From detecting serious illnesses early, to reducing the administrative burden on doctors, health experts say artificial intelligence promises to change the way medicine is practised.

But if AI is the future of medicine, it is also part of the past in Ottawa, according to a leading artificial intelligence expert based at The Ottawa Hospital. Ottawa researchers have pioneered some predictive tools that are the foundation of AI.

That puts Ottawa’s researchers and health providers at the forefront of what many say is the coming AI revolution in health care, according to Dr. Doug Manuel, who heads the TD Artificial Intelligence in Medicine (AIM) hub at The Ottawa Hospital.

Manuel notes that Ottawa is well known for an algorithm that has helped doctors around the world accurately predict — for more than 30 years — which patients showing up in emergency departments with injured ankles have broken bones and need X-rays.

The so-called Ottawa Ankle Rules were published in 1992 by a team of doctors at the emergency department of the then Ottawa Civic Hospital, led by Dr. lan Stiell. Clinical testing over the years has confirmed that the rules accurately help doctors separate sprained ankles from broken ankles, reducing wait times and the use of X-ray machines. The rules essentially say that an X-ray is only necessary if there is pain in one of two specific zones of the foot and the patient can’t bear weight.

Manuel calls development and use of that predictive algorithm and others a strength of Ottawa researchers and clinicians. They are also foundations of artificial intelligence.

“This is our strength, we have a deep understanding of what it takes to get algorithms into practice. It is in our DNA, we have been doing it for so long.”

What is changing now is the torrent of data that is increasingly available to create predictive algorithms and other forms of AI in order to assist health-care providers with decision making and, crucially, to predict who is at most risk long before they get there.

Last week, doctors and researchers working in the field described AI as a potential game-changer for the way medicine is practised. They were part of a session organized by the Ontario Medical Association. Among them were researchers working on projects that use AI to predict who will develop Type 2 diabetes and which patients will suffer from delirium.

“The promise of what AI can deliver to the system is something that has galvanized people more than anything I have seen (in recent years),” said Dr. Fahad Razak, an internist and Canada Research Chair in Healthcare Data and Analytics at the University of Toronto.

Razak said AI hasn’t yet significantly influenced the way care is delivered at patient bedsides, but he expects things to change in the next five years. Among projects already underway are the use of artificial intelligence to predict who will develop Type 2 diabetes and AI to predict which patients are at highest risk to be transferred to intensive care and have their conditions deteriorate.

It is already being used by doctors at the University of Ottawa Heart Institute, says Dr. Peter Liu, who is the chief scientific officer and vice-president of research. Numerous other applications are in the research stage.

AI is being used to help doctors make more accurate diagnoses, more quickly, based on information gained from some diagnostic imaging of the heart. A computer can take the huge amounts of information available through imaging to quickly make a determination about whether a patient is in danger because of narrowing of the arteries, for example, and needs immediate treatment.

“The computer is able to take this huge amount of information. It can integrate it pretty quickly and be able to give us some sense of severity of the disease and help in terms of evaluation,” he said.

Work is underway to expand that application of AI in cardiology, among other projects. They include a project by Dr. Chris Sun, of the heart institute and uOttawa’s Telfer School of Management. He is using big data to help make sure defibrillators are placed in parts of cities where people are more likely to experience cardiac arrest. His work looks at risk predictors for cardiac arrest and placement of automatic defibrillators in Boston, said Liu. The work found a 60 per cent mismatch, according to Liu, meaning in the majority of cases defibrillators were not in the places they were most likely to be needed. He is now doing similar research in Ottawa.

Researchers and physicians say AI has significant promise as a key tool in preventative medicine as well as a way to focus resources where they are most needed.

Dr. Ibukun Abejirinde, a scientist at Women’s College Hospital in Toronto who is working on an AI diabetes prediction and prevention project, said AI can also potentially be used to predict who is going to develop cancer, Parkinson’s disease and other serious illnesses.

“That is what we foresee this work is going to evolve into. We are very excited about what the next two to three years will bring.” She said AI should enable the better planning and use of resources.

Dr. Andrew Park, president of the Ontario Medical Association, called AI a potential game-changer for Ontario health care.

“Through early detection of diseases and complications, we can save more lives. And we can make our health system more efficient and resilient.”

Abejirinde said there are issues to overcome, including patient concerns about the use of AI and questions about potential implicit bias built into algorithms, and who might be benefiting from the information. She said those concerns need to be studied carefully and data thoroughly validated.

Many doctors say they especially look forward to ways AI can help reduce the burden of administrative work that often keeps doctors’ eyes glued to a computer screen during patient appointments, and has driven some physicians out of practice.

Ottawa’s Manuel said physicians are being “buried by paperwork and forms.” He said the administrative burden on doctors is contributing to severe mental health issues and driving some away. AI, he said, has the potential to take some of that burden off doctors’ shoulders and “make medicine become more human for physicians.”

Among questions at the OMA’s session on artificial intelligence was whether AI could some day replace doctors. In response, OMA president Park, who described AI as a tool to “augment how we make decisions”, quoted the head of the American Medical Association.

“AI won’t replace doctors. But doctors who use AI will replace doctors who don’t.”

Liu of the Heart Institute acknowledged there is a worry about AI replacing cardiologists and other doctors.

“It is really helping us to be better cardiologists.”