It’s time to move beyond curiosity: AI in Canadian Healthcare

An AI tool capable of reducing unexpected deaths in hospitals by 26%? Tell me more.?

A recent article in the Canadian Medical Association Journal (CMAJ) highlighted that machine learning-based early warning systems for patient deterioration are linked to a lower risk of non-palliative death. Read all about it here: Clinical evaluation of a machine learning–based early warning system for patient deterioration | CMAJ

This stuff is pretty neat. AI algorithms can analyze vast amounts of medical data, including patient history, imaging studies, and lab results, to identify patterns that may indicate the early stages of diseases such as cancer, heart disease, and diabetes. AI can facilitate timely interventions that have the real potential to save lives.

So why aren’t we seeing more AI in Canadian healthcare?

Canada seems to be quite risk-averse on this front. While there has been some experimentation with AI-powered tools, these initiatives have generally been limited in scope or conducted as pilot projects. When we compare where Canada is at with other countries, there is a widening gap.

AI has numerous applications in healthcare, including clinical decision support (CDS), radiology, robot-assisted surgery, nursing assistance, administrative workflows, fraud detection, dosage error reduction, connected electronic medical records, clinical trial participation, improved diagnostic accuracy, radiologic image interpretation, and cybersecurity. We have already seen other countries successfully adopt this technology:

Singapore

Singapore's electronic health record system is now AI-driven. It allows for the analysis of patient data, enabling predictive analytics to identify at-risk patients and streamline care. This adoption of AI is part of the country’s broader initiative to use advanced technologies to improve public services, including healthcare.

Another notable application is the use of AI in medical imaging and diagnostics. The Singapore National Eye Centre developed an AI system that screens for diabetic retinopathy, glaucoma, and macular degeneration. This tool has improved the speed and accuracy of diagnosis.

United Kingdom

The UK has been at the forefront of AI integration in healthcare through initiatives like the National Health Service (NHS) AI Lab (2019). This lab focuses on accelerating the safe adoption of AI technologies to improve patient care. One significant AI application is in medical imaging, where AI algorithms are used to detect conditions like cancer, and heart disease earlier and more accurately than traditional methods.

Many other countries come to mind, including Saudi Arabia, Israel, the United States, and China. What stands out to me is that these nations are enhancing access to services while significantly reducing administrative burdens and inefficiencies.

So what do we need to do to move beyond curiosity?

Digital Foundation

Because AI systems need to be ‘trained’ using data garnered from digital health systems, a mature digital foundation is necessary to develop and implement data-hungry AI solutions.

Canadian healthcare providers may need to strengthen their digital foundations. Healthcare data is often stored in disparate systems across provinces, hospitals, and clinics, often using different standards and formats. This fragmentation may limit the ability to aggregate and analyze large datasets needed for AI.

From what I have seen in my work, radiology may be one of the most suitable areas for AI adoption in Canadian hospitals and clinics. This is largely due to the extensive data standardization that has taken place in the field. Since the early 1990s, key elements such as data standards, interoperability, and data collection practices—foundations for big data analysis and AI—have been well-established. Accepted data formats are in place.

In many of the countries noted above, AI is already being used in radiology due to the abundance of standardized historical data available for AI to learn from.

The Human Factor

The human factor is key in bringing AI into the healthcare sector.? Countries that already have strong science and technology-focused education systems will have an advantage. Canada needs to create incentives for attracting top talent. We have some of the leading research organizations focusing on AI. However, there is a lack of focus on translating the technical to the clinical at scale.

Workforce readiness is also a key factor. Healthcare professionals may not be adequately trained to work with AI tools or understand how to integrate them into clinical practice. Building AI literacy and trust among healthcare workers is essential.

And of course, speaking of trust, Canada has data privacy regulations that may create some constraints on how healthcare data can be collected, stored, and shared. Navigating these regulations while implementing AI-driven solutions has the potential to pose challenges.

Doctors are already concerned about this, with a very recent survey finding that more than three-quarters (77%) of Canadian doctors agree that AI use should be subject to government or medical association oversight. 81% called for a formal legal framework to regulate AI in medicine.

While this is the case, the same study found that the majority of doctors surveyed are enthusiastic about integrating AI into their practice, citing benefits such as the potential reduction of medical errors. Canada: Physicians and AI Report 2024 (medscape.com)

The move beyond curiosity

How do we transform the Canadian healthcare system in ways we have not even thought about yet? Ultimately, to truly drive the business impact of AI across any sector, we need transformers—professional service providers and innovators who bridge the gap between technology, customers, and real-world business use cases.

Beyond curiosity, we need the courage to embrace new ideas and concepts.?