Supporting AI Deployment with a Toolkit for Implementers

By
Kristie Jones
-
0

By Mario Voltolina

From Neuromancer, to HAL 9000, to Rosey the Robot, artificial intelligence (AI) has long captured the human imagination. But this fast-growing field has already made the leap from science fiction to computing reality. Not only is AI already impacting health care, its usage is growing — prompting many health care organizations to wonder how they can plan for, leverage and implement AI. With the right tools, these organizations can take the first steps on their AI journey.

AI has been defined as, “any current or future machine learning approach to predictive analytics, decision-support systems and/or automated decision making.”[i] That is, AI is computer technology that can mimic and support human intelligence. While the AI capabilities imagined by science fiction remain beyond our reach for now, its current applications can benefit Canadians’ health in three primary ways:

  • Improving the effectiveness, efficiency and safety of health care delivery;
  • Providing insights to inform both disease prevention and policies addressing broader population health determinants; and
  • Underpinning the discovery and development of new diagnostic tools and treatments.

In practical terms, an AI chatbot might respond to routine patient questions and triage their symptoms. Computer vision might process and interpret diagnostic images to detect the presence of tumours. Or an AI algorithm might assist with scheduling hospital resources based on predicted need.

Canada’s AI roots run deep. In 2017, Canada became the first country to develop a National AI Strategy through the Canadian Institute for Advanced Research (CIFAR); CIFAR subsequently initiated the AI for Health (AI4H) Task Force in 2020, publishing a report calling for action in the AI4H space in several key areas:

  • Establishing an AI4H “infostructure” to enable secure and responsible access to data
  • Accelerating the development and deployment of safe AI4H applications
  • Developing an AI4H strategy.

Moreover, 20 years of electronic medical record (EMR) adoption and consolidated health data sets in certain jurisdictions have contributed to good quality and comprehensive Canadian health data. These robust datasets mean that AI algorithms can produce more relevant, accurate and reliable outputs.

While Canada is well-positioned to act as a global leader in AI, lack of familiarity and comfort with AI in the health sector can lead to a “watch and wait” approach with respect to AI training, experimentation and operationalization. Small to medium-sized health care delivery organizations in particular may feel their enthusiasm outpaced by uncertainty about AI best practices.

After all, new opportunities bring new considerations. Health care delivery organizations need to understand and manage the new concerns that AI brings with it, such as issues of transparency, explainability and bias. For newcomers to the space, the path to AI deployment can feel daunting.

To help address these barriers and accelerate the adoption of AI enabled technologies, Canada Health Infoway (Infoway) has developed the “Toolkit for Implementers of Artificial Intelligence in Health Care.” Created in collaboration with INQ Law, a leading law firm and data consultancy with deep expertise in AI risk management, this toolkit serves as a “nuts and bolts” primer for health care delivery organizations just beginning to implement AI. It is intended for a range of roles and professionals, including AI project leads, CXOs, clinicians and other SMEs, IT leads and those involved with legal, privacy, security and compliance services.

Six modules — downloadable as one document or individually — cover key themes for health care delivery organizations to consider when beginning work with AI, from identifying strategic opportunities, to understanding emerging AI regulations, to guidance on AI governance. Checklists, best practices and case studies explore common AI considerations such as explainability, transparency and bias, and additional resources can be found throughout the toolkit’s bibliography and annexes.

Learn more and sign up to receive the toolkit. Webinars held throughout the winter of 2021/2022 will explore the toolkit in further detail. AI implementers can also contribute their suggestions to refine future editions of the toolkit and contribute additional Canadian AI case studies and use cases.

With its strong AI talent and knowledge, and the opportunities afforded by our single-payer health system, Canada can take its place as a global leader in AI for health care. As a country, we’ve laid the right groundwork — with the right tools and the right roadmap, Canadian health care organizations can begin their journey towards an AI future.

Mario Voltolina is CTO and Executive Vice President, Innovative Technologies, at Canada Health Infoway.

[i] Canadian Institute for Advanced Research. Building a Learning Health System for Canadians: Report of the Artificial Intelligence for Health Task Force. July 2020. <https://cifar.ca/wp-content/uploads/2020/11/AI4Health-report-ENG-10-F.pdf> Accessed 25 Oct 2021.