Latest / Becker’s Healthcare Digital Health + Health IT / Dr. Clara Lin, CMIO and VP of Digital Health & Informatics at Seattle Children’s

Transcript

Dr. Clara Lin, CMIO and VP of Digital Health & Informatics at Seattle Children’s · Becker’s Healthcare Digital Health + Health IT

• 0:00
  Exciting things are happening at Becker's Healthcare. Stay
• 0:04
  ahead of industry trends with the new Becker's
• 0:06
  CFO plus Revenue Cycle podcast,
• 0:08
  your go to source for insights from top
• 0:11
  healthcare finance leaders. Tune in wherever you get
• 0:14
  your podcasts.
• 0:15
  And don't miss the tenth annual health IT
• 0:18
  plus digital health plus RCM conference, happening September
• 0:22
  30 to 10/03/2025
• 0:24
  in Chicago.
• 0:25
  Join thousands of
• 0:27
  executives, engage with industry leaders, and explore the
• 0:30
  future of health care innovation.
• 0:32
  Learn more about our upcoming events at beckershospitalreview.com.
• 0:36
  See you there.
• 0:38
  This is Laura Dierda with the Becker's Healthcare
• 0:40
  podcast. I'm thrilled today to be joined by
• 0:42
  doctor Clara Lynn, vice president of digital health
• 0:44
  and informatics as well as chief medical information
• 0:47
  officer at Seattle Children's. Doctor Lynn, it's a
• 0:49
  pleasure to have you on the podcast today.
• 0:51
  Hi, Laura. Thank you so much. Glad to
• 0:53
  be back.
• 0:54
  Absolutely. And, you know, it's always so fascinating
• 0:56
  to talk with you because Seattle Children's really
• 0:58
  is doing some innovative things and, is on
• 1:01
  the forefront of a lot of what's happening
• 1:02
  with digital technologies, AI, and health care and
• 1:04
  more. And I know you recently partnered with
• 1:06
  Google on some technology developments as well, so
• 1:09
  I'm excited to learn more about that. But
• 1:11
  before we dig in, for those who
• 1:13
  may be newer to the podcast, can you
• 1:16
  tell us a little bit more about yourself
• 1:18
  and Seattle Children's? What makes it unique?
• 1:21
  Yeah.
• 1:22
  Seattle Children's is a coronary
• 1:25
  referral hospital in the Pacific Northwest.
• 1:28
  One of the things that we're super proud
• 1:29
  of of among other things is that we
• 1:31
  are ranked as one of the top 10
• 1:34
  children's hospitals in the country by US News
• 1:36
  and World Report, and we are unique in
• 1:39
  that we serve a very large region. We
• 1:41
  cover the Washington State, Alaska, Montana, and Idaho
• 1:45
  region,
• 1:46
  as the referrals center for kids in this
• 1:48
  area, and we're also the only freestanding children's
• 1:51
  hospital in in the state.
• 1:53
  Well, that's fantastic. And, you know, could you,
• 1:56
  dive in and tell us a little bit
• 1:57
  about the accomplishment that you're most proud of
• 1:59
  from the last year?
• 2:01
  Yeah. I think,
• 2:03
  we've done so many cool things. We really
• 2:05
  try to keep Seattle Children's as a leader
• 2:08
  in innovation, particularly when it comes to pediatric
• 2:11
  health care and be an advocate for for
• 2:13
  kids when it comes to,
• 2:15
  the the national and and global trend of
• 2:18
  developing
• 2:19
  and innovative technology.
• 2:21
  But that being said, if we go back
• 2:22
  to sort of the bread and butter, the
• 2:24
  really basic stuff, some of the things that
• 2:26
  we're super proud of are, of course, our
• 2:28
  quality of care,
• 2:29
  the fact that we are we've been
• 2:31
  ranked and certified by US News World Report
• 2:34
  as one of the top 10 children's hospitals
• 2:36
  as I mentioned earlier. But, also,
• 2:38
  we did our class survey for the first
• 2:40
  time in three years in 2024.
• 2:43
  The last time that we did it was
• 2:44
  in 2021
• 2:45
  right after we implemented
• 2:47
  a new EHR in the middle of a
• 2:49
  pandemic.
• 2:49
  And then this is the first time since
• 2:51
  we've come out of that sort of stabilization
• 2:53
  phase and moving into more of an optimization
• 2:55
  and innovation phase.
• 2:57
  This is the first time we've done a
• 2:58
  class survey, and we're so proud that our
• 3:01
  both our physician and our nursing
• 3:03
  EHR satisfaction score went up significantly.
• 3:07
  My role as the chief medical information officer
• 3:09
  is of course with our providers and making
• 3:11
  sure that,
• 3:13
  that the experience with health IT in general
• 3:15
  is a good one. And the physician and
• 3:18
  provider
• 3:19
  class,
• 3:20
  EHR satisfaction score went up by 30 points
• 3:23
  in the three years, and that's just something
• 3:24
  that that we're so proud of. I mean,
• 3:26
  there is so much AI, so much cool
• 3:28
  stuff out there, but really the bread and
• 3:30
  butter things to make it
• 3:32
  a good environment for our clinicians to work
• 3:34
  in, to make it not a,
• 3:38
  cumbersome some experience for them to provide care
• 3:40
  to to the kids in the region. That's
• 3:42
  really what my job is, and, really excited
• 3:45
  to see that that outcome.
• 3:47
  That's great to hear. And, you know, when
• 3:49
  you look at those changes you've made and
• 3:50
  the transformation that brought you to this point,
• 3:53
  where you're seeing such an increase in clinician
• 3:55
  satisfaction,
• 3:57
  what what did you do? What can you
• 3:58
  attribute to some of those things? Is it
• 4:00
  just the advanced technology, or are there other
• 4:02
  ways that you have really been able to
• 4:04
  connect with the,
• 4:06
  clinical leaders in in team and,
• 4:09
  improve that satisfaction that that you have with
• 4:11
  the EHR?
• 4:13
  Yeah. I think it's it's a little bit
• 4:15
  of both.
• 4:16
  AI and innovation
• 4:18
  is definitely top of our mind. It's top
• 4:20
  of everyone's mind. But really the basic stuff
• 4:23
  the going back to the basics the bread
• 4:25
  and butter stuff is also really really important
• 4:28
  making sure that our EHR
• 4:30
  is optimized in from an informatics perspective,
• 4:33
  from a usability perspective, from just a supporting
• 4:38
  and designing workflows that make sense. I think
• 4:41
  all of those things helping helping every role
• 4:43
  practice at the top of their license, helping
• 4:45
  supporting our physicians and our providers to do
• 4:48
  what they're trained to do to help,
• 4:50
  provide the best quality of care. I think
• 4:52
  all going back to the basics is what
• 4:54
  we really focused the last three years,
• 4:57
  on doing,
• 4:58
  and, and it's really paying off in in
• 5:00
  that in that survey result.
• 5:03
  That's helpful to know. And and really, definitely
• 5:05
  centering to have an understanding of just, you
• 5:07
  know, getting back to those basics and understanding
• 5:10
  what is gonna be most helpful and beneficial
• 5:12
  for the clinicians and bringing that forward. I'm
• 5:14
  curious. Where do you see some of the
• 5:16
  growth opportunities for your team in the next
• 5:17
  twelve months or so?
• 5:20
  Yeah. I think the from a from a
• 5:22
  health IT perspective,
• 5:23
  like everybody
• 5:25
  else, the biggest growth opportunity is, of course,
• 5:27
  artificial intelligence and what it can offer,
• 5:30
  and not just the the the potential that
• 5:32
  it can bring to the care that we
• 5:34
  deliver and the experience of our our,
• 5:37
  workforce, but it's also
• 5:39
  in terms of how we can
• 5:42
  help our internal workforce skill set grow when
• 5:45
  it comes to artificial intelligence. And we're not
• 5:47
  just talking about the IT teams growing and
• 5:49
  not just talking about giving, you know, cool
• 5:52
  generative AI to our our every, you know,
• 5:55
  employee to to write emails or make PowerPoints.
• 5:58
  That that is great. But I think what
• 6:01
  another really big opportunity here is we can
• 6:04
  help our workforce become more comfortable
• 6:07
  in using artificial intelligence, and they can have
• 6:09
  the tool built and the framework that they
• 6:12
  need in order to, and kind of a
• 6:14
  mental model that they that we can provide
• 6:16
  them to evaluate
• 6:17
  AI technology when they're interacting with it on
• 6:20
  a day to day basis. So evaluating it
• 6:21
  critically. Am I using it right? Is this
• 6:23
  tool fishy?
• 6:25
  You know, do I feel like there's a
• 6:26
  concern here? How do I do how how
• 6:29
  do I use this particular tool on my
• 6:31
  desktop, you know, responsibly? I think those are
• 6:33
  that's that's where I think we need as
• 6:35
  an organization and as an industry need to
• 6:37
  grow in is to start to disseminate this
• 6:40
  type of,
• 6:41
  this mental model, this framework to help them
• 6:43
  critically evaluate these technologies,
• 6:45
  on a regular basis.
• 6:48
  Got it. And that and it's fascinating to
• 6:49
  hear it. And really cool,
• 6:52
  to see some of those the AI technology
• 6:54
  come into
• 6:55
  the day to day use spaces,
• 6:57
  and understanding what the workforce needs, supporting them
• 7:00
  the right way, but then also of scaling
• 7:01
  them into a a place where it,
• 7:03
  really becomes most valuable and useful. I'm wondering
• 7:06
  too, could you tell us a little bit
• 7:08
  about some of the things that you've been
• 7:09
  working on with Google and,
• 7:10
  just, you know, early what that experience has
• 7:13
  been like and and the outcomes?
• 7:15
  Yeah. So that was a that was a
• 7:16
  really cool experience. So,
• 7:18
  about a year ago when,
• 7:21
  artificial intelligence and particularly generative AI
• 7:24
  became sort of the talk of the town,
• 7:26
  Seattle Children's, like everyone else, started to think
• 7:28
  about sort of where our use cases are,
• 7:30
  where our needs are. So we're not we're
• 7:32
  not really going out there to look for
• 7:34
  technology and then trying to retrofit it into
• 7:36
  our environment. We're really starting with what the
• 7:38
  frontline users are asking for and what their
• 7:41
  where their needs are. And so,
• 7:43
  we
• 7:44
  we came up with a list of use
• 7:46
  cases that we thought would be,
• 7:49
  would be very beneficial to the organization from
• 7:51
  an operational efficiency standpoint, from an improving the
• 7:54
  quality of the care perspective.
• 7:57
  And this particular one that we developed with
• 8:00
  in partnership with Google,
• 8:02
  ended up being sort of a proof of
• 8:04
  concept for us, an early use case that
• 8:07
  we wanted to see if it's possible in
• 8:10
  a in a sort of fail fast model.
• 8:12
  Can we get this technology
• 8:14
  using Google Gemini,
• 8:17
  to to a state where we think it's
• 8:19
  clinically acceptable and from a workflow's perspective,
• 8:22
  that is usable and from a technology standpoint
• 8:24
  that is scalable.
• 8:26
  And so we can of the four or
• 8:28
  five, use cases that we ended up coming
• 8:30
  up with, we you know, for each of
• 8:32
  the use case, we partner with a different
• 8:34
  different
• 8:35
  team outside of Seattle Children's, and this particular
• 8:37
  one was with Google.
• 8:39
  And, what we've done is I I need
• 8:41
  to take us back a little bit further.
• 8:43
  So in 2010,
• 8:44
  our,
• 8:45
  Seattle Children's created this group called the clinical
• 8:48
  effectiveness program, which is under our center of
• 8:51
  quality.
• 8:52
  And what the clinical effectiveness program does is
• 8:54
  that it looks at all of the medical
• 8:56
  literature out there. There's millions upon millions. Right?
• 8:59
  The medical literature doubles every 70 something days
• 9:01
  these days. And the clinical effectiveness team sort
• 9:04
  of gathers the most up to date information
• 9:07
  and the guidelines that are out there for
• 9:10
  each and every disease across the 70 diseases
• 9:12
  or conditions that that we have on our
• 9:14
  catalog. And for each one of these condition,
• 9:16
  they develop
• 9:17
  a guideline, a pathway
• 9:19
  that we think is the standard provides the
• 9:21
  best of care in a standardized fashion. And
• 9:24
  so now that we have these 70 something
• 9:26
  conditions
• 9:26
  sort of created and
• 9:29
  provided a guideline and a pathway to to
• 9:32
  our clinicians, we can standardize the best of
• 9:33
  care for every kid regardless of which provider
• 9:36
  you see at Seattle Children's. And so that's
• 9:38
  kind of the basis that the shoulders of
• 9:40
  the giants that we're standing on. So this
• 9:42
  particular tool that we created with Google called
• 9:45
  Pathway Assistant
• 9:46
  sort of is an evolution to that. So
• 9:48
  we took these thousands of pages of really
• 9:51
  great content
• 9:53
  developed by the clinical effectiveness team under our
• 9:56
  center of quality,
• 9:57
  and we layered basically Google Gemini on top
• 10:00
  of it. And it's not also just the
• 10:02
  straight of Gemini that we can see on
• 10:04
  online, but it's really a Gemini that's been
• 10:07
  fine tuned specifically to these,
• 10:10
  these pathways that we see our children's has
• 10:12
  developed. So it's only referring to these pathways.
• 10:15
  We did it purely through prompt and fine
• 10:17
  tuning of Gemini.
• 10:19
  And, right out the box, it had a
• 10:21
  decent
• 10:22
  accuracy, but not good enough for clinical use.
• 10:25
  And we kept tuning fine tuning and prompting
• 10:27
  and changing, you know, the the prompts and
• 10:29
  and making sure that that Gemini understands what
• 10:32
  we were trying to get out of it.
• 10:33
  And the end product is a conversational
• 10:36
  chatbot that can ask you questions if your
• 10:38
  prompt is not specific enough, and then it
• 10:40
  gives you an answer back. So I can
• 10:42
  ask prompt the the chatbot, for example, the
• 10:44
  pathway assistant, for example. Hey. I have a
• 10:46
  four year old in front of me in
• 10:48
  the emergency room with an asthma exacerbation.
• 10:51
  What should I do next? The chatbot will
• 10:53
  basically ask you questions about the patient. And
• 10:56
  then in the end, it doesn't tell you
• 10:58
  what you should do next even though that
• 10:59
  was the question I asked as a physician.
• 11:01
  But what it will do is it will
• 11:02
  collate all of that evidence based information
• 11:05
  from the pathways that we've created
• 11:08
  over the last fifteen years,
• 11:10
  by these, you know, dozens of medical experts
• 11:12
  here at Seattle Children's. And then they present
• 11:14
  that information to the request or to the
• 11:17
  to the person that's chatting with and says,
• 11:19
  this is here all that information.
• 11:21
  Here is the page of guideline that where
• 11:24
  I'm referencing from, and here's what we the
• 11:26
  guideline suggests your next step of evaluation should
• 11:29
  be. And so the human ultimately is still
• 11:31
  the one that's making the decision,
• 11:33
  and ultimately deciding if this is the information
• 11:36
  that is most appropriate for the patient in
• 11:38
  front of them. But we using pathway assistant
• 11:40
  effectively shrunk the information searching time to just
• 11:44
  seconds for for the clinician.
• 11:47
  Wow. That's fascinating. You know? And really amazing
• 11:50
  what you're able to do bringing all that
• 11:52
  information,
• 11:53
  in in data and,
• 11:55
  pathways together to then, you know, in the
• 11:58
  moment, find what will be a recommendation,
• 12:01
  that is personalized to the patient. That's really
• 12:03
  cool. And and seems like it has a
• 12:06
  lot of benefits for clinicians as well as
• 12:08
  patients in the future.
• 12:10
  Oh, yeah. We're we're really excited. And and
• 12:12
  one of the things that that we use
• 12:15
  this also as a proof of concept to
• 12:16
  see, can generative AI and large language models
• 12:19
  cat catalog internal documents like this in that
• 12:23
  sort of a breadth. And also, it wasn't
• 12:25
  just text because we also fit part of
• 12:27
  the a lot of the guidelines are flowcharts
• 12:30
  and and spreadsheets and images and, you know,
• 12:34
  references and and things like that. And so
• 12:36
  it was really,
• 12:37
  a multimodal ingestion into the large language model.
• 12:40
  And so this was the proof of concept
• 12:42
  that we needed to show that, you know,
• 12:44
  it can
• 12:45
  technologies like this can accommodate
• 12:47
  our needs,
• 12:48
  beyond sort of the the the the pathway
• 12:51
  what what the pathway assistant itself can do.
• 12:54
  It's fascinating to hear. Thank you so much
• 12:56
  for sharing with us that,
• 12:58
  the technology and the innovation journey you've been
• 13:00
  on. I I think it meet makes for
• 13:02
  a lot of huge opportunities in the future.
• 13:04
  We'll just keep an eye on how things
• 13:06
  continue to evolve and grow.
• 13:08
  I'm curious. You know, we talked a lot
• 13:10
  about some of the cool things that are
• 13:11
  happening, but what are the challenges that you're
• 13:13
  anticipating, especially with technology and AI as well?
• 13:17
  Yeah. So I've been thinking about that a
• 13:19
  lot lately.
• 13:21
  I mentioned earlier that we selected
• 13:24
  a small handful, three or four,
• 13:27
  AI solutions that we developed ourselves,
• 13:30
  starting over the last year with with different,
• 13:34
  different teams and different people that we collaborated
• 13:36
  with.
• 13:37
  This was just one of one of the
• 13:39
  three or four that we developed. But then
• 13:41
  at the same time, the organization
• 13:44
  hasn't slowed down. Right? Like and and the
• 13:46
  industry didn't slow down. And so so the
• 13:48
  industry is booming, including our own EHR our
• 13:51
  EHR epic. You know, they're developing a ton
• 13:53
  of AI AI solutions. And then out there,
• 13:56
  there's also a lot more off the shelf
• 13:58
  products that a lot of our teams are
• 14:00
  evaluating and trying to incorporate into their workflows
• 14:03
  and into into the organization.
• 14:05
  So this is only a growing technology, and
• 14:07
  it's growing quite rapidly. So the biggest challenge
• 14:10
  is how do we support the technology
• 14:12
  from an infrastructure standpoint
• 14:14
  all the way up to governance.
• 14:17
  How do we make sure that we are
• 14:19
  evaluating,
• 14:20
  you know, implementing,
• 14:22
  scaling,
• 14:23
  and then maintaining
• 14:24
  that growing portfolio,
• 14:26
  whether it's something that's purchased or licensed off
• 14:29
  the shelf, or something that we've developed,
• 14:31
  upfront. And so so that is what I've
• 14:34
  been thinking about a lot and really trying
• 14:36
  to figure out because I think that ultimately
• 14:39
  will be the challenge, not just for Seattle
• 14:40
  Children's, but everybody out there in health care
• 14:43
  IT,
• 14:44
  that is seeing sort of a rapid expansion
• 14:46
  of their their AI portfolio right now.
• 14:50
  That's such a great point. And I I
• 14:51
  think that is a lot of organizations that
• 14:53
  are are really truly seeing the
• 14:55
  broad leaps in their AI technologies and portfolios
• 14:58
  or or pilots that they're working on. And
• 14:59
  then as you mentioned, just trying to figure
• 15:01
  out, you know, what's gonna make the most
• 15:03
  sense and, and keep that a growing portfolio
• 15:06
  under control is,
• 15:07
  easier said than done, I think. Yeah. For
• 15:10
  sure. Fantastic. Well, before we wrap up here,
• 15:13
  could you talk a little bit about the
• 15:14
  future? What is the number one thing that
• 15:16
  you're doing right now to set Seattle Children's
• 15:19
  up for long term success?
• 15:21
  I think that's exactly it. Is is how
• 15:23
  do I how do we, at Seattle Children's,
• 15:26
  set up an ecosystem
• 15:28
  so that we can accommodate this rapidly
• 15:32
  evolving field? Because we don't one is that
• 15:34
  we don't know
• 15:35
  we don't know where this is all going
• 15:36
  right now. It's such a fast moving train.
• 15:38
  It's super exciting.
• 15:39
  Everybody's on board. Most people are on board.
• 15:43
  But how do we how do we support
• 15:45
  that? Right? And so what I'm hoping to
• 15:48
  do now is to sort of set the
• 15:50
  organization up, not just from a people and
• 15:53
  skill set perspective,
• 15:55
  but also getting,
• 15:57
  like I said mentioned earlier, getting the workforce
• 15:59
  to a point where they can evaluate
• 16:02
  critically their experience with artificial intelligence
• 16:06
  on a day to day click by click
• 16:08
  basis
• 16:09
  so that, so that they know,
• 16:11
  hey, that just seems fishy or, hey, that's
• 16:13
  a great opportunity for me. So, you know,
• 16:16
  that that to me is is a really
• 16:18
  critical
• 16:18
  step in in democratizing,
• 16:21
  I guess, this technology to to every every
• 16:24
  area of the workforce. Because I I think
• 16:25
  it's only a matter of time that we're
• 16:27
  gonna start to see AI incorporated in everything
• 16:29
  that we use.
• 16:30
  So I think that that is in and
• 16:32
  that is, sort of what we're what we're
• 16:34
  working on next is how do we set
• 16:35
  up our workforce, both structured teams and also
• 16:38
  this AI literacy
• 16:40
  to to prepare us for the future.
• 16:43
  Absolutely. That it is amazing to hear. Doctor
• 16:46
  Lin, thank you so much for joining us
• 16:47
  on the podcast today. This has been a
• 16:49
  really fantastic conversation, and I look forward to
• 16:51
  connecting with you again soon.
• 16:53
  Thank you so much, Laura.