Technology
Leveraging AI to provide better health care for Canadians
September 20, 2024
Share
Xiaodan Zhu is a Mitchell Professor and expert in machine learning, natural language processing, and AI, and is working on large language models at the core of the new digital health care platform as the Queen’s lead for the CareAI project.
While training more doctors and nurses is fundamental to relieving Canada’s current health care crisis, technology can also play an important role, such as in telemedicine and patient contact centres – digital platforms that can manage interactions between patients and health care providers via emails, phone calls, and chats. The increasing use of these tools in private and public practice is a good example of how technology can expand health care capacity.
However, these platforms still demand human-power in patient intake and management, and therefore continue to be vulnerable to traditional labour shortages while facing a growing demand for digital health services.
Part of the solution may be found in artificial intelligence (AI), as AI can help address vulnerabilities and scale up patient contact centres by automating and streamlining workflows in digital health care platforms. This possibility motivated Queen’s and industry partners to jointly develop an AI-driven commercial platform to manage both virtual and human agents for telemedicine and patient contact centres in Canada and the U.S.
The CareAI project is led by Calabrio Canada, a workforce performance company, in collaboration with AI-driven solutions developer ORX and tech-enabled health care provider WELL Health Technologies. The initiative secured $14 million from , Canada’s global innovation cluster for digital technologies, which will be matched with an additional $30 million investment from industry partners.
“By partnering with our industry collaborators and DIGITAL, we are excited to research emerging generative AI models, particularly large language models (LLMs), to help transform health care patient contact centres. This is made possible by recent significant advancements in AI technologies for processing both unstructured and structured data,” says (Electrical and Computer Engineering), AI lead at the and the Queen’s lead for the CareAI project.
CareAI will address challenges such as the need for reliable reporting mechanisms, reducing errors in data entry, and providing always-on services such as patient scheduling and nursing triage.
The new platform will consist of two modules, the Unity Agent and the Workforce Unity Platform, that can be deployed together or individually. The Unity Agent will automate routine administrative tasks, engage patients through improved communications and support clinical decision making by providing actionable insights based on collected data. The Workforce Unity will facilitate the integration of the Unity Agent and human resources by integrating different aspects of patient care and administrative processes, from assessing medical history to follow-up after a treatment.
The goal is to provide both patients and health care providers with world-class experiences, while also reducing operational costs. At the same time, CareAI will address challenges such as the need for reliable reporting mechanisms, reducing errors in data entry, and providing always-on services such as patient scheduling and nursing triage.
Queen’s will provide academic and research support to the industry partners, ensuring the project benefits from the latest scientific research and best practices in healthcare and AI. A Mitchell Professor and expert in machine learning, natural language processing, and AI, Dr. Zhu has been working on large language models (LLM). Based on large amounts of data, this type of AI application is capable of understanding and generating human-like language and will be at the core of the new digital health care platforms.
“By combining the strengths of generative AI and human experts – what we know as human-in-the-loop technologies – we aim to deliver effective and robust solutions, and we will explore trustworthy machine-learning models that adhere to privacy requirements,” says Dr. Zhu, who is also a faculty affiliate at the .
While this is a made-in-Canada solution, the team expects that CareAI will also be applicable to other countries, and could have an impact in other health care-related markets, including insurance. The goal is to have a pilot platform available for commercialization in 2025.
To learn more about CareAI, .
