MONTREAL — The Quebec health-care system's reliance on fax machines, handwritten notes and paper files is standing in the way of the province fully harnessing artificial intelligence, according to a recent report.
And the network's ability to overcome its outdated practices could determine whether it becomes a global leader in health-care AI or languishes behind other jurisdictions, a digital health expert says.
Chief among the obstacles hindering Quebec's adoption of AI is the prevalence of "outdated or inadequate" tools in health-care facilities, says a commission charged with advising the government on issues related to science and technology.
In a report published last week, the commission pointed to the multitude of independent databases within the health network that lack robust metadata — built-in descriptors that help software navigate and interpret information. Without this data, the province can't deploy AI at scale because to do so computers would need to pull from vast, interconnected pools of uniform information, the commission said.
It also found that, in many instances, health professionals continue to rely on paper files, handwritten notes and scanned documents.
In a statement, Quebec's Health Department called modernizing the health network a "colossal task" involving "more than 10,000 technological systems supporting the work of hundreds of thousands of people."
Work is underway, it said, to "optimize workflows and reduce, if not eliminate, the use of paper and fax machines." It pointed to efforts to digitize health records in select establishments, end the use of faxes to send prescriptions to pharmacies, and create a digital appointment-booking platform.
But the commission encouraged Quebec to go further. John Kildea, a McGill University medical physics professor and leader of the Quebec SmartCare Consortium, agrees.
Kildea, whose group advocates for better medical data sharing, said "if Quebec is not able to step up soon … and to address the shortcomings of the data access in the health-care system, I think we risk falling behind."
That could mean missing out on some of the potential benefits of artificial intelligence, which the commission said could include the automation of administrative tasks, diagnostic assistance, prognostic modelling and illness prediction.
Analog data systems aren't only a Quebec issue: "the problem is universal" among health networks in Canada, explained Alexandra Greenhill, a Vancouver-based physician and co-editor of "AI in Clinical Medicine: A Practical Guide for Healthcare Professionals," recently published by Wiley-Blackwell.
Canadian health care, she said, is "two decades now behind all the other sectors of our society and economy" in digitization efforts. "That has to change."
Effectively upgrading health systems will require an expert workforce. In its report, the Quebec commission said the province will have to invest in IT expertise to prime the health network for AI. Among its recommendations are more funding for IT training and more emphasis on IT in university medical programs. It also said the province should prioritize the acquisition of user-friendly digital technology.
"Most physicians feel that the boat has sailed and they will not understand" AI, Greenhill said. "And so we have to invest in upskilling of the workforce so that they are ready to start using these amazing tools."
A final challenge for Quebec as it equips its health network with AI will be to ensure the technology's application is responsible and accessible. The commission warned that artificial intelligence could exacerbate inequalities in health care by alienating patients who are unfamiliar with digital technology, such as seniors, low-income people and rural residents.
If these populations are unable to use the digital tools that power AI solutions they could become under-represented in medical data, thus skewing AI results in such a way that ignores their needs, the commission said.
Greenhill and Kildea said health networks must privilege human input in AI integration to avoid such a scenario.
"If the government goes ahead and tries to implement" the commission's recommendations "by themselves, without involving the people who it matters to the most," Kildea warned, "they're in danger of achieving a solution that doesn't work for everyone."
He's nevertheless optimistic Quebec will be able to successfully mobilize AI tools in the health network and said the province is well poised to become a trailblazer in the field.
"If the government and the (Department) of Health are willing to step forward and to work on this, I think we have a chance to leapfrog" other jurisdictions, he said.
This report by The Canadian Press was first published Aug. 9, 2023.
Thomas MacDonald, The Canadian Press
