Latest / Becker’s Healthcare Podcast / Trust Is Not a Feature: The New Rules of Clinical AI Governance and Validation

Transcript

Trust Is Not a Feature: The New Rules of Clinical AI Governance and Validation · Becker’s Healthcare Podcast

• 0:00
  - Well, hello, everyone. This is Erica Spicer Mason
• 0:02
  - with Becker's Healthcare.
• 0:04
  - Thank you so much for tuning in to
• 0:05
  - the Becker's Healthcare podcast series.
• 0:08
  - Today, we're going to talk about clinical
• 0:10
  - AI governance.
• 0:12
  - And joining me for this conversation today, we
• 0:14
  - have doctor Sheila Bond, the head of clinical
• 0:17
  - AI at Wolters Kluwer Health.
• 0:19
  - Doctor Bond, welcome. Thank you so much for
• 0:21
  - being here today.
• 0:22
  - Thank you. Pleasure to be here. Well, doctor
• 0:25
  - Bond, it's so great to have you with
• 0:27
  - us today. And I thought to get us
• 0:28
  - started, it would be great if you could
• 0:30
  - just share a a little bit more about
• 0:32
  - yourself, your path in health care, and how
• 0:34
  - does your work as head of clinical AI
• 0:37
  - at Wolters Kluwer
• 0:38
  - look day to day? What does it involve?
• 0:41
  - Sure.
• 0:42
  - It's been I mean, one, it's just such
• 0:44
  - an exciting era
• 0:46
  - for medicine.
• 0:47
  - I've been in medicine for
• 0:49
  - twenty two years now, so taking care of
• 0:52
  - patients either at Mass General or Brigham
• 0:54
  - for that time period.
• 0:56
  - And
• 0:57
  - for the past,
• 0:59
  - fourteen years or so, alongside of that, I've
• 1:01
  - been in clinical decision support, usually at the
• 1:04
  - intersection of content and technology.
• 1:07
  - So I have really thought deeply
• 1:09
  - about how the words that we create in
• 1:12
  - medicine, whether that's through an evidence based medicine
• 1:14
  - reference
• 1:15
  - or tool,
• 1:17
  - or how we teach students on the words,
• 1:20
  - making sure that the information that we give
• 1:22
  - to the next generation is the right information
• 1:25
  - for patient care and always encouraging them to
• 1:28
  - think critically.
• 1:29
  - So this has been just a really exciting
• 1:31
  - era for me. I think it's just a
• 1:33
  - shift in how we're gonna create medical information
• 1:35
  - for generations to come. So
• 1:38
  - I feel really honored to, you know, simply
• 1:41
  - shepherd that in. And the teams that I
• 1:43
  - work with are
• 1:45
  - amazing in terms of the dedication and passion
• 1:48
  - that they have for this mission.
• 1:50
  - As the head of clinical AI, my role
• 1:53
  - is really to
• 1:55
  - help them create the the vision of what's
• 1:57
  - next for how we deliver information,
• 2:00
  - whether that's through technology or through the words
• 2:02
  - that we write on the page.
• 2:04
  - It's also getting clinicians and engineers
• 2:07
  - together
• 2:07
  - to create that vision that takes both of
• 2:09
  - their combined
• 2:11
  - expertise
• 2:11
  - to figure out how to use generative AI
• 2:14
  - to remain trustworthy
• 2:16
  - and, to create not only the system
• 2:19
  - that creates information and presents it in a
• 2:22
  - new way, but also how we validate it,
• 2:24
  - how we govern it, and how we hold
• 2:26
  - ourselves accountable.
• 2:27
  - And it's, you know, exciting from start to
• 2:30
  - finish, and it probably taps into every element,
• 2:33
  - of every skill that I've accumulated over the
• 2:36
  - years.
• 2:37
  - Yeah. Doctor Bond, what I appreciate the overview,
• 2:40
  - by the way, and what an exciting role
• 2:42
  - you've stepped into as you mentioned. You know,
• 2:44
  - the fact that it's combining so much of
• 2:46
  - what you've already done in health care and
• 2:48
  - is navigating a lot of uncertainty ahead. It
• 2:50
  - sounds like a very exciting place to be.
• 2:53
  - And I wanna go a little bit deeper
• 2:54
  - on something that you mentioned earlier in your
• 2:56
  - response, and that was this
• 2:58
  - real dedication that I know your teams at
• 3:00
  - Walter's Cloura Health have when it comes to
• 3:03
  - clinical AI and making
• 3:05
  - sure that the content that is surfaced at
• 3:07
  - the point of care is trusted.
• 3:10
  - Trust comes up constantly in conversations when we're
• 3:13
  - talking about clinical AI, but it can definitely
• 3:15
  - mean different things to different people, different stakeholders
• 3:19
  - in the care journey.
• 3:20
  - So from your point of view, when a
• 3:22
  - clinician or a health system says that they
• 3:24
  - trust a tool, what do you think that's
• 3:26
  - really resting on?
• 3:28
  - That's that's a great question. I think the
• 3:29
  - era has really come back around to trust
• 3:32
  - and what it means specifically. So that's one
• 3:34
  - of the questions that I hear
• 3:36
  - a lot these
• 3:37
  - days. It's not a feeling. Right? And I
• 3:39
  - think the way we have defined it is
• 3:41
  - one,
• 3:42
  - there's governance.
• 3:43
  - There are people who oversee the system, how
• 3:46
  - it's designed,
• 3:47
  - and know how it's designed so they can,
• 3:50
  - a, create what needs to be created, and
• 3:53
  - b, measure outcomes
• 3:54
  - and measure how it functions and display that
• 3:57
  - to others. So the fundamental tenants of trust
• 4:00
  - for us are, one, having a governance system.
• 4:03
  - Two, at this point in time, you know,
• 4:05
  - generative AI is still early,
• 4:08
  - in its history.
• 4:09
  - So we make sure that, it is grounded
• 4:12
  - in content that is worthy of our trust,
• 4:15
  - already. And I think that's one thing that
• 4:17
  - people forget is
• 4:18
  - that there has to be a source of
• 4:20
  - truth. Right? Generative AI is very good at
• 4:22
  - assembling words in the right order and having
• 4:24
  - them sound great, but making sure that they
• 4:27
  - are actually tied to a source
• 4:29
  - that is trustworthy itself is important.
• 4:32
  - And then the validation and the data that
• 4:34
  - we produce
• 4:35
  - is also, I think, really core to trust.
• 4:37
  - Right? Does the system perform as expected? How
• 4:40
  - do we know? How do we measure that?
• 4:41
  - And are we willing to share it with
• 4:43
  - you?
• 4:44
  - And then ultimately,
• 4:45
  - you know, who is
• 4:47
  - accountable for this? When something goes wrong, who
• 4:50
  - do you call? How do you discuss it?
• 4:51
  - How do people act, in that circumstance?
• 4:54
  - You know, it is a new technology, so
• 4:56
  - it won't be perfect, but we do have
• 4:58
  - ways and means of introducing things into medicine
• 5:01
  - when they're new. And I think those are
• 5:03
  - the central tenets. Right? That there is accountability,
• 5:05
  - governance,
• 5:07
  - measurements, and the ability to act and improve.
• 5:10
  - And I think the other thing that is
• 5:12
  - really important
• 5:14
  - is that, you know, technology took hold of
• 5:16
  - this early on, and a lot of the
• 5:18
  - measurements that you see out there are really
• 5:20
  - coming from the technical world.
• 5:22
  - But for us in health care, right, the
• 5:24
  - end user
• 5:25
  - is in health care. And usually,
• 5:28
  - well, the information that is acted upon is
• 5:30
  - for a patient. So I think we all
• 5:32
  - collectively as a community who is bringing this
• 5:34
  - together needs to remember that that's the end
• 5:37
  - game in this, and the measurements are not
• 5:39
  - just technical measurements of progress
• 5:41
  - for an AI system, but the measurements that
• 5:44
  - we wanna put forth, really demonstrate that this
• 5:48
  - is enough to act upon and the right
• 5:50
  - guardrails and safety mechanisms are in place for
• 5:53
  - use for with the patient
• 5:55
  - as the endpoint in mind.
• 5:58
  - It's a really helpful framework, doctor Bond, and
• 6:00
  - I appreciate you elaborating on those key pillars,
• 6:04
  - that kind of are building this this concept
• 6:07
  - of trust,
• 6:08
  - and not just a concept, but something that
• 6:10
  - is measurable, actionable.
• 6:13
  - And I wanna go a little bit deeper
• 6:14
  - too on governance. You had mentioned governance a
• 6:17
  - few times now, and this is another term
• 6:19
  - that can be used kind of broadly just
• 6:21
  - like the concept of trust. So
• 6:24
  - inside an organization,
• 6:25
  - what does AI governance look like when it's
• 6:27
  - working well?
• 6:29
  - I'd love to know,
• 6:30
  - you know, from your where you stand, who
• 6:32
  - tends to own it, and how do decisions
• 6:34
  - actually get made in a way that's effective?
• 6:37
  - That's a wonderful question as well.
• 6:40
  - And it really forces you and I think
• 6:42
  - a healthy way to break down what quality
• 6:44
  - is so that other people can understand it,
• 6:47
  - see it, and measure it. So it's a
• 6:48
  - good forcing function.
• 6:50
  - Our governance is really cross
• 6:52
  - disciplinary. I think, our clinical and our clinical
• 6:55
  - AI team set the standard for what good
• 6:58
  - is.
• 6:59
  - Everybody, right, needs to understand that so that
• 7:01
  - we can measure it. That includes our product
• 7:03
  - teams and our technology teams, And all of
• 7:06
  - us, right, are responsible for understanding the industry
• 7:09
  - out there. So technology does need to understand
• 7:11
  - a bit of health care and health care
• 7:12
  - does need to understand
• 7:14
  - a bit of technology.
• 7:16
  - So I think for the most part, you
• 7:17
  - know, in terms of the organization working well,
• 7:20
  - it's all of us collectively
• 7:22
  - understanding the same thing, having the same vision,
• 7:25
  - and predefining what good is and how we're
• 7:27
  - going to measure that, building the system, and
• 7:30
  - then building the tools to make those measurements,
• 7:33
  - looking at them all together, right, and understanding,
• 7:36
  - whether we've met our goals or not. And
• 7:39
  - if we have or if we haven't,
• 7:41
  - is there anything unexpected that we didn't foresee
• 7:44
  - that has come up? And then I think
• 7:46
  - one of the most challenging things is sort
• 7:48
  - of shifting modes. It's shifting from sort of
• 7:51
  - creating
• 7:51
  - into measuring.
• 7:53
  - And then before release, you know, technology is
• 7:55
  - known for fast release cycles. But once you
• 7:58
  - get into the elements
• 8:00
  - of
• 8:01
  - understanding and managing risk, it's a different framework.
• 8:04
  - Meaning, you know, you you work until that
• 8:07
  - risk is gone or otherwise mitigated or settled
• 8:10
  - in some other way.
• 8:11
  - And I think that's,
• 8:13
  - sort of a mark of the era is
• 8:15
  - a, you know, we all have to be
• 8:16
  - able to share a vision, communicate to one
• 8:19
  - another because there's no one
• 8:21
  - domain, so to speak, that's gonna be able
• 8:23
  - to do this themselves. It's gonna take an
• 8:25
  - alliance between health care and technology.
• 8:29
  - So it is that ability to, you know,
• 8:31
  - break something down so that,
• 8:33
  - folks you don't usually work with can understand
• 8:35
  - it, to measure it, to measure it to
• 8:37
  - a predefinition,
• 8:38
  - and then,
• 8:40
  - to look at it and say, have we
• 8:41
  - done enough? And if we haven't, you no
• 8:43
  - longer work to a time. Right? You work
• 8:45
  - until that risk is gone. Mhmm.
• 8:48
  - Yeah. Yeah. Doctor Bond, I appreciate you, defining
• 8:51
  - governance.
• 8:52
  - You know, we we are fresh off of
• 8:54
  - our IT and revenue cycle conference at Becker's,
• 8:56
  - and I have to say that the way
• 8:58
  - that you just described it aligns so closely
• 9:01
  - with what I know a lot of organizations
• 9:03
  - are are striving for.
• 9:05
  - But to your point, it is challenging to
• 9:07
  - put into action, and also there's no one
• 9:10
  - way to do governance. So it seems to
• 9:13
  - certainly be a challenge for most organizations.
• 9:16
  - And I also wanna follow-up with
• 9:18
  - on this idea of risk here
• 9:21
  - and talk a little bit more about validation.
• 9:23
  - I know validation tends to get framed as
• 9:25
  - something that happens before a new tool or
• 9:28
  - technology goes live.
• 9:30
  - But how do you think about it once
• 9:32
  - a tool is already in clinical use? What's
• 9:35
  - hardest to measure at that stage?
• 9:38
  - Great question as well. I think that speaks
• 9:40
  - to sort of the accountability
• 9:42
  - pillar
• 9:43
  - that I talked about earlier.
• 9:46
  - It's not only, right, what you do before
• 9:48
  - you design your system and how you measure
• 9:50
  - it when it's at the gate,
• 9:52
  - but, you know, that post monitoring surveillance
• 9:55
  - is very important.
• 9:57
  - So looking at, you know, whatever your system
• 10:00
  - is designed to do, right, looking at how
• 10:02
  - it fares in the real world, at this
• 10:04
  - point in time, I think for most clinical
• 10:06
  - AI systems, it's understanding the types of questions
• 10:08
  - that will come in, what your system's output
• 10:11
  - is.
• 10:12
  - And we're always really sure that we've got
• 10:14
  - expert eyes on the sampling of what is
• 10:16
  - out there in the industry
• 10:18
  - just because of how, precious
• 10:21
  - it is. The hardest thing, of course, is
• 10:24
  - that, you know, this is a technology that
• 10:26
  - isn't going to be perfect for a while.
• 10:29
  - So measuring how,
• 10:31
  - you do in the real world isn't an
• 10:33
  - easy matter. And even so with some of
• 10:35
  - our traditional medical references, right, you can write
• 10:38
  - something to perfection in that state,
• 10:41
  - but still someone might not not understand it
• 10:43
  - well enough. Right? So the hardest part is,
• 10:45
  - you know, the gap between,
• 10:47
  - you know, doing everything you possibly can to
• 10:49
  - ensure that, you're delivering
• 10:52
  - what should be worthy of trust.
• 10:55
  - But then, you know, the actions that someone
• 10:57
  - takes, or how someone interprets that information
• 11:00
  - is,
• 11:01
  - you know, always the thing one of the
• 11:03
  - things that, will keep you up at night.
• 11:05
  - That's sad. Right? I think, you know, when
• 11:07
  - you're talking about governance,
• 11:09
  - it probably, you know, there's the industry level
• 11:12
  - governance and then there is also,
• 11:14
  - you know, what we are responsible for as
• 11:16
  - a community of medicine and as an industry,
• 11:19
  - which I think is to educate our users
• 11:21
  - on what AI literacy is, and how to
• 11:24
  - use these tools
• 11:26
  - both individually and then, you know, that's now
• 11:28
  - part of medical school education as well, coming
• 11:32
  - more and more
• 11:33
  - AI literate. So I think those things go
• 11:35
  - hand in hand with the governance of the
• 11:37
  - system.
• 11:38
  - And as always, it's just really heartening to
• 11:40
  - be part of the community of medicine and
• 11:43
  - see that they do respond to that call.
• 11:45
  - Right? I've been all over the world talking
• 11:47
  - about clinical AI.
• 11:48
  - And whether it's, you know, in Europe or
• 11:51
  - Asia or
• 11:53
  - The United States,
• 11:54
  - I think everyone is starting to speak the
• 11:57
  - same language and taking on, the responsibility that
• 12:00
  - the community of medicine does. So it's been
• 12:01
  - really heartening to see folks step up in
• 12:04
  - the way that they usually do and heartening
• 12:05
  - to see that, you know, across the world,
• 12:08
  - there's the layer of of medicine and how
• 12:10
  - much people care no matter what region they're
• 12:12
  - from that we get this right. So
• 12:15
  - I appreciate the global perspective there, doctor Bond.
• 12:18
  - And so interesting too to hear how medical
• 12:21
  - curricula are already changing to support AI in
• 12:25
  - clinical use.
• 12:26
  - And kind of in that vein, I'm curious
• 12:28
  - whether we're talking about clinical AI use or
• 12:31
  - governance or even validation.
• 12:34
  - What questions are you hearing from clinical and
• 12:37
  - IT leaders now that maybe you weren't hearing
• 12:39
  - even just one or two years ago?
• 12:42
  - I think the questions are more pointed and
• 12:45
  - precise now. I think we're sort of beyond
• 12:47
  - the wow factor,
• 12:49
  - and, we're beyond the usability factor and sort
• 12:52
  - of the fun factor and the novelty factor
• 12:54
  - of this, and we're into
• 12:56
  - how are we actually gonna integrate it, and
• 12:58
  - what does it take to get it right,
• 13:00
  - and what does it take to scale.
• 13:03
  - So, you know, as you talked about at
• 13:05
  - the beginning of the interview and trust, people
• 13:07
  - are now asking
• 13:08
  - not just for the who and the how
• 13:10
  - or the system design, but really the specific
• 13:12
  - data and starting to drill down into what
• 13:15
  - really matters.
• 13:16
  - So I've been one, you know, we've been
• 13:18
  - working on that foundationally for so long. I'm
• 13:21
  - really thrilled to hear those questions coming back
• 13:24
  - from the industry itself and from, you know,
• 13:27
  - academic medical centers
• 13:29
  - as to, you know, what really the foundation
• 13:31
  - should like should look like, and what should
• 13:34
  - you measure to make sure that,
• 13:36
  - you know, whatever we build now is gonna
• 13:38
  - be the foundation for the future.
• 13:40
  - So I'm really happy to hear that people
• 13:43
  - are educating
• 13:44
  - themselves,
• 13:45
  - looking at the data, reviewing the data. Right?
• 13:48
  - There are, you know, chats on LinkedIn
• 13:50
  - about,
• 13:51
  - benchmarks in medicine and how best we should
• 13:54
  - measure things.
• 13:55
  - So I'm just very glad to see that
• 13:58
  - those questions are coming up and people are
• 14:00
  - asking for,
• 14:02
  - what precisely
• 14:03
  - one should measure. And in that regard, you
• 14:05
  - know, I think
• 14:07
  - for us, one of the most important things
• 14:10
  - is to make sure and,
• 14:12
  - this is really interesting because,
• 14:15
  - as I said, I was just in Asia
• 14:16
  - and Singapore and in China, and, you know,
• 14:19
  - we were all coming to the same point
• 14:21
  - about, you know, what of all of this
• 14:22
  - should be measured
• 14:24
  - as we're starting to grow and to scale.
• 14:27
  - And making sure
• 14:29
  - that,
• 14:30
  - when this technology is creating words that will
• 14:33
  - impact patient care,
• 14:34
  - that there is,
• 14:36
  - no unsupported leaps. Right? And how do we
• 14:38
  - start to quantify
• 14:40
  - inference, right, is sort of this point of
• 14:43
  - intersection that we were at across, you know,
• 14:45
  - three different countries. So it was very interesting.
• 14:47
  - It's not like a benchmark or a safety
• 14:49
  - score or
• 14:50
  - something gimmicky.
• 14:51
  - But the intersection that we were at was,
• 14:53
  - you know, for years, we have,
• 14:56
  - and decades really, in the field of evidence
• 14:58
  - based medicine made sure that, we have a
• 15:01
  - system for
• 15:03
  - analyzing and appraising the medical literature to make
• 15:06
  - sure it's the right information to act on
• 15:08
  - for patient care.
• 15:09
  - And we have that strong foundation. So we
• 15:11
  - now don't want to
• 15:13
  - create a new foundation based on ease of
• 15:15
  - use
• 15:16
  - and or just information that is plausible or
• 15:19
  - unsupported, and that technology can so easily do
• 15:22
  - that.
• 15:23
  - What we came around to, I think, in
• 15:24
  - discussion was
• 15:26
  - the ability to quantify
• 15:28
  - and measure whether your system is taking a
• 15:30
  - leap and making a recommendation for patient care
• 15:33
  - or are you being disciplined? And, you know,
• 15:35
  - even though it might not be as
• 15:37
  - personally pleasing, right, to to get a more,
• 15:41
  - challenging answer to your question, that may be
• 15:43
  - the right thing to do, and that's what
• 15:45
  - we wanna measure.
• 15:47
  - Yeah. Well, it it's been great hearing doctor
• 15:49
  - Bond about this increased
• 15:51
  - curiosity and more targeted questions that you're hearing
• 15:54
  - from clinicians and leaders alike over the over
• 15:56
  - just the last two years, really.
• 15:59
  - But as you look ahead,
• 16:00
  - you know, when you consider where clinical AI
• 16:03
  - may be heading and even just three to
• 16:05
  - five years from now, is there anything on
• 16:07
  - your mind that we haven't yet touched on
• 16:09
  - that you wanted to highlight or any final
• 16:11
  - thoughts you'd like to leave our listeners with?
• 16:14
  - Sure. Probably two things. I think first, I
• 16:17
  - can't underscore enough. I said it probably throughout
• 16:20
  - this interview, but
• 16:21
  - right now, we're foundation building again. Right? And,
• 16:25
  - making sure that we get this foundation right
• 16:28
  - is just mission critical for us because I
• 16:30
  - think it really will be the foundation
• 16:33
  - for the health care information that we use
• 16:36
  - in EHRs, in ambient systems,
• 16:38
  - and for many, many, many things to grow.
• 16:41
  - So don't rush into it. You know, make
• 16:43
  - sure that foundation is worthy
• 16:45
  - of the decisions that it's going to support.
• 16:48
  - That's number one. And number two is
• 16:52
  - the whole reason to do that is that
• 16:54
  - this opportunity for health care is absolutely enormous.
• 16:57
  - Right? When you think
• 16:58
  - about the power of generative AI, it's not
• 17:00
  - really to summarize and retrieve things. It really
• 17:03
  - is to transform things so that people can
• 17:06
  - understand them better.
• 17:07
  - And throughout my career in medicine and twenty
• 17:10
  - two years, that really is the most important
• 17:13
  - thing is to make sure that your patients
• 17:15
  - understand. Right? Every patient journey is unique.
• 17:19
  - And I remember starting in medicine, it was
• 17:21
  - entirely different field.
• 17:24
  - You used to take care of your patient
• 17:26
  - in the outpatient setting, in the inpatient setting.
• 17:29
  - Sometimes you would meet them in the emergency
• 17:30
  - department
• 17:31
  - and walk them up to the room. And
• 17:33
  - you knew who they were holistically as a
• 17:35
  - person, so you knew how to connect with
• 17:37
  - them or how to
• 17:38
  - communicate with them. And, you know, in those
• 17:41
  - little moments, you might tell them what would
• 17:43
  - happen next. And that was such a huge
• 17:45
  - part of taking care of a person. And
• 17:47
  - now as medicine has changed and grown, right,
• 17:50
  - someone else does those things. Someone else takes
• 17:52
  - a patient to the ED. Someone else walks
• 17:54
  - them from point a to point b.
• 17:57
  - And that ability to explain is lost
• 18:00
  - in some ways, or that moment to moment
• 18:02
  - care where you're helping someone understand what is
• 18:05
  - next is lost.
• 18:06
  - And I think the really enormous opportunity that
• 18:09
  - we have
• 18:10
  - for health care workflow
• 18:12
  - is to bring that back in. Right? So
• 18:15
  - folks these days talk about workflow like it's
• 18:17
  - an efficiency thing or like workflow as a
• 18:19
  - destination and everything in it is fixed.
• 18:22
  - But I think the worst thing we could
• 18:23
  - do for health care in many ways is
• 18:25
  - to stuff more information into that old workflow.
• 18:28
  - Right? That workflow was designed for the past.
• 18:31
  - And if I'm to sort of reenvision a
• 18:33
  - new new workflow, right, at the center of
• 18:35
  - it is probably using this technology so that
• 18:38
  - other people can understand,
• 18:40
  - what's happening to them again and for its
• 18:43
  - transformative potential.
• 18:45
  - So, you know, the reason we invest so
• 18:47
  - much in our foundation building
• 18:49
  - is so that we can realize that. Right?
• 18:51
  - So that we can take what is trusted
• 18:53
  - information, so that we can take
• 18:55
  - information that we know doesn't have any unsupported
• 18:57
  - or crazy leaps in it and then change
• 18:59
  - it. Right? So that someone else can understand
• 19:01
  - in a way that is helpful to for
• 19:03
  - them.
• 19:04
  - And to me, that's sort of the ultimate.
• 19:07
  - You know, when you think about medical
• 19:09
  - information, it's sort of locked in writing, and
• 19:11
  - that's a really elite form of communication
• 19:14
  - that takes a lot of years, right, to
• 19:16
  - understand. And for medical information,
• 19:18
  - even more so. Right? So if we can
• 19:20
  - use our brains and our expertise and everything
• 19:22
  - we learned, instead of unlock that, I think
• 19:24
  - the future is
• 19:26
  - is bright.
• 19:28
  - Well, doctor Bond, it's been such a pleasure
• 19:31
  - talking to you. Your passion for your work
• 19:33
  - comes comes through so clearly just as we're
• 19:36
  - speaking to one another. So I wanna thank
• 19:38
  - you for for bringing that to our conversation
• 19:41
  - and also
• 19:42
  - helping our listeners
• 19:43
  - see what is possible when that foundation
• 19:46
  - is
• 19:47
  - gotten right from the beginning. So,
• 19:49
  - again, appreciate your time today and sharing your
• 19:51
  - thought leadership with us. Very much appreciated.
• 19:54
  - Thank you. Always a pleasure to be here.
• 19:56
  - Wonderful questions. And as I said, I could
• 19:57
  - talk about this forever. So thank you.
• 19:59
  - Maybe we'll do a part two.
• 20:03
  - Well, thank you again, doctor Bond. And, of
• 20:04
  - course, we also wanna thank our podcast sponsor
• 20:07
  - today, Wolters Kluwer.
• 20:09
  - Listeners, be sure to tune in to more
• 20:10
  - podcasts from Becker's Healthcare by visiting our podcast
• 20:13
  - page at beckershospitalreview.com.