Skip to main content

Blog

Learn About Our Meetup

5000+ Members

MEETUPS

LEARN, CONNECT, SHARE

Join our meetup, learn, connect, share, and get to know your Toronto AI community. 

JOB POSTINGS

INDEED POSTINGS

Browse through the latest deep learning, ai, machine learning postings from Indeed for the GTA.

CONTACT

CONNECT WITH US

Are you looking to sponsor space, be a speaker, or volunteer, feel free to give us a shout.

Vector Institute kicks off series of Pathfinder Projects focused on health AI adoption

Machine learning being deployed at St. Michael’s Hospital to provide early warning for patients at risk of transfer to intensive care unit (ICU)

Toronto – Today, the Vector Institute, an independent, not-for-profit research institute focused on leading machine and deep learning, announced the launch of the first in a series of Pathfinder Projects to implement AI-assisted technologies in the health sector.

“These projects will showcase the positive outcomes that can be achieved if we leverage the power of AI in the health sector,” says Dr. Garth Gibson, Vector’s Vice President and CEO.

The first Pathfinder Project will support St. Michael’s Hospital in Toronto. Led by Dr. Amol Verma, an internist and clinician-scientist, and Dr. Muhammad Mamdani and their team at the Li Ka Shing Centre for Healthcare Analytics Research and Training (LKS-CHART) at St. Michael’s, the goal is to test and refine an AI-based early warning system for the hospital’s general internal medicine (GIM) unit where patients receive hospital care. About one out of every 13 patients in the GIM unit are critically ill and will ultimately need to be transferred to the intensive care unit (ICU) or will succumb to their illness in hospital. However, predicting which patients are likely to need ICU-level care is often difficult: that’s where AI comes in.

The system will use AI to process regular feeds of health data and predict when a patient needs to be transferred to the ICU. Accurately predicting when patients need to be transferred 12 to 24 hours earlier may allow more time for potentially life-saving early-intervention care, decreasing rates of cardiac arrest and mortality.

Pathfinder Projects are small-scale efforts designed to produce results in 12 to 18 months that guide future research and technology adoption. With technical and resource support from the Vector Institute, they each bring together a multidisciplinary research team to tackle an important health care problem or opportunity using machine learning and AI more broadly.  Each project was chosen for its potential to help identify a “path” through which world-class machine learning research can be translated into widespread benefits for patients.

“The Vector Institute and our faculty members are very interested in contributing to health AI implementation; we want to improve outcomes for patients and lower costs for providers,” says Dr. Alison Paprica, Vector’s VP, Health Strategy and Partnerships. “It’s our hope that these Pathfinder Projects inspire more teams within the health care system to focus on moving high quality health AI research into practice.”  

 

About the Vector Institute

The Vector Institute is an independent, not-for-profit corporation dedicated to advancing artificial intelligence, excelling in machine and deep learning. The Vector Institute’s vision is to drive excellence and leadership in Canada’s knowledge, creation, and use of AI to foster economic growth and improve the lives of Canadians.

The Vector Institute is funded by the Government of Ontario, the Government of Canada through the Pan-Canadian AI Strategy administered by CIFAR, and industry sponsors from across the Canadian economy.

 

Early Warning System for General Internal Medicine

Dr. Muhammad Mamdani, director of the Li Ka Shing Centre for Healthcare Analytics Research and Training (LKS-CHART) at St. Michael’s Hospital in Toronto work with the reality that one out of every 13 internal medicine patients at St. Michael’s Hospital in Toronto ends up either transferred to the ICU or succumbs to their illness. Further complicating matters, it is often difficult for doctors to predict the deterioration. “If you’re sick enough to need a hospital you’re pretty sick,” explains internist physician and research scientist Dr. Amol Verma, the team’s clinical lead. “If you get so sick that you need life supporting therapies, that’s when you need to be in an intensive care unit.”  

Working in their favour was the abundant data available to their team – hospitals are awash in patient data. Yet, despite all the expertise of the hospital’s clinicians, sorting through it all in a timely manner was next to impossible. “By the time that we realize these patients have a problem, we typically have about three hours to react,” says Dr. Mamdani.

“And three hours typically isn’t enough time,” adds Dr. Verma.

Though challenging, the problem was not insurmountable; Dr. Mamdani has been working with large data sets—what we now call big data—for over 20 years and the Li Ka Shing Centre specializes in data analytics. Given the large number of inputs that needed to be sorted, he and his team turned to machine learning. A subfield of artificial intelligence, machine learning is ideal for finding structure, patterns, and trends in large sets of data.

The team wrote an algorithm and trained it with anonymized health records from previous internal medicine patients at the hospital. As the algorithm takes in new data from current patients, it compares it against that of over 20,000 other previous cases, he explains.

The Early Warning System for General Internal Medicine (Dr. Mamdani admits their creation could probably use a catchier title) uses a predictive risk model to make medical recommendations. A “smart” computer system, it calculates the risk of a patient getting sicker and requiring transfer to the ICU.  “When it reaches a certain threshold it alerts the medical team,” he says. “Their glucose levels are high, that goes into our lab system, this information gets fed into an algorithm along with others, and it says, ‘Huh, there might be a problem here.’”

Only a handful of institutions across the globe are currently trying similar approaches to health care. Drs. Verma and Mamdani and their team will be among the first to test the quality and effects of its predictive power in a clinical trial. “It’s way more accurate than we’re used to with traditional methods,” he says. “It’s pretty powerful.”

Recognizing the trailblazing nature of the team’s work and its potential to produce direct, positive health care outcomes for patients, the Vector Institute is providing operational and technical research support, to maximize its impact.

With proof of concept in hand, the next step is integrating it into existing hospital systems, a task Dr. Mamdani admits is easier said than done. “How do you get it to work in a way that it provides physicians with helpful and meaningful information in an environment where alarms go off regularly?”

Both Dr. Mamdani and Dr. Verma agree that the key to their success, as well as what makes their project so unique, is the access to frontline care workers. They bring doctors, nurses and patients together to first find out what they need and then work within those constraints. “It’s much more effective to tailor an algorithm to clinicians’ workflows than to tailor clinician behaviour to how an algorithm performs.”

 

Early Warning System for General Internal Medicine is the first of a series of Pathfinder Projects identified and supported by the Vector Institute.

By Ian Gromely