UOHI’s patient population. Given the challenges in accessing data, the model will first be built using publicly accessible international data, and then adapted to the UOHI’s own datasets.
Medical professionals at UOHI perform many surgeries, which requires complex operating room scheduling. Enter Professor Sun’s third research objective: to create an optimization model to automate operating room scheduling for cardiac surgery. This tool would consider the many factors involved, while also balancing fairness and efficiency, to ensure that different patient groups are treated equitably. This tool would also improve efficiency by preventing staff shortages and reducing the time required to plan operating room schedules.
Such tools, once implemented, could improve patient outcomes, reduce the burden on hospitals, combat inequity in cardiac health, and lessen the negative impact on first responders who attend to emergency cardiac events. They would also demonstrate the value of creating direct links between data systems and allowing greater access to health data across our health systems.
Better health outcomes for everyone
For Professor Sun, working in the field of healthcare analytics has highlighted the many inequities in our health systems. “It’s important to incorporate our knowledge of health disparities in everything,” he said. “Specifically in cardiac health, there’s very prominent evidence of disparities across sex and gender; that’s something to consider when building these predictive models and decision models.”
Moreover, even though age, disability, socio-economic status, race, and ethnicity are some of the factors that can contribute to health inequities, patient files often lack this information. Data is one of the main challenges in healthcare research: it is difficult to access, often limited to one institution, and usually lacks equity-related information, such as patient race and ethnicity. If we don’t collect data on the factors that lead to health disparities, how can we eliminate them?
Data issues are widespread in healthcare around the world but are particularly challenging in Canada. Each province or territory manages their own health system, and data is rarely shared between institutions, let alone across the separate systems.
To help combat health inequities from a healthcare analytics perspective, it is essential to take steps to address biases while creating automated tools. One of the challenges to implementing AI models in healthcare is the existence of implicit bias. If biased humans are building the models, how can we ensure they aren’t building the biases into the models?
As a data scientist, Professor Sun is acutely aware of this issue and is actively working to identify underlying problems in order to solve them. Being aware of the problem is a great start, but it’s far from sufficient. The work of combatting implicit bias requires interdisciplinary collaboration.
One of ways Professor Sun aims to do this is through international collaborations that enrich and diversify the data and expertise needed to develop these models. One of his collaborators, Dr. Gong, echoes this importance: “Montefiore is a very diverse hospital. Half of our patients are Hispanic or Black; we service the underserved. That brings a certain diversity to the patient population, and to the research, which is important if we want this tool to be generalizable, scalable to other parts of the world.”
Furthermore, Professor Sun intends to publish one optimization model as an open-source tool so that smaller or differently focused health centres can also benefit from its use. This form of knowledge dissemination and public sharing of this resource will also help prevent the data siloing that is so prevalent in healthcare.
Purposefully integrating diversity and inclusion into the design of the research is a testament to Professor Sun’s strong desire to build tools and products that will be useful to multiple segments of the population, as well as his efforts to reduce inequitable health outcomes. Hopefully, these measures will also encourage healthcare providers to collect more thorough data on the factors that contribute to inequities in healthcare.
Leading by example
Professor Sun has a strong desire to contribute to the improvement of healthcare systems and the health of Canadians, as evidenced by his research and collaborations. In demonstrating the planning and effort invested in building fair models and in making his research publicly available, he is raising awareness of existing challenges in healthcare as well as showing the world ways to overcome them. His work illustrates the value of collecting and sharing diverse patient data and is likely to reduce resistance to the implementation of AI-supported models, which will lead to positive outcomes for a greater number of people.