Latest / Becker’s Healthcare Digital Health + Health IT / Scaling AI and Simplifying Healthcare Data for Better Clinical Care with Edwina Bhaskaran

Transcript

Scaling AI and Simplifying Healthcare Data for Better Clinical Care with Edwina Bhaskaran · Becker’s Healthcare Digital Health + Health IT

• 0:00
  - What we wanna do with AI is help
• 0:01
  - clinicians go from being reactive
• 0:04
  - to proactive.
• 0:05
  - At Philips, we can recognize patterns in monitoring
• 0:08
  - data across populations
• 0:10
  - and apply those insights to specific patients.
• 0:14
  - Safe and responsible AI can support sharper insights,
• 0:18
  - pave the way for faster diagnosis,
• 0:21
  - and help streamline care.
• 0:23
  - That can benefit clinicians,
• 0:25
  - it can benefit health systems. And most importantly,
• 0:28
  - it can benefit patients and their families.
• 0:33
  - Phillips.
• 0:34
  - Hello, everyone. This is Mackenzie Bean with Becker's
• 0:37
  - Healthcare. Thanks for tuning in to the Becker's
• 0:39
  - Healthcare podcast series.
• 0:41
  - I'm so thrilled for our conversation today with
• 0:43
  - Edwina Beskaron, who is the inaugural chief clinical
• 0:46
  - systems and informatics officer at Mayo Clinic. Edwina,
• 0:50
  - thanks for joining us on the podcast today.
• 0:52
  - How are you? Good. Thank you, Mackenzie, for
• 0:54
  - having me. Yeah. A pleasure to have you
• 0:55
  - here. I I know you're no stranger to
• 0:57
  - the Becker's podcast, but I'd love to have
• 0:59
  - you introduce yourself and share a little bit
• 1:01
  - more about your role for anyone who might
• 1:03
  - not be familiar.
• 1:04
  - Sure. Thank you.
• 1:05
  - I have the privilege of supporting 65,000
• 1:07
  - clinical end users of our core clinical systems,
• 1:10
  - the EHR being one of them. I am
• 1:13
  - a nurse by training and by heart. I
• 1:14
  - picked up some informatics skills along the way,
• 1:17
  - and I am a relentless optimizer by choice.
• 1:20
  - So that's a little bit about me.
• 1:22
  - My role is inaugural at the Mayo Clinic.
• 1:25
  - It really bridges the process technology
• 1:28
  - and people part of it and making sure
• 1:30
  - our clinical systems meet the needs of our
• 1:32
  - clinical practice. Mhmm. I love that relentless optimizer.
• 1:36
  - I think we're gonna talk a lot about
• 1:37
  - a lot of optimizing today.
• 1:40
  - You know, we're at this time in health
• 1:41
  - care where we're seeing health systems facing a
• 1:43
  - growing stream of patient data with the rise
• 1:45
  - of virtual digital care, whether it's bedside monitoring,
• 1:48
  - ambulatory devices,
• 1:50
  - home care.
• 1:51
  - Where is that data creating actual insights for
• 1:53
  - clinicians at Mayo Clinic? And where is it
• 1:55
  - perhaps still
• 1:57
  - overwhelming or burden rather than insightful?
• 1:59
  - Yeah. I think that's a really interesting question
• 2:02
  - because I think with great data gives great
• 2:05
  - options Mhmm. And opportunities.
• 2:07
  - So,
• 2:09
  - specifically around the areas of personalization.
• 2:12
  - Mhmm. So, we now have the ability to
• 2:14
  - personalize views
• 2:16
  - for our clinician
• 2:17
  - in very different ways. And what it brings
• 2:20
  - to to surface is it gives the ability
• 2:22
  - to make sure that our clinicians have the
• 2:24
  - right data, but it also creates an interface
• 2:27
  - that is very specific
• 2:29
  - to
• 2:30
  - set set individual. So supporting
• 2:32
  - them in terms of adoption of new tools
• 2:35
  - and technologies
• 2:36
  - is a consideration when you have hyper personalized
• 2:39
  - views into into the data. So that's one
• 2:42
  - piece that I would say. I'd say the
• 2:44
  - other thing that,
• 2:46
  - we are still finding is that some of
• 2:48
  - the tools that are coming out,
• 2:51
  - are great at summarization,
• 2:53
  - but it still is not solving the fragmentation
• 2:56
  - that lies underneath
• 2:58
  - as a result. So I think the the...
• 3:00
  - One of the things that we have to
• 3:02
  - start thinking about is
• 3:04
  - simplification over summarization. So where do we need
• 3:06
  - to simplify
• 3:08
  - the the data inputs, the the documentation
• 3:11
  - so that we get better summation tools, better
• 3:14
  - views into our data?
• 3:16
  - Because otherwise, we are just
• 3:18
  - layering
• 3:19
  - a user interface over bad processes. Mhmm. Which
• 3:23
  - is not effective
• 3:24
  - at the end of the day.
• 3:25
  - You mentioned that fragmentation. Let's go a bit
• 3:28
  - deeper there.
• 3:29
  - When any clinician is trying to piece together
• 3:31
  - that diagnosis and all the various data points
• 3:33
  - they have at their disposal. That is across
• 3:35
  - information from radiology, cardiology, pathology, other specialties.
• 3:40
  - Where are you seeing that fragmentation creating the
• 3:42
  - most friction at Mayo Clinic? How do you
• 3:44
  - bring together those data sources to address some
• 3:47
  - of that? I think I read that your
• 3:49
  - teams are processing billions of data pieces a
• 3:52
  - day. It's a lot of data. Right. Yeah.
• 3:54
  - So where we see a lot of that
• 3:56
  - fragmentation is actually in the way that we
• 3:58
  - process
• 3:59
  - outside materials today.
• 4:01
  - And so,
• 4:02
  - being a destination medical center, we rely heavily
• 4:05
  - on outside materials from other organizations.
• 4:08
  - And so from a fragmentation
• 4:10
  - perspective, while
• 4:12
  - our internal systems speak to each other and
• 4:14
  - we have an integrated solution,
• 4:16
  - when we need to see patients that are
• 4:18
  - bringing to us lots of outside records
• 4:21
  - and those records are fragmented,
• 4:23
  - I think that's where we have opportunities.
• 4:26
  - Personally,
• 4:27
  - I also help take care of my parents
• 4:28
  - who are 86 and 83.
• 4:31
  - Just two weeks ago, I had to
• 4:33
  - go and grab a disc,
• 4:36
  - a CD disc for them that they actually
• 4:38
  - grabbed for for me... For us and then
• 4:41
  - had to go and purchase a DVD drive
• 4:44
  - from Amazon Mhmm. To upload their records. And
• 4:47
  - this wasn't at our at our organization, but
• 4:49
  - it was another organization. So we still have
• 4:51
  - a lot of archaic processes
• 4:54
  - that we need to think through in terms
• 4:56
  - of that
• 4:57
  - organization to organization
• 4:59
  - flow and how can we do a better
• 5:01
  - job of standardizing some of the... That interoperability.
• 5:04
  - Mhmm. So... Yep. Yeah. Real real real world
• 5:08
  - example. I think many people can relate to
• 5:09
  - that feeling of them being like, wait a
• 5:11
  - minute. My laptop doesn't even have a disc
• 5:13
  - drive anymore. How do I get these records?
• 5:15
  - Yeah. And then not to mention, I had
• 5:16
  - to explain to my son what a disc
• 5:19
  - is.
• 5:21
  - That... Yep. We can all relate to that
• 5:23
  - very much.
• 5:25
  - Let's talk about that patient journey. You know,
• 5:28
  - you mentioned
• 5:29
  - helping your parents get medical records. I mean,
• 5:31
  - patients across the country are moving in and
• 5:32
  - out of outpatient settings, home care, the hospital.
• 5:36
  - Where does some of the clinical context
• 5:38
  - most often get lost during those transitions?
• 5:41
  - What would it take for care teams to
• 5:43
  - have that full view of the patient journey
• 5:45
  - and the data that comes with it? Yeah.
• 5:47
  - I think that's an excellent question. Again, I'm
• 5:49
  - probably gonna use my personal experience more so
• 5:51
  - than my professional one. So oftentimes, we rely
• 5:54
  - on our patients or their caregivers to be
• 5:56
  - the relayers of information.
• 5:58
  - And so I think we have great tools.
• 6:01
  - We have patient portals,
• 6:02
  - but they're... They only give a certain story,
• 6:06
  - that is still reliant on the patient to
• 6:09
  - transcribe
• 6:10
  - that. What does it mean and that that
• 6:12
  - clinical personal context to them? So I think
• 6:15
  - where we get... Where we miss that handoff
• 6:18
  - is that clinician to clinician experience of this
• 6:21
  - is what's important to me from a patient
• 6:22
  - perspective, but also from a clinical perspective. These
• 6:25
  - are the things
• 6:26
  - that we,
• 6:27
  - are looking for for that said patient.
• 6:29
  - So it's it's not just about here are
• 6:31
  - the active problems that we're managing. Here's the
• 6:34
  - active medications. I think that only tells
• 6:36
  - a portion of the story. But the clinical
• 6:39
  - context is missing in the subjective narrative, which
• 6:42
  - we have reams and reams of data and
• 6:44
  - and notes. And while it's vis... It's visible,
• 6:47
  - it's not already... It's not always contextually aware.
• 6:51
  - Mhmm. So really thinking through those handoffs more.
• 6:55
  - Let's shift gears a bit from the data
• 6:57
  - piece to
• 6:59
  - AI. Because I understand that you really oversee
• 7:01
  - Gen AI deployment across the system. Is that
• 7:04
  - accurate?
• 7:05
  - Yes. Yeah. I partner with... We have several
• 7:07
  - programs at Mayo Clinic, but I partner with
• 7:09
  - most of our enterprise programs. Yeah. Wonderful.
• 7:12
  - So with so many AI use cases competing
• 7:15
  - for attention
• 7:16
  - and growing into technology advancing every single day,
• 7:19
  - How do you distinguish between what might be
• 7:21
  - a compelling pilot to something that truly is
• 7:23
  - ready for daily clinical care?
• 7:26
  - So we have,
• 7:27
  - a decentralized
• 7:28
  - innovation model at Mayo Clinic. And one of
• 7:31
  - the things that I, in particular, look for
• 7:33
  - is sustained adoption,
• 7:35
  - that thirty, sixty, and ninety day mark. So,
• 7:38
  - there's often lots of great excitement over that
• 7:41
  - initial,
• 7:42
  - technology at that thirty day mark. But when
• 7:44
  - you start seeing sustained adoption at that sixty
• 7:47
  - and ninety day model, then you probably have
• 7:50
  - something that is worthy
• 7:52
  - of increasing scalability
• 7:53
  - to the broader enterprise. So right now, we're
• 7:56
  - really targeting areas where we are seeing
• 7:59
  - really good adoption outcomes, and then we're also
• 8:02
  - seeing places where we're not seeing expected outcomes
• 8:04
  - that we would like. Mhmm. So really, what,
• 8:06
  - what does that look like over the ninety
• 8:08
  - day mark? And then that's your sign, integrating
• 8:10
  - because the sustained adoption will also tell you
• 8:12
  - how integrated is it in the workflow. Mhmm.
• 8:15
  - So previously I used to talk about, well,
• 8:17
  - if it's integrated into workflow, you know, this
• 8:19
  - is something that
• 8:19
  - will
• 8:21
  - will lend itself well. But I actually think
• 8:23
  - it's the adoption metric that is the... Is
• 8:26
  - what we should be really paying attention to.
• 8:28
  - If it's not efficient and effective and natural
• 8:30
  - in the workflow, they're not gonna use it
• 8:31
  - at the end of the day. Yep. Then
• 8:33
  - can you talk a bit about what does
• 8:34
  - the responsible scaling look like? What lessons have
• 8:37
  - you learned in that realm? So I think
• 8:39
  - the the challenge is that the technology is
• 8:41
  - advancing at a pace that we can't keep
• 8:43
  - up with. And so
• 8:45
  - it's not just about deployment. It's about maintenance
• 8:48
  - and monitoring and decommissioning
• 8:50
  - when we are not seeing sustained adoption. And
• 8:53
  - that is the challenge that we have with
• 8:55
  - us right now
• 8:56
  - is we see,
• 8:58
  - a lot of really great ideas,
• 9:00
  - that are coming
• 9:01
  - through, but the challenge is is when are
• 9:03
  - those ideas
• 9:05
  - beyond the ninety days? How do we have
• 9:07
  - that organizational discipline to decommission?
• 9:10
  - And who's monitoring that? Because oftentimes,
• 9:13
  - our technical teams are on to the next
• 9:16
  - next thing, and so we have to create
• 9:18
  - that rigor,
• 9:20
  - to to make sure that we are not
• 9:21
  - cluttering our clinical systems ecosystems.
• 9:24
  - To assess, perhaps sunset a pilot and then
• 9:27
  - move on to something that's higher value. Correct.
• 9:29
  - Makes sense.
• 9:30
  - Let's now look at
• 9:32
  - the partner relationships.
• 9:34
  - We're seeing so many new technologies, and I
• 9:37
  - think health systems are increasingly recognizing
• 9:39
  - that those relationships with health care companies are
• 9:42
  - moving beyond one time purchases to, in fact,
• 9:45
  - longer term relationships.
• 9:47
  - What should a partner... Partner, excuse me, remain
• 9:50
  - accountable for after a go live? What do
• 9:53
  - you look to them to provide in a
• 9:54
  - relationship that is genuinely valuable over time?
• 9:57
  - So I think it actually goes to the
• 9:59
  - heart of sustained adoption. Mhmm. So I am
• 10:02
  - looking for a partner that will work with
• 10:04
  - us to understand
• 10:06
  - where are those metrics that we need,
• 10:10
  - and where is that usage data in different
• 10:12
  - forms and and formats for all levels of
• 10:14
  - the organization. So somebody who's working and partnering
• 10:17
  - with that and
• 10:19
  - not afraid to have those tough conversations
• 10:22
  - of, okay. We're not seeing adoption
• 10:24
  - as we should be. Let's have a conversation
• 10:27
  - about why, and let's be brave enough to
• 10:29
  - decommission
• 10:30
  - if we need to. I think the other
• 10:32
  - piece,
• 10:33
  - in terms of the relationship is I'm looking
• 10:35
  - for a partner who is open to feedback
• 10:39
  - and willing to listen to our frontline clinical
• 10:41
  - clinical staff to be able to develop what's
• 10:44
  - most needed for their workflow and their most
• 10:47
  - pressing needs.
• 10:48
  - Not something that's done in a vacuum that
• 10:51
  - is done through
• 10:52
  - software vendors or the the development arm, but
• 10:56
  - really listening to the clinicians
• 10:58
  - of the health care community. Mhmm. And having
• 11:01
  - them shape the innovation and where all of
• 11:03
  - this work is going. Right. And have them
• 11:05
  - engaged
• 11:05
  - early in their development. Not, I've already developed
• 11:08
  - something, and I need you to try and
• 11:10
  - test it out for me. I I want
• 11:12
  - a partner who's willing to work with us
• 11:13
  - further upstream. Mhmm. And really help them be
• 11:16
  - co developers or designers versus testers. Correct. That
• 11:19
  - makes sense. The last thing I wanna ask
• 11:21
  - you about is diagnostics, which I think is
• 11:24
  - another interesting space that we're seeing quickly evolve
• 11:26
  - in health care. What do you see as
• 11:28
  - a future of integrated diagnostics and the the
• 11:31
  - field in general? Where do you see that
• 11:33
  - going across Mayo Clinic in the next five
• 11:35
  - years? Yeah. Integrated diagnostics have a very crucial
• 11:39
  - role because they are the heart of our
• 11:41
  - objective data in health care. Mhmm. And so
• 11:44
  - as we're thinking about
• 11:46
  - new and novel cures at Mayo Clinic, we
• 11:49
  - have a very strong
• 11:50
  - research shield. And one of the things that
• 11:53
  - our research organization is focused on is preCURE.
• 11:56
  - So the role of diagnostics in helping us
• 11:59
  - see the invisible,
• 12:01
  - I think is very important as we look
• 12:03
  - at the technology that is emerging and how
• 12:05
  - can we help our patients
• 12:07
  - know what's coming down the pipeline for them,
• 12:10
  - and catch those diseases much earlier in the
• 12:12
  - process. Mhmm.
• 12:14
  - A very exciting time and so impactful for
• 12:16
  - patient care. Well, Edwina, thank you so much
• 12:18
  - for joining us on the podcast today. It's
• 12:20
  - been, a pleasure speaking with you, so we
• 12:23
  - appreciate your time. I'd also like to thank
• 12:25
  - our podcast sponsor, Philips. Listeners, you can tune
• 12:28
  - in to more podcasts at Becker's Healthcare by
• 12:30
  - visiting our podcast page at beckershospitalreview.com.
• 12:34
  - Thank you, Mackenzie.