Latest / Becker’s Payer Issues Podcast / From Detection to Prevention: AI’s Role in Payment Integrity

Transcript

From Detection to Prevention: AI’s Role in Payment Integrity · Becker’s Payer Issues Podcast

• 0:00
  Hi, everyone. This is Lucas Voss with Becker's
• 0:02
  Healthcare. Thanks so much for tuning in to
• 0:03
  the Becker's Healthcare podcast series, and welcome to
• 0:06
  today's episode from detection to prevention,
• 0:09
  AI's role in payment
• 0:11
  integrity. And I'm very excited
• 0:14
  to welcome Steve Southern. He's the senior vice
• 0:16
  president information systems at Cirrus
• 0:18
  for our conversation today. Steve has close to
• 0:20
  thirty years
• 0:22
  of experience in IT and joined Cirrus in
• 0:24
  1996.
• 0:25
  He's been passionate
• 0:26
  and very involved in proactively building efficiencies and
• 0:29
  leveraging technology to keep health care organizations ahead
• 0:32
  of the curve. Steve, I did not call
• 0:34
  you old. Welcome to the podcast.
• 0:38
  Thank you. Thank you, Lucas. Thank you very
• 0:40
  much for the introduction. Yeah. Thirty years, time
• 0:43
  flies when you're having fun as they say.
• 0:44
  Right? Exactly. I'm so excited to have you
• 0:47
  just because of all of the experience that
• 0:49
  you bring to our conversation today. And and
• 0:51
  I wanna start off with
• 0:53
  putting things into perspective
• 0:55
  for our audience.
• 0:57
  As we're looking
• 0:59
  at our topic today, how do you define
• 1:02
  payment integrity today, and how is that definition
• 1:06
  evolved
• 1:07
  as AI has become more embedded in payer
• 1:09
  operations?
• 1:11
  Sure. That's a great question.
• 1:13
  In today's environment, I would say payment integrity
• 1:15
  really spans the entire claim life cycle. So
• 1:19
  end to end set of processes from claim
• 1:22
  submittal
• 1:23
  all the way to final payment for in
• 1:25
  a prepayment setting
• 1:26
  as well as the traditional
• 1:28
  post payment reviews.
• 1:31
  Along the way, there are many tools, analytics,
• 1:33
  technologies out there used to ensure the health
• 1:36
  care payments are getting paid or getting adjudicated
• 1:39
  paid accurately compliant
• 1:41
  and timely.
• 1:43
  AI is obviously at the forefront of all
• 1:45
  of that. Just about every organization's
• 1:48
  strategy and road map
• 1:50
  are intertwined, wrapped around that, if you will,
• 1:52
  these days. So I feel we're really just
• 1:54
  starting to scratch the surface with the impacts
• 1:56
  that AI can and will have down the
• 1:59
  road.
• 1:59
  We're seeing we're already seeing some major impacts
• 2:02
  with process improvements,
• 2:04
  automation, obviously.
• 2:06
  However, I do think that some organizations
• 2:08
  might be taking a slightly more conservative approach
• 2:11
  or slower pace, let's say, to
• 2:14
  to roll some of these things out just
• 2:15
  due to governance, regulatory,
• 2:18
  data use challenges, just making sure those those
• 2:21
  boxes are checked.
• 2:22
  And you've seen
• 2:24
  so many trends evolve over your time in
• 2:27
  health care IT, and there's certainly
• 2:29
  a lot of change that's happening right now,
• 2:32
  especially 2025 and then now obviously in 2026.
• 2:36
  From your work with these organizations across the
• 2:38
  country,
• 2:39
  how has the integration of AI and machine
• 2:42
  learning
• 2:43
  changed the strategic approach then to payment integrity
• 2:47
  for leaders? And is there anything that surprises
• 2:49
  you about that process?
• 2:51
  Yeah. Sure. I think the biggest change over
• 2:53
  the last year or so has been the
• 2:55
  shift from a traditional post payment or a
• 2:57
  pay and chase model
• 2:59
  to,
• 3:00
  prioritization
• 3:01
  of of prepayment solutions.
• 3:03
  So as I mentioned previously,
• 3:06
  audits that are happening during the payment cycle
• 3:08
  as opposed to after,
• 3:11
  AI has really helped fuel this due to
• 3:13
  the acceleration of development tools and the speed
• 3:15
  at which some of these solutions sophisticated solutions,
• 3:17
  if you will, can be deployed.
• 3:19
  I would say I'm most surprised by the
• 3:21
  rapid adoption rate that we've already experienced.
• 3:24
  Everybody is using these tools. They're everywhere. You're
• 3:26
  hearing about them
• 3:28
  constantly.
• 3:29
  Also,
• 3:30
  on the flip side of that really is
• 3:32
  the significant amount of resources and effort required
• 3:35
  for oversight and governance.
• 3:38
  This is an area that sometimes can lag
• 3:40
  behind the development cycle,
• 3:42
  but it really cannot be minimized or ignored.
• 3:45
  You know, the question today is not is
• 3:47
  no longer are you using AI, it's how
• 3:50
  are you using AI and what are the
• 3:52
  guardrails and framework that you have established around
• 3:55
  it.
• 3:56
  Yeah. And you mentioned the the fast adoption
• 3:58
  rates and everybody sort of jumping
• 4:01
  to or on that trend, so to speak,
• 4:03
  but that doesn't necessarily mean there is an
• 4:05
  impact right away for a lot of organizations.
• 4:08
  Where are you seeing the most meaningful impact
• 4:10
  for AI today? And is there an example
• 4:12
  that sticks out to you
• 4:14
  where AI really surfaced
• 4:16
  patterns that traditional rule based systems
• 4:19
  that we've touched on likely would have missed?
• 4:23
  Definitely.
• 4:24
  The most meaningful impact I'm seeing today is
• 4:26
  the ability to convert
• 4:28
  and process very large,
• 4:30
  complicated,
• 4:32
  unstructured
• 4:32
  documents and and even datasets into
• 4:35
  useful structured,
• 4:37
  easily machine readable data, which can be processed
• 4:40
  much more efficiently and accurately.
• 4:43
  So I guess
• 4:45
  an example I would say is so traditional
• 4:47
  rules based systems
• 4:49
  usually only detect specific patterns. Right? Programmatic,
• 4:53
  deterministic
• 4:54
  algorithms that, you know, humans have have developed
• 4:58
  as where on the flip side, where machine
• 5:00
  learning and large language models are able to
• 5:02
  analyze ambiguous notes, narrative patterns across these very
• 5:05
  large datasets that those traditional rules would have
• 5:08
  missed.
• 5:09
  A great example of this is medical records.
• 5:12
  So these documents are typically very large,
• 5:15
  many, many pages, sometimes hundreds, if not thousands
• 5:18
  of pages
• 5:19
  in these documents, very cumbersome to handle,
• 5:22
  challenging to process, and for, you know, humans
• 5:25
  to to review.
• 5:27
  So using a model to summarize that information,
• 5:30
  extract key data points can really assist clinical
• 5:32
  auditors
• 5:34
  by increasing their efficiency, accuracy, consistency
• 5:37
  to go through this information.
• 5:39
  Yeah. And we're seeing already seeing that impact
• 5:42
  on organizations,
• 5:43
  I think. I I wanted to come back
• 5:45
  to the governance piece that you've mentioned and
• 5:48
  sort of a couple of related topics in
• 5:50
  regards to governance because I think this is
• 5:51
  a very important topic.
• 5:53
  Again, you've talked about it. Automation can improve
• 5:55
  speed and consistency. It's very crucial. But there
• 5:58
  is concern around fairness, false positives, provider trust.
• 6:01
  Again, we're sort of getting into that governance
• 6:03
  territory here. How can organizations
• 6:06
  balance efficiency
• 6:07
  with that clinical accuracy and credibility at scale,
• 6:10
  really?
• 6:12
  Yeah. I I think finding this balance is
• 6:14
  really one of the most urgent challenges that
• 6:16
  we face today in the payment integrity space.
• 6:19
  So just as with the development tools and
• 6:21
  the technology,
• 6:22
  the appropriate resources and prioritization
• 6:25
  must be dedicated
• 6:26
  specifically to these areas
• 6:29
  these areas as in, you know, the governance,
• 6:31
  security,
• 6:32
  those things. So
• 6:34
  building a strong governance program,
• 6:36
  transparent
• 6:37
  controls,
• 6:38
  always keeping humans in the loop,
• 6:40
  and continuous
• 6:41
  modeling,
• 6:42
  not all not only model monitoring, but, you
• 6:44
  know, solution monitoring to make sure, you know,
• 6:47
  of model drift and the things that are
• 6:50
  of concern are are at the forefront.
• 6:53
  I would say those are some of the
• 6:55
  keys the the main keys, I would say,
• 6:57
  to ensure trust and credibility in this space.
• 7:00
  Now turning some of these topics that you've
• 7:03
  mentioned into strategic elements. Right?
• 7:07
  What should payer IT and operations leaders be
• 7:10
  doing right now
• 7:12
  to prepare for that next phase of AI
• 7:14
  driven payment integrity?
• 7:17
  Sure. I can share a couple of, at
• 7:18
  least, points that I think are important here.
• 7:21
  So
• 7:22
  IT leaders should be assessing and improving their
• 7:25
  teams,
• 7:26
  teams from a staffing standpoint,
• 7:29
  experience,
• 7:30
  technology,
• 7:31
  governance, and data infrastructure,
• 7:33
  as we just talked about a little bit.
• 7:35
  AI success really hinges on high quality, accessible,
• 7:39
  interoperable data. So data is key here.
• 7:42
  So having a clear, concise, consolidated data foundation,
• 7:45
  again, is the most important thing when it
• 7:48
  comes to AI.
• 7:50
  Preparing to support the shift to prepaid solutions
• 7:53
  by proactively considering some of the core systems,
• 7:56
  the foundational
• 7:58
  applications and and data and workflows
• 8:00
  that would support that.
• 8:02
  Considering
• 8:04
  integrated ecosystems
• 8:05
  and prioritizing
• 8:06
  API integrations with partners over, you know,
• 8:10
  more traditional fragmented point to point solutions.
• 8:14
  And finally, building equity and fairness testing into
• 8:17
  model governance,
• 8:19
  always maintaining a human in the loop,
• 8:21
  creating clear and transparent documentation along with defensible
• 8:25
  policies and procedures to back everything up.
• 8:28
  The future of AI driven payment integrity isn't
• 8:31
  really just about more automation. It's about trusted
• 8:34
  transparent
• 8:36
  human aligned intelligence that's embedded across the entire
• 8:39
  claim life cycle.
• 8:41
  Three decades of experience
• 8:42
  in fifteen minutes.
• 8:44
  Steve, thanks so much for for being here.
• 8:47
  Thanks so much for your time. It's so
• 8:48
  great to have you.
• 8:50
  You bet. Hey. It's it's truly exciting time
• 8:52
  to be working in this industry and at
• 8:53
  the intersection of payment integrity and technology.
• 8:57
  We're watching AI reshape not just health care,
• 9:00
  but nearly every industry,
• 9:02
  and the pace of innovation is just accelerating
• 9:04
  daily.
• 9:05
  So the work we're doing now will define
• 9:07
  the next generation of accuracy, fairness, efficiency in
• 9:10
  the health care space. And I'm just grateful
• 9:12
  to be a part of the transformation and
• 9:14
  working with the Saris team and our partners.
• 9:16
  I also appreciate the opportunity to collaborate with
• 9:19
  Becker's and share these conversations, so I appreciate
• 9:21
  you having me.
• 9:23
  Absolutely. It's great to have you. Thanks for
• 9:24
  your time. And we also want to thank
• 9:26
  our podcast sponsor, Cerus. You can tune in
• 9:28
  to more podcasts from Becker's Healthcare by visiting
• 9:31
  our podcast page at beckershospitalreview.com.