Latest / Becker’s Healthcare Digital Health + Health IT / Dr. Rebecca G. Mishuris, Chief Medical Information Officer and VP at Mass General Brigham

Transcript

Dr. Rebecca G. Mishuris, Chief Medical Information Officer and VP at Mass General Brigham · 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 executives, engage with industry leaders,
• 0:29
  and explore the future of health care innovation.
• 0:32
  Learn more about our upcoming events at beckershospitalreview.com.
• 0:36
  See you there.
• 0:38
  This is Laura Dierda with the Becker's Health
• 0:40
  Care podcast. I'm thrilled today to be joined
• 0:42
  by doctor Rebecca Mashouris, chief medical information officer
• 0:45
  and vice president at Mass General Brigham. Doctor
• 0:47
  Mashouris, it's a pleasure to have you on
• 0:49
  the podcast today.
• 0:50
  Thanks very much for having me. Now I'm
• 0:53
  looking forward to our conversation and really digging
• 0:55
  deep into some of the cool things that
• 0:56
  you're doing at Mass General Brigham as well
• 0:58
  as how you're thinking about the future. But
• 1:00
  before we dive into my questions, can you
• 1:01
  tell us a little bit more about Mass
• 1:03
  General Brigham and what makes it unique?
• 1:06
  Sure.
• 1:07
  Mass General Brigham is a,
• 1:09
  health system,
• 1:11
  primarily located in Massachusetts, but also up in
• 1:14
  New Hampshire and a little bit into Maine.
• 1:17
  And we are made up of two anchor
• 1:19
  academic medical centers, the Mass General Hospital and
• 1:22
  Brigham Women's Hospital,
• 1:24
  and then a number of community hospitals and
• 1:27
  a very large number of ambulatory care sites.
• 1:31
  We are,
• 1:32
  the largest employer in Massachusetts,
• 1:35
  health care or otherwise,
• 1:38
  and,
• 1:38
  you know, really aim to serve the communities
• 1:41
  in which we exist.
• 1:43
  I think in terms of uniqueness,
• 1:45
  we bring together
• 1:47
  the patient care,
• 1:48
  the research,
• 1:50
  and the training and education aspects of health
• 1:52
  care. And that's really important and and really
• 1:55
  critical to our mission of,
• 1:57
  serving the communities in which,
• 2:00
  we
• 2:01
  we sit.
• 2:02
  Absolutely. That's fantastic to hear. You know, what
• 2:05
  a great opportunity
• 2:06
  to make a difference for a large patient
• 2:08
  population that you have there that are serving
• 2:10
  in the Boston area. Now,
• 2:12
  could you tell us a little bit about
• 2:14
  an accomplishment from the last year that you're
• 2:15
  most proud of?
• 2:17
  Sure. I think, I'm probably most proud of
• 2:20
  the way in which we have brought
• 2:23
  a new technology
• 2:25
  to bear against clinician burnout.
• 2:28
  That new technology is ambient documentation,
• 2:30
  the use of generative AI,
• 2:33
  to help draft a clinical note,
• 2:36
  from a recording of a of a visit
• 2:37
  between a patient and their,
• 2:40
  health care provider.
• 2:42
  And not just having brought that, to our
• 2:45
  our providers,
• 2:46
  but the way in which we did it
• 2:48
  in a very robust,
• 2:50
  methodical way in terms of
• 2:53
  understanding
• 2:53
  the technology,
• 2:54
  evaluating the technology,
• 2:56
  piloting the technology, and then bringing it to
• 2:59
  scale.
• 3:00
  And part of that intentionality
• 3:02
  was really around the fact that this is
• 3:04
  generative AI,
• 3:05
  and really has only been on the on
• 3:08
  kind of the in the market, for the
• 3:10
  last two plus years and and, you know,
• 3:12
  really in the market even less than that.
• 3:15
  And it's a technology that clearly has a
• 3:17
  role moving forward in health care, not just
• 3:20
  for ambient documentation, but for lots of different
• 3:23
  purposes.
• 3:24
  But it is a technology that we have
• 3:26
  not seen the likes of before in health
• 3:28
  care or otherwise
• 3:30
  in that it is
• 3:31
  built off of an immense amount of data,
• 3:34
  in that the algorithms
• 3:36
  are,
• 3:37
  you know, built by people but then are
• 3:39
  built to evolve over time as the data
• 3:42
  evolves, as the understanding
• 3:45
  evolves. Right? They're they're meant to learn.
• 3:48
  And that is very different than a lot
• 3:50
  of the static technology we've had in health
• 3:52
  care to date.
• 3:53
  And so we undertook an approach that was,
• 3:57
  modeled after actually clinical trials. So it's a,
• 3:59
  you know, zero to four clinical trials. We've
• 4:02
  taken a clinical trial informed approach to the
• 4:05
  evaluation and implementation of generative AI,
• 4:08
  really kind of first applying that at scale,
• 4:10
  to ambient documentation
• 4:12
  where we sought first to understand the technology,
• 4:16
  understand the safety of the technology, use a
• 4:18
  very small cohort of about 20 physicians,
• 4:22
  to evaluate the safety
• 4:24
  and workflows with ambient documentation,
• 4:27
  having seen that it was safe, right, no
• 4:29
  rampant hallucination,
• 4:31
  wasn't making things up, from the visit,
• 4:34
  and and seeing sort of the enthusiasm of
• 4:36
  those first twenty providers in terms of the
• 4:38
  workflow and the impact it was having on
• 4:40
  them, very quickly moved to a, what became
• 4:43
  a fairly large pilot or initial implementation, if
• 4:46
  you will, because it ended up being about
• 4:48
  800 physicians and advanced practice providers across our
• 4:52
  system, all care settings, all sites, all specialties,
• 4:56
  to really try and understand the impact that
• 4:58
  the technology was having on our providers.
• 5:01
  You could very quickly see that this technology
• 5:04
  was transforming the way in which we interact
• 5:08
  with our patients
• 5:09
  that ends up being a note.
• 5:11
  But it it was about more than just
• 5:13
  writing the note. It was actually a technology
• 5:16
  that was putting us back in the room
• 5:18
  with our patients and that's really unlike any
• 5:20
  other technology in healthcare. Right? A lot of
• 5:22
  times technology will put up barriers between people
• 5:25
  rather than take those barriers down.
• 5:28
  And so we,
• 5:29
  evaluated about 20 different measures
• 5:32
  of impact of ambient documentation and from that,
• 5:36
  particularly because of the impact it was having
• 5:38
  on clinician burnout with a 20% absolute reduction
• 5:41
  in clinician burnout,
• 5:43
  have now moved to scaling
• 5:45
  the technology across our system.
• 5:48
  There is nothing else in health care that
• 5:50
  has been able to
• 5:51
  reduce burnout to that extent.
• 5:54
  I don't necessarily
• 5:56
  expect,
• 5:57
  everyone to to experience that degree of impact.
• 6:01
  But even if we were to impact burnout
• 6:04
  by 5%, that would still be a huge
• 6:06
  amount at scale.
• 6:08
  And so, you know, the technology
• 6:11
  really you know, it it takes away the
• 6:14
  documentation
• 6:16
  aspect of things while you're with the patient.
• 6:18
  And so providers report 80% of providers report
• 6:22
  that they're spending
• 6:23
  more time looking at their patient. They're spending
• 6:26
  more time paying attention to their patients in
• 6:28
  the course of a visit. And 60% of
• 6:30
  our providers reported that they were more likely
• 6:33
  to extend the length of their clinical career.
• 6:35
  And then, as I mentioned, 20%
• 6:37
  reduction in providers reporting burnout symptoms.
• 6:41
  And so, you know, again, really kind of
• 6:44
  impressive
• 6:45
  impact of the technology, but I think that
• 6:47
  the way that our team went about evaluating
• 6:49
  and implementing and bringing the technology to scale
• 6:52
  is actually,
• 6:54
  the the most important thing to me. And
• 6:56
  the fact that we've been able to impact
• 6:58
  so many,
• 7:00
  providers' lives and, quite honestly, patients' lives as
• 7:02
  well. Patients notice that their doctors are no
• 7:04
  longer flapping away at the keyboard during their
• 7:06
  visit.
• 7:08
  And when I hear stories from my providers,
• 7:10
  like, I can go to my child's gymnastics
• 7:12
  meet on Saturday morning instead of writing my
• 7:14
  notes, or I can sit down to dinner
• 7:17
  with my family now instead of finishing my
• 7:19
  notes, that is huge, huge impact.
• 7:23
  Well, that's amazing to hear. You know? And,
• 7:25
  certainly, it's so inspiring to have that personal
• 7:27
  touch to it, not only from how it
• 7:29
  impacts patients and can really improve the patient
• 7:32
  care and experience, but then for the clinicians
• 7:35
  as well and the physicians who very much
• 7:37
  need to have that ability to, be there
• 7:40
  for their families
• 7:41
  and to fill their buckets in that space
• 7:43
  as well.
• 7:44
  Just knowing you have that kind of support
• 7:46
  from technology to,
• 7:48
  you know, do that, I can imagine, is
• 7:50
  really gratifying.
• 7:51
  I'm curious. I I know a lot of
• 7:53
  health systems are going down this path of
• 7:56
  trying to use this type of technology,
• 7:59
  and and figure out, you know, who the
• 8:00
  best partners are for them and what the
• 8:03
  best investments are for their organization.
• 8:07
  What types of considerations did you make when
• 8:09
  you were evaluating different opportunities, different partners, and
• 8:12
  technologies, and that before making a decision on
• 8:15
  how to move forward with, what you're currently
• 8:17
  using today?
• 8:19
  So, you know, first things first was, could
• 8:22
  it write a good note? Right? Could it
• 8:23
  draft a good note? And so we looked
• 8:25
  at the actual functionality of the technology
• 8:28
  at the current at that current time. Right?
• 8:31
  And that that had a lot to do
• 8:33
  with the,
• 8:34
  robustness and sophistication of the algorithms that were
• 8:37
  you know, the the large language models that
• 8:39
  were underlying that. And,
• 8:41
  so that was that was step one. Step
• 8:43
  two was really paying attention to the workflow,
• 8:45
  and looking for
• 8:47
  opportunities
• 8:48
  with, applications that would integrate into our existing
• 8:51
  workflows, and that meant the ability to integrate
• 8:54
  with our electronic health plan.
• 8:56
  And then, you know, I think last but
• 8:58
  certainly not least was really trying to get
• 9:00
  an understanding of where that vendor was going
• 9:03
  in the future because I don't think the
• 9:06
  the end of this is writing or drafting
• 9:08
  a a clinical note. Right? There are so
• 9:10
  many things you can do on top of
• 9:12
  that large language model and that recording of
• 9:15
  a visit.
• 9:16
  Everything from,
• 9:17
  helping me place an order. Right? If I
• 9:19
  mentioned during the visit that we're going to
• 9:21
  start you on metformin a thousand milligrams once
• 9:24
  a day, can it queue up that order
• 9:26
  for me,
• 9:28
  to helping me with my billing and coding
• 9:31
  to some point in the future, maybe even
• 9:33
  helping me with diagnostic,
• 9:36
  and and clinical reasoning and decision making.
• 9:39
  You know, if if I'm seeing a patient
• 9:41
  for something and I go through kind of
• 9:43
  my list of things that I wanna ask
• 9:44
  them about to try and rule in or
• 9:46
  rule out possible etiologies or or treatment options,
• 9:50
  could that large language model prompt me if
• 9:52
  I forgot to ask something, right, or,
• 9:55
  alert me to something that I wasn't aware
• 9:58
  of in the chart that might change my
• 10:00
  thinking. We're not there yet,
• 10:02
  but we were looking for vendors who were
• 10:04
  who really kind of had that forethought and
• 10:06
  foresight,
• 10:07
  to to see what was coming and and
• 10:09
  be able to build towards that.
• 10:11
  That's amazing to hear. Thank you for going
• 10:13
  a little bit deeper in there. Now I'm
• 10:15
  curious. Looking ahead, what do you see as
• 10:17
  some of the biggest opportunities for growth over
• 10:19
  the next twelve months or so?
• 10:22
  Oh, gosh. There's huge opportunities here. You know,
• 10:24
  I think we have some of the biggest
• 10:26
  challenges right now in health care from,
• 10:29
  clinician burnout to the financial stress in health
• 10:32
  care,
• 10:33
  to just quality and safety of the care
• 10:35
  that we're delivering and included in that the
• 10:37
  equity of the care that we're delivering.
• 10:39
  And so, you know, I think we continue
• 10:41
  to see big opportunities in those spaces for
• 10:44
  the technologies we have, but also really excited
• 10:47
  about what we're seeing coming from,
• 10:50
  the generative AI space,
• 10:52
  and opportunities
• 10:54
  that that provides that is significant again, significantly
• 10:57
  different,
• 10:58
  from what we've had available in the past.
• 11:00
  Generative AI is clearly not going to be
• 11:02
  the right tool for every problem that we
• 11:05
  encounter in health care, But I think some
• 11:07
  of our sickier problems,
• 11:09
  right, it could have real impact,
• 11:12
  moving forward.
• 11:14
  That's great to hear. And, you know, really
• 11:16
  cool to have that type of vision in
• 11:18
  what's possible over the next few months and
• 11:20
  and see, you know, where things are headed.
• 11:22
  Now I know we've talked a lot about
• 11:24
  the cool things and opportunities, especially with some
• 11:27
  of the AI and other technologies, but I'm
• 11:30
  curious as well. What are some of the
• 11:31
  big challenges that you're anticipating?
• 11:35
  So, you know, I think when it comes
• 11:36
  to the technology, one of the biggest challenges
• 11:39
  we're facing right now is,
• 11:41
  how do we evaluate these technologies? Right? Do
• 11:44
  we continue to use, and I think we
• 11:46
  should,
• 11:47
  our clinical trials informed approach and, how robust,
• 11:50
  you know, should the robustness of that change
• 11:52
  based off of the risk of the application?
• 11:55
  Should it change based off of the risk
• 11:57
  of the use case?
• 11:59
  And then as part of that, we have
• 12:00
  phase four, right, the last phase is actually
• 12:02
  the monitoring phase.
• 12:04
  And I think that is the biggest opportunity,
• 12:08
  in the technology space actually is to ensure
• 12:11
  that we have tools
• 12:12
  that allow us to do the sufficient monitoring
• 12:16
  of these applications moving forward.
• 12:18
  Again, in somewhat of a risk based approach
• 12:20
  where,
• 12:21
  you know, some of the applications, some of
• 12:24
  the use cases may be fairly low risk.
• 12:26
  Think about a chatbot that is giving you
• 12:28
  directions to get from one place to another.
• 12:31
  Fairly low risk when it comes to health
• 12:32
  care,
• 12:34
  versus,
• 12:34
  some of the tools I was describing that
• 12:36
  don't quite exist yet, right, but may very
• 12:38
  well in the in the kind of near
• 12:39
  term future that are helping us with clinical
• 12:42
  decision making. Those are really high risk when
• 12:45
  it comes to to outcomes,
• 12:47
  and and quality of care. And so we
• 12:49
  need monitoring tools that are aligned with that
• 12:52
  risk evaluation, that risk assessment
• 12:55
  that allow us to monitor these technologies moving
• 12:57
  forward. And in some spaces, actually in quite
• 12:59
  a few spaces, the technology actually doesn't exist
• 13:02
  yet, to do that monitoring. And so really,
• 13:05
  looking forward to to seeing the evolution of
• 13:07
  the applications themselves and also,
• 13:10
  in our in the the applications that are
• 13:12
  going to monitor those applications,
• 13:14
  so that we can get in start to
• 13:15
  get into some of these higher risk but
• 13:17
  maybe higher reward spaces as well.
• 13:20
  That's helpful for to understand. And, you know,
• 13:23
  it seems like a really,
• 13:25
  great way to look at things and make
• 13:27
  sure that, you know, you're
• 13:29
  seeing the challenges, but also developing solutions and
• 13:32
  able to overcome them moving forward. Before we
• 13:35
  wrap up here, I'm curious. What is the
• 13:37
  number one thing that you can do right
• 13:39
  now to set yourself up for success in
• 13:41
  the future?
• 13:43
  I think, you know, one thing we're doing
• 13:44
  right now is actually pulling together the multidisciplinary
• 13:49
  team that needs to move and drive this
• 13:52
  forward.
• 13:53
  And so, you know, a year and a
• 13:54
  half, two years ago, we set up an
• 13:56
  AI governance committee really to establish what our
• 13:59
  rules of the road were, establish a responsible
• 14:01
  use of AI,
• 14:03
  framework for the organization.
• 14:05
  And now it's about what,
• 14:07
  groups of people do we need to bring
• 14:09
  together in order to realize,
• 14:11
  and bring to life these technologies. And so
• 14:14
  that is much more about kind of the
• 14:16
  the technologist that
• 14:18
  with the data infrastructure, with data analytics,
• 14:22
  with the rest of the organization that brings
• 14:25
  the use cases,
• 14:26
  for these for these, technologies. And so bringing
• 14:30
  those people together into,
• 14:32
  a multidisciplinary
• 14:34
  team that actually will drive all of this
• 14:36
  forward, I I think, is the biggest thing
• 14:38
  we can do at this point, in addition
• 14:40
  to working you know, continuing to work with
• 14:43
  the vendors of these technologies to really understand
• 14:46
  how they work, to evaluate them within the
• 14:48
  context of our own health care system,
• 14:51
  and as I mentioned, you know, develop the
• 14:52
  tools to monitor them going forward.
• 14:56
  That's amazing to hear. Doctor. Macheras, thank you
• 14:58
  so much for joining us on the podcast
• 14:59
  today. This has been a really fun conversation,
• 15:01
  and I look forward to connecting with you
• 15:03
  again soon. Thanks very much for having me.