Latest / Becker’s Healthcare Digital Health + Health IT / Dr. Yasir Tarabichi, Chief Medical Informatics Officer at Ovatient and Chief Health AI Officer at MetroHealth

Transcript

Dr. Yasir Tarabichi, Chief Medical Informatics Officer at Ovatient and Chief Health AI Officer at MetroHealth · 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 Dierdo with the Becker's Healthcare
• 0:40
  podcast. I'm thrilled today to be joined by
• 0:42
  doctor Yasser Terabishi,
• 0:44
  chief medical informatics officer at Ovation and chief
• 0:47
  health AI officer at MetroHealth. Doctor Terabishi, it's
• 0:49
  a pleasure to have you on the podcast
• 0:50
  today. Thank you. It's a pleasure to be
• 0:52
  here.
• 0:53
  Absolutely. Now I'm excited to have you here
• 0:55
  because I know there's so many innovative and
• 0:57
  cool things that you're doing, especially on the
• 0:59
  digital health side at, MetroHealth system. But before
• 1:02
  we dive into that, can you tell us
• 1:03
  a little bit more about yourself and the
• 1:06
  organization? What makes it unique?
• 1:09
  Sure. So, I'll start with my role at
• 1:11
  MetroHealth. So I'm chief health AI officer at
• 1:13
  MetroHealth.
• 1:14
  And in in that role, it's really to
• 1:17
  guide the responsible
• 1:19
  implementation
• 1:20
  and strategic planning around AI initiatives within within
• 1:23
  the organization.
• 1:24
  And for reference, MetroHealth is Cleveland's safety net
• 1:27
  health care system that serves a diverse patient
• 1:29
  population
• 1:30
  in Cleveland and and beyond.
• 1:33
  The other hat that I wear is I
• 1:34
  am CMIO
• 1:35
  of Ovation. And Ovation is actually a virtual
• 1:38
  care start up that's spun out of MetroHealth
• 1:41
  and our partners
• 1:42
  in South Carolina, namely the Medical University of
• 1:45
  South Carolina.
• 1:47
  The goal behind this platform is to create
• 1:49
  a single solution that integrates into several different
• 1:53
  partner settings where patients can get integrated
• 1:57
  primary and urgent care. And so what that
• 1:59
  means specifically is we can take care of
• 2:02
  connecting patients
• 2:03
  to on the ground brick and mortar solutions
• 2:06
  at the partner organization and their home organization
• 2:08
  in a way that other point solutions that
• 2:10
  exist today do not.
• 2:13
  That's really fantastic to hear. You know, cool
• 2:15
  to hear about that kind of collaboration
• 2:17
  between,
• 2:19
  MetroHealth and,
• 2:21
  and USC. It certainly seems like a great
• 2:23
  opportunity
• 2:24
  to, work on a a digital technology that
• 2:27
  is really meaningful and beneficial for patients and
• 2:30
  can make a big difference for the health
• 2:31
  system as well in getting for the right
• 2:33
  people at the right places more quickly. How
• 2:35
  did that partnership come about?
• 2:37
  Actually, it's been, several years in the making,
• 2:39
  mostly a collegial,
• 2:41
  friendship
• 2:43
  and collaboration
• 2:44
  that started off at the highest levels of
• 2:46
  each organization,
• 2:47
  namely the executives at the time. And it
• 2:50
  was a it was a bit of a
• 2:51
  natural, I think, collaboration.
• 2:53
  So MetroHealth
• 2:55
  is known for being very innovative in the
• 2:58
  digital and informatics space. And I would say
• 3:00
  we're scrappy innovators. We we generally punch above
• 3:04
  our weights for for the organization that that
• 3:06
  we're
• 3:06
  at, the, you know, the size of our
• 3:09
  organization
• 3:10
  and and the fact that we're a safety
• 3:11
  net system that's typically more resource resource strapped
• 3:14
  than the average,
• 3:15
  you know, health care system.
• 3:18
  And MUSC really
• 3:19
  similar patient population that they serve.
• 3:23
  They also are
• 3:24
  relatively well known in the virtual care space.
• 3:27
  They have a very strong,
• 3:30
  asynchronous
• 3:31
  care program, actually, which is nationally recognized. And
• 3:34
  so the leadership on that side felt that,
• 3:37
  the leadership on this side
• 3:39
  could work together to achieve
• 3:41
  a share a common goal, really, a shared
• 3:44
  platform here between both organizations.
• 3:47
  And I think the the really cool thing
• 3:49
  about that opportunity
• 3:50
  is, you know, if you bring two organizations
• 3:53
  together that are trying to solve virtual care,
• 3:56
  integrated virtual care within their systems,
• 3:58
  When we start comparing the problems we're trying
• 4:01
  to solve,
• 4:02
  we start to align on shared problems and
• 4:05
  start to actually co solution.
• 4:08
  And what that does is create a platform
• 4:10
  or an environment where we start solving problems
• 4:13
  and providing solutions,
• 4:15
  that are much more scalable.
• 4:17
  Right? So we're pouring in resources together, co
• 4:19
  solutioning, collaborating,
• 4:21
  and we're creating a platform and a system
• 4:23
  that we can take to other partners. And
• 4:25
  so that's that's kind of our opportunity there
• 4:27
  to scale this out to a to a
• 4:29
  higher, you know, to a higher level across
• 4:31
  different states and different
• 4:33
  organizations.
• 4:34
  Our solution is built
• 4:36
  on the Epic platform and takes advantage of
• 4:39
  all of the things that the Epic platform
• 4:40
  provides. And we've built on top of that
• 4:42
  with Epic and in addition,
• 4:44
  internally with our developers
• 4:47
  to kind of complete that experience so that
• 4:49
  it really is as integrated and seamless as
• 4:51
  can be even though it it's truly a
• 4:54
  separate
• 4:55
  entity.
• 4:56
  That's fascinating to hear. Thank you so much
• 4:57
  for explaining that to us. Now I'm curious.
• 5:00
  What is the accomplishment that you're most proud
• 5:02
  of from the last year?
• 5:04
  Right. So I'll say, you know, I'll I'll
• 5:06
  use my two hats here to to go
• 5:08
  through these.
• 5:09
  When I think about my c m I
• 5:11
  CMIO hat innovation, so we have successfully launched
• 5:15
  a virtual care service for for patients that
• 5:18
  are in Ohio that are attributed to the
• 5:21
  MetroHealth system, both established and new.
• 5:24
  We've we've had, at this point, well over
• 5:27
  10 providers that we've hired on. We've seen
• 5:30
  thousands of patients, and we've had just fantastic
• 5:33
  results, raving reviews,
• 5:35
  a lot of emphasis on the ease of
• 5:37
  access,
• 5:39
  lot of a lot of good positive feedback
• 5:41
  on the integration.
• 5:43
  So, you know, I'll I'll give you an
• 5:44
  example.
• 5:45
  You have a fever.
• 5:47
  You call into our services.
• 5:49
  We figure out that you really ought to
• 5:50
  get swabbed for COVID,
• 5:52
  and flu, and we can go from a
• 5:55
  virtual visit
• 5:56
  to a swab in a brick and mortar
• 5:58
  facility that's a MetroHealth facility,
• 6:01
  to the prescription
• 6:03
  and even delivery, home delivery,
• 6:05
  for instance, Paxlovid,
• 6:07
  if if you had COVID,
• 6:09
  within several hours,
• 6:11
  which is just amazing compared to the usual
• 6:13
  experience of going into an express care, an
• 6:15
  emergency room and waiting for much longer.
• 6:18
  And so it's been really rewarding to see
• 6:20
  that integration
• 6:21
  play out.
• 6:22
  The other, you know, so I'd say that's
• 6:24
  a really big win we've had over the
• 6:26
  last year that's
• 6:28
  developed and evolved. And actually, we're hoping to
• 6:30
  go live in South Carolina
• 6:32
  shortly
• 6:33
  actually to to match the level of service
• 6:35
  that we've achieved in Ohio.
• 6:37
  And so that's been a phenomenal
• 6:39
  opportunity,
• 6:40
  lots of growth potential,
• 6:42
  for Ovation and also MetroHealth and MUSC in
• 6:45
  that collaboration. So that's really coming to fruition.
• 6:48
  We also have several other partners that are
• 6:49
  lined up thinking about, you know, launching in
• 6:52
  other states and other systems. So I think
• 6:53
  that the value of what we're developing
• 6:56
  has been well received.
• 6:58
  As far as the AI work,
• 7:00
  that I do at MetroHealth, so,
• 7:02
  you know, I'm a physician informaticist first. I'm
• 7:04
  a pulmonary critical care doc at MetroHealth.
• 7:07
  And
• 7:08
  my growth into the AI space really came
• 7:11
  from several implementations
• 7:13
  of AI solutions that maybe didn't always have
• 7:15
  something to do with my specialty,
• 7:18
  but got a lot of attention for being
• 7:19
  really pragmatic, responsible, ways to
• 7:23
  validate, vet, and implement a solution.
• 7:26
  And the takeaway from that has always been,
• 7:28
  you know, there's a large responsibility
• 7:31
  in implementing these solutions, especially in in very
• 7:34
  diverse, different environments where these models may not
• 7:36
  have been validated or developed.
• 7:39
  We always want to make sure we are
• 7:40
  improving,
• 7:42
  care for all of our patients. We wanna
• 7:43
  make sure that we are not
• 7:45
  worsening disparities that we know exist in all
• 7:48
  aspects of care.
• 7:50
  And, you know, beyond that,
• 7:52
  thinking strategically about how do we bring these
• 7:54
  solutions in, you know, how do we make
• 7:56
  sure that they're actually having an impact? And
• 7:58
  I'll say that a lot of our learning
• 8:00
  is consistent with with others and that the
• 8:02
  vast majority of the value
• 8:05
  of any AI or tech implementation for that
• 8:07
  matter
• 8:08
  is more to do with the people in
• 8:10
  the process and the why and less to
• 8:12
  do with the actual solution. So we've been
• 8:14
  building predictive models and analytics for really decades
• 8:17
  at this point in medicine.
• 8:19
  And our opportunity has really been harnessing them
• 8:21
  in a way that makes sense
• 8:23
  at the point of care
• 8:25
  and measuring and understanding that impact, both in
• 8:27
  terms of positive impact and also unanticipated
• 8:30
  consequences.
• 8:32
  And there are several examples that actually that
• 8:34
  we have published.
• 8:35
  You can find in the literature today if
• 8:37
  you looked.
• 8:39
  So, you know, we've taken that framework and
• 8:41
  we've built governance on top of that. So
• 8:43
  we have healthy,
• 8:45
  multidisciplinary
• 8:47
  AI governance in our in the form of
• 8:49
  our AI advisory committee.
• 8:51
  And I'll say one thing that I thought
• 8:53
  we've done relatively well is not jump onto
• 8:57
  the wagon too quickly. And so we did
• 8:59
  not want to be at the extreme cutting
• 9:01
  edge of these solutions.
• 9:03
  You know, we don't have that much flex
• 9:05
  money to spend on
• 9:07
  experimental concepts.
• 9:08
  And so we waited just long enough to
• 9:11
  know that, for instance, ambient AI needs to
• 9:13
  happen, and we already know who the top
• 9:15
  players in the space are. And so we've
• 9:16
  made a selection now for for our first
• 9:18
  partner in the ambient AI space without having
• 9:21
  to do six or seven pilots like other
• 9:23
  sites were were doing.
• 9:25
  And, you know, from the the the rest
• 9:28
  of the generative AI perspective,
• 9:30
  we've just partnered with,
• 9:33
  a relatively mature but still evolving
• 9:36
  vendor
• 9:37
  who will be
• 9:38
  conducting general advice summarization
• 9:41
  within the electronic health records. So it's not
• 9:44
  a native solution,
• 9:45
  but we found an opportunity where we could
• 9:47
  partner and collaborate and build
• 9:50
  for our own, you know, our own purposes
• 9:52
  in a relatively,
• 9:53
  I would say, affordable way that both feeds
• 9:57
  what we hope to do with the core
• 9:58
  product as it exists and also gives us
• 10:00
  opportunity
• 10:01
  to develop what we think we need as,
• 10:04
  you know, as MetroHealth individually
• 10:06
  or independently.
• 10:08
  That's great to hear. And really exciting on
• 10:10
  both ends, especially as you mentioned, just building
• 10:13
  out that AI infrastructure and how you're able
• 10:14
  to,
• 10:15
  understand the advisory council, looking at the ways
• 10:18
  that, you know, it will be most meaningful
• 10:21
  to bring AI and generative AI into,
• 10:23
  the broader organization, I think, is something that
• 10:26
  so many hospitals and health systems are still
• 10:28
  trying to work through today and, you know,
• 10:30
  much a position like yourself and looking at,
• 10:32
  you know, what partnerships makes the most sense,
• 10:34
  what products then you know, what applications are
• 10:36
  gonna be most meaningful for them. Given this
• 10:39
  background you you've, shared with us in terms
• 10:41
  of the digital transformation,
• 10:43
  where do you see some of the big
• 10:44
  growth opportunities over the next twelve months or
• 10:46
  so?
• 10:47
  I I think it's leading into exactly those
• 10:49
  spaces. So if we focus on the generative
• 10:52
  AI applications,
• 10:54
  I'll say that there's, you know, there's a
• 10:56
  lot that we know that this technology can
• 10:58
  do. It's probably strongest in terms of information
• 11:01
  retrieval.
• 11:04
  The opportunities beyond that, I think, you know,
• 11:06
  there's a lot of questions that come my
• 11:08
  way as chief health officer, like why can't
• 11:11
  we just put the entire electronic health record
• 11:13
  into, you know, a large language model? Like
• 11:15
  why can't you just copy and paste once
• 11:16
  all trying to chat
• 11:18
  in a protected setting, obviously? And those of
• 11:20
  us that live in the space know it's
• 11:21
  a lot more complicated. So what information we
• 11:23
  submit to a large language model, the context,
• 11:26
  the prompting.
• 11:27
  Right. That's that's the hard part. That's the
• 11:29
  part that really requires a lot of diligence
• 11:31
  and oversight.
• 11:33
  And so,
• 11:34
  you know, we've gotten to a point now
• 11:36
  where we, I would say, as a community,
• 11:38
  have started to understand where this technology does
• 11:41
  well, where it does not. I think the
• 11:43
  obvious parts are, like I said,
• 11:45
  information retrieval. So what happened to this patient?
• 11:47
  You know, in this last hospital encounter, for
• 11:50
  instance, or write me a discharge summary. I
• 11:51
  think those are all fairly achievable things.
• 11:55
  What we'd like to see more of and
• 11:57
  I think the literature starting sport,
• 11:59
  it's more of that kind of physician's copilot
• 12:02
  where if I can identify,
• 12:04
  I'll use the pulmonary and critical care example
• 12:07
  here, but I can identify when somebody looks
• 12:09
  like they could have an infection so that
• 12:11
  I can determine that they may or may
• 12:12
  not be septic.
• 12:14
  That's information that can come from rich notes,
• 12:16
  that can come from imaging, for instance, that
• 12:18
  isn't immediately available in the electronic health record
• 12:21
  today. So I think that there's opportunities to
• 12:24
  develop phenotypes from the data
• 12:26
  that can serve both our operational and clinical
• 12:29
  need and then beyond that, actually, our quality
• 12:33
  focus. And so being able to capture or,
• 12:36
  you know, identify
• 12:38
  potentially
• 12:39
  reversible or preventable
• 12:41
  conditions early in the state.
• 12:43
  I think that's gonna be really crucial. And
• 12:45
  another example, for instance, is someone might have
• 12:48
  a Foley catheter or a line that does
• 12:50
  not need to be there. It's a lot
• 12:51
  more difficult to identify that
• 12:54
  in real time when you're not trying to
• 12:57
  analyze
• 12:58
  context through documentation or a note.
• 13:02
  So I'm excited about that prospect, and that's
• 13:04
  the kind of development that we're going to
• 13:06
  be working on. In addition to those things,
• 13:08
  we always have to focus on
• 13:11
  what can potentially improve our bottom line. So
• 13:13
  if we're paying for this expensive technology, then
• 13:16
  naturally after that, start asking, can we can
• 13:18
  we use the technology for appropriate billing? Right.
• 13:20
  So we don't miss diagnosis
• 13:22
  to make sure that we're submitting the right
• 13:24
  information for for CDI, for instance.
• 13:27
  We're also using the technology for, prior authorizations,
• 13:31
  for example. So that does help us
• 13:34
  solve a unique health care problem faster.
• 13:37
  Although I'm usually
• 13:39
  really nervous about using technology
• 13:41
  to solve broken processes.
• 13:44
  Obviously, as a single system, we're not going
• 13:46
  to be able to solve a lot of
• 13:47
  these problems. But if they can help us
• 13:49
  reduce the effort
• 13:50
  and get the patients the outcomes that we
• 13:52
  need, I think that's that's fair game at
• 13:54
  this point.
• 13:56
  Absolutely. That sounds great.
• 13:58
  And really truly will be cool when you
• 14:00
  have that system in place to use the
• 14:02
  technology, use the AI,
• 14:05
  and get the most out of,
• 14:06
  the team members that you have in in
• 14:08
  in really truly elevating,
• 14:10
  throughout the space. So that's really cool to
• 14:12
  hear about, and thank you for sharing with
• 14:14
  us. I I know there's a lot of
• 14:16
  opportunities, especially with technology and AI, but I
• 14:18
  can imagine some big challenges as well. What
• 14:20
  are you anticipating for the next year or
• 14:22
  so? Anything that could be potential roadblocks for
• 14:24
  you?
• 14:25
  Yeah. I'll I'll do the the two hat
• 14:27
  thing now. So,
• 14:28
  when I when I think about my role
• 14:30
  in in virtual care and CMI of Invesha,
• 14:33
  I think
• 14:35
  I'll just say that the the telehealth extension,
• 14:37
  while reassuring
• 14:38
  and, you know,
• 14:41
  definitely a positive move,
• 14:43
  only gives us so much runway to build
• 14:46
  our our plat to continue to build our
• 14:48
  platform and solution
• 14:50
  in a comfortable way.
• 14:52
  So I think most of us have now
• 14:54
  reached this conclusion that virtual care is here
• 14:57
  to stay. It's a necessary component. It's it
• 15:00
  obviously
• 15:01
  works best at the right patient time and
• 15:03
  place. It's not for all care settings.
• 15:05
  It's really can you right size it?
• 15:07
  Patients expect
• 15:09
  it. Clinicians enjoy it. At least, you know,
• 15:11
  based on the surveys we've done internally,
• 15:14
  it does provide more access. It likely is
• 15:17
  similar or better quality in terms of keeping
• 15:19
  patients engaged to their health care
• 15:21
  journey. And so I would love to see
• 15:24
  more of a realization of that,
• 15:26
  again, at the kind of the policy level.
• 15:28
  Right. To continue to fund that mechanism. Because,
• 15:31
  again, it's I think it's it's here to
• 15:33
  stay. It's not going to go back to
• 15:35
  COVID levels
• 15:36
  anytime soon. But there is a time and
• 15:38
  a place for it. And I think we
• 15:39
  need to acknowledge that
• 15:41
  when I put on my C H I
• 15:43
  O hat.
• 15:44
  The biggest concern that I have
• 15:47
  continues to be spending money in the wrong
• 15:49
  places and on the wrong solutions.
• 15:52
  I still think we're figuring these things out,
• 15:54
  so we are still treading lightly. We are
• 15:56
  trying to, at least when we commit to
• 15:59
  partners in the space,
• 16:00
  come in in a way where we have
• 16:02
  leeway. You know, our EHR vendors developing solutions
• 16:05
  and opportunities,
• 16:07
  when it comes to generative AI specifically
• 16:10
  that start to conflict.
• 16:12
  I'm really concerned about
• 16:14
  AI scope creep. So when we pick a
• 16:16
  solution for a particular pain point or if
• 16:19
  it's even part of a larger platform,
• 16:21
  it starts to grow in terms of scope.
• 16:23
  So, you know, first, it's summarizing notes. Next,
• 16:26
  it's listening to your conversation. Well, hold on.
• 16:28
  I already have an AMP and AI solution.
• 16:30
  These things are starting to bump into each
• 16:32
  other. So I am worried about the overall
• 16:34
  spend.
• 16:35
  From the clinician perspective,
• 16:38
  I think there's a lot of assumption
• 16:40
  about the fact that all of this is
• 16:42
  going to make clinicians' lives easier. But it's
• 16:44
  still out. You know, the question is still
• 16:47
  out there as to whether or not this
• 16:50
  really decreases
• 16:51
  a provider's cognitive burden or if it just
• 16:53
  kind of changes the level of work that
• 16:56
  they have to do. Right? So
• 16:57
  writing a note versus
• 16:59
  reviewing a note that's been written by a
• 17:01
  generative AI solution,
• 17:03
  looking for hallucinations,
• 17:04
  being always worried about something that crept in
• 17:07
  there that's not supposed to be in there
• 17:09
  or even worse, potentially something that was that
• 17:11
  was missed. Right? So none of these solutions
• 17:13
  are perfect. And so we have to be
• 17:15
  really careful about
• 17:16
  how much more work we're throwing on our
• 17:18
  care team and making assumptions
• 17:20
  about whether or not we're truly, you know,
• 17:23
  truly enhancing
• 17:25
  their workload.
• 17:26
  So, yeah, I think that those are those
• 17:28
  are the things that I don't know that
• 17:30
  they keep me up at night, but they're
• 17:31
  definitely concerns that I have as we step
• 17:34
  more, you know, into the space.
• 17:36
  Absolutely. I think, definitely, it makes a lot
• 17:39
  of sense and, you know, especially in bringing
• 17:41
  in AI,
• 17:42
  you know,
• 17:43
  not really a blueprint of how that should
• 17:45
  all work. And so I think,
• 17:47
  it seems like you've got a a great
• 17:48
  approach and strategy to just keep an eye
• 17:50
  on that and understanding the the risks as
• 17:52
  well. Before we wrap up here with our
• 17:54
  last minute, what is the number one thing
• 17:56
  you're doing right now to set your organization
• 17:58
  up for long term
• 18:00
  success? I think we've we've echoed,
• 18:02
  a lot of these points here, but it's
• 18:04
  it's it's worth rehashing.
• 18:05
  So making sure we have a solid
• 18:08
  AI governance
• 18:09
  strategy and plan, making sure we have the
• 18:11
  right people, a diverse set of experiences,
• 18:14
  at the table,
• 18:15
  and making sure that we upscale our clinicians
• 18:18
  and what these technologies can do and what
• 18:21
  they can't do. And also empowering them. Right.
• 18:23
  To speak up when it's not working the
• 18:25
  way that it's supposed to be working.
• 18:28
  So continuing down that road,
• 18:30
  being really careful about
• 18:32
  the the life cycle of all of these
• 18:34
  solutions, meaning
• 18:36
  being ready from a financial
• 18:39
  responsibility perspective
• 18:41
  and from even just an overall governance perspective
• 18:43
  to say this tool or this solution
• 18:46
  is not working. We need to decommission it.
• 18:49
  I think there's this natural assumption of more
• 18:50
  technology is always good. So we have to
• 18:53
  keep balancing that out. Right. To make sure
• 18:54
  that we're actually literally getting what we pay
• 18:57
  for and making sure that we're not making
• 18:59
  anything harder than it has to be.
• 19:02
  So those are the things that I focused
• 19:03
  on, at least from the AI perspective.
• 19:07
  From the
• 19:08
  virtual care
• 19:10
  perspective,
• 19:11
  I think we're just keeping our heads down
• 19:13
  and
• 19:14
  trying to build a solid solution
• 19:17
  or I should say building a solid solution
• 19:19
  and a product
• 19:21
  that really scales well. And the most important
• 19:23
  thing I think that that I've learned in
• 19:25
  that space is you can build the most
• 19:27
  exquisite piece of technology and platform that is,
• 19:30
  to a high degree,
• 19:32
  seamlessly integrated into others, EHRs or solutions.
• 19:35
  But really, what's driving
• 19:37
  our
• 19:38
  energy
• 19:39
  and driving our positive results, it's the people
• 19:42
  that we've hired.
• 19:43
  And so we've
• 19:44
  continued to, you know, make sure that people
• 19:48
  understand people that we bring out, understand our
• 19:50
  culture, that they're good fits for our culture.
• 19:52
  And I think a lot of the positive
• 19:54
  responses we've gotten from from patients of our
• 19:57
  services really come from the stellar care that
• 19:59
  they've that they've had independent of how easy
• 20:02
  and, you know, seamless the integration of the
• 20:04
  virtual care experience truly was.
• 20:08
  That's amazing to hear. Doctor Terabishi, thank you
• 20:10
  so much for joining us on the podcast
• 20:11
  today. This has been such a fun and
• 20:13
  informative conversation, and I look forward to connecting
• 20:15
  with you again soon. Yeah. Thank you for
• 20:17
  having me.