Imagine how much time would be freed up if technology could read for you. That’s what Dr. Joseph Ross Mitchell, recently appointed as the inaugural Alberta Health Services Chair in AI In Health at the University of Alberta, is working on.
“To use this new technology is a way to help make things more efficient for physicians so they can spend less time worrying about all the data that bombards them from different sources, and spend more time thinking about and communicating with the patient,” says Mitchell.
The AI techniques will use natural language processing to comb through text-based sources of information, so that clinicians and researchers can easily search through texts and spend more facetime with patients. But, creating AI tools for the health-care field doesn’t come without challenges, like having a knowledgeable team to support the research and finding funding. Yet, Mitchell notes that Alberta has all its cards in place to be a hub for AI.
“The thing that intrigued me about the job was the potential of what is likely to be available and what we can probably do in Edmonton that would be difficult to do anywhere else.”
Dr. Mitchell left Alberta and his role as professor of biomedical engineering, radiology and clinical neurosciences at the University of Calgary in 2011 to work as a professor of radiology at a Mayo Clinic campus in Arizona. He then took on the role as the inaugural artificial intelligence officer at the H. Lee Moffitt Cancer Center and Research Institute in Tampa before making the move back to Alberta in 2021.
“I was well aware of the folks in Edmonton and I knew the quality of the faculty that they had and the quality of the work that they were doing and was happy to be a part of it.” An important asset when working with AI in health care is having access to population-level health data. Over the last two years, Alberta has been rolling out Connect Care, a province-wide system that gives AHS healthcare providers, partnering facilities and patients a central access point for patient information.
“In artificial intelligence to train models, you need lots of data,” says Mitchell. “The more data, the better. And one of the biggest issues is that the data needs to represent the entire population, it can’t just be specific groups. Otherwise, you get bias in your algorithms.”
Mitchell notes it’s also crucial to have the support of a health-care provider and the assistance of a skilled university. Alberta has both. AHS has an AI team dedicated to developing and implementing AI into the field, and the University of Alberta has a strong computer science and engineering program. The campus also is home to Amii (Alberta Machine Intelligence Institute), a world class AI group of which Mitchell is a fellow.
“But, if you want to build a program, you have to find alternate ways to fund it,” says Mitchell. “And there are all sorts of several really interesting groups and agencies in Alberta that are interested in funding this kind of work.”
How long does it take for the algorithm to be built and implemented into the healthcare system? Mitchell says the rule of thumb is to overestimate what will be accomplished in the next two years and underestimate what will be accomplished in 10 years.
“These are the five major significant barriers that you have to address to even have a hope of tackling this, but for Alberta to do all five is just incredible,” says Mitchell. “What has been built in Alberta is incredibly rare and incredibly valuable.”
This article appears in the May 2022 issue of Edify