Latest / Becker’s Healthcare Podcast / AI Governance and Digital Access at Mount Sinai Health System with Girish N. Nadkarni and Dr. Nicholas Gavin

Transcript

AI Governance and Digital Access at Mount Sinai Health System with Girish N. Nadkarni and Dr. Nicholas Gavin · Becker’s Healthcare Podcast

• 0:01
  This is where health care leadership comes together.
• 0:04
  Becker's sixteenth annual meeting brings more than 3,500
• 0:07
  hospital and health system executives and nearly 800
• 0:10
  speakers to Chicago
• 0:12
  April.
• 0:14
  This year's event includes keynote conversations with Dallas
• 0:17
  Cowboys legend Troy Aikman and former president George
• 0:20
  w Bush. For the agenda and event details,
• 0:22
  visit beckershospitalreview.com
• 0:25
  and click on the events tab in the
• 0:26
  upper right. We're looking forward to hosting you
• 0:29
  in Chicago.
• 0:31
  This is Laura Dierdahl with the Becker's Healthcare
• 0:34
  podcast. I'm thrilled today to be joined by
• 0:36
  Girish Nadkarni,
• 0:37
  chair of the Windreich department of artificial intelligence
• 0:40
  and human health as well as director of
• 0:42
  the Hassold Plattner Institute for Digital Health and
• 0:44
  chief AI officer at Mount Sinai Health System,
• 0:47
  and doctor Nicholas Gavin, vice president and chief
• 0:49
  clinical innovation officer at Mount Sinai,
• 0:52
  to talk about some of the great things
• 0:53
  that they're doing, in AI, technology, and more.
• 0:56
  So, Girish,
• 0:57
  doctor Gavin, thank you so much for being
• 0:59
  here today. Thanks for having us. Thanks for
• 1:01
  having us.
• 1:02
  Excellent. Well, I'm excited for our conversation. But
• 1:05
  before we dive in, I'd love to have
• 1:06
  you both just briefly toss a little bit,
• 1:09
  about yourselves, your background, maybe, you know,
• 1:12
  twenty seconds or so. Girish, can we start
• 1:14
  with you?
• 1:14
  Yeah. Sure. My name is Girish. I'm a
• 1:16
  clinician and clinical informaticist.
• 1:18
  I'm the chair of the department of AI
• 1:20
  and the midlife department of AI and human
• 1:21
  health as well as chief AI officer.
• 1:24
  By background, I'm a clinician,
• 1:26
  internal nephrologist,
• 1:28
  but also have a lot of experience
• 1:31
  in clinical informatics,
• 1:32
  AI implementation,
• 1:34
  and, deployment.
• 1:37
  Fantastic.
• 1:38
  So glad to have that perspective. And doctor
• 1:40
  Gavin?
• 1:41
  I'm Nick Gavin. I'm the health system's inaugural
• 1:44
  chief clinical innovation officer. By training, I'm an
• 1:46
  emergency physician and clinical informaticist.
• 1:50
  We're both representing Mount Sinai Health System in
• 1:52
  New York, which is a large large
• 1:54
  academic health system,
• 1:56
  serving mostly New York City, but some, presence
• 1:59
  in Westchester, Long Island, and New Jersey.
• 2:02
  We have, eight acute care hospitals,
• 2:05
  hundreds of clinics, and, run many multi specialty
• 2:08
  group practices.
• 2:10
  Fantastic. And, you know, given the breadth and
• 2:13
  depth that you oversee, I can imagine there's
• 2:15
  a lot going on at any given time.
• 2:16
  But when you think about the last year
• 2:18
  or so, can you tell me about a
• 2:19
  project or initiative that you led and have
• 2:21
  been working on? What did you do, and
• 2:22
  what were the results?
• 2:24
  Sure. So I'll start.
• 2:26
  We
• 2:27
  have been really thinking about our digital front
• 2:29
  door in the health system,
• 2:31
  and one of the projects that we led
• 2:33
  over the last year was really expanding our
• 2:35
  access to asynchronous
• 2:37
  care.
• 2:38
  This is a a care modality
• 2:40
  that some health systems have leveraged,
• 2:43
  Kaiser,
• 2:44
  on the West Coast, Baylor Scott and White
• 2:46
  in in Texas.
• 2:48
  And this is a care modality that I
• 2:50
  think is in its infancy in terms of
• 2:52
  how it can be leveraged.
• 2:54
  What do we mean by asynchronous
• 2:56
  care? So asynchronous care basically says we we
• 3:00
  don't aren't dependent
• 3:02
  on a model where a patient has to
• 3:04
  find time to get on a video visit
• 3:06
  or go to a clinic in order to
• 3:08
  receive care.
• 3:09
  They can respond to a survey. They can
• 3:12
  provide wearable data.
• 3:13
  They can,
• 3:14
  answer questions,
• 3:16
  that we may have of them,
• 3:18
  asynchronously, and, ultimately, we'd give them
• 3:21
  anticipatory guidance, prescription medications,
• 3:25
  and really provide continuous care,
• 3:29
  not requiring,
• 3:30
  an in in office visit or even, even
• 3:33
  a video visit. We've seen a lot of
• 3:35
  our world shift in this direction,
• 3:38
  and people want to be able to,
• 3:40
  to do things from from the palm of
• 3:42
  their hand.
• 3:44
  And we've started to experiment with this at
• 3:46
  Mount Sinai. So we've expanded our asynchronous care
• 3:49
  often offerings.
• 3:51
  I really think asynchronous care today
• 3:53
  is where telemedicine was in 2017.
• 3:57
  The incentives aren't necessarily there yet, but we're
• 3:59
  starting to see,
• 4:01
  patients adopt this type of care.
• 4:04
  We're we're expanding on this in three different
• 4:06
  areas. One is in acute care. So,
• 4:09
  young women with urinary tract infection, for instance.
• 4:12
  Do they need to have a video visit
• 4:14
  or an in person visit,
• 4:15
  or can we shift this care to the
• 4:17
  most efficient,
• 4:20
  most,
• 4:20
  and most cost effective way of delivering that
• 4:23
  type of care?
• 4:25
  We were also experimenting with this around preventive
• 4:27
  care,
• 4:29
  particularly,
• 4:31
  prep medications for HIV
• 4:33
  prevention,
• 4:34
  and then chronic condition management. So,
• 4:37
  insulin regimen,
• 4:39
  GLP one titration. We're really interested in shifting
• 4:41
  that care out of the practice and into
• 4:44
  the digital space.
• 4:46
  This is one way we're trying to alleviate
• 4:48
  our huge,
• 4:50
  huge access challenge, within a large academic health
• 4:52
  system.
• 4:54
  The one thing that we also focused on
• 4:56
  over the last year was setting up appropriate
• 4:59
  governance structures
• 5:01
  for AI and digital health.
• 5:03
  Now AI and digital health are also intertwined
• 5:05
  with each of them. We sort of harmonize
• 5:08
  and streamline all of the government structures in
• 5:11
  order to make sure that it becomes more
• 5:14
  of an enabler and a force multiplier
• 5:17
  rather than a blocker. So we divided it
• 5:18
  up by sort of personas
• 5:20
  in terms of whether it's patient facing, clinician
• 5:23
  facing, workforce facing,
• 5:25
  researcher facing. And all of that gets rolled
• 5:27
  up into an executive committee where the highest,
• 5:32
  echelons of our health systems sit in order
• 5:34
  to make decision
• 5:35
  with rapid speed. So one of the things
• 5:38
  that we accomplished over the last year was
• 5:40
  to get the governance and guardrails up. The
• 5:42
  other thing that we set up also was
• 5:44
  sort of an AI assurance lab in order
• 5:46
  to make sure that AI is safe, effective,
• 5:49
  responsible so that anything that gets deployed into
• 5:51
  clinical care is assured and and made sure
• 5:54
  that we,
• 5:56
  it it and, you know, there's no bias,
• 5:58
  there's no hallucinations, etcetera, etcetera. So I think
• 6:01
  that was one thing that we were very
• 6:02
  proud of last year.
• 6:04
  That's amazing to hear. And I know that
• 6:06
  is no small task to set up those
• 6:07
  governance structures. And then for the asynchronous care
• 6:10
  too, what a a great benefit to the
• 6:11
  community. But, again, it takes change management. It
• 6:14
  takes a good amount of effort on the
• 6:16
  technology and people's side to bring that about
• 6:18
  and and get it done in efficient and
• 6:19
  effective way possible.
• 6:21
  I'm curious. When you look ahead to the
• 6:23
  next year or so, what are some of
• 6:24
  the big priorities and headwinds that you're focused
• 6:26
  on for 2026?
• 6:28
  Obviously, a lot of the the headwinds that
• 6:31
  our health system among many others across the
• 6:33
  country are facing is,
• 6:35
  all of the people who will be uninsured
• 6:38
  secondary to federal policy shifting people off of
• 6:42
  Medicaid roles and,
• 6:44
  reducing coverage through,
• 6:47
  ACA coverage programs. So we're doing everything we
• 6:50
  can to support patients
• 6:52
  in staying active on, on their Medicaid
• 6:56
  and making sure that people have continuity of
• 6:58
  care. But that's the biggest headwind that we'll
• 7:01
  face going into this year.
• 7:04
  Yeah. I completely agree with Nick. The regulatory
• 7:07
  and reimbursement changes around the biggest segments that
• 7:09
  we face this year. And sort of it's
• 7:11
  our goal as a leading health system in
• 7:13
  order to make sure that our community and
• 7:15
  our patients are well cared.
• 7:16
  That makes a lot of sense. You know,
• 7:18
  it's a huge challenge and concern, I know,
• 7:20
  for many hospitals and systems across the country
• 7:22
  as they're looking at what it will mean
• 7:24
  already with some of the the ACA changes
• 7:26
  and then into Medicaid for next year. So,
• 7:29
  you know, when you're looking at,
• 7:31
  the ways that you're preparing today or how,
• 7:33
  you know, you're you're getting ready for that
• 7:35
  type of reality, what does it look like
• 7:36
  from a technology perspective? Are there things that,
• 7:39
  you know, you have been able to put
• 7:41
  in place or or planning to to start
• 7:43
  digging into deeper as, you know, access to
• 7:46
  care and those kinds of things I can
• 7:48
  imagine are going to be, a really big
• 7:50
  focus in the next couple years?
• 7:52
  Yeah. I think what we're trying to do
• 7:56
  on the access front,
• 7:57
  which I I think is has to be
• 8:00
  our holy grail right now. Right? We're we're
• 8:02
  in an, hyper, hyper competitive market.
• 8:06
  We need to start leveraging our data
• 8:08
  to be really
• 8:10
  precise about matching patient needs
• 8:13
  with the supply
• 8:15
  of the physicians and providers that we have.
• 8:18
  We've sort of left it up to patients
• 8:21
  a lot of the time, and I think
• 8:22
  there's,
• 8:23
  more of a role that we need to
• 8:25
  play actively
• 8:27
  in getting patients
• 8:28
  matched to the right resource, making sure that
• 8:31
  those patients that are getting
• 8:33
  seen in our specialty clinics in particular
• 8:35
  really need to be seen in those practices.
• 8:38
  The efficient use of specialty
• 8:41
  time, real estate energy,
• 8:44
  should be a real singular focus for us,
• 8:47
  and we're thinking about how to leverage technology,
• 8:49
  particularly
• 8:51
  generative artificial intelligence
• 8:52
  to to really highlight where our opportunities are
• 8:55
  and where our gaps are.
• 8:57
  So I'd like to just add on to
• 8:59
  that. I think generative AI, I mean, patients
• 9:01
  with the chatbot provide a unique opportunity
• 9:04
  in order to augment care and improve access
• 9:06
  and provide shift care from reactive to proactive
• 9:10
  at a scale that has never been done
• 9:11
  before. Now, obviously, this needs to be safe
• 9:14
  and effective and responsible.
• 9:16
  But I think that, you know, just
• 9:18
  two advancing,
• 9:20
  trends of generative AI capabilities
• 9:23
  improving as well as the access issues that
• 9:25
  we are running into provide a unique opportunity
• 9:29
  in order to, you know, understand patient preferences,
• 9:32
  match them to the correct care,
• 9:35
  escalate if need be, but also think about,
• 9:38
  like, augmenting and maybe automating some of the
• 9:41
  routine
• 9:43
  administrative
• 9:44
  work that patients go through the pre authorization,
• 9:47
  insurance checking, etcetera, etcetera, in order to make
• 9:49
  the patient experience much better.
• 9:52
  That makes a lot of sense, and it's
• 9:53
  really helpful to understand how you're thinking through
• 9:55
  all of those things and and changing dynamics,
• 9:58
  on an executive level. Now, what do you
• 10:00
  think the hardest thing you'll have to do
• 10:02
  in the coming year will be?
• 10:04
  The the the constant challenge of the next
• 10:06
  year is gonna be where to prioritize and
• 10:09
  where to where to focus.
• 10:10
  We are being pitched different solutions
• 10:13
  every single day,
• 10:14
  and,
• 10:15
  I think it's really, really important to focus
• 10:18
  on execution
• 10:19
  and focus on execution around core
• 10:22
  problems for the organization
• 10:24
  as opposed to getting distracted by by shiny
• 10:27
  objects in the periphery.
• 10:29
  That's gonna be our challenge over the next
• 10:31
  year.
• 10:32
  Yeah. I completely agree with that. It's important
• 10:35
  to start with the why we are doing
• 10:36
  this, other than we find a point solution
• 10:39
  that solves a very new
• 10:40
  niche, very narrow niche,
• 10:42
  case case. And also sort of keep the
• 10:44
  patients at the center of all of this
• 10:46
  and figure out what's the best, most scalable
• 10:49
  way to help them safely and effectively. That
• 10:51
  would be the most important challenge coming in
• 10:53
  with '26.
• 10:54
  Because technology sort of keeps on improving,
• 10:57
  but we need to, like, focus and, like,
• 10:59
  make sure that we're helping patients other than
• 11:01
  just being tech building technology for technology.
• 11:05
  That makes a lot of sense. And I
• 11:06
  think, you know, it's always a helpful reminder.
• 11:08
  It seems like it should be easy, but
• 11:09
  I know it never is, especially as you
• 11:11
  said when there's, you know, every single day,
• 11:13
  new companies coming at you and new pitches
• 11:15
  coming forward.
• 11:16
  And speaking of looking into the future, I'm
• 11:19
  curious, how are you thinking about organizational growth
• 11:21
  in the next couple years or so? What
• 11:23
  does that look like, especially given some of
• 11:25
  the different factors and dynamics we talked about
• 11:27
  today?
• 11:28
  So I feel strongly that we,
• 11:31
  have an opportunity
• 11:32
  to step into the space of being
• 11:35
  the most convenient health system in our in
• 11:37
  our market. No one has really stepped forward
• 11:40
  to say we are,
• 11:42
  going to own
• 11:44
  being the most consumer oriented
• 11:47
  health system in New York City, and that's
• 11:49
  something that we're we're trying to advance internally
• 11:52
  as,
• 11:53
  as a concept.
• 11:54
  Patients are,
• 11:56
  increasingly demanding
• 11:58
  about,
• 11:59
  their health care services and want it
• 12:02
  to be on their terms
• 12:05
  on their in their time.
• 12:08
  And we need to we need to really
• 12:10
  meet people where they're at just as other
• 12:12
  industries
• 12:13
  have done.
• 12:15
  We've been a laggard,
• 12:16
  in this arena.
• 12:18
  We really, really have the opportunity with the
• 12:20
  the assets that we have within the health
• 12:22
  system to to leap ahead. So that's, something
• 12:25
  I'm hoping for to be part of our
• 12:27
  growth strategy in the year ahead.
• 12:29
  Yeah. And,
• 12:31
  completely agree with Nick. And to add to
• 12:33
  that, I I think we also have a
• 12:34
  unique opportunity of being a learning health system
• 12:36
  in which we,
• 12:38
  you know, learn
• 12:39
  from experiences, we learn from data, and we
• 12:42
  learn from our own actions in order to
• 12:45
  improve the next iteration of what we're doing.
• 12:47
  And what that helps is do sort of
• 12:49
  create a virtual cycle of,
• 12:51
  learning, acting, doing, and then further learning in
• 12:54
  order to improve,
• 12:55
  patient experience, improve outcomes, and sort of the
• 12:59
  consumer oriented piece in the learning health system
• 13:01
  goes in hand in hand because, ultimately, it's
• 13:03
  all about learning
• 13:05
  what we should do next and listening to
• 13:07
  patients and improving the experience.
• 13:09
  I
• 13:10
  love that. That's such a a clear way
• 13:12
  to really add a ton of value to
• 13:14
  the organization overall, but then improving the care,
• 13:16
  improving experience, and and really growing the,
• 13:19
  reputation of everything that you're doing at Mount
• 13:21
  Sinai. So I really appreciate your time today.
• 13:24
  Thank you so much, Girish and Nick. It's
• 13:25
  been a pleasure speaking with you both, and
• 13:27
  I look forward to seeing you as well
• 13:28
  at our annual meeting in April. I know,
• 13:31
  we'll be speaking on panels with us, and
• 13:32
  so that'll be a ton of fun to
• 13:33
  keep this conversation going. Thank you so much.
• 13:36
  You.