AI Scribes from Evaluation to Adoption

This transcript was generated by AI and may contain minor errors.

Dr. Onil Bhattacharyya: So you shave off 30 minutes of the worst time of your day. It means a lot actually.

Katie Bryski: Hello and welcome to Digital Health in Canada, the Digital Health Canada Podcast. I’m Katie Bryski.

Shelagh Maloney: And I’m Shelagh Maloney.

Katie Bryski: Among AI clinical use cases, one in particular is catching attention. AI scribes have been piloted in multiple provinces and territories. But what’s the verdict so far? Today we bring in two leaders to explore the data, the potential impacts, and the balance between hype and hope.

We are thrilled to welcome to the podcast Dr. Onil Bhattacharyya, Frigon Blau, chair in Family Medicine Research and Director, the Women’s College Hospital Institute for Health Systems Solutions and Virtual Care, which you may know as WIHV. And Simon Hagens, an independent consultant with expertise in evaluation and analytics. Welcome to you both and thank you so much for joining us.

Shelagh Maloney: And Simon, we should congratulate you on your recent appointment as an adjunct professor at the University of Ottawa.

Simon Hagens: Thanks. Yeah, it’s been great jumping into that world.

Shelagh Maloney: Well, uh, so I just wanna say welcome as well. We’re really thrilled to have you. AI scribes, as Katie said, is a really hot topic right now.

Before we jump into that, and if you’re familiar with our podcasts and our listeners know that, the first question we always ask is we ask people to just tell us a little bit about their career journey. This is always something that people are interested about and frankly, we always have some very circuitous routes, mostly. So Onil, why don’t we start with you?

Dr. Onil Bhattacharyya: Sure. I’m maybe less circuitous than some, but I, uh, studied family medicine and as I was finishing my studies, I had a sense that clinical care is about a lot of micro problem solving and that I was working in a health system that was kind of broken. And so the idea of just like working in a system that I didn’t think was very good was not that gratifying.

And so the opportunity to like do a PhD and try to learn more about health systems and how to redesign care, I think got me excited, you know, I was excited about that prospect. And then did a PhD in health services research. Did a postdoc at the Harvard School of Public Health. Got interested in, uh, social enterprises and startups and sort of got me into a wide range of things, but essentially an interest in care integration and then an interest in how to design health services and finally technology and how it can make care better.

So, you know, mostly an academic, but moving really from the more focus on care and people with complex needs and more and more into technology and how it can make care closer to what I hoped it might have been 35 years ago.

Katie Bryski: I think that desire to improve the system is another theme that we’ve heard through many of the journeys that our leaders have been on.

And Simon, why don’t you tell us about your career path?

Simon Hagens: Sure. I dropped into healthcare by accident. Uh, ended up working in a community health centre because I was into community gardening and just found a, a love of the sector while I was there. And, and I would recommend to anybody to go and work at a community health centre because it’s a place where you meet amazing people doing amazing things.

But also they’re small organizations, so they’ll let young people do any number of strange jobs. So I took on a job there as a community garden coordinator, ended up a few years later leading an EMR implementation. That launched my journey in digital health. I spent 20 years at Canada Health Infoway, which was a wonderful place to work, and I got to spend some great time with you folks. And was thrilled to have the opportunity to support the launch of the AI Scribe Investment Program.

So, learned a lot about AI Scribes and, and now as a researcher and taking a real interest in watching that adoption cycle start to play out.

Katie Bryski: I think community gardener to EMR implementer is maybe the best pivot we’ve had thus far in the podcast.

Simon Hagens: Yeah, no, there was not a lot of intersection there. There was. I was either in the garden or I was in the server room, but I was usually one of those two places for a couple of years there.

Katie Bryski: So either way, you were deep in the weeds.

Simon Hagens: Nailed it, Katie.

Katie Bryski: Okay. Well, we could go down quite a path, but we are here to talk about AI scribes and Onil. I would love to hear more about the studies that you have undertaken recently.

Can you set the stage for us and tell us a little bit about the studies themselves, what you looked at? What findings particularly stood out to you?

Dr. Onil Bhattacharyya: So we started looking at this in October, 2024. Administrative burden in primary care was a big deal at the time, and the idea that technology has made our job worse instead of better.

And that AI scribes is maybe a spot where we’re starting to see some change. And so that was kind of the impetus and the Ministry of Health funded the work. And so we were working with, uh, some industrial engineers. Enid Montague is one of my colleagues in the U of T Department of Mechanical Industrial Engineering.

So the first step was to look at what is the impact of the scribe on documentation time. So we looked at doctors and simulation doing, uh, scribe, no scribe on the same cases, and we found a four-minute reduction in documentation time. So we thought, okay, this is promising. Then we looked at six different scribes.

We created quality metrics. At the time, there’s only six on the market, uh, in Canada. We looked at quality metrics for all six scribes and compared them, they were all pretty good, but some more significantly better than others. And we chose the top three and we rolled them out to 152 docs. And so in that next phase, we, uh, had a pre-post survey, and in the survey they said that it had reduced after hours administrative burden by about three hours.

So that, you know, it looks like a modest number, but in the qualitative stuff was very striking. I found the magic of family medicine again. I get home in time to have dinner with my kids. I’m volunteering and I was thinking about retiring, but I feel like I can hang on a bit longer. So things like that.

So we had presented that to folks at the ministry. They got very excited with those results, even though they’re relatively preliminary. And then they launched into a provincial procurement process. We also shared this with Simon and folks at Infoway and think they’d also commissioned a systematic review at the time.

But I think this was, can it local data that was promising, and I think Simon can comment on this, but I think contributed to the, the idea to roll out, you know, do a large procurement scribes across the country. So that was the first question of does it work? So yes, we think it works. The next question is, does it scale?

And so when we look at the Infoway program, 10,000 people signed up in two weeks, so definitely scales when you give away free licenses. Then on the Ontario side, with the Supply Ontario program, basically you can pay for the, you have a preferential rate and a very good, solid contract. So it’s in, you know, it’s a good contract that providers can feel comfortable with.

But Supply Ontario is new to them. Family doctors don’t procure anything, they just buy it. And so this was a sort of different way of working. We’re about six months out, 200 individuals or entities have signed up, but I don’t how many licenses they’ve distributed. We’re still working on that. So it’s interesting when you ask the question, does it scale?

You know, obviously giving people a free license is a mode of scale, but maybe not a sustainable one. And then giving people a preferential contract with a lower price is probably more sustainable, but has, you know, grown much more slowly.

Simon Hagens: Love what Onil has just been talking about, it gets me excited about the role that evaluation is now starting to play in digital health, more so than it did when I started my career 27 years ago, when it was really hard to encourage evaluation at in those pilot studies that Onil was so lucky to be able to, to execute around AI scribes. And then, into implementation and that measurement through the implementation process becomes so critical, to be able to course correct and get to the outcomes we’re looking for.

It’s great to see that it’s unfolded that way and, uh, something that to me is striking – the time period you’re talking about. So you told a story about a series of pilot setting, like what’s the total time span that, that, that work took to take?

Dr. Onil Bhattacharyya: So that first phase, you know, it was pitched to us in September.

We started in October. We finished all the simulation type work in February, January, rolled out the study, basically finished everything by June. So timelines were short, and we’re hoping to get those even shorter in the future. In that

Simon Hagens: timescale that you’re talking about, like from pilot study, lots of interest scaling activities like two or three years, which is a timeframe that I haven’t seen in, in the digital health world in terms of that speed from idea to, you know, mid-market penetration.

Shelagh Maloney: So, you know, Simon, I was just gonna mention that ’cause I know, certainly we recognize, first of all, in, in digital health and technology, because it’s moving so quickly, often the evaluations aren’t done in the first place or they’re done, but they’re done years later and the technology’s either moved on, it hasn’t scaled because of all the issues in Canada that we know, that our failure to scale in a lot of different areas.

So I’m curiou,s and Onil, maybe you can talk about this as well, so I know difference between take up and AI scribe compared to like physicians using rs, which was at least a decade.

And the AI scribe the number that’s in my head, and you can correct if it’s wrong, it’s like 40% of clinicians in Ontario. So took it up very quickly, and maybe that’s a little bit too generous a number, but the takeup has been very fast and the evaluation that came with it has also, as Simon pointed out very quick.

Do you think this is like a turning point, or is just this AI scribe is an exception because it’s AI?

Dr. Onil Bhattacharyya: I actually think it’s a bit of an exception. I think it’s web-based tool. It requires no integration. It’s like a function that people are really keen on ’cause it helps them do something that they’re really annoyed about doing and it’s pretty effortless to implement.

So there’s a lot of like, you know, it’s not, I won’t say it’s a completely a special case, but I don’t know that we’re gonna find that many in the short term that are this much of a slam dunk. Now what we might see is. AI functions on the backend of an EMR, right? Like Epic keeps changing things on the backend.

I admit there’s gonna be seamless implementation that we’ll see in the hospital space. In the primary care space, I’m less clear, but I do think this has been very quick uptake and mostly people feeling pretty, like it’s pretty safe.

Simon Hagens: I wonder if I could just a, a, a curiosity Onil, ’cause I’m, what I’m looking at in the trajectory in the long term is trying to figure out can it maintain this momentum and so.

I think there’s probably more data we’ll see in the next year, but that notion of like, are people using it occasionally or is it really embedded in their practice and are we seeing fully embedded in practice in a large portion of the population? That’s my curiosity to know if this is actually stuck and I’m curious if your results are giving you a sense of that sort of ratio of who’s a light user versus like we’re doing this.

Dr. Onil Bhattacharyya: The data from Infoway actually is very instructive in this respect. When you took the first wave, that was 10,000 people who signed up. And then of them, like, you know, when you got to October, you’ve got 10,000 licenses out there, 8,600 that are, you know, used at at least once and then 6,100 active users.

So that’s like 60% of the people who are using it more than 10 times a month. But 10 times a month is not a very high bar ’cause you’re gonna see hundreds of patients a month. And so I think it, it’s, it is variable. And then Infoway took away licenses from people who weren’t using them, gave them to other people.

And we’ve moved on to, I think there’s like 12,000 licenses out there, I think 6,900 active users. So you can keep giving it to people, but your number of active users. Is not tracking the same way as you get more licenses out there. I don’t know that everyone will use it all of the time. And you know, Shelagh, you mentioned that 40%.

So everyone is kind of struggling to figure out what the right number is. Ontario gave licenses to 12% of family doctors. And then we have the 200 entities that have registered with the Supply Ontario Vendor of Record. So it might get us to 40, I think it’s more in the 30 range, but when you look at an adoption curve, that’s taking us into the early adopters. So to your original question, Simon, I don’t know, I think we might hit a plateau in the next year or so and maybe something else needs to shift before we see, you know, 50, 60% penetration.

Katie Bryski: So if we’re in the early adopters phase of diffusion. What factors do you see influencing that curve from here?

Dr. Onil Bhattacharyya: My guess is a subset of people are gonna love it and then pay for it. And they’ll sign up with, in Ontario, they’ll sign up with the vendor of record. We’re gonna be able to track all those conversions, you know, basically it’s not, you’re not gonna get to a hundred percent, even in the next two years, I think you’re gonna get to maybe 40, and then the free licenses will go up, you know, and then you’ll drop to 30 or something, and then maybe you’ll sit around there.

That’s my guess. You know, other things may change. Maybe if it’s embedded in an EMR, more people will use it. But that’s my, my initial take.

Simon Hagens: And if I could just suggest that’s starting to look a lot more like the usual diffusion of adoption curves that we’ve seen with other technology. So it had a really good, uh, launch and a really good, uh, head start on others.

But that gravity of workflow, reimbursement, policy change that are all these really important things for the solutions that are now broadly spread across their health system. It, it all took those pieces kind of coming together. You know, they needed to sort of fit into clinical practice in a way that scribes certainly can, but are they there today?

Probably not quite yet. So there’s time and work before this is a really integrated part of practice. It would be my assessment from the data.

Dr. Onil Bhattacharyya: And you know, if I could just add what you know, when we went from 150 docs that said they saved three hours a week of after hours documentation, when we went to a thousand people answering a survey, that number dropped to 1.6 hours.

That’s common. These are different people with different practice profiles so that you get a bit of a dilution of that effect and the overall benefit. What we’re seeing essentially is you have a subset of people who document well and they’re saving time because you spend a lot of time documenting.

Then you have a subset of people who don’t document well and don’t spend a lot of time. And their notes are getting more complete. We are just, uh, beginning a pre-post six months of pre-post data on quality of note six months before you, uh, AI scribe and six months after. So we’ll get that in the next couple months, but we’ll see what, you know, what does it do to note completion.

Katie Bryski: That’s a fascinating distinction, and I think it leads me into something I was curious about because I, I saw the three hours per week time savings, and you’re right, it sounds modest on the surface, but I was hoping you could help us see that bigger picture: what the cumulative effect of those three hours really could mean for clinicians, and for the health system?

Dr. Onil Bhattacharyya: So the thing about those three hours, let’s say you work four or five days a week, that’s at least maybe 30 to 45 minutes between 5:00 and 7:00 PM the time of the day where you hate yourself ’cause you’re so slow and you just wanna get out of there and you just wanna get home in time for dinner with your kids.

And it’s actually the worst time of the day. So you shave off 30 minutes of the worst time of your day. It means a lot actually. That’s why we see that disparity between a numerical number that does not huge in qualitative data. That’s very impressive.

Simon Hagens: And Onil made a really important point about those three hours saved.

And then the larger real world sample where you see some of those, those numbers goes down. And that’s just the way the the world works. The clinical trial world is a more controlled environment. It’s a, you’re gonna see better outcomes than you generally will in the real world. So that’s where that. That evaluation all along that adoption curve, through those early adopters, through those early majority into the late majority, is so important because the, the context changes.

You’re, you’re starting to deal with people that are dealing with more challenging patient populations that are dealing with more challenging conditions to deliver care in, that have a whole lot of other things that they’re trying to manage. And playing with a new tool isn’t on their list. So to me, that.

Real world, long-term evaluation is a really interesting thing for us to spend time on. And it gets into some of those things around things like data quality. ’cause as we’ve talked about, there’s, there’s the, the early benefits and getting, getting some of that pajama time back is a massive benefit. Um, but I actually think that’s the, the smaller part of the benefit.

And we talked, touched a little bit. About quality of data. And so maybe I can just speak a little bit about how important that is. I think one of the big findings over the, the, the sort of outcome of the pandemic and the Pan-Canadian Health Data Strategy and some of the really important health policy work that happened in Canada over the last five years was a recognition that lack of sharing of information, lack of collaboration across care settings, that’s, that’s using information has been a real issue for our health systems coordination of care and related to that access to care are a huge challenge and data is at the heart of that, right? So we know that interoperability is lacking and clinicians can’t access or share the information they need.

And often that’s ’cause the point of service systems don’t talk to each other. What AI scribe offers as Onil noted, is an opportunity to better structure that data and even to figure out how to better put that data in places that it needs to be, and sharing it with other clinicians is one important use case.

Giving it to patients is another very important use case. So this notion of getting a better structured, higher quality capture of the note and making that then available for other really important purposes, that’s delivery of care, that’s managing the healthcare system, that’s driving innovation. Those are all the opportunities that are downstream of AI scribes.

So we gotta do that implementation right, so that clinicians see immediate value, but let’s not lose sight of the fact that this generates a data opportunity, which is too big to miss out on.

Shelagh Maloney: And maybe just to add on there, you know, we talked about the importance of evaluation and it feels as though people are taking more heed.

So I think increasingly the awareness that evaluation is important is really gaining that traction. I remember hearing, uh, Dr. Amol Verma do a presentation and talking about randomized clinical trials, so a little bit different but in, but in health. And there were a number of them. And then he said the ones that are involved in AIthere was like 85 or something like that out of tens of thousands and then, and there’s one in Canada.

So I’m curious about what your thoughts around evaluation in general and. Evaluation in the AI sphere in particular?

Dr. Onil Bhattacharyya: Yeah, so I mean, definitely these are emerging product categories and so I think evaluation is essential. Do we always need RCTs? I’m not a hundred percent sure. Like it is true. Like, you know, the a, the data on AI scribes out of the US is actually much more modest.

They’re using timestamp in a, in Epic rather than survey data. But they also, the, the way they practice is different. So there may be different things going on there. You know, I’ve been a little less bullish than some just now on AI scribes and uptake. But when you take AI scribe and you layer on robotic process automation with AI scribe as the interface, and ambient listening as the interface, and all kinds of downstream work that’s enabled.

That is probably gonna be game changing and that will see additional uptake.

Simon Hagens: I think that’s such an important point that like many digital health technologies, sort of the initial version might have sort of a productivity gain with it, but that’s not necessarily where the largest value is gonna come from.

And I think about EMRs like in the early days of EMRs, is like, okay, we, we need to get rid of some of these. The paper and it’s, there’s some of those efficiencies, but in the long run it’s a quality of care play. It’s all about quality of data and being able to access right information and all those kinds of things.

And, and I think in the AI space, it’s gonna be a similar thing where the AI scribes provide some near term value. But in the longer term, how well are they integrated into clinical workflows? How well do they support all of these other interactions you have in the healthcare system? And that’s just gonna take years to mature.

’cause it’s not just technology is workflows and reimbursement and all these other factors. So like many technologies, the value proposition matures over time.

Katie Bryski: Okay. But then Simon, I have a question. Because I hear you, I agree with you. It takes time for benefits to accrue. It takes time to see how value is realized and manifests.

But technology is moving so fast, right? Like the AI of today is not the AI of even a year or two ago. So how do you see evaluation evolving in that context, right? Where, yes, it takes time to see what the value is, but the technology is accelerating so quickly, like it may be a very different game by then.

Dr. Onil Bhattacharyya: So our approach to this, we’re trying to build out a kinda living lab concept where we look at. Primary care workflows and simulation and also in, uh, a handful of partner clinics. And here we’re essentially trying to look at what are the jobs to be done in primary care? And we map out those jobs and who’s doing what.

And then anytime we add technology and we think about ecosystems, not just one tool or ’cause some tools have eight functions. Like the current AI scribes of a year ago now have AI decision support. Robotic process automation, they’ll call your patient for a follow up phone call. They just have like eight or nine functions.

So you could imagine all the functionality of a given tool or a suite of tools, and then map it against the jobs to be done in primary care. And then you start to see the cumulative impact on care in general. And so, you know, RCTs are often looking at like one benefit. What is the impact on X? So here I think we wanna look at the impact of ecosystems.

On a wide range of jobs to be done. So that’s kind of my evaluative lens, starting with simulation and then going into real world settings.

Simon Hagens: Yeah. So what one other topic as we were talking about diffusion curves with ai, that I thought an interesting one is around diagnostic imaging, because diagnostic imaging artificial intelligence was pretty mature a decade ago.

There were good clinical trials demonstrating its efficacy and. A decade later, clinical trials showing real, real promise in terms of the tool’s ability to diagnose an image. The really interesting thing that we’re starting to see in evaluations around AI for diagnostic imaging is that. The outcomes in terms of are we actually better at catching disease?

Do we have less false positives? Are we actually improving patient outcomes? Is much more of a nuanced question. And so that decade of experience of starting to use those AI tools, we’re learning a lot about that diffusion. We’re learning that. In fact, how that tool is applied in the clinical process is massively important.

So we see an AI a couple of different ways that DI can be supported. You can have it as a triage tool. It can just flag the most important ones and stick them in the see this nail pile. It can be an assistant hand in hand with a clinician saying, we’re both looking at the same thing. What do you see?

What do you see? Or it can be a blind partner. Those three options all generate different outcomes and it, and it looks like there’s good ways to implement as a blind partner that that actually does have clinical outcomes that that improve. But the other ones are more mixed. And so some of those really important questions about how do we need to use this tool so it actually generates value, I think are yet to be seen and AI’s ability to capture the right note is good.

Does that generate the outcome we want, which is a better patient experience, a better provider experience? Those things turn out to be much more complicated than, yeah, if it generates a good note than it obviously is gonna generate good outcomes. Turns out there’s a lot more things that happen in the middle there, and so to me the AI for diagnostic imaging is a really important use case for us to look at and say.

What do we need to do as policymakers, as researchers, as leaders in healthcare that take a, a technology that’s pretty smart and turn it into an intervention that actually improves outcomes in the healthcare system?

Dr. Onil Bhattacharyya: I guess the thing is like so, scribes, we’ve talked about it as a distinct tool, but I think, you know, ultimately what it does is it opens up ambient listening as an interface and as an interface.

It’s way better than typing on a computer. And I think what we’ll end up seeing is many, many more voice-based interactions. For me, it’s opening up that whole space. And so we started with the most basic function, but it’s gonna take us to other places.

Shelagh Maloney: Can I just ask one question as a, now we’re hearing about AI scribes for patients. What’s the hype around that, and do you see that coming anytime soon?

Dr. Onil Bhattacharyya: You know, recently I was on a plane, uh, with a, someone talking about my work. And I, the person next to me was a patient who had prostate cancer, and he was saying how he records all of his interactions with his physician on his phone.

And he said recently he was fiddling with his phone and then his doctor noticed and said, Hey, are you recording this? He said, yeah, I’ve been doing it for 14 years. How am I gonna remember these conversations? I’ve got prostate cancer. I’m stressed out. How can I remember everything we’ve said? And what was the clinician reaction?

So he was shocked, but he took it. He took it. I think our expectations will shift and I think ultimately this is gonna be seen as reasonable and normal. But it’s funny ’cause he did it in a sort of stealth guerilla way for the last 14 years.

Simon Hagens: Absolutely. And the patient context is evolving rapidly.

Patients have the opportunity to conceivably bring their own scribe into the room, as we’ve discussed. They also have the opportunity to maybe get that download of the scribe. The other really important things that have been happening in the patient environment that created a bit of a potential tsunami are patients now have a lot more access to their own data.

So I, in the last couple of years, have been able to download a bunch of my patient information, test results and things like that. I had to go to four different portals to get them, but I could get my hands on them. The other thing that’s happening is that consumers are starting to use AI tools for a lot of purposes, including asking healthcare questions.

And the last month we’ve seen major announcements from big AI companies that they’re getting into the health space. So they’ll be offering products that would allow me as a patient to go and download my results, bring in whatever other contextual data that I want to inform my care and get an opinion from ChatGPT.

Now, that is gonna be appealing to a certain portion of the population. How big is that portion of the population? That’s to be seen, but it could be a really material effect on healthcare in the coming years.

Katie Bryski: I guess I just wanna stay mindful of some of the potential risks though, right? Both from like, you know, generative AI hallucinates, it makes errors.

Um, I think there are well-documented like privacy data governance, data sovereignty considerations. So yes, agree, opportunity, but what, what should people be keeping in mind from like a risk mitigation, risk tolerance point of view as they look to embrace that opportunity?

Simon Hagens: Absolutely. So, so to be clear, there are absolutely significant potential risks, and these are brand new products.

And so assessing those risks is really challenging. So consumers need to assess those choices for themselves essentially in in the world today. Which realistically they do all the time. Apple and Google have an absurd amount of information a about me. They do a variety of things with it, and I’ve made a personal decision as a consumer that being able to see myself as a blue dot on a map is worth the convenience for that.

I think one of the big challenges we have in the Canadian healthcare system is access to care is really challenging to get in and see your providers, and so that kind of market force of an access barrier. Causes people to make different decisions about what they will do to access care. So the interesting thing about this is it’s no longer a regulatory question.

Those tools are gonna be available. People are gonna be able to get their data. The market will do what the market does. And you know, and Onil’s anecdote about, uh, a person in an exam room that’s been recording things for years is a bit of a window into what clinicians might be encountering in that.

I don’t know what it’s gonna bring. I don’t think anybody does, but it’s certainly something that we should be watching and studying, uh, with great interest to make sure it, it benefits rather than harms.

Shelagh Maloney: You know, in our last episode we had Dr. Anderson Chuck and Dr. Fahad Razak talking about health data as a national asset, and then we can pull that data and really, you know, take a global leadership role in terms of.

Using data to drive innovation and, and so the thing that I think about is if ChatGPT’s getting all my health data, the system isn’t getting it. And so it’s an interesting tension and conundrum, right? It’s about what I need to know, what I wanna know as, you know, somebody who has a diagnosis of something versus the greater good.

And, and so it’s a great question and what would be interesting to see what comes out of that, Katie, you want you ask your digital health question?

Katie Bryski: All right. So as it’s Digital Health Canada’s 50th anniversary, we’ve been ending our conversations with a consistent question. For all of our guests, we’ve been asking: “In one word, what does digital health mean to you?”

So Onil, if you had to pick one word to sum it all up, what would it be?

Dr. Onil Bhattacharyya: At its best, I would say seamless.

Shelagh Maloney: Okay. Simon, your turn. One word.

Simon Hagens: So I’m gonna go with potential. I’ve been in digital updates for a long time, so I’ve seen a lot of value and a lot of impact and a lot of things that are fantastic.

But I’ve also seen a lot of opportunity left on the table. So I think that in digital health there is a lot of potential there today and a confluence of different kinds of inputs that make the, the moment we have today a real opportunity for digital health to go further. So potential it is, for me.

Katie Bryski: And I think it’s brought us full circle.

You know, we talked Onil at the beginning about easing technology to make care better, and I think we’ve also heard that technology in that way makes care more human as well and facilitates that relationship. And I wanna thank you again for your time today.

Dr. Onil Bhattacharyya: Thank you all.

Simon Hagens: Awesome.

Katie Bryski: You know, every time we do an episode on ai, we get such a different perspective on the topic and such a different kind of slice of life in what’s happening in in AI right now.

Shelagh Maloney: Yeah, you, you’re right. And I really appreciated the comments that both Simon and Onil had around adoption and scalability.

Because we’ve talked about those two topics, but in different contexts and the nuances of all of the. Other things that are not just the adoption of the technology, you know, the regulations, the pricing, the marketing, the, the change management, the workflow integration. Not new topics, but again, a different slice or lens on them from that perspective.

Katie Bryski: And I think such a good example of how the right storytelling makes such a difference as well, and sense-making with these sorts of topics. Like I’ll, I’ll harken back to that three hours a week. Which on the face of it, it doesn’t sound like much, right? But the way Onil described it, it’s a critical 35 to 45 minutes of your day.

Thinking of the worst 45 minutes of your workday, if you could just cut that, how much of a difference to your quality of life would that make? And then you really start to see the benefit, and the valu,e and the improvement of the care experience for the physician side and also for the patient. So I think that communication and that humanizing of some of this data is so critical when you’re doing adoptions like this.

Shelagh Maloney: Yeah, I, I couldn’t agree more. It’s interesting ’cause that I was having a conversation with, I can’t even remember who it was, but we talked about the importance of data, but the importance of stories. You know, stories trump data every time, and you can see data and it gives you a, a one picture. But then you have those anecdotal or those stories that accompany the data and, and if we have both of them together, all the better.

And you know, I think one of the things for me was the whole conversation about. The patient experience, and I know it came out in the studies around physicians reporting that it improved their relationship with their patients, because they weren’t typing on a computer that they were, because it’s ambient listening.

They had more face-to-face and eye contact. Now I’m curious about what the patient’s perspective. So that was the physician-attributed patient relationship piece. I’m assuming it’s the same, but thata really an interesting point that came out of that.

Katie Bryski: I think I might have read something about this actually. I think I read a presentation. That may have been posted somewhere. And I think they did ask patients, and I think they did say that they felt their provider was more engaged. They felt their body language was better. So I think there has been some investigation into how that comes across from the patient’s point of view.

Shelagh Maloney: You know, it’s interesting, so we had one of our conferences, it was in BC and it was a patient with kidney disease, and he said, yeah, he says, I go to ChatGPT because there are no bad questions. I take as much time as I want. And I have enough information about me and my numbers and my studies to know what I need to ask my nephrologist about and what I can trust and what’s clearly a hallucination.

So it was a really interesting take from the patient’s perspective about why he’s going to ChatGPT for his care because I get 14 minutes every quarter with my nephrologist. So it was kind of an interesting perspective.

Katie Bryski: It speaks to Simon’s points about access to care and AI potentially being a way to mitigate some of those barriers.

But it really seems to me like, beyond AI, it just speaks to those systemic inequities and systemic barriers. There will always be people who either can’t use AI, don’t want to use AI, and I just worry, I guess, that if we default to that being a way you get care, if you can’t get into your primary care provider, the people that may be left behind.

We barely have digital health literacy figured out, wait until there’s AI health literacy.

Shelagh Maloney: And then, you know, coming full circle with Onil’s first comment about how he came into recognizing that technology can improve healthcare and Simon’s, you know, word to describe digital health is around potential. The importance of evaluation is really underscored when we’re trying to evaluate these new, disruptive, potentially expensive technologies.

So I think that, you know, “Yay, evaluation.”

Katie Bryski: Perfect note to end on. We will be back next month with ta brand new episode right here on Digital Health in Canada. Until then, you can check out many resources on Digital Health Canada’s website. Thanks for listening, and we’ll see you next time right here on Digital Health in Canada, the Digital Health Canada podcast.

Thank you for listening to today’s episode. Digital Health Canada members can continue the conversation online in the community hub. Visit digital health canada.com to learn more. Be sure to subscribe to the podcast to get new episodes as soon as they’re available and tell a friend if you like the show.

We’ll see you next month. Stay connected, get inspired, and be empowered.