Latest / Becker’s Healthcare Digital Health + Health IT / Dr. Justin Coran, PhD, MPH, Former Chief Analytics Officer and Adjunct Professor

Transcript

Dr. Justin Coran, PhD, MPH, Former Chief Analytics Officer and Adjunct Professor · 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 Dierda with the Becker's Healthcare
• 0:40
  podcast. I'm thrilled today to be joined by
• 0:42
  doctor Justin Koren, who is an analytics executive.
• 0:46
  Justin, it's a pleasure to have you on
• 0:47
  the podcast today.
• 0:49
  Thank you so much for having me. Exciting
• 0:51
  for this discussion.
• 0:52
  Absolutely. Well, Justin, I am excited that we
• 0:55
  have you here as well because I know
• 0:56
  you've had a lot of experience at a
• 0:58
  variety of health systems across the country and
• 1:00
  certainly
• 1:01
  in-depth knowledge on analytics
• 1:03
  in the IT side. And so this will
• 1:05
  be a great perspective to share with our
• 1:07
  audience. I I know things are moving so
• 1:09
  rapidly, especially in the artificial intelligence space. So
• 1:12
  I think I without further ado, let's get
• 1:14
  started here.
• 1:16
  Before we begin, could you introduce yourself and
• 1:18
  tell us a little bit more about your
• 1:19
  background?
• 1:20
  Absolutely.
• 1:22
  So, I am, as you stated earlier, an
• 1:24
  analytics executive
• 1:26
  and, been a former chief analytics officer in
• 1:29
  health care systems. I'm also a data scientist
• 1:31
  and innovator with over twenty years of experience
• 1:35
  in the health care industry.
• 1:36
  In 2024,
• 1:38
  I won the AI 100 award, which is
• 1:40
  given to the top 100 executives in the
• 1:43
  country that conduct AI ML
• 1:46
  operations
• 1:47
  within their companies.
• 1:48
  And,
• 1:49
  overall, I've held multiple adjunct professor positions in
• 1:52
  medical schools. And I I would classify myself
• 1:55
  as, you know, just having this ongoing passion
• 1:58
  for solving complex problems
• 2:01
  and integrating applied analytics into company's operations.
• 2:05
  Well, that's great to hear. You you know,
• 2:07
  and and certainly,
• 2:08
  really a lot of experience I know you
• 2:11
  you bring.
• 2:12
  So I guess we'll dive right in. As
• 2:14
  an analytics executive, what are the most successful
• 2:16
  projects you've led at Health Systems that yielded
• 2:19
  the best results?
• 2:21
  Sure. Actually, one of my earlier projects is
• 2:23
  still yielding,
• 2:25
  results. And at this point, it's gone global.
• 2:27
  So I created a machine learning
• 2:30
  algorithm to help the risk risk stratify patients,
• 2:34
  in the population health space and within value
• 2:36
  based care. And what that algorithm did is
• 2:39
  it found the patients that were at the
• 2:42
  greatest need of care.
• 2:43
  And when facing limited human resources in a
• 2:46
  large health system and when they have a
• 2:48
  lot of patients,
• 2:50
  under a population health umbrella.
• 2:53
  The question is, how do you get the
• 2:54
  sickest patients to the doctors as fast as
• 2:57
  possible?
• 2:58
  And so, when I conducted that work earlier
• 3:00
  on in my career,
• 3:02
  those patients overall, that proportion of patients were
• 3:05
  responsible for about $1,200,000,000
• 3:08
  in spend.
• 3:09
  And by using that algorithm and being able
• 3:12
  to bring the sickest patients into care and
• 3:15
  to the attention of care coordinators,
• 3:18
  we were able to save $200,000,000,000
• 3:21
  A lot of that money
• 3:22
  was in the employee health care plan within
• 3:25
  that health system, which meant there was a
• 3:27
  huge return on investment.
• 3:29
  And now,
• 3:30
  multiple health systems across the world use
• 3:33
  that algorithm or a customized version of it
• 3:36
  today to help with their population health efforts
• 3:40
  Some of my other work that I've done,
• 3:42
  is,
• 3:43
  several work in ai and ml space and
• 3:45
  detecting undiagnosed disease And how can we find
• 3:48
  the patients
• 3:50
  that either are currently under care and undiagnosed
• 3:53
  and get those patients to primary care or
• 3:56
  to the appropriate specialist,
• 3:58
  so that we can address
• 3:59
  those variations in care or that potential undiagnosed
• 4:03
  disease as fast as possible.
• 4:06
  Other things that I've done more recently is
• 4:08
  a lot of technology
• 4:09
  consolidation,
• 4:10
  building modern cloud infrastructures
• 4:13
  and really enabling
• 4:14
  what I call data and analytic governance,
• 4:17
  within health care systems and other organizations.
• 4:19
  And that really sets the foundation for being
• 4:22
  able to do applied analytics or even moving
• 4:25
  into the artificial intelligence and machine learning world,
• 4:28
  being able to integrate those type of products
• 4:30
  with success into a company's ongoing operations.
• 4:34
  That's amazing to hear. You know, it it
• 4:36
  really cool to understand that real world example
• 4:39
  of not only using the AI algorithms to,
• 4:42
  provide better care and more targeted care, but
• 4:44
  then also
• 4:46
  realize those types of savings. I mean, that's
• 4:48
  significant, especially in a time right now where
• 4:50
  so many hospitals and health systems are tightening
• 4:52
  their belts and and seeing thinner margins than
• 4:54
  before, being able to have the right kind
• 4:57
  of expertise in,
• 4:59
  AI models in place, it seems like is
• 5:00
  a huge, huge benefit.
• 5:02
  It's critical in in today's world. And and
• 5:04
  as we know, health care costs are very
• 5:07
  hard to control,
• 5:08
  and seem to continue to go up year
• 5:10
  over year. And so being able to leverage
• 5:13
  a company's information to make better clinical decisions
• 5:16
  or even better administrative decisions if we're not
• 5:18
  just talking about clinical applications,
• 5:21
  It's gonna be critical over the next five
• 5:23
  years,
• 5:23
  as health systems become more analytically mature and
• 5:26
  start looking at different enabling technologies to achieve
• 5:29
  efficiencies in their core business.
• 5:32
  Absolutely. That's such a great point. Now I'm
• 5:34
  curious looking into that future. Where do you
• 5:36
  see ai and machine learning headed in health
• 5:39
  care? What are some of those best opportunities
• 5:41
  for hospitals? And what should they watch out
• 5:43
  for?
• 5:44
  So I think we're gonna see multiple categories
• 5:47
  of AI and ML across the company. And
• 5:49
  then let me define that a little bit
• 5:50
  better. So category number one is the question
• 5:53
  of where is AI and ML going, let's
• 5:55
  say, within the clinical space, and to aid
• 5:58
  clinical operations,
• 5:59
  physicians, nurses, and other health care workers.
• 6:03
  The biggest boom we're seeing right now is
• 6:05
  the integration of AI and ML within electronic
• 6:08
  health record systems.
• 6:10
  Definitely, we're seeing Epic as one of the
• 6:11
  leaders here where they've got over three dozen
• 6:14
  types of AI products that they're bringing into
• 6:17
  the Epic ecosystem.
• 6:18
  Oracle has announced that they're doing some very
• 6:20
  similar things on the integration of AI within
• 6:23
  their electronic health record product.
• 6:26
  And then also within that clinical domain, it's
• 6:28
  not just EHRs that are benefiting from the
• 6:31
  integration of AI and ML, but we're already
• 6:33
  seeing medical devices
• 6:35
  utilize it more frequently than they ever have
• 6:37
  before. And what's interesting in the medical device
• 6:40
  space is MRIs were utilizing machine learning technologies
• 6:44
  for decades at this point. And so some
• 6:46
  of the earliest machine learning models were actually
• 6:48
  used in radiology and in the construction
• 6:51
  of different types of radiological
• 6:52
  equipment. But we're now seeing AI and ML
• 6:55
  and medical devices
• 6:57
  that are related to cardiology,
• 6:59
  neurology. Of course, radiology
• 7:02
  is still a pioneer in the space. And
• 7:04
  then we're even seeing some examples in pathology.
• 7:07
  And that's just a few to name many,
• 7:09
  of the different innovations that are occurring across
• 7:11
  the country.
• 7:12
  Also,
• 7:13
  so if we talk a little bit about
• 7:14
  the clinical domain, then we have the administrative
• 7:16
  domain. And that's a really an interesting one
• 7:18
  because I think a lot of times in
• 7:20
  health care, we first gravitate towards the clinical
• 7:22
  side and how can we first help our
• 7:24
  physicians and our nurses.
• 7:26
  But on the administrative side, we're now seeing
• 7:29
  a rise in AI that can have what
• 7:31
  what some of the data scientists and practitioners
• 7:34
  call a force multiplier effect. And what that
• 7:36
  means is we're looking at being able to
• 7:38
  have people be more efficient in the work
• 7:41
  that they're doing day to day. And if
• 7:43
  you can utilize an AI tool to do,
• 7:46
  let's say, two or three times the amount
• 7:48
  of work but not have to include any
• 7:50
  additional hours in your day, then AI didn't
• 7:53
  replace people. What AI is really doing is
• 7:55
  taking the existing work staff and
• 7:59
  multiplying the amount of work they can do
• 8:01
  with the same amount of effort. And then
• 8:02
  that's what we mean by a force multiplier.
• 8:05
  Some of this we're seeing within tools like
• 8:06
  Microsoft Copilot,
• 8:08
  where, you know, in traditional day, you might
• 8:10
  have a person that would have to sit
• 8:11
  in a board meeting or in other critical
• 8:13
  committee meetings within an organization,
• 8:15
  and they would take notes or potentially transcribe
• 8:18
  the meeting minutes. Now we have ambient technology
• 8:21
  that can just listen and automatically
• 8:23
  transcribe
• 8:24
  those meetings for us. That type of technology
• 8:26
  is also existing on the clinical side where
• 8:29
  we can have ambient technology listening to a
• 8:31
  physician patient encounter and then helping that doctor
• 8:34
  transcribe that activity directly into
• 8:37
  documentation within the electronic health record system.
• 8:40
  Also, on the administrative side, I think we're
• 8:42
  gonna see a lot of innovation in revenue
• 8:44
  cycle management in the finance area, especially within
• 8:47
  the automation space.
• 8:49
  And then somewhere, let's call it maybe five
• 8:52
  years down the road because there's only about
• 8:54
  three dozen health systems nationwide that are really
• 8:57
  doing this today. But we're gonna start seeing
• 8:59
  a lot of in house innovation take off
• 9:02
  where a health system or even another company
• 9:04
  could even be outside the health industry Is
• 9:07
  starting to build their own ai models their
• 9:09
  own ai agents or modifying a large language
• 9:12
  model that is specific and customized to health
• 9:15
  care
• 9:15
  And we definitely see we're gonna think we're
• 9:17
  gonna see a huge Proliferation
• 9:20
  in that type of work, you know anywhere
• 9:22
  from five to ten years from now And
• 9:24
  the reason that could take a long time
• 9:25
  is it requires a company to have a
• 9:28
  very strong data foundation
• 9:30
  and an analytic foundation, a governance program in
• 9:33
  place where you can have standards and policies,
• 9:36
  your technology and your environments that house your
• 9:38
  data needs to be consolidated
• 9:40
  so that if you're training an AI model,
• 9:43
  or an application that it can source that
• 9:46
  data from the smallest number of locations possible
• 9:49
  in order to speed the development,
• 9:51
  of that product. And there's a lot of
• 9:53
  health systems that are still doing a lot
• 9:54
  of work on their foundation. But as they
• 9:57
  catch up, to let's say the, you know,
• 9:59
  the top 20, top 36 health systems that
• 10:01
  might already be doing this today,
• 10:03
  we're gonna start seeing that exponential effect again
• 10:06
  of in house talented teams
• 10:09
  being able to customize these products alongside doctors,
• 10:12
  nurses, and health care workers.
• 10:15
  That was fascinating to hear. You know? And
• 10:16
  and, so truly, truly important to have that
• 10:19
  right mix of, you know, outside expertise that
• 10:22
  you're working with as well as building those
• 10:23
  in house teams to meet the need needs
• 10:26
  of the organization overall. And I'm curious as
• 10:28
  we get closer and closer to that space
• 10:30
  of of,
• 10:32
  more AI, more of those models, more technology
• 10:34
  that's supporting the in person teams,
• 10:38
  What does that look like for the IT
• 10:40
  leaders? How are their teams transforming? What type
• 10:42
  of talent or or skill sets do they
• 10:44
  need to bring in in order to prepare
• 10:46
  for that future?
• 10:47
  Sure. So data engineering is gonna be critical.
• 10:50
  And, you know, sometimes that comes into,
• 10:53
  a buy versus a build type strategy. Do
• 10:55
  you partner with a third party that can
• 10:56
  really help with your data management and and
• 10:59
  data engineering environments? Think of, like, cloud based
• 11:01
  environments and going from potentially fragmented data centers
• 11:05
  or multiple,
• 11:06
  technology platforms into just one just to enable
• 11:09
  that work to be done.
• 11:11
  Regarding
• 11:12
  inside talent and I think this is another
• 11:14
  reason why health care is gonna be a
• 11:15
  little bit slower to adopt
• 11:17
  a build strategy or even being able to
• 11:20
  innovate at innovate at scale is because you
• 11:22
  really need to have the right mix of
• 11:24
  talent. And that means the combination
• 11:27
  of engineers,
• 11:28
  data scientists, analysts, and potentially even software developers
• 11:32
  all being able to work together with a
• 11:34
  common goal,
• 11:35
  for that organization. And historically,
• 11:37
  most health systems,
• 11:39
  especially if they're not an academic health system,
• 11:42
  because we do see this quite a bit
• 11:43
  in academic health systems,
• 11:45
  don't have that same mix of talent or
• 11:48
  that mix of expertise in house today. The
• 11:50
  other challenge is this field is growing exponentially,
• 11:53
  where one of my key pieces of advice
• 11:56
  to any analytic executive, but this can even,
• 11:59
  be a piece of advice to CIOs and
• 12:01
  CTOs,
• 12:02
  is to master the art of self directed
• 12:05
  learning
• 12:06
  because the field is evolving so quick year
• 12:08
  over year. And that could be based on
• 12:11
  technologies that are advancing high performance computing.
• 12:14
  This could be brand new programming languages or
• 12:16
  innovations within popular programming languages such as Python
• 12:20
  and even just the availability of large language
• 12:22
  models. You know, several years ago, there was
• 12:24
  only maybe a dozen. Today, there's over 350.
• 12:27
  And being able to be that strong self
• 12:30
  directed learner will allow you to stay ahead
• 12:32
  of the curve and to really be able
• 12:34
  to understand what's coming down in the pipeline.
• 12:37
  But even more important is how do you
• 12:38
  filter out some of that noise? Where if
• 12:40
  you're being bombarded with, you know, 100 to
• 12:42
  200 new ideas a day, how do you
• 12:45
  figure out what the top 10 or top
• 12:47
  20 most important are that you should focus
• 12:49
  your time on and then really start thinking
• 12:51
  of what does that mean for my operations
• 12:53
  or even for my company?
• 12:55
  Got it. That makes a lot of sense.
• 12:57
  And, you know, it is really helpful specifics
• 12:59
  in terms of looking at how different organizations
• 13:01
  are thinking about this and going about this.
• 13:03
  Obviously, short and long term,
• 13:05
  you know, looks different for everybody, but to
• 13:07
  have that kind of skill set and understanding
• 13:09
  of what's possible and how quickly things are
• 13:11
  changing in the technology side is really, really
• 13:13
  helpful. Before we wrap up here, I'm curious,
• 13:15
  what advice do you have for emerging health
• 13:17
  care leaders, especially in the analytics field?
• 13:21
  Yep. Specific to analytics executives, one of the
• 13:24
  top pieces of advice that I could provide
• 13:27
  is partner very closely
• 13:30
  with your chief information officers, your chief technology
• 13:33
  officers,
• 13:34
  and even your your vice presence of IT.
• 13:37
  Analytic operations
• 13:38
  cannot reach its full potential if analytic leaders
• 13:42
  operate in a silo.
• 13:43
  It requires
• 13:45
  collaboration
• 13:46
  across the company, but even stronger collaboration
• 13:49
  between
• 13:50
  your IT department operations and your analytic operations.
• 13:54
  And that's specific no matter where analytics sits.
• 13:57
  In some comp in many companies, analytics sit
• 13:59
  in the IT department, and and you might
• 14:01
  have a natural hierarchical relationship
• 14:03
  to those leaders, your CIO, your CTO, or
• 14:06
  other VPs.
• 14:07
  But there are other companies where analytics sits
• 14:10
  outside of IT.
• 14:11
  And
• 14:12
  it is critical that even in those business
• 14:15
  models
• 14:16
  that they partner very closely, they should have
• 14:18
  a seat at the table both in the
• 14:20
  governance
• 14:21
  program of of how you're governing data and
• 14:24
  using analytic products. But even in the innovation
• 14:26
  space and the technology space of what technology
• 14:29
  do we need in the company to enable,
• 14:32
  the analytics strategy, and what does that look
• 14:35
  like if we do wanna create something new
• 14:37
  in house within the company?
• 14:40
  Those members should have a seat at the
• 14:41
  table during all of those discussions and be
• 14:44
  very close partners.
• 14:47
  That's amazing to hear. Justin, thank you so
• 14:48
  much for joining us on the podcast today.
• 14:50
  This has been a really amazing discussion, and
• 14:52
  I look forward to connecting with you again
• 14:53
  soon.
• 14:54
  Thank you very much for having me. It
• 14:56
  was a pleasure.