Healthcare is among the industries in the forefront of AI for All, Canada's new national AI strategy, which Canadian Prime Minister Mark Carney launched on June 4. While Canada has world-class talent and one of the fastest-growing digital sectors in the G7, it is among the slowest countries to adopt AI at scale, according to the prime minister's announcement. "This gap risks undermining public trust, driving Canadian talent and startups abroad, and leaving critical parts of our AI ecosystem under foreign control." 

AI can shorten emergency room wait times and make small businesses more competitive, Carney added. "That's why we need an ambitious new strategy: AI for All. We will build trust so that all Canadians are empowered to use this technology safely and with confidence."

In keeping with its overarching mission, AI for All included the launch of the first "AI Missions Program," which is intended "to accelerate the adoption of AI in diagnostics, patient care, and system efficiency, delivering faster, better care for Canadians while strengthening Canada's life sciences and health innovation ecosystem."

Two projects now underway show the power of health AI as it moves from policy to practice. One is building the national data infrastructure that makes health AI possible, and the other is deploying AI tools at the bedsides of young and vulnerable patients. 

Building the Infrastructure: Vital

Health AI requires vast amounts of diverse, comprehensive, and privacy-protected data. That's the premise behind Vital, a pan-Canadian health data platform co-led by Fahad Razak, MD, an internist, and Amol Verma, MD, an assistant professor in general internal medicine, both at St. Michael's Hospital and the University of Toronto.

Vital builds on a decade of work achieved with its predecessor, GEMINI, the largest multi-institutional hospital data-sharing network in Canada. That public sector platform now holds data on more than 3 million Canadian hospitalizations and covers approximately 60% of Ontario's hospital patients, Razak told Medscape News Canada. 

As the federal government was developing its health AI strategy, "it recognized the incredible value of this kind of data in making AI safe and accessible in the health sector for Canadians," he said. As a result, in partnership with the provinces, "the government made the largest investment in Canadian history to create this next generation of our platform."

Vital expands the data-sharing model nationally, connecting more than 160 hospitals across Ontario, Alberta, and Quebec that serve more than 20 million Canadians. There are plans to extend the model to additional provinces and territories.

Countries like Denmark and the United Kingdom have also invested in ways to harness health data at scale, and Canada has been losing ground, Razak said. "Certain kinds of innovation are moving to other countries. For example, the large-scale randomized trials that Sanofi is doing around the most protective new vaccines for influenza or respiratory syncytial virus are being run in Denmark instead of Canada." 

In addition, "the United Kingdom now allows their data to be used in cancer screening, and so hundreds of thousands of people there are getting early cancer screening using the newest technologies," he said. "We don't get that here in Canada at scale."

Vital is designed to change that situation by giving researchers the infrastructure to run clinical trials, test AI algorithms, and reach patients who have historically been left out. Cancer trials and heart disease trials, for example, "systematically underrepresent women, less-educated groups, smaller communities, northern communities, and indigenous communities," Razak said. "Vital can use digital health records to identify every eligible patient across the country and offer them the chance to participate in a trial, regardless of geography."

That capability also may mitigate bias, a key concern with health AI. Algorithms trained on narrow, homogeneous populations can fail when deployed in diverse ones, and Canada "is perhaps the most diverse high-income society on Earth," Razak said. "So, one of the important things about a structure like Vital's is that it gives us the chance to take a tool that's promising and revalidate it in the populations it's being deployed in to make sure that it can work well, free of bias."

"Healthcare is the sector where AI stands to have the greatest benefits," Verma added. "If implemented well, AI will help people live longer and healthier lives, help healthcare workers avoid burnout, and help make our healthcare system more accessible."

At scale, he told Medscape News Canada, "Vital will allow health AI technologies to be developed by and for Canadians and allow us to evaluate health AI technologies in Canadian settings to ensure they are safe and effective before they are used."

Deploying AI at the Bedside: SKAI

While Vital is building infrastructure, SKAI (SickKids Artificial Intelligence) at Toronto's Hospital for Sick Children is putting AI to work with the aim of getting young patients to diagnosis, treatment, and when appropriate, discharge, as fast as possible. 

"SKAI represents one of the world's first enterprise-wide AI initiatives that functions in two ways," Devin Singh, MD, co-lead of SKAI and an emergency physician at SickKids, told Medscape News Canada. "It serves as a deployment arm for AI technologies across the hospital and it also governs AI use, defining what responsible, ethical, and equitable AI is and writing the policies and procedures to implement it.

This summer, SKAI is testing a new use of the technology with the launch of what Singh says is a "world-first" clinical trial. After a patient goes through emergency department triage, AI models will process their data in real time and, when appropriate, automatically order a test for appendicitis before a physician even sees the child. 

With this approach, he said, "as an emergency physician, I know that when I'm focusing on the patient in front of me, the kids who are in the waiting room are also starting to get the care and diagnostics they need. When I get to see them, I already have a diagnosis of appendicitis in hand. I can accelerate them to a surgical intervention or antibiotics, and this dramatically cuts down the time a patient is sitting, waiting for this to happen."

An AI tool will never catch every case of appendicitis, but AI is good at capturing 20% to 40% of cases, Singh continued. "So, we're going to use that tool in the trial to catch the low-hanging fruit." The other patients will go through the normal workflow; they'll be seen by a clinician and receive the standard of care.

Other SKAI tools are also coming online, including a model predicting individualized patient wait times in the emergency department, an ICU model that can flag impending clinical deterioration before a cardiac arrest occurs, and a precision medicine model that combines genetic and socioeconomic data to identify children who might benefit from genetic testing.

"There's a groundswell of AI solutions, and to become a truly AI-powered children's hospital, we need to harness that groundswell in safe and responsible ways," Greg Kennedy, director of AI operations at SickKids, told Medscape News Canada. "Our central AI program is looking at those tools, building them out in more robust ways, and scaling them to all the areas in the organization that can benefit from that solution."

What underlies the program is a commitment to something more fundamental than technology, Singh added. "If we keep our motives tuned to genuinely trying to do what's best for our patients, and if we keep that human spirit the entire time, we will see that responsible, equitable AI is what makes it to the bedside. And If we get it right, AI can actually make healthcare more human."

Razak declared being a co-lead for Vital and a recipient of federal funding for the initiative. Verma is a part-time employee of Ontario Health and holds a minority interest in Signal1, a health AI company, for co-inventing an early warning system for patient deterioration. Singh and Kennedy reported no relevant financial relationships. 

Marilynn Larkin, MA, is an award-winning medical writer and editor based in New York City whose work has appeared in numerous publications, including Medscape Medical News and its sister publication MDedge, The Lancet (where she was a contributing editor), and Reuters Health.