Canadian healthcare workers using AI for assistance

Imagine receiving a notification that saves a life. This is what happened to Dr. Yuna Lee during her shift at St. Michael's Hospital in Toronto, thanks to CHARTWatch — an AI-powered early-warning system. The system alerted Dr. Lee about a patient at high risk, leading to timely interventions that uncovered a serious medical issue.

Dr. Lee's reaction? "That was very surprising... CHARTWatch is amazing."

What's CHARTWatch all about?

Developed by the hospital's own data science team, CHARTWatch analyzes hundreds of patient data points, providing hourly risk scores. Its implementation has led to a significant 26% reduction in the relative risk of death among non-palliative patients. This is a testament to the system's effectiveness in enhancing patient care.

Despite the global acclaim of Canadian AI research, the integration of AI in healthcare has been gradual. As Bo Wang, Chief AI Scientist at Toronto's University Health Network, notes - AI tools in Canada's medical system are still rare. The challenges? Limited investment in necessary computing power, privacy concerns, and potential data biases.

Thanks to a significant donation from businessman Li Ka-Shing, St. Michael's has been experimenting with about 50 AI solutions. These range from nurse assignment tools in the emergency department to systems that provide wait-time estimates and synthesize complex medical records for quick review.

Also worth noting is that the development process at St. Michael's involved continuous interaction with end users, ensuring the AI models remain effective and accurate.

Meanwhile, Health Canada has been cautious, approving only AI devices with locked algorithms to ensure safety and reliability. In contrast, across the border, the U.S. FDA has been more progressive, approving a variety of AI-enabled medical devices, particularly in radiology.

Down the road

Large Language Models (LLM) for healthcare, like those developed by Google and Microsoft, are turning into sophisticated AI solutions for medical inquiries and clinical notes. However, there's a growing concern about leaving these powerful tools in the hands of corporate giants who might prioritize profits over accessibility.

As we move forward, the integration of AI in healthcare presents a delicate balance between innovation and regulation. While the technology promises to enhance patient care and ease the workload of healthcare professionals, it's crucial to navigate the challenges carefully to ensure that AI remains a boon, not a bane, in the medical world.