Lifesaving AI and an H5N1 mystery
Public Policy Forum
Good Policy. Better Canada | Bonnes politiques. Meilleur Canada
Here’s what we’re following in the world of health?security this week, including the lifesaving AI tool being used at a Toronto hospital, rising vaccine hesitancy and the H5N1 case that has some experts worried and upset.?
Sponsored by:
AI life saver
An artificial intelligence-driven monitoring tool in place at one of Toronto’s busiest hospitals is being credited with helping to decrease unexpected deaths by 26 percent among hospitalized patients, according to a new study. The tool is called CHARTwatch, and it was implemented at St. Michael’s Hospital in downtown Toronto.??
CHARTwatch measures dozens of inputs from a patient’s medical record (vital signs, heart rate, blood pressure, as well as daily lab results), monitors for changes and makes a dynamic prediction about whether the patient’s state will deteriorate. The results from this project, albeit a limited one, are encouraging, given its real-world impact.??
Dr. Amol Verma, the study’s lead author, a clinician-scientist at St. Michael’s, and Temerty professor of AI research and education in medicine at University of Toronto, told PPF via email that “as with all technologies in health care that show promise in a single setting, it would be ideal to see a multi-centre study evaluating the effects of CHARTwatch across multiple hospitals."?
“For AI technologies to work across different settings they need to be developed in large, diverse datasets,” Verma added. His team has established GEMINI, a data-sharing network across 30 Ontario hospitals and is working to develop AI tools that can be applied widely.?
But integrating AI into workflows has its challenges. “A big part of our work was co-designing CHARTwatch with front-line clinical staff and hospital leaders,”?Verma said, which helped with uptake and ensured proper integration into routine process of care.??
Honesty helped, too. “An important component of training for staff is to transparently inform them about the tool’s capability,”?Verma said. “It is not perfect. We expect that it will miss cases and that some of its alerts will be false, so we remind clinicians to use their own judgement in addition to the tool’s predictions.”?CHARTwatch is “designed to supplement their clinical expertise, not replace it.”
Fall Lecture 2024:?This year’s?much-anticipated PPF Fall Lecture features five incredible?speakers on?Canada-U.S. relations. Hear from?Kelly Craft?and?Gordon Giffin,?former U.S. Ambassadors to Canada,?Clifford Young, President of IPSOS U.S. Public Affairs,?Janice?Stein,?Belzberg Professor of Conflict Management and the Founding Director of the Munk?School of Global Affairs & Public Policy, and?Edward Greenspon,?PPF?President and CEO. Join us in Ottawa on Oct. 24.?
Vaccine (dis)approval??
Health Canada approved Moderna’s latest COVID-19 vaccine, which targets the KP.2 variant. Another vaccine, from Pfizer, still awaits approval in Canada?—?though both have already been approved in the U.S.
But, now that the vaccine is approved, will anyone take it? Last September, Health Canada approved vaccines targeting the XBB1.5 variant of the virus, and as of July 12, about 7.3 million Canadians (or 18.2 percent of the total population) had received it. The proportion of American adults who say they’ve had a COVID vaccine in the last year isn’t much better: just 22.5 percent.?
And a new poll in the U.S. suggests those uptake numbers may not improve this year. The national poll from the Ohio State University Wexner Medical Center found that more than a third (37 percent) of Americans who’ve received a vaccine in the past don’t plan to do so again this year, whether for the flu, COVID, pneumococcal or RSV. The same poll found that just over half (56 percent) of adults say they’ve had a flu shot this year or plan to get one, while only 43 percent have said they’ve received or plan to get a COVID shot.?
领英推荐
Already this year, multiple measles outbreaks in the U.S. have erupted in communities with low vaccine rates. And, depending on the outcome of the November election, promoting vaccine uptake in the U.S. may become even more difficult. In June, Republican presidential nominee Donald Trump vowed that, if re-elected, he “will not give one penny to any school that has a vaccine mandate or a mask mandate.”
Mpox ‘not under control’
Director-General of the Africa Centres for Disease Control (CDC), Dr. Jean Kaseya, said that “mpox is not under control in Africa. We still have this increase of cases that is becoming worrying for all of us.”?
Indeed, there has been a recent jump in mpox-related deaths —?up another 14 this past week, and 107 the week before. According to the latest figures from the Africa CDC, there have been over 29,000 reported cases of mpox this year and 738 deaths.?
Meanwhile, the push to get more vaccines to Central Africa continues. The WHO has now added the MVA-BN vaccine to its prequalification list (which ensures vaccines are safe and effective) — the first mpox vaccine to ever be on it. This should make the vaccine easier for countries to procure. But that vaccine is still only technically for people over 18 (though the WHO has also said that it can be used ‘off-label’ in infants, children, adolescents and in pregnant and immunocompromised people in outbreak settings where the benefits outweigh the potential risks). Getting a vaccine to the many children who are disproportionately impacted by the mpox outbreak (60 percent of the cases in the DRC have been in kids under 10) still poses a significant challenge.?
Japan said that it will donate three million doses of the LC16 vaccine, the bulk of a total 3.6 million doses pledged from western nations. LC16 is designed for children and is administered in a single dose, much like the smallpox vaccine, theoretically making it a cheap and efficient option.
H5N1 confusion
The number of human cases of H5N1 in the U.S. remains at 14, but questions still swirl about the latest case in an infected person from Missouri. While it’s emerged that a close contact of that individual also became ill on the same day — raising the possibility of the first known person-to-person transmission of the virus —?key details about that second individual remain unknown, which has angered some experts.?
The CDC has only been able to partially sequence the virus’s genome from that Missouri case, but that partial analysis has reportedly revealed two mutations not yet seen in human cases, both of which were on the hemagglutinin protein. One of the mutations has been identified in fewer than 1 percent of infected dairy cattle samples, and as one expert told The New Scientist, it can diminish antibodies’?ability to recognize and neutralize the virus. “This mutation is not changing our assessment of the risk that this virus will take off in humans, per se,”?Jesse Bloom at the Fred Hutchinson Cancer Center in Washington told the magazine. “It is just showing that we need to think carefully about which vaccine candidates we want to have ready in case that should happen.”
“We don’t really understand yet what's happening with H5N1. We’ve been dealing with this since April, but big questions remain about how much infection is in cattle and why it continues to spread within farms,” epidemiologist Michael Osterholm told Think Global Health. Put another way: H5N1 is “not causing a human pandemic right now, which is great,” Michael Tisza, a bioinformatics scientist at Baylor College of Medicine in Houston told the MIT Technology Review. “But it is a virus of pandemic potential.”
MORE FROM PPF: Read our latest report on health, Exposed: How Canada can close its health security gaps
AMR: Good news, bad news
A new study published in The Lancet has forecasted a significant rise in deaths associated with antimicrobial resistance (AMR —?when bacteria, viruses, fungi and other parasites are unresponsive to antimicrobial medicines) by 2050.?
The study estimated that 4.71 million deaths worldwide were associated with bacterial AMR, including?1.14 million deaths that were directly attributable to it. The study forecasts that without additional measures, in 2050 those numbers will have increased to 8.22 million deaths associated with bacterial AMR, with 1.91 being directly attributed. That falls far from the 10 percent reduction target The Lancet called for by 2030.
But the study also had some good news. It found that there was a greater than 50 percent reduction in AMR mortality for children under 5 between 1990 and 2021, which the study authors say deserves “special attention.” In its forecast, the study expects there to be continuous reductions in AMR mortality in children under 5 years (even as the AMR mortality in adult age groups likely increases).?
Email subscribers read this first. Sign up and get this weekly newsletter delivered straight to your inbox.
This newsletter is produced by journalists at PPF Media. It maintains complete editorial independence.