Latest / Becker’s Healthcare Podcast / Scaling Ambient AI to Reduce Clinician Burnout and Improve Patient Experience with Dr. Albert Villarin

Transcript

Scaling Ambient AI to Reduce Clinician Burnout and Improve Patient Experience with Dr. Albert Villarin · Becker’s Healthcare Podcast

• 0:02
  - Every health system executive right now is being
• 0:04
  - asked to do something that has no useful
• 0:06
  - precedent,
• 0:07
  - make durable technology decisions in an environment that
• 0:10
  - looks meaningfully different every six months.
• 0:13
  - The pace isn't slowing,
• 0:14
  - the vendor landscape isn't getting simpler, and the
• 0:17
  - cost of missteps
• 0:18
  - keeps rising.
• 0:19
  - Becker's eleventh annual IT and Revenue Cycle Conference,
• 0:23
  - The Future of AI and Digital Health, exists
• 0:25
  - for exactly this challenge.
• 0:27
  - Here, you'll find more than 2,500 executive level
• 0:30
  - attendees,
• 0:31
  - 600 plus speakers, and four days of sessions
• 0:34
  - built around the questions leaders are actually wrestling
• 0:36
  - with.
• 0:37
  - How to scale AI responsibly,
• 0:39
  - how to protect increasingly complex systems,
• 0:42
  - how to wring more performance out of the
• 0:44
  - revenue cycle, and how to make tech investments
• 0:46
  - that hold up over time.
• 0:48
  - Join us September fourteenth through the seventeenth at
• 0:51
  - the Hilton Chicago.
• 0:53
  - You can register on our events page at
• 0:55
  - beckershospitalreview.com
• 0:57
  - by clicking on the events tab in the
• 0:59
  - upper right.
• 1:00
  - This is Laura Dirda with the Becker's Healthcare
• 1:03
  - podcast.
• 1:03
  - I'm thrilled today to be joined by doctor
• 1:05
  - Albert Valerin,
• 1:06
  - vice president and chief medical informatics officer at
• 1:09
  - Northwell North. Doctor Valerin, it's a pleasure to
• 1:11
  - have you on the podcast today. Thank you
• 1:13
  - for the opportunity.
• 1:15
  - Absolutely. Now I'm excited for our conversation. I
• 1:17
  - know we're gonna talk a lot about Ambient
• 1:19
  - AI and some of the cool things you're
• 1:20
  - doing at Northwell. But before we dive in,
• 1:22
  - can you introduce yourself and just tell us
• 1:24
  - a little bit more about the health system?
• 1:26
  - Sure. Al Valerin,
• 1:28
  - emergency physician
• 1:29
  - by training for thirty years and, chief medical
• 1:32
  - information officer at multiple health centers
• 1:35
  - for the last
• 1:36
  - twenty. Joined Northwell back in 2010
• 1:39
  - and stayed there through 2019. And then I
• 1:41
  - joined
• 1:42
  - Nuvance Health
• 1:43
  - at which is a collaboration
• 1:45
  - of
• 1:46
  - Western Connecticut
• 1:47
  - and Poughkeepsie Hospital, which they joined forces in
• 1:50
  - 2018.
• 1:51
  - I've been
• 1:53
  - with Nuvance Health for the last seven years,
• 1:55
  - and now we just joined
• 1:57
  - back Northwell again after a merger acquisition.
• 2:00
  - And we're working through the,
• 2:02
  - work the fine tuning of the the revision
• 2:05
  - of that structure.
• 2:07
  - But it's a pleasure to be here, and,
• 2:09
  - part of our our goal here is to
• 2:12
  - talk about the the strategy in terms
• 2:14
  - of what we want to
• 2:16
  - bring from a technological perspective
• 2:18
  - to improve health care overall.
• 2:21
  - I love it. I I think what a
• 2:23
  - a journey you've had and certainly a great
• 2:25
  - background in both the clinical side as well
• 2:27
  - as technology to really come to this moment
• 2:29
  - where so much of technology is now integrating
• 2:31
  - into clinical care and really making a big
• 2:34
  - difference in transforming care delivery.
• 2:36
  - From your perspective,
• 2:37
  - what made scaling Ambient AI documentation across
• 2:41
  - Northwell different from piloting at a single site?
• 2:44
  - Could you just tell us a little bit
• 2:45
  - more about your strategy for Ambient at Northwell
• 2:48
  - and and how you've been able to roll
• 2:49
  - things out over the last few years?
• 2:52
  - Of course. So our,
• 2:53
  - integration was with our own Northwell
• 2:56
  - region. So we have a Northwell north region.
• 2:59
  - There's multiple other regions. They're going through different
• 3:02
  - transitions.
• 3:03
  - We currently have a,
• 3:05
  - unique integrated
• 3:07
  - ambient listening model within our EMR
• 3:09
  - and allowed really,
• 3:11
  - efficient,
• 3:12
  - adoption models there.
• 3:14
  - We had two
• 3:16
  - clear leaders in terms of physician informaticists
• 3:18
  - who led the way in our ambulatory sites.
• 3:22
  - Doctor Ka Shah and doctor,
• 3:24
  - Anthony D'Ambrosio,
• 3:26
  - who were on the ground every day processing
• 3:29
  - and evaluating
• 3:30
  - the the pilot work that we did for
• 3:32
  - several months. And we started our work about,
• 3:35
  - oh, about eighteen months ago in facilitating and
• 3:38
  - understanding the best practice
• 3:40
  - prior to deployment.
• 3:42
  - And once we had
• 3:43
  - a good sense and understanding of how the
• 3:46
  - workflow should be and an efficient workflow for
• 3:50
  - a quick adoption,
• 3:51
  - then we started to roll it out across
• 3:53
  - our inventory sites. And that took about six
• 3:55
  - months or so to complete that process.
• 3:57
  - And then we move into our,
• 3:59
  - subspecialist
• 4:00
  - sites on the inventory side,
• 4:03
  - and that completed this,
• 4:05
  - this past spring.
• 4:07
  - We started then with our emergency departments. And,
• 4:10
  - of course, the workflow there is a little
• 4:11
  - bit different. Ambient listening with the noise and
• 4:14
  - the ED,
• 4:15
  - created a little bit of a, a hurdle.
• 4:18
  - We overcame that by,
• 4:20
  - changing the workflow and utilizing more advanced technology.
• 4:23
  - And now we're,
• 4:25
  - on our way to complete
• 4:27
  - our acute
• 4:28
  - rollout by the end of summer and then
• 4:30
  - work into, eventually the nursing staff,
• 4:33
  - by the end of the year.
• 4:35
  - Oh, wow. That's incredible. And I think, obviously,
• 4:38
  - being able to do so many different settings
• 4:40
  - as you mentioned, it's not quite the same
• 4:42
  - in each different one and then for the
• 4:44
  - different clinical specialists as well.
• 4:46
  - From your perspective in going through that time,
• 4:49
  - you know, what were some of the things
• 4:50
  - that you learned and and,
• 4:52
  - you know, advice you have, I guess, for
• 4:54
  - health system leaders, especially right now who are
• 4:57
  - going through a similar process and trying to
• 4:58
  - figure out how to more broadly scale this
• 5:01
  - across the system and and,
• 5:03
  - leaders that they're using it with on the
• 5:05
  - clinical side?
• 5:07
  - That's an excellent question because, you know, every
• 5:10
  - every activation is different depending on the people,
• 5:13
  - process, and technology,
• 5:15
  - being deployed. From from our perspective,
• 5:18
  - it isn't
• 5:19
  - an IT project. It's a it's an initiative.
• 5:22
  - Initiative led by folks
• 5:25
  - who have a vision.
• 5:26
  - Several of them being,
• 5:28
  - a chief information officer, Jeff Hook,
• 5:31
  - as well as our chief physician executive,
• 5:35
  - Chris Larocque.
• 5:36
  - And both of them led the under helped
• 5:39
  - the understanding across our network to say, this
• 5:41
  - is an important project.
• 5:43
  - We're gonna help you understand
• 5:45
  - the reasons for it from adoption,
• 5:48
  - through facilitation, efficiencies, etcetera,
• 5:50
  - but have it have it integrate with a
• 5:52
  - couple of things.
• 5:54
  - From our chief physician executive,
• 5:57
  - wellness was a major factor in support of
• 6:00
  - the project.
• 6:01
  - So from his perspective,
• 6:03
  - he's gonna help us support adoption
• 6:06
  - of the technology, which allowed us to then
• 6:10
  - turn around enhanced
• 6:12
  - experience for the clinicians.
• 6:14
  - Automated documentation
• 6:16
  - released the burden of health care from the
• 6:19
  - clinicians
• 6:19
  - and allowed them to take care of the
• 6:21
  - patient.
• 6:22
  - From a,
• 6:24
  - ROI perspective
• 6:26
  - and,
• 6:27
  - initiations
• 6:28
  - of future innovations,
• 6:30
  - the chief information officer,
• 6:32
  - Jeff Hook, brought in the the the goals.
• 6:35
  - Work with the board, work with our leaders
• 6:38
  - to help
• 6:39
  - facilitate that adoption across all aspects,
• 6:42
  - operations,
• 6:43
  - finance,
• 6:44
  - and clinical.
• 6:45
  - And then it goes down to our IT
• 6:47
  - department too. And,
• 6:49
  - our our lead,
• 6:51
  - manager over the project, Kimberly Evangelista,
• 6:54
  - She was the,
• 6:56
  - you know, the center of this project and
• 6:59
  - led her team and helped coordinate really the
• 7:02
  - project success over the last two years from
• 7:05
  - the initial idea of we need to do
• 7:07
  - this to where we are today.
• 7:09
  - So it it it took a lot of
• 7:10
  - different people, a lot of different levels
• 7:13
  - to understand
• 7:14
  - what the goal was and help
• 7:17
  - educate
• 7:18
  - and deploy
• 7:19
  - a successful,
• 7:20
  - AI adoption.
• 7:22
  - That makes a lot of sense, and it's
• 7:23
  - helpful to understand exactly the infrastructure you put
• 7:26
  - in place, especially on the people side in
• 7:28
  - order to make this happen. Now I know
• 7:30
  - for many, looking at clinician burnout is a
• 7:33
  - huge challenge and in bringing in this technology
• 7:35
  - is something that I'm gonna support a solution
• 7:38
  - there in in many ways in terms of
• 7:40
  - helping the clinicians,
• 7:42
  - give time back in their day and really
• 7:43
  - feel like they're able to connect with their
• 7:45
  - patients on a more personal level as well.
• 7:48
  - Could you tell us a little bit about
• 7:49
  - how you're thinking about burnout when bringing in
• 7:52
  - Ambient AI documentation to the health system and
• 7:55
  - figuring out validating,
• 7:56
  - that the burnout was real and then, you
• 7:59
  - know, not just early adopter enthusiasm as you
• 8:02
  - went through the process?
• 8:04
  - Great question. I think we started
• 8:06
  - several years prior to the go live of
• 8:09
  - AI and the ambient aspects.
• 8:11
  - Because we we knew,
• 8:13
  - I would say, back in,
• 8:15
  - 2022
• 8:18
  - around the time, you know, the whole COVID
• 8:20
  - and people were just quite disenchanted
• 8:23
  - with, health care, both from a clinical perspective
• 8:27
  - as well as a patient perspective. A lot
• 8:29
  - went on back then after in the COVID
• 8:31
  - times. And
• 8:32
  - what we tried to do is find ways,
• 8:34
  - easy wins
• 8:35
  - to provide
• 8:37
  - an understanding
• 8:38
  - for our clinicians.
• 8:39
  - But first, we went out and asked them,
• 8:42
  - what is the biggest aspect of burden for
• 8:44
  - you? And the EMR, electronic medical record,
• 8:48
  - was that factor. It was difficult to document
• 8:52
  - and were and the workflow was not as
• 8:53
  - efficient as should be. So with the work
• 8:56
  - from a clinical perspective and a technology perspective,
• 9:00
  - the collaboration
• 9:01
  - there between the
• 9:03
  - chief physician executive and the chief information officer
• 9:06
  - working together
• 9:08
  - with myself
• 9:09
  - and the teams, we facilitated a road map
• 9:12
  - on where we can go and find innovative
• 9:15
  - technology to to facilitate this.
• 9:18
  - First, we asked the clinicians,
• 9:20
  - the score is very high, that EMR needed
• 9:23
  - to have some revisions to be done. So
• 9:26
  - that was our baseline.
• 9:27
  - And then we measured every six months. We
• 9:30
  - took a pulse every six months as we
• 9:32
  - put in different parts of the EMR,
• 9:35
  - between
• 9:36
  - workflow enhancements in the message center, evaluating the
• 9:40
  - best efficiencies
• 9:42
  - for patients contacting physicians. I mean, just
• 9:45
  - going everywhere we could to understand
• 9:48
  - and fix
• 9:49
  - the burden and removing,
• 9:52
  - moving to automation
• 9:53
  - for as much as possible.
• 9:55
  - And then you fast forward to two years
• 9:57
  - ago when the whole AI,
• 10:00
  - influence started happening. It's like we said, okay.
• 10:02
  - Fine. Let's go out and do an RFP
• 10:04
  - and look what was out there at the
• 10:06
  - time what was best. And we chose an
• 10:08
  - integrated
• 10:09
  - process within our EMR
• 10:12
  - to facilitate,
• 10:13
  - ambient listing documentation.
• 10:15
  - Went through a whole pilot testing with,
• 10:18
  - about, 25,
• 10:20
  - clinicians,
• 10:21
  - fine tuned it, and then made it, a
• 10:24
  - part of practice.
• 10:25
  - But understanding that
• 10:28
  - all along the way, the communication is the
• 10:30
  - most important part.
• 10:32
  - We reminded people at every meeting, every process,
• 10:35
  - every business interaction,
• 10:36
  - once a month to say, we are bringing
• 10:38
  - this to you, clinicians,
• 10:41
  - supporting the adoption, but also allowing them to
• 10:44
  - give us feedback
• 10:45
  - all the way through. So at the end
• 10:47
  - product
• 10:48
  - was really
• 10:49
  - something that was a mutually created platform that
• 10:53
  - was adopted and
• 10:55
  - enjoyed by by all clinicians throughout our network.
• 10:59
  - And that's amazing to hear. And I think
• 11:01
  - having the clinicians be part of that process,
• 11:03
  - it seems like,
• 11:04
  - from, you know, early stages, it was really
• 11:07
  - helpful and beneficial. And then fine tuning,
• 11:11
  - everything that goes into
• 11:13
  - the workflows and bringing technology in,
• 11:15
  - is is really cool to understand and hear
• 11:18
  - about. How do you really then, go about
• 11:21
  - measuring
• 11:22
  - that reduction in burnout? What do you do
• 11:24
  - in order to make sure that you're getting
• 11:27
  - the right return on the investment beyond clinicians
• 11:29
  - just saying that they're happier or anecdotal evidence?
• 11:32
  - Is there anything that you're doing to,
• 11:35
  - you know, really see what that ROI is?
• 11:38
  - Sure. A year ago, a little more than
• 11:40
  - a year ago, we decided to make our
• 11:44
  - annual
• 11:45
  - KPIs
• 11:47
  - leveraging
• 11:48
  - our goal. Our goal is to reduce the
• 11:50
  - time to document
• 11:52
  - in our EMR. So therefore, we said, okay.
• 11:55
  - That'll be part of our KPI for the
• 11:57
  - year. 20% of our our
• 11:59
  - skill our our goals this year, we based
• 12:02
  - on our ability to bring this to fruition.
• 12:04
  - So we set the bar very high.
• 12:07
  - We want to reduce
• 12:09
  - one to three minutes off the chart. And
• 12:11
  - at the end of the year last year,
• 12:13
  - we're able to show that every clinician that
• 12:16
  - was on AMBA listening was able to reduce,
• 12:20
  - on the average,
• 12:21
  - their charting time
• 12:23
  - to about almost just under three minutes.
• 12:26
  - So that way, we gave back that time
• 12:28
  - across the network in terms of hours per
• 12:31
  - month back to the clinicians so they can
• 12:33
  - utilize it, not to see more patients, but
• 12:36
  - to be more efficient in
• 12:38
  - evaluating the patients that they have.
• 12:40
  - And turn it around, just face the patient
• 12:42
  - so that the ambulances working in the background
• 12:45
  - when the clinicians now are back to delivering
• 12:48
  - the care we all, as clinicians, want to
• 12:52
  - give to our patients. That respect,
• 12:54
  - that integration, that interaction
• 12:56
  - that we remember
• 12:58
  - what drives us from the emotional aspect and
• 13:01
  - why we wanna deliver care for our patients.
• 13:03
  - So it was a win win for the
• 13:05
  - clinicians
• 13:06
  - as well as from our perspective
• 13:08
  - in delivering better innovation,
• 13:11
  - for our network.
• 13:13
  - Got it. That's helpful to understand, you know,
• 13:15
  - and a great kind of metric to think
• 13:16
  - about and way to look at reducing burnout.
• 13:20
  - Now I'm curious, how did ambient listening move
• 13:22
  - the needle on patient satisfaction scores as well?
• 13:26
  - We started looking at,
• 13:28
  - the impact
• 13:29
  - as we turn on
• 13:31
  - each,
• 13:32
  - area
• 13:33
  - of the of the ambient listening. And we
• 13:36
  - started noticing that
• 13:38
  - when you run a report,
• 13:40
  - the higher the adoption was,
• 13:43
  - meaning the more pay more charge you did
• 13:46
  - by AI or ambient listening,
• 13:48
  - the better the scores were for from the
• 13:51
  - press gaining perspective. So we did a correlation
• 13:54
  - of the two and noticed that
• 13:56
  - clinicians that were in the
• 13:59
  - 75 to 85 percentile use
• 14:02
  - of their charting
• 14:04
  - had patient scores that were higher
• 14:06
  - by a large factor. So we did an
• 14:09
  - r square analysis, and it showed across the
• 14:11
  - board,
• 14:12
  - anyone who's the higher you use it, the
• 14:15
  - more positive the responses were from the patients.
• 14:18
  - And that is a ROI that was
• 14:22
  - kind of far about, I hope we could
• 14:23
  - achieve this, and we show that it could
• 14:26
  - be achieved by just giving the patients
• 14:29
  - the respect they deserved
• 14:31
  - and having having a conversation with them rather
• 14:33
  - having a a conversation with your keyboard.
• 14:36
  - That turnaround
• 14:38
  - made all the difference
• 14:40
  - for the patients to say, well, I wanna
• 14:42
  - go see that clinician because they are listening
• 14:45
  - to me now. They're understanding me now. They're
• 14:48
  - having the ability to focus on me.
• 14:51
  - And that's
• 14:53
  - what they have missed for years since we
• 14:55
  - went through the process of going from paper
• 14:57
  - to an EMR. And now we utilize technology
• 15:00
  - to go back to
• 15:02
  - that relationship
• 15:04
  - between patient and clinician. But, yes, our prescreen
• 15:07
  - scores for those areas
• 15:09
  - have improved, and we use that as a
• 15:12
  - leverage point to bring better adoption to clinicians
• 15:15
  - who want to also come on board.
• 15:17
  - As time went through, the positive experience by
• 15:20
  - the clinicians in terms of less,
• 15:24
  - demand,
• 15:25
  - well, increased wellness,
• 15:26
  - and noting the scores have gone up have
• 15:29
  - significantly
• 15:30
  - improved the, and facilitated the adoption by our
• 15:33
  - clinicians across the board.
• 15:35
  - I love that. And I I think it's
• 15:37
  - so, critical that message you had there that,
• 15:39
  - you know, you're able to get back to
• 15:41
  - that relationship with patients. They feel like they've
• 15:43
  - got the attention of their clinicians and have
• 15:46
  - a better experience overall.
• 15:48
  - Now what are you most excited about for
• 15:49
  - the future?
• 15:51
  - Well, I think, obviously, AI is here to
• 15:53
  - stay, but I I fully believe that the
• 15:57
  - patients will be influencing
• 15:59
  - health care in the near future in a
• 16:01
  - greater way.
• 16:02
  - We're, you know, creating,
• 16:05
  - patient AI bots that will monitor them, monitor
• 16:08
  - their EMR,
• 16:09
  - monitor
• 16:10
  - their their life. You'll have your own personal
• 16:14
  - assistant that does a lot of these factors,
• 16:16
  - and then that will be the bridge or
• 16:19
  - the communication into your EMR for the future.
• 16:23
  - The best part about this is the efficiency
• 16:26
  - and ease of,
• 16:29
  - communicating
• 16:30
  - your needs to your clinicians and vice versa.
• 16:33
  - Once this becomes part and parcel of a
• 16:36
  - full integrated process,
• 16:38
  - we'll be
• 16:39
  - having an AI in the cloud for
• 16:42
  - health care
• 16:43
  - overall, whether it be developed by the EMRs
• 16:46
  - or developed by the health care systems,
• 16:48
  - we're gonna be facilitating
• 16:50
  - efficiencies,
• 16:51
  - efficiency of care,
• 16:52
  - and
• 16:53
  - automation. So that is not only the burden
• 16:56
  - of
• 16:57
  - the scheduler
• 16:58
  - for an appointment or the patient
• 17:00
  - for the medications
• 17:01
  - or the clinician
• 17:03
  - for following up lab tests. It'll be an
• 17:05
  - automated process.
• 17:07
  - It's a game changer, but it's also
• 17:10
  - a tool. Just like we went from paper
• 17:12
  - to EMRs,
• 17:13
  - had to adopt, we're going in that and
• 17:16
  - that took eighteen years. Going from,
• 17:20
  - no AI to
• 17:22
  - deployment of AI took about eighteen months of
• 17:25
  - demand, but we're only at the beginning of
• 17:28
  - understanding how we use it.
• 17:30
  - We haven't seen a large tremendous
• 17:33
  - global enterprise health care ROI yet, but we
• 17:36
  - will
• 17:37
  - as we become better stewards
• 17:39
  - of the technology
• 17:41
  - as well as building that trust.
• 17:43
  - And I think that's the next big step.
• 17:46
  - I'm excited that it's here, but I'm also
• 17:48
  - excited that we can facilitate
• 17:50
  - trust around it to make the
• 17:53
  - clinical practice of medicine
• 17:56
  - a a a better place
• 17:58
  - to enjoy.
• 17:59
  - AMA had the,
• 18:02
  - wherewithal to create the joy in medicine
• 18:04
  - award, which is a fantastic way of evaluating
• 18:07
  - how we want to bring back that joy
• 18:10
  - of medicine at the bedside.
• 18:12
  - We're utilizing AI
• 18:14
  - to support that goal.
• 18:16
  - And and in the future, I think it'll
• 18:18
  - be an important part of not only current
• 18:21
  - state clinical practice, but training.
• 18:24
  - I'm thrilled to know that we are also
• 18:27
  - training our residents
• 18:29
  - and have a clinical informatics rotation
• 18:31
  - that brings in AI into their education.
• 18:35
  - Because now we have a tool just like
• 18:36
  - a stethoscope needs to be understood, how to
• 18:39
  - practice it, how to use it, and what
• 18:42
  - results come from it, AI is the same
• 18:44
  - way. It's another
• 18:46
  - facilitated
• 18:47
  - tool that health care now can utilize
• 18:50
  - and and deliver
• 18:51
  - more advanced
• 18:53
  - care.
• 18:54
  - And it's our
• 18:55
  - responsibility
• 18:56
  - to train those into the future to bring
• 18:59
  - that better practice through new technologies such as
• 19:01
  - AI.
• 19:02
  - Absolutely. Wow. I love that. Doctor Villaren, thank
• 19:06
  - you so much for joining us on the
• 19:07
  - podcast today. This has been such a a
• 19:09
  - fun conversation. I I think we got through
• 19:12
  - a lot when we're, like, thinking about ambient
• 19:14
  - documentation and then really exciting to hear about
• 19:16
  - the future. And I look forward to keeping
• 19:18
  - this conversation going. Thank you, Laura. Appreciate the
• 19:21
  - time, and thank you for having me.