← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Scaling AI Innovation at Seattle Children's with Dr. Zafar Chaudry
00:24:05
Loading summary
Transcript23 lines
- [00:02]
A
This is where healthcare leadership comes together. Becker's 16th annual meeting brings more than 3,500 hospital and health system executives and nearly 800 speakers to Chicago, April 13th through the 16th. This year's event includes keynote conversations with Dallas Cowboys legend Troy Aikman and former President George W. Bush. For the agenda and event details, visit Beckershospitalreview.com and click on the Events tab in the upper right. We're looking forward to hosting you in Chicago.
- [00:32]
B
This is Laura Dearda with the Beckers Healthcare Podcast. I'm thrilled today to be joined by Dr. Zafar Choudary, Senior Vice President and Chief Digital Officer, as well as Chief AI and Information Officer at Seattle Children's. Dr. Choudhury, thank you so much for being here today.
- [00:46]
C
Thanks for having me. My pleasure.
- [00:48]
B
Absolutely. Well, I know we're going to talk a lot about technology and AI and how you're looking at the future, but before we dive into that, can you tell us a little bit more about yourself and Seattle Children's?
- [00:58]
C
Yeah. So I've been running the IT shop here at Seattle Children's. It's been about eight and a half years, but I've been in the healthcare IT industry. This is year 40, and I actually started my journey in health as a physician. So I'm one of those folks that has defected to the dark side over many years and now fully embedded on the technology side.
- [01:25]
B
Absolutely. That's fantastic to hear. I know you've got a lot going on, especially when it comes to AI in some really advanced things that you're doing there at Seattle Children's. And so I'm excited to dig in. Can you tell us a little bit more about some of the things you've been doing in the last year? What's an initiative you're most proud of? What did you do and what were the results?
- [01:44]
C
Yeah, So I think AI, obviously, 2026. Huge buzz around AI. We've been doing AI for quite a few years now, and it just seems to get better with the more data you feed into the engine. So about 18 months ago, we started a refresh of our data warehouse. So the question was, how do you get the right information into the right place? Because that's the information you're going to need to build your AI models so that they're accurate and are beneficial to an organization. So we moved from an older platform, which was IBM Netezza, on to a full Google Cloud platform for all of our data analytics. And our data lake is fed with 171 systems at present. And that's where all of our data resides. And then from that data we've been working with stakeholders. So I truly believe that the use cases for AI are focused on what do the end users need, not necessarily what I think they need. And so one of our most successful programs, we use that Google platform, we use Gemini AI, we use Vertex AI. So of course we did the training mandatory for everyone in AI in our organization. We set up AI user policy. We have an AI review board that reviews all the intake for use cases that our stakeholders are asking for. And then out of that, what came about was our first agentic AI model, which we call pathway assistance. What we did is we took many, many, many years worth of knowledge around clinical pathways to treat kids that we had developed. And we actually freely give those pathways away to hospitals that want to use them around the world to. And we took those pathways, 6,7000 pages of PDFs validated by clinicians over the years, and we fed it into an agent, which we call pathway assistant. And that agent was able to amalgamate that information and then create an agentic AI agent that allows people to query it and question it and have a dialogue with it. So if you're a, a new physician and you're trying to learn how to practice medicine and pediatrics, you can tell the agent about your patient, the age, the weight, the symptoms, what's going on with the patient. And the agent will look at all the information it has access to and start guiding that physician through the care pathway for that patient. It'll tell the physician what to do next, what tests to run, ask for the results of the test, then step through the diagnosis, the differential, the treatment plan. And that helps newly trained physicians learn how to practice. For seasoned physicians, it's a validation tool. It makes sure that they are doing the right thing for the right diagnosis, the right condition. And that's been very, very accurate and very, very successful. And that then becomes your friend, know that phone a friend thing, and you're able to query it and get the answers that you need. So that's one example of an agentic AI that we've built. Another example is a translation assistant. So we have a patient population that is varied. We tested English to Spanish, translation of instructions, discharge instructions, et cetera, post op instructions. That's been hugely successful. We did find though, that AI agents translating to other languages like Ukrainian and Somali, weren't as accurate. And so we didn't pursue other languages. So we're doing English to Spanish and Spanish to English. That's another example of a use case that has been valuable not only to our clinical staff, but also to our patient population. So those are some of the things we've been working on. There's many more, of course, there's more in the pipeline than we can probably afford to do or our resource to do in one year. But we stepping through multiple use cases, the third one I'll throw out there is ambient AI. We've agreed, we've set up a partnership with the Bridge and we have launched the ambient AI transcription for all of our physicians. We bought an enterprise license and that's currently being rolled out with huge success. Great feedback from our clinicians in the test phase. Many tell us that they would never allow us to take this tool away from them. So that's another win that we've had in the AI space.
- [06:52]
B
That's amazing to hear and I think especially really cool to understand how you're using AI. Looking at AI agents, looking at some of the diagnoses and validations as well as then getting into the translation as well. And of course, you know, ambient technology is something that has become so much more prevalent, especially in the last 12 months within many organizations and echoing those same sentiments, that clinicians are finding it extremely beneficial in time savings and also for their relationships with patients and satisfaction with their jobs overall. So I think that's amazing to hear. And you know, my thing I'm curious about is for any one of these use cases, whether it's ambient or some of the agents and more, how do you make sure that they are continuing to work at the level that they're supposed to? How do you measure that success and then continue monitoring them to make sure that there aren't unwanted changes as time goes on.
- [07:50]
C
So the success with these use cases has to come from the partnership between IT and the clinical teams. So we may be the folks that engineer the back end of these tools and these agents, but on the front end there are many doctors involved. So for our pathway Assistant, we had 50 physicians work on this project, each of which dedicated about 100 hours of their time. So that was in the build design, test, post test phase. And the same group of physicians, predominantly in my informatics team, are always testing and checking if new pathways are added. Those pathways are validated, tested, torture tested, and then released into the agent. So it's a continuous PDSA cycle. But you don't do that with IT people, you do that with the people that use the tool on a day in, day out basis. And the people are Volunteering their time because they see the value of the tool for them in their day to day work. And so that's the partnership you really have to build. Because if it's only driven by it because it's the next shiny toy, you're not going to get adoption and you're certainly not going to get the most important thing in AI, which is sustainability of solutions. Because it's easy to create something shiny and then people won't use it after a few weeks, it's hard to get them to continually use it, provide feedback. If there's technical tweaks, the team can do any technical tweaks, but it is a PDSA cycle.
- [09:32]
B
Got it. That makes a lot of sense and you know, it is really helpful context for how these different teams and departments work together in order to make sure that the technology is functioning in the most beneficial way possible. Now, looking ahead into 2020, what are some of the big priorities and headwinds that you're most focused on?
- [09:50]
C
So in 2026, of course we've got to make sure that the costs that we have in providing these support services and IT are optimized and continue to remain optimized. We're going to continue to do some application rationalization because we have a lot of different systems and we need to validate whether those systems are still of use in 2026. We have consolidated and outsourced most recently all of our infrastructure. So all of our servers, storage, backup, disaster recovery for all of our systems is now a service contract that we don't run, that's outsourced to Rackspace and that's been very successful because that helps us deal with our technology debt in the future and also allows us to have service level based infrastructure which is important if you're trying to grow in the AI space because the demands on infrastructure won't go down, they will continue to go up. And at the same time, I like to almost every year take a look at my organizational structure in IT and make sure that that is still optimized and fits the the needs of the customers I serve, which are the end users, the clinicians, the patients, the parents, caregivers. We'll also continue to build tools in our patient facing app. We have a patient and parent advisory group. We meet with them monthly and they tell us the things that we should build that they need to help them navigate our health system. So we'll continue to innovate in the digital space at the front end and that will also continue this year. And the other thing we'll do is we'll continue to track any regulatory changes, federal changes that impact us. The thing that will always keep us up at night moving through 2026 will be cybersecurity. So we will continue to keep our eye on that and also continue to optimize tools in that space. There is also use of AI around cybersecurity to protect our organization. And we'll continue to educate and train our end users in the cybersecurity risks and pitfalls that are very easy to fall into.
- [12:19]
B
That's helpful to understand. You know, I know so many different areas that you touched on are so critical, but it was cool to hear you talking about some of the different things that you're doing in order to meet the needs of patients, especially, you know, with some of the feedback you're getting in the advisory group. Is there anything you've been able to build because of their feedback that you know, has been particularly successful? Any success stories you can share with us?
- [12:48]
C
Yeah, so I think one of the really interesting ones was around surgical notification. So the, the use case was that when a patient goes into surgery, the parent usually accompanies their child and then waits for their child whilst the child is in surgery. And as you know, there are different types of surgeries, some are short, some are extremely long. But when you are a worried patient, you don't necessarily know if you should leave the surgical suite because you don't know when your child is coming out of surgery or if there are any updates about your child in surgery. So the feedback we had from our patients was the parents of our patients was that I can't leave the surgical suite and sometimes my kids in surgery for 4, 5, 6, 7 hours and I'm afraid to leave because something might happen and I wouldn't know because they wouldn't be able to find me. And what that means for me is I don't even have the opportunity to go and get a coffee or even get something to eat. So could you build us something that could keep us informed and then allow us to not be stuck in one place? And so within our digital front door app, we added a component that actually keeps that parent or caregiver updated on where their child is in what stage in the surgical suite. And it's code based. So we issue them a code they can see on their app or even on the the big board in surgery and it will tell them whether the child has been taken in, is an intake, is in surgery, is in post op, is in recovery, where they are exactly which then frees the parent or caregiver to go out and, yeah, get that small sandwich to eat so that they can continue to not be anxious and then come back in time to welcome their child back from surgery. So that's a really small thing, but it's a big thing to someone who is super worried about their child and doesn't want to move from one location.
- [15:03]
B
Absolutely. That's such a powerful example of the ways that technology can provide a big difference within the care journey. And truly, truly amazing. I can imagine for the families to be able to have that peace of mind when a child is going through surgery, which is huge. Now, I'm curious, what do you think the hardest thing you'll have to do in the coming year will be?
- [15:28]
C
I think the hardest thing for us in 26 will be deciphering the impact on our health system around the big bill that was approved. And then for me in it, trying to figure out what that means financially to my budget because I have to right size my budget, balance my books, and still continue to provide the services that our stakeholders and our patients and our parents need. And that's not always easy in a climate when costs are actually going up. You know, as you know, we've been talking about AI today, but those tools cost money, infrastructure costs money, labor and skill sets cost money, and those costs aren't coming down. So this year we'll have to be very, very careful around continuing to innovate, but also continuing to do more for less and be fiscally responsible because we need to do that so that we can keep our lights on and doors open.
- [16:34]
B
That's such a good point. And I think especially when you look at some of the financial pressures, I know every dollar in healthcare is precious. And investing in technology while critical, especially when you're trying to make sure you've got the right experience in quality, inefficien, you know, certainly requires an amount of value that it has to bring back into the system. And so when you're making these decisions, in having that fiscal responsibility in mind, what do you see as being rising to the top, I guess, for your the next year? And how do you work with different companies and vendors in your team as well to make sure that they're making the right decisions given those limited resources?
- [17:17]
C
Yeah. So I think the most important thing for any health IT shop is to validate everything that they are actually running on a daily basis. You know, we have over a thousand different applications that we support and run at Seattle Children's. And the question I always ask my team is, are we sure those are the right tools and they are the right software applications that we should continue to use? And that's not a question we answer. We have to go and ask the stakeholders who use those tools. And if there are tools that really aren't being used or can be archived, can be deprecated, can be consolidated into newer tools that we've purchased, then we need to undertake that exercise. So rationalization, year on year is always key. And you do that rationalization in partnership with the actual users of those systems. Also in parallel, we're going to have many, many tools and subscriptions come up this year, as we do every year. And we will. We have a vendor management team in it. We work with our supply chain and we always fight back. So when a renewal comes through, we look at the renewal, we see how much it has gone up, and we then go back and forth and fight the good fight to make sure that the increase isn't too taxing on our budget budgetary model. And so that's, I would say sometimes that's the fun part of going back and forth and arguing till you get to something that's amicable for both sides because, you know, vendor relationships are partnerships. So we do that consistently. And I have a team that pushes back on renewals and makes sure that we're renewing at the right rate so that we can actually afford it. And then going through the health system and making sure that we've inventoried all of our assets. Asset management can sometimes be a problem. Make sure that we have taken in all the equipment that needs to be replaced and making sure that any equipment that's end of life is either deprecated or replaced. And if you continually do that, then that you'll find tools, pieces of hardware that you don't need. And if you can decommission those pieces of hardware, that also ends up saving money because you don't renew on maintenance and all of maintenance and support. So then you have to do a good job around your IT asset management portfolio, otherwise you'll have massive amounts of hardware sprawl that becomes unmanageable. So the team will also spend the year going through all of those lists and making sure that firstly we can find everything and secondly making sure everything should stay in place or be deprecated.
- [20:08]
B
That's helpful to understand. Thank you for laying out that process so clearly. And I think goes back to something else you mentioned at the beginning of our conversation. Looking at decommissioning those things that aren't in use anymore and managing that technical debt in a different way, especially when you think about that technical debt, how does your team approach those conversations or really efficiently managing some of those things and so you can cut back on what isn't in use.
- [20:35]
C
Yeah. So I think the interesting thing about working in health and health, it is a lot of the work that we do isn't technical. It's very much communication. So once you've identified assets, the next step is to go and validate the asset with the person who has the asset or who we think has the asset, and then validate whether that use case that they have the asset for is still a valid use case or not. We can take the example of laptops. How many times do you find out that a user has been issued more than one laptop in your organization? And then if they have go and validate whether they truly have two or three, or whether or not two of them are no longer functioning and they never gave it back to us for decommissioning. So it's sort of a detective work in addition to lots of communication. And my job is to communicate to the health system about all these different initiatives that we have and ask my user community for help because we can't do it alone. So sometimes we have amnesty programs. Hey, if you have any of our equipment that you didn't give back, let us know. And people come forward and say, yeah, I forgot to tell you this wasn't working, I'll hand it back and then we can decommission it. So my job is to communicate at system level the different initiatives we have. Then the team on the ground will go and interact round in our different facilities, look at things, find things, either refresh things or deprecate them all together. And that's a, that's a continuous cycle that we will undertake during the year because the more you can communicate your challenges to your end users, the more helpful they can actually be in partnering with you to, to deal with the challenges ahead.
- [22:37]
B
I love that. Thank you again for digging a bit deeper there. Now, before we wrap up, I wanted to ask you about growth. Where do you see some of the best opportunities for organizational growth in the next couple of years?
- [22:47]
C
But certainly for us in pediatrics, we are seeing growth in terms of patient demand for our services. So we will definitely be expanding across the region that we serve in terms of access to clinics, access to out of hours services that our parents and kids need, and, you know, more streamlined ways in getting the appointments and secondary and tertiary care that parents need for their kiddos in the Pacific Northwest. So it's a multi pronged approach. Some growth, some new clinics, some new ambulatory centers. And we will continue to do that through 26 so that we can meet the growing capacity that we have here in the Pacific Northwest.
- [23:38]
B
I love that. Thank you so much, Dr. Chowdhury, for everything that you have talked through with us today. This has been such a fascinating and valuable. You know, I'm excited to see you as well at our annual meeting. I know you'll be there speaking on a panel and just always excited to have you as part of what we're doing here at Becker. So thank you.
- [23:57]
C
Well, thank you. Thank you for spending some time with me. Always a pleasure to chat with you, Laura.