Inside a bustling ward at St. Michael’s Hospital in downtown Toronto, one of Shirley Bell’s patients was suffering from a cat bite and a fever but otherwise seemed fine, but an alert from an AI-based early warning system revealed his condition was more serious than it appeared.
The nursing team would typically check blood test results around noon, but the technology alerted them a few hours before the results were due in. The alert indicated the patient’s white blood cell count was “extremely high,” recalled Bell, a clinical nurse educator in the hospital’s integrated medicine program.
The cause turned out to be cellulitis, a bacterial skin infection that, if not treated immediately, can lead to extensive tissue damage, amputation, and even death. Bell said the patient was quickly put on antibiotics to avert the worst-case scenario, thanks in large part to his team’s proprietary AI technology, ChartWatch.
“There are many other scenarios where a patient’s condition can be detected earlier, nurses can be alerted earlier and interventions can be made earlier,” she said. “This does not replace the bedside nurse, but actually enhances nursing care.”
A year-and-a-half-long study on ChartWatch, published Monday in the Journal of the Canadian Medical Association, found that use of the AI system reduced unexpected deaths among hospitalized patients by 26 per cent.
“We’re pleased that lives are being saved,” said co-author Dr. Muhammad Mamdani, vice president of data science and advanced analytics at Unity Health Toronto and director of the Centre for AI Research and Education in Medicine at the University of Toronto’s Temerti Faculty of Medicine.
‘A silver lining’
The researchers looked at more than 13,000 admissions to St Michael’s General Medicine ward – an 84-bed ward that treats some of the hospital’s most complex patients – and compared the tool’s impact on that patient population with thousands of admissions to other specialties.
“Over the same time period, we saw no change in these unexpected deaths in other units of our hospital that were not using ChartWatch,” said lead author Dr. Amol Verma, a clinician-scientist at St. Michael’s Hospital, one of three sites in Toronto’s Unity Health hospital network, and the Temelti Professor of AI Research and Education in Medicine at the University of Toronto.
“That was a promising sign.”
The Unity Health AI team began developing Chartwatch in 2017, based on a suggestion from staff that predicting death and serious illness was a key area where machine learning could have a positive impact.
Verma said the technology underwent several years of rigorous development and testing before being introduced in October 2020.
“ChartWatch measures about 100 inputs from (patient) medical records that are currently routinely collected in the course of care delivery,” he explained, “that is, the patient’s vital signs, heart rate, blood pressure, all the lab test results that are done every day.”
The tool works in the background alongside clinical teams, monitoring changes in a patient’s medical record and “dynamically predicting on an hourly basis whether that patient’s condition is likely to worsen in the future,” Verma told CBC News.
That could mean a patient’s condition worsens, requires intensive care or is on the brink of death, giving doctors and nurses a chance to intervene.
In some cases, these interventions include increasing the level of treatment to save a patient’s life or providing early palliative care in situations where the patient cannot be saved.
In both cases, the researchers said, ChartWatch appears to complement clinicians’ own judgment and help improve outcomes for frail patients and avoid sudden, potentially preventable deaths.
AI on the rise in healthcare
In recent years, artificial intelligence has attracted a great deal of attention and backlash beyond its use in the medical field.
From controversy over the use of machine learning software to churn out academic papers to concerns about AI’s ability to create realistic audio and video content that mimics real-life celebrities, politicians, or ordinary citizens, there are many reasons to be cautious about this emerging technology.
Verma said he’s long been wary of them, but he stressed that in health care, such tools have great potential to supplement traditional bedside care and combat staffing shortages plaguing Canada’s health-care system.
Artificial intelligence, often called the future of medicine, is already being rolled out in Canadian hospitals, but AI is far from perfect, and some are concerned about the costs involved.
Many of these efforts are still in their early stages: Various research teams, including private companies, are exploring how AI can be used to detect cancer early, one study suggests that simply listening to a person’s voice could warn of high blood pressure, and other research suggests that scanning brain patterns could detect signs of concussions.
Verma emphasized that ChartWatch is noteworthy because it has successfully saved real patients’ lives.
“There are very few AI technologies that have actually been deployed in clinical practice. To our knowledge, this is one of the first in Canada to be deployed in our hospitals to care for patients every day,” he said.
A real-world look at the impact of AI on healthcare
The St. Michael’s-based study also has limitations. It was conducted during the COVID-19 pandemic, at a time when the health care system was facing unusual challenges. The team acknowledged that the urban hospital’s patient population is unique, with many complex patients including those experiencing homelessness, addictions, and overlapping health issues.
“Our study was not a randomized controlled trial across multiple hospitals. It was within one organization, one department,” Verma said. “So I think we need to explore its use in different settings before we can say this tool can be used broadly everywhere.”
Dr. John Jose Nunez, a psychiatrist and researcher at the University of British Columbia who was not involved in the study, agreed that the study needs to be replicated elsewhere to better understand how well ChartWatch works in other settings, and he added that any use of new AI technology also requires consideration of patient privacy.
Still, he praised the research team for providing a “real-world” example of how machine learning can improve patient care.
“I really think of AI tools as becoming another team member on the clinical care team,” he said.
The Unity Health team hopes to see their technology deployed more broadly in the future, both within and outside their Toronto-based hospital network.
Mamdani, Unity Health’s vice-president of data science, said much of this work is being done through GEMINI, Canada’s largest hospital data-sharing network for research and analysis.
He said more than 30 hospitals across Ontario are working together, providing the opportunity to test ChartWatch and other AI tools in different clinical settings and hospitals.
“This will just lay the groundwork for expanding these facilities beyond our premises,” Mamdani said.
