This article originally appeared in the Fall 2022 issue of the Ontario Medical Review magazine.
AI and machine learning touch every aspect of medicine, from doctor's office to human resources
The technological future of medicine is in harnessing data so physicians can provide better care. Health-care providers in Ontario’s medical system, from general practitioner offices to surgical suites, hospital administration to system-wide planning, see a revolution unfolding in artificial intelligence and machine learning tools that use data.
Dr. Amol Verma sees two possible versions of a data-driven future. The general internist and health services researcher at St. Michael’s Hospital and assistant professor at the University of Toronto said now is the perfect time for the medical community to choose a path that leads to a more sustainable, humanistic and compassionate system.
“My most hopeful vision for health care 10 to 15 years from now is that clinical decisions will be guided by smart digital technologies, some of which will be based on artificial intelligence tools that help clinicians access the right information to treat their patient at the right time,” Dr. Verma said.
Those tools would reduce the daily grind of filling out forms, documentation and billing that lead to burnout, he said. Physicians would spend more time with patients and be better decision-makers.
“There is also a less optimistic version of the future where increasing focus on data collection and digital solutions leads to more workload and documentation demands,” he said. Dr. Verma adds that involving clinicians, patients, family members and caregivers in developing new technologies would make a brighter future more likely.
Dr. Verma is co-lead for GEMINI, a data collection and analysis study based at St. Michael’s Hospital. The project aims to improve the quality of care delivered in general internal medicine units, where older patients often come in with a number of co-existing conditions.
Researchers saw an opportunity to harness data from hospital electronic medical records, including the rooms patients stayed in, the doctors who cared for them and the tests, treatments and medications they received. GEMINI now receives data from more than 30 Ontario hospitals and provides customized feedback and insights on practice to 600 physicians. It has become one of Canada’s largest digital health platforms for research.
GEMINI found that most patients hospitalized with stroke receive a heart echocardiogram and ultrasound. But in about 90 per cent of cases, that echocardiogram doesn’t change the patient’s care and requires two extra days in hospital, on average. That suggests an opportunity to reduce echocardiograms and length of patient stay, said Dr. Verma.
His team has also developed an artificial intelligence solution that will use EMR data to identify hospital patients with delirium and predict their risk of developing delirium in the future.
“I’m hopeful that it will allow us to identify the areas in hospitals, or the specific hospitals, that have higher delirium rates, where we can target delirium prevention strategies and resources,” he said.
This is an example of an AI tool that reduces the burden on the physician, using data already available to prevent a condition that debilitates patients, said Dr. Verma.
Physicians are on the front line of developing these tech solutions. “A typical clinical practice will undoubtedly be touched by this data revolution and the average
doctor of tomorrow, or frankly of today, needs to be data-savvy in a way that we didn’t need to be a decade ago,” he added.
(Left to Right) Dr. Muhammad Mamdani, vice-president of data science and advanced analytics at Unity Health Toronto; Dr. Chandi Chandrasena, chief medical officer at OntarioMD; Dr. Amol Verma, co-lead for GEMINI, a hospital data and analytics network based at St. Michael’s Hospital, Unity Health Toronto, discuss the importance of advancing technology in medicine.
Dr. Muhammad Mamdani, vice-president of data science and advanced analytics at Unity Health Toronto, sees the power of physician involvement in future technology.
Unity Health’s St. Michael’s Hospital uses artificial intelligence and machine learning in a number of areas and new technology projects begin with problems and issues raised by front-line clinicians, said Dr. Mamdani, who is a doctor of pharmacy. Those end users commit to working through solutions with the hospital’s team of 30 data scientists and then champion the new technology in their departments.
St. Michael’s uses AI and machine learning from bedside to administrative suite. The hospital’s CHARTWatch system is an example of artificial intelligence’s direct benefit to clinicians and patients.
“It monitors all of our patients and it predicts within the next 48 hours if this patient is going to die or need ICU transfer,” Dr. Mamdani said. That allows the clinician to identify patients at higher risk who need intervention to prevent deterioration.
"We want the best for our patients…Without this type of information, without data about what we do right and what we do wrong, it’s very difficult to improve." — Dr. Teodor Grantcharov
Nursing staffing is another major challenge for the hospital. Machine learning can help predict the pool required for nurse resource teams, which relieve nurses who take leave, call in sick or go on vacation. “That’s going to minimize costs and maximize the number of nurses that you need for efficient care that is optimal for patients,” he added.
About 70 per cent of the problems hospital staff bring to his team eventually find a solution, said Dr. Mamdani. And others will be solvable in the future with more precise data. One such future solution will be emergency department flow.
“Can we actually say ‘Hey this patient is going to be waiting for eight hours...Maybe they can go home, come back in the morning.’ We can schedule them in at 9 a.m. and save them the stress and the anxiety of being in the emergency department when they don’t need to be.”
But the data needed for that to happen is still “messy,” Dr. Mamdani said. He believes as more structured and modern patient records are developed and data collection is refined, technology will change how patients are managed in emergency departments.
Tech solutions get buy-in from health-care teams quickly when they prove their value. Take the OR Black Box, similar to the familiar aviation black box, invented by Dr. Teodor Grantcharov, professor of surgery at the University of Toronto and Keenan Chair in Surgery at St. Michael’s Hospital.
Dr. Grantcharov said the biggest supporters of the technology, which monitors and collects a wealth of data during surgery, started off being resistant or anxious.
“The same individuals become…the biggest supporters because we’re all competitive people. We want the best for our patients…Without this type of information, without data about what we do right and what we do wrong, it’s very difficult to improve,” he said.
He compares the data, including video and audio records, to tools commonly used by coaches to improve athlete performance.
Dr. Teodor Grantcharov, professor of surgery at the University of Toronto and Keenan Chair in Surgery at St. Michael’s Hospital, invented the OR Black Box that monitors and collects data during surgery to measure and improve performance.
A randomized trial of 18 surgical residents, who received structured feedback and coaching based on black box data, showed a reduction in adverse events and cut their errors in half compared to surgeons who received conventional teaching and subjective feedback, Dr. Grantcharov said.
The surgical black box is being used in hospitals in Toronto and Ottawa, and Queen’s University recently signed an agreement to use it to measure and improve performance of medical residents in general surgery.
Dr. Grantcharov’s black box has also grown in sophistication since its introduction in 2015. It started as a piece of hardware that looked like a DVD recorder but now it is largely based on software, integrated into a hospital’s information systems. Its scope has also grown from non-invasive surgery to any surgery.
A new generation of the black box being launched this year will provide data in real time. “Getting this information before the patient has left the operating room…will guide us in post-operative management,” said Dr. Grantcharov.
Technology is changing fast for all aspects of health care, including family practice. COVID-19 added rocket fuel to the trend of virtual medicine, said Dr. Chandi Chandrasena, chief medical officer for OntarioMD, a wholly owned subsidiary of the Ontario Medical Association that supports physicians in adopting digital health technology.
Dr. Chandrasena said doctors are now figuring out the right balance of virtual and in-office care. Every new technology has to be examined for pros and cons, she said. Virtual appointment platforms offer flexibility in terms of not having to be in the office at specific times for both doctors and patients. But patients who don’t have fast or private internet access can be disadvantaged.
Wearable health-monitoring devices and applications offer promise for the future, but the wearables many patients buy now are not designed for clinical use and may produce unreliable results. “I really do think that is something for the future of health care. We need to learn how to harness that,” she said.
Dr. Chandrasena said the wealth of data available on patients will be key going forward. Besides the data available in the patient’s electronic medical record, there is hospital, home care, lab and, eventually, wearable device data. “How do we integrate that data…because we have all these different systems?” she asked.
System integration and improving the functionality of electronic records and related data tools are what family physicians and other specialists need to cut down on administration and burnout, added Dr. Chandrasena.
She said she can’t understand why in 2022, she still needs to type everything the patient tells her, pull up requisitions, manually fill them out and fax or print them. An AI tool could write the physician note, bring up the lab requisition and send it through a portal to the lab. There are some companies developing these types of AI scribes, but they’re still in their infancy.
From the wider perspective of the entire provincial medical system, technology is coming that will help overall population health. “It’s important to know where we have to put our resources. It is important to reduce costs,” Dr. Chandrasena said.
That’s where the OMA is targeting its technology efforts, with machine learning that is assessing provincewide human resource needs. The project, which is poised to be presented to stakeholders, including the provincial government, this fall, uses information from medical billing to analyze all physician-patient interactions. Data flows from hospitals, drug benefits and physicians’ offices.
That information can be used to predict future needs for primary and hospital care and what specialties will be needed in specific regions, based on the pattern of doctor billing. Data on doctors’ ages can also help predict retirement and practice-reduction attrition.
The OMA project, dubbed PRIME (for Physician Resources Integrated Model), will help universities, residency programs and health-system policymakers determine where skills gaps are and will be. Developing projects such as GEMINI, PRIME, CHARTWatch and OR Black Box requires not just buy-in from end users, but also plenty of financial and institutional support. Researchers report that government, private donors, hospitals and universities in Ontario are stepping up.
Dr. Mamdani said federal and provincial governments are investing in artificial intelligence strategies. “One of the core fundamental things we need is for hospitals to say, ‘this is a priority for us.’ We have a very progressive, visionary CEO who has declared artificial intelligence to be a strategic pillar for Unity Health,” he said.
Dr. Verma added the timing is right for health-care data research. “I think policymakers, health system leaders and other decision-makers are very aware of the importance and value of data and that awareness has grown in leaps and bounds over the last decade,” he said.
Katherine Kerr is an Edmonton-based writer.