Loading summary
Transcript30 lines
- [00:00]
A
Hi, everyone. This is Erica Spicer Mason with Becker's Healthcare. Thank you so much for tuning into the Becker's Healthcare podcast series. So today we're going to talk about how AI is transforming clinical care and healthcare delivery. And joining me for this conversation is Dr. Jared Saul, Chief Medical Officer at Amazon web services, and Dr. Heather Chait, AI ecosystem lead at Philips. Dr. Saul, Dr. Che, welcome to the podcast. Thank you both so much for being here today.
- [00:30]
B
Thank you, Erica.
- [00:31]
A
I'm really glad to have you both on the line. And before we get into kind of the meat of our conversation today, I wanted to give you both a chance to share a little bit more about yourselves, your work in healthcare, and also how your work intersects with AI. Dr. Saul, do you want to get us started?
- [00:47]
C
Sure, yeah. Thank you, Erica. Thanks for having me. So I'm Jared Saul. I'm the Chief Medical Officer at Amazon Web Services and started my career actually as a practicing radiologist for some time and also a person with a real interest in technology. When I was training, the Internet just kind of became a thing. And so I was rather smitten with that. In fact, so much so I spent a lot of time thinking about different types of companies that could leverage the Internet. And I started a few. And eventually that led me into about a decade and a half of creating companies and investing before I joined Amazon. And I've been here for about seven and a half years, and it's really a privileged position where I really get to see how technology is impacting care delivery across the spectrum for both healthcare and life sciences. And obviously, these days, every conversation involves the discussion of AI and how it can or is impacting healthcare. So I'm really looking forward to the discussion today and talking to Heather about our collaborative work with Philips.
- [01:51]
A
Wow, thanks, Dr. Saul. And likewise, I really appreciate you sharing a bit more about your background in radiology. Kind of this smittenness that you have with technology, it really teases up nicely for a great conversation. Dr. Chait would love to hear more about you as well.
- [02:06]
B
Jared, great to see you. And Erica, nice to be with you. I'm Heather Chait and I lead AI Clinical AI Ecosystems at Philips. My role is about how we consider building partnerships with innovative companies that are advancing AI in healthcare so that we can bring the right solutions to patients and clinicians faster. And on background, I spent 17 years practicing medicine as a reconstructive foot and ankle surgeon. And that experience really shapes how I look at AI. I had an opportunity to be involved in several early stage Startups that transacted to life sciences companies. So I've always had an interest in how we can advance clinical care, and now I get to do that for looking at how AI makes a difference in how we treat our patients.
- [02:53]
A
Oh, fantastic. Dr. Chait, great learning more about you and that 17 year practice in reconstructive foot and ankle surgery. So I know you're coming at this from a very, a very honed clinical angle as well. So that is actually where I wanted to start our conversation today. In the clinical realm, AI is usually discussed in terms of its potential to accelerate diagnosis and also reduce uncertainties. So, Dr. Chait, from your perspective, how is AI already starting to impact speed and accuracy of clinical decision making, and what does that mean for both clinicians and also patients?
- [03:29]
B
Well, Erica, AI is already having a measurable difference for both clinicians and patients. And a powerful example of that is in documentation. Tools like ambient listening free physicians from the keyboard so they can actually look their patients in the eye again. And that kind of change has even convinced some physicians to delay retirement with comments from users like, I feel like my doctor is listening. Patients have felt truly engaged with their physicians again. And it's reasonable with this measure of success to expect ambient listening capabilities to permeate additional clinical specialties in the near term. And in imaging, AI is speeding up tasks like scan interpretation and measurements, which mean faster results for patients and more time for clinicians to focus on complex medical decision making. At the core, it's about shifting the burden away from administrative work and toward real patient care and engagement.
- [04:28]
A
Yeah, great examples, Dr. Chait. It's really compelling when we hear some of these success stories about ambient technology convincing physicians to delay retirements or just have a better work life balance or sentiment toward work. And interesting to hear how it's also speeding up scan interpretations. Dr. Saul, what would you add here?
- [04:50]
C
Yeah, you know, I think in terms of clinical decision making, we're still in early innings here. But to Heather's point, there's a lot of very interesting stuff happening. I'd say one of the, one of the challenges a modern clinician faces is there's just so much information, right? Both on an individual patient basis, what's their history that might be scattered across many systems and many different databases, but also the rate of new research and clinical care pathways and recommendations. It's a lot to keep up with. And so one area where AI is exquisitely helpful is to help summarize and contextualize some of that data. And if you do that at the point of care you can imagine, it really can take some of the load off of the position and result in more timely and up to date thinking and diagnostics or therapeutic decision making. So there's a big role for AI there and we're seeing that happen now, particularly on patient summarization and providing appropriate context to the physician, whether that's a doctor who's treating a patient in a primary care setting or somebody who's interpreting a radiology examination or a pathology examination.
- [06:01]
A
Yeah. Dr. Sall, I appreciate you highlighting this important piece. The ability to summarize and contextualize the massive amounts of data we know providers are inundated with. So just a quick follow up question on that. You know, how does the collaboration between Philips and AWS support or advance some of this project, some of this progress, whether that's in the data space or some of the areas that Dr. Chait mentioned as well?
- [06:25]
C
Yeah, I think we'll talk a little bit more about this. But you know, to really get the benefits of AI, you need to have some fundamental data posture that integrates well across data sets and can power AI algorithms. And so this, there's a lot of work we've been doing with Phillips over many years to this end. And in fact, if you look just in the last 12 months, about 34 million patient exams have been processed exclusively on the cloud with Philips. And about 134 petabytes of data sit in the cloud representing a billion images and patient records. So you think about the scale of that data and the ability of us to host, secure and then leverage that data to inform these algorithms. And so that's a little bit less exciting perhaps in some of the specific applications. But this foundational work is really essential to drive this next wave of technology. And that's where we've been concentrating a lot of efforts over the past many years.
- [07:23]
A
I think that's a really important point to highlight, Dr. Saul, kind of how this is a foundational approach for enabling greater advancements in this space. I'm hearing more and more about the importance of that in AI conversations. But Dr. Chait, want to check in with you if there's anything you'd like to add on this.
- [07:41]
B
Yes, at Philips we bring the clinical expertise and build the solutions tailored for care delivery. And AWS provides the secure scalable cloud foundation. So together we can deliver innovations faster and with the trust clinicians and patients expect. A couple of these Examples include on AWS, HealthSuite Imaging Cloud Packs, delivers high speed remote reading, integrated reporting and AI orchestrated workflows so that we can scale on demand. And at the department level, our radiology operations Command center uses AWS connectivity to link technologists with remote imaging experts in real time for advanced scanning oversight. So our partnership with AWS is really strong and has helped to develop and advance our own products.
- [08:33]
A
It's so great to hear about the strength of this partnership and I really appreciate you both sharing more about it. And I want to go a little bit deeper on kind of how healthcare leaders and providers are managing this really, this constant wave of new technologies. You know, it's, I imagine that's overwhelming to some degree, especially as organizations are also trying to manage this data overload that we've touched on and other operational complexities. So, Dr. Chait, I want to stick with you here. From your perspective, what are the biggest barriers that providers face today in adopting new tools? And where do you see AI having the greatest potential to make a meaningful impact?
- [09:12]
B
Well, the most pressing barriers that we're seeing, and these have been confirmed by the Phillips Future Health Index, which is an annual survey that's provided us valuable information for over a decade on what's top of mind across those surveyed, which spans 16 countries, 1900 providers, and approximately 16,000 patients. What has this shown us? It's confirmed that the two biggest barriers we see today are clinical data being heavily siloed and trust in AI. And to address the clinical data, as Jared talked about a little bit earlier, clinical data is messy, it's in many silos, and it makes it hard for clinicians to get a full picture of the patient, especially at that point of care where it's most critical for the insights. This slows down care and leads to frustration and delays. The good news is AI and cloud technologies are beginning to connect those dots and span the silos to allow information to flow freely at the point of care. And the second barrier being trust. Our research has shown that nearly 80% of clinicians are optimistic about the role AI will play in healthcare delivery, but patients tend to be more cautious. Yet what we've also learned is that when providers reassure patients that AI is there to support, not replace them, their confidence rises. And over time, as AI reduces clinician burnout and allows for more patient engagement, we expect that trust to continue.
- [10:45]
A
Thanks, Dr. Chait. It's so interesting what you're saying here. That simple act of providers communicating to patients the purpose of the tool seems to help to foster that trust and reinforce its purpose. So interesting to hear some of those findings from the Philips index. Dr. Saul, what would you Add in terms of barriers or areas where you're seeing the greatest potential for AI.
- [11:08]
C
I agree with Heather. This foundational challenge around the fragmented and siloed data sets or poor data quality is significant. In fact, McKenzie did a survey and found about 32%, about a third of organizations found that combining their different data systems was the hardest part of using AI. So, as we talked about, a lot of important work has to happen at that level to really unlock the capabilities of AI. I'd say the second barrier is a bit of a skills gap. Right. There's not a lot of AI expertise in health systems. And while most health systems, about 75% believe that AI can reshape the industry, only 6% have established a real strategy around this. And I think the lack of expertise and the newness of the technology, it's going to take a little bit of time to figure out how to penetrate these systems and get people up to speed and comfortable, and to Heather's point, get patients comfortable with these technologies. Another challenge is things are changing so fast. Every time you open a paper or look at the headlines, there's a new model and there's new startups and there's new capabilities. And I think that sometimes that hyperdynamic market can slow adoption because people want to wait and see how things settle out, or they need to bake into their decision making that they might be operating on shorter timelines than healthcare typically is comfortable. So maybe you need to adopt the solution knowing that in 12, 18, 24 months, you want to be able to have the flexibility and the agility to move to something else if there's a better, newer, more capable technology that's entered the market. And then we spoke about the integration complexity. And so that kind of goes with the data silos, but those are some of the ones that come up a lot when we speak to customers.
- [12:57]
A
Yeah, thank you, Dr. Saul. And I know you mentioned some of these concepts around flexibility, agility, and those words are often used in the context of cloud infrastructure. And I know this is an area of expertise for aws, so I wanted to go a little bit further on this cloud infrastructure. It's increasingly positioned really, as a key Enabler for scaling AI and healthcare. So wondering, Dr. Saul, if you could share just a few examples of where combining AI with cloud capabilities is already making a difference in care delivery and what were the factors that were critical for sustaining that momentum?
- [13:32]
C
Yeah, so, good question. And look, you know, again, this goes back to some fundamentals that predate AI, I suppose, but are extra Relevant now and that, that is that to, to work in this industry, this very sensitive industry, you have to have a scalable, very resilient, very secure infrastructure. It has to operate with relative cost efficiency, you know, and be responsive. And so these are areas where the cloud really excels. And you're seeing this payoff with. So think about imaging that's being processed with an AI algorithm. You have to get the image up to the cloud or in front of the algorithm, you have to spin up the GPUs, you need to run the algorithm and produce the results. So doing that in a sub second or very low latency manner and having sort of unlimited scale depending on the demands at any given moment are areas where the cloud is incredibly helpful. In fact, I don't think there's any other way to do it. So some examples, we have a cloud based training and model development and testing studio called Amazon SageMaker and Philips Health Suite's been using AWS to develop algorithms across all of their products. And so that's one area where I think the power of the cloud has really played a very important role. And then I think about our clinical decision support work. And we have a data store called Amazon AWS HealthLake where you can store clinical data using FHIR format and apply algorithms to it. And so that's increasingly being adopted to facilitate and solve some of the challenges we spoke to earlier. So, you know, I think the most important thing is really again having the right foundational elements in place so that you can unlock the full potential of these AI capabilities that are coming so rapidly.
- [15:23]
A
Fantastic examples. Thank you so much. Dr. Saul and Dr. Chait wanted to check in with you too. Any examples that come up from your end and, and where you sit at Phillips, Anything that you'd like to share in this realm?
- [15:36]
B
Well, I think Jarrod has laid a great foundation, Erika. And an example of what we've done together with NYU and digital pathology has been truly transformative and it speaks to the power of the partnership and a willing health system to undergo change. Now, traditionally, pathology has been based on glass slides. And when Jared and I were training, we spent time in the hospital basement. We probably broke a few slides as we carried them from the trays to the microscope. But in our changing world, you cannot apply AI to glass. So the first step for NYU was digitization. And Philips, as a leader in the digital pathology space, worked with AWS and NYU to transform their department with cloud based capabilities, scanning and image management. And the results speak for themselves. In less than a year, 100% of their surgical pathology cases are digitized and 95% are now signed out out digitally every day. With more than 90 pathologists having adopted the digital workflow in under a year, that's not just an AI story. It's about faster collaboration, easier second opinions, and more integrated diagnostics. And that same model of combining AI and cloud is expanding into other areas like cardiology and radiology, where seamless workflows are just as critical. And Philips Integrated Diagnostics is truly integral in driving multidisciplinary care. And this is done through our partner, AWS.
- [17:12]
A
Thanks, Dr. Chait. So interesting to hear how this is expanding to other areas of healthcare and really that rapid adoption rate. I think you said in less than a year, 100% of surgical pathology cases digitized. I mean, that's, that's really incredible. I feel like I could keep talking to you both about this for quite some time, but I know we are winding down on our time together, so I wanted to be sure to get some of your closing thoughts. You know, is there anything we didn't touch on or any final takeaways that you'd like to share with our audience, including how you see AI shaping care in the years ahead? Dr. Chait, maybe if you'd like to get us started on closing thoughts.
- [17:49]
B
Well, Erica and Jared, it's been a real pleasure to join this conversation. And if I have to leave one thought, it's that AI is already making healthcare feel more human again. And the tools that provide ambient listening and automated reporting are reshaping the landscape to enable physicians to truly practice medicine again. And AI is bringing the right information to the point of care to help clinicians make faster decisions with more confidence. And I believe we'll see this continue to gain momentum at Philips together with aws. Our commitment is simple, better care for more people. That's where the future is headed and it is an exciting journey to be part of such a Great note to.
- [18:31]
A
End on, Dr. Chait. Thank you, Dr. Saul. I'll pass it over to you for any flavor. Final closing thoughts.
- [18:36]
C
Yeah, echoing Heather's sentiment, it's been a great conversation. Real pleasure speaking with you both. I'll just leave with this one thought that, you know, the future of healthcare AI really is more than just individual tools and algorithms. It's really how do we create an integrated ecosystem where in every aspect of care delivery and the patient experience, the physician experience, you know, from diagnosis all the way through treatment. And follow up, how do we use AI to improve care, deliver it better, make it more accessible, make it more affordable. Healthcare professionals focus on what they do best, which is caring for patients and less on administrative tasks. As we've talked about in this conversation, there's a lot of promising work happening across all those dimensions and it's really exciting to think about where this is going to lead to in the coming years.
- [19:27]
A
Well, Dr. Saul, Dr. Chait, it's been a pleasure learning from you both today not only about AWS and and Philips partnership but what you're seeing in the industry and this hope that you have in AI scaling, more integrated care and also as you said, Dr. Chait, making healthcare feel more human again. So thank you again for making time for Beckers today. It's been a pleasure having you both.
- [19:50]
C
Thank you.
- [19:50]
B
Thank you.
- [19:52]
A
And we'd also like to thank our podcast sponsor for today AWS listeners. Please be sure to tune into more podcasts from Becker's Healthcare by visiting our podcast page@beckershospitalreview.com.