Latest / Becker’s Healthcare Podcast / Unlocking Specialty Drug Access: Pharmacy Strategy Meets AI Automation

Transcript

Unlocking Specialty Drug Access: Pharmacy Strategy Meets AI Automation · Becker’s Healthcare Podcast

• 0:00
  Hello, everyone. This is Erica Spicer Mason with
• 0:02
  Becker's Healthcare. Thank you so much for tuning
• 0:04
  in to the Becker's Healthcare podcast series today.
• 0:08
  So today, I'm joined by Will Yin, the
• 0:10
  CEO of Mandolin.
• 0:11
  And Will is going to talk to us
• 0:13
  about unlocking specialty drug access and where pharmacy
• 0:16
  strategy is intersecting with AI automation.
• 0:20
  Will, welcome to the podcast. I'm excited to
• 0:22
  have this conversation today.
• 0:24
  Thanks, Erica.
• 0:25
  Happy to be here. I also as an
• 0:27
  avid reader of Becker's newsletters, it's awesome to
• 0:29
  be on the podcast.
• 0:31
  Oh, well, we're so happy to have you,
• 0:32
  and thanks for being a reader as well.
• 0:35
  But to get our conversation started, I thought
• 0:37
  we could start a little bit high level
• 0:39
  here. So can you share us share with
• 0:41
  us a little bit about Mandolin's origin and
• 0:44
  your work in health care and particularly how
• 0:46
  you came to focus on specialty drug workflows
• 0:48
  and automation?
• 0:50
  Yeah. Sure.
• 0:52
  So
• 0:53
  a family history of Alzheimer's was my initial
• 0:56
  introduction to health care. And because of that,
• 0:58
  I spent my entire early life thinking that
• 1:00
  I was gonna go into academia,
• 1:02
  doing research around these incredibly exciting and life
• 1:05
  changing therapies for neurologic disease, for autoimmune conditions,
• 1:08
  for cancer.
• 1:10
  Very quickly, though, what I realized and ultimately
• 1:13
  what was incredibly disheartening was that the bottleneck
• 1:15
  to real patient impact
• 1:17
  is not in more science.
• 1:19
  There's a wealth of incredibly accomplished and incredibly
• 1:21
  smart individuals doing great groundbreaking science on important
• 1:25
  diseases in academia.
• 1:27
  The bottleneck is in a broken American health
• 1:29
  care system.
• 1:30
  And so that's what led me to swing
• 1:32
  the other way. Drop out of Stanford undergrad
• 1:34
  and start mandolin.
• 1:36
  The fact is that so much of what
• 1:38
  slows down patient access are these repetitive, mundane,
• 1:41
  administrative tasks
• 1:43
  that lead to errors and mistakes,
• 1:45
  lead to building up of bad debt for
• 1:47
  systems and health care providers, and ultimately lead
• 1:50
  to patients
• 1:51
  not receiving therapies that could change and save
• 1:53
  their lives.
• 1:55
  At the same time, I was seeing this
• 1:56
  incredibly exciting
• 1:58
  convergence of large language models and agents,
• 2:01
  and that was exactly the premise on which
• 2:03
  we started Mandolin. I mean, we only started
• 2:04
  the company about a year and a half
• 2:06
  ago, but in that time, we've grown to
• 2:07
  partner with some of the largest health care
• 2:09
  organizations
• 2:10
  in The US, totaling $10,000,000,000
• 2:13
  a year of annual drug spend to close
• 2:16
  the gap between patients and groundbreaking science that
• 2:19
  could change and save their lives.
• 2:22
  The last thing that I'll say here, by
• 2:23
  the way, is that most industries that benefit
• 2:25
  from automation
• 2:26
  are not as complex as the infrastructure to
• 2:29
  deliver specialty drugs specifically.
• 2:32
  That's because most industries market participants, they want
• 2:34
  the system to work more efficiently. Right?
• 2:36
  They're just limited by the challenges of clunky
• 2:40
  systems of record or maybe data fragmentation. But
• 2:43
  in the specialty drug market specifically,
• 2:46
  you've got a much more broken process,
• 2:48
  which means far more opportunity, frankly, because that's
• 2:51
  one of the most
• 2:53
  or rather one of the most important market
• 2:54
  participants, in other words, payers, they actually benefit
• 2:57
  from the complexity.
• 2:59
  For them, the complexity and confusion in the
• 3:02
  system is a feature, not a bug.
• 3:04
  So the system itself has much more upside
• 3:07
  around agentic automation and ultimately empowering health care
• 3:10
  providers
• 3:11
  and their patients
• 3:12
  than many other industries.
• 3:15
  Well, it's great to learn a little bit
• 3:16
  more about your background, what led you to
• 3:18
  where you are today. And you're in a
• 3:21
  very exciting space right now, the specialty,
• 3:24
  pharmacy space. And you've you've already touched on
• 3:27
  really how we're seeing not only the spend
• 3:29
  of this industry grow, but complexity to your
• 3:32
  point about the payer piece, complexity is continuing
• 3:35
  to accelerate as well.
• 3:37
  And you also touched on some of those
• 3:39
  administrative and operational
• 3:41
  hurdles. So I'd love to go there next
• 3:43
  and talk a little bit more about some
• 3:44
  of those barriers.
• 3:46
  So what are you seeing as the biggest
• 3:48
  admin or ops bottlenecks in hospitals and health
• 3:51
  system pharmacies right now, especially when it comes
• 3:53
  to those key areas like access, benefits,
• 3:57
  prior auths, and revenue cycle?
• 4:00
  Sure.
• 4:01
  Well, in a system as broken and inefficient
• 4:04
  as the specialty medication space, there's
• 4:07
  actually several bottlenecks.
• 4:08
  So first,
• 4:10
  staffing is one huge bottleneck.
• 4:12
  The industry is growing incredibly quickly, and as
• 4:15
  the cost of screwing up the paperwork grows,
• 4:17
  the risk of mistakes grows.
• 4:19
  It's just enormously challenging to hire, train, and
• 4:22
  retain diligent workers.
• 4:24
  Of course, changes to side of care requirements,
• 4:26
  that's another big bottleneck,
• 4:28
  with site neutral payment reforms and so on
• 4:30
  and so forth.
• 4:32
  And documentation is another massive bottleneck. I mean,
• 4:34
  the way that that the system continues to
• 4:36
  evolve has led to a wide range of
• 4:38
  expertise at the customers that we work with,
• 4:41
  where there's little consistency across the entire organization
• 4:44
  on the right process for them to follow
• 4:46
  in processing paperwork.
• 4:47
  It often takes us months of effort,
• 4:50
  for us or or our customers to excavate
• 4:52
  the process that back offices are using.
• 4:55
  And it's always a deep surprise when leadership
• 4:57
  sees the actual steps that are being taken.
• 4:59
  The last big bottleneck that I'll mention is
• 5:02
  the proliferation and profusion of technologies, portals, and
• 5:06
  SOPs.
• 5:07
  On the technology side, you've got faxes to
• 5:10
  ePortals, different EHRs, specialty specific EHRs, separate logins
• 5:14
  for PA portals
• 5:16
  or bridge program enrollments,
• 5:18
  RCM data,
• 5:19
  living in alternate systems, then the EHR itself.
• 5:23
  And then you've got rapidly evolving policies and
• 5:25
  protocols for every payer.
• 5:28
  In short, you've got this massive body of
• 5:30
  expertise that's impossible for anyone to master
• 5:33
  without having worked in the industry for decades.
• 5:37
  All of these bottlenecks lead to real pains
• 5:39
  and challenges, not just for the hospital system,
• 5:42
  but also all of its employees and ultimately
• 5:44
  for their patients,
• 5:45
  which I I hope we'll get a chance
• 5:47
  to talk about more.
• 5:48
  Yeah. Absolutely.
• 5:50
  And I appreciate you outlining so clearly some
• 5:52
  of those bottlenecks,
• 5:53
  staffing documentation,
• 5:55
  just the sheer proliferation of technology and platforms.
• 5:59
  Sounds like a really challenging environment, and I
• 6:01
  understand
• 6:01
  to address some of that, Mandolin is using,
• 6:04
  AI agents to help automate tasks like intake,
• 6:07
  benefit checks, prior auths, claim statusing.
• 6:11
  So in your view, what are the strategic
• 6:13
  implications for health system pharmacies of adopting this
• 6:16
  kind of automation? Would love to know more
• 6:18
  about that.
• 6:20
  Great question.
• 6:22
  I think that we can start to think
• 6:24
  about implications for AI based changes in perhaps
• 6:28
  two different time scales.
• 6:30
  The first is in the near term where
• 6:32
  the market is heterogeneously
• 6:33
  adopting AI and the comparison
• 6:36
  is often between an AI powered workforce
• 6:39
  performing against how work has historically been done.
• 6:42
  The second is gonna be once AI is
• 6:45
  the norm.
• 6:46
  So,
• 6:46
  you know, I live in San Francisco. Most
• 6:48
  people here take self driving cars as our
• 6:51
  taxis, basically.
• 6:53
  That's only happened over the past few years.
• 6:55
  On the one hand, you have the near
• 6:57
  term change where,
• 6:59
  we're in these cars comparing the experience to,
• 7:01
  say, an Uber ride with a human driver.
• 7:04
  But then there's, of course, also the deeper
• 7:06
  implications
• 7:07
  of long term what the world can look
• 7:10
  like when all the vehicles are self driving
• 7:13
  and you don't need things like, say, traffic
• 7:15
  lights anymore?
• 7:17
  So I'll talk about the near term first.
• 7:18
  So in the near term, the biggest implications
• 7:20
  of AI are on those simple, repetitive, mundane
• 7:24
  processes. So these are, for example, data entry,
• 7:27
  processing tasks,
• 7:29
  and, you know, the the the benefit of
• 7:31
  AI in in in those regimes is that
• 7:33
  we end up seeing significantly lower denial rates
• 7:35
  and faster turnaround times.
• 7:38
  Most of the bad debt increases that we've
• 7:39
  seen at health systems are driven by denials
• 7:42
  caused by preventable errors with claims.
• 7:45
  As an example, with with high cost specialty
• 7:47
  drugs, we've seen single errors costing hundreds of
• 7:50
  thousands of dollars on a denial.
• 7:53
  And keep in mind that the more expensive
• 7:54
  the drug is,
• 7:55
  the higher the scrutiny from the payer side.
• 7:58
  These are clinical requirements, side of care requirements,
• 8:00
  these requirements change all the time.
• 8:02
  As you extrapolate this across a system and
• 8:05
  factor in the explosive increase in biologics,
• 8:09
  you're looking at tens of billions of dollars
• 8:11
  in losses attributable to paperwork errors.
• 8:14
  And, of course,
• 8:15
  there's ultimately massive impact on patient care, which
• 8:18
  is the most important. Studies have shown that
• 8:21
  patients benefit significantly from having continuity of care
• 8:24
  when they stay within the four walls of
• 8:25
  an IDN.
• 8:27
  Longer term,
• 8:28
  staffing remains a big challenge. Patient volumes are
• 8:31
  only gonna continue to scale.
• 8:33
  What does what does that mean when, to
• 8:35
  grow capacity quickly without needing to spend time
• 8:37
  and resources hiring pharmacy techs?
• 8:39
  That's what AI enables us to do to
• 8:41
  scale operations like software.
• 8:44
  You just did a fantastic job outlining,
• 8:48
  really,
• 8:49
  three key strategic implications
• 8:51
  for health system pharmacies. I mean, you mentioned
• 8:53
  the ability to address some of those simple,
• 8:55
  repetitive, mundane tasks that staff have on their
• 8:58
  plates. Sounds like there are really huge financial
• 9:00
  implications
• 9:01
  of automating some of these features. And then,
• 9:03
  of course, that all trickles down into the
• 9:05
  patient care experience
• 9:07
  as well as staffing ability.
• 9:10
  So,
• 9:11
  again, appreciate you outlining all of that. And
• 9:13
  from your experience working with health systems, I'm
• 9:15
  sure you've seen them come across some adoption
• 9:19
  or change management hurdles as they're embracing automation
• 9:22
  in these ways despite knowing all of these
• 9:24
  value
• 9:25
  points.
• 9:26
  So I'd love to know how
• 9:28
  you've seen
• 9:30
  organizations get through those change management hurdles and
• 9:32
  how pharmacy leaders should really prepare to support
• 9:34
  their teams during that process.
• 9:39
  Yeah. Good question. So
• 9:42
  at core,
• 9:44
  I think that there's been
• 9:46
  frankly pretty shameful history of tech companies swaggering
• 9:50
  into the healthcare space and making outrageous claims.
• 9:53
  And also assuming that health systems have teams
• 9:56
  of eager world class software engineers with nothing
• 9:59
  else to do, sitting around waiting to implement,
• 10:03
  a software company's product.
• 10:04
  And that's just been a failure mode that
• 10:07
  has caused enormous amounts of pain and scar
• 10:09
  tissue that we, as a young tech business,
• 10:13
  have had to traverse.
• 10:15
  Secondly, of course, health systems just tend to
• 10:17
  move far more slowly. There's complex orchestration
• 10:21
  challenges of having lots of stakeholders trying to
• 10:23
  make a big decision.
• 10:25
  And then of course there are also things
• 10:26
  like reverification season and budgeting season. So,
• 10:30
  they end up moving more slowly relative to
• 10:32
  some of the private entities that we worked,
• 10:35
  with on the AIC home infusion and independent
• 10:37
  specialty pharmacy side.
• 10:40
  In terms of preparation,
• 10:42
  something we've noticed with our clients,
• 10:45
  is that the teams that get the most
• 10:47
  value the fastest
• 10:49
  are those with well documented SOPs for their
• 10:51
  patient access workflows.
• 10:53
  If we sit down with your team and
• 10:54
  hear, say, 10 different answers to the same
• 10:56
  question about how your team handles a specific
• 10:58
  edge case, that's something we need to spend
• 11:01
  time ironing out upfront
• 11:03
  that leads to potentially longer time to value.
• 11:07
  Fantastic. Well, thank you.
• 11:09
  And I know that we are winding down
• 11:11
  on our time together in this conversation. But
• 11:14
  since you are you know, you mentioned you're
• 11:16
  in San Francisco, you're in the Silicon Valley
• 11:18
  area, you're at the forefront of innovation and
• 11:21
  ongoing,
• 11:22
  advancements with AI automation, especially in the specialty
• 11:25
  pharmacy space. So as you look ahead, what
• 11:28
  do you see as the next frontier for
• 11:30
  specialty infusion or pharmacy operations
• 11:33
  when it comes to these AI driven workflows,
• 11:35
  analytics, and value?
• 11:36
  And what do you think leaders should be
• 11:38
  thinking about right now?
• 11:42
  I I love this question.
• 11:44
  So
• 11:45
  when we look at where health systems want
• 11:47
  solutions,
• 11:48
  they always gravitate immediately towards automating prior auths
• 11:51
  on the pharmacy side or, say, reducing PA
• 11:53
  denials.
• 11:54
  And as much as these things feel like
• 11:57
  discrete problems, they're really a symptom of confusions,
• 12:01
  errors, ambiguities, and inconsistent SOPs
• 12:04
  upstream all over the organization.
• 12:08
  I think that as intake and other processes
• 12:10
  get documented,
• 12:11
  codified, and converted towards repeatable
• 12:14
  automated agentic processes, the opportunities for solving these
• 12:17
  hugely expensive and complex problems
• 12:19
  become much clearer
• 12:21
  and unlocks much more significant opportunities for value
• 12:24
  creation.
• 12:25
  One massive area of opportunity that we've seen
• 12:28
  is is in bad debt reduction.
• 12:30
  We all know hospitals are in dire financial
• 12:33
  situations.
• 12:34
  I
• 12:35
  believe Kodiak reported earlier this year that found,
• 12:38
  most hospitals are writing off something like 3%
• 12:41
  of net revenue in denials. That's nearly the
• 12:43
  same amount of margin they have, so it's
• 12:46
  ultimately just canceling out.
• 12:49
  As an as another example, we we spoke
• 12:51
  with a large IDN in in the Northwest
• 12:53
  that had massive write offs for
• 12:56
  just lack of prior auth or identifying that
• 12:58
  a prior auth was necessary.
• 13:00
  Could be upwards of 750
• 13:01
  k for a single therapy.
• 13:03
  And these are due to simple errors, ultimately,
• 13:07
  an incorrect j code,
• 13:09
  a lack of medical necessity.
• 13:11
  These numbers add up to tens of millions
• 13:14
  of dollars in preventable bad debt write offs.
• 13:17
  By operating as a central command on every
• 13:20
  script that runs through a health system, we
• 13:22
  can make sure that we're not only avoiding
• 13:24
  those missed revenue opportunities, but we also
• 13:27
  optimize for, optimal site of care,
• 13:30
  wherever possible, both for the system as well
• 13:33
  as ultimately for the patient.
• 13:36
  Fantastic. Will, it's been so great learning from
• 13:39
  you today where you see
• 13:41
  AI automation technology in the specialty pharmacy space
• 13:45
  headed, where it's at right now, some hurdles
• 13:47
  that you're seeing organizations
• 13:49
  tackle when it comes to change management and
• 13:51
  adopting this tech.
• 13:53
  So, again, I I just wanna thank you
• 13:54
  for being part of the podcast today and
• 13:56
  for sharing your insights with us.
• 13:59
  Of course. Thanks. Thanks for having me. I
• 14:01
  think the last thing that I'll say here
• 14:03
  is, you know, it's an incredibly exciting time
• 14:06
  for,
• 14:07
  the industry. Right? There's these groundbreaking
• 14:10
  life saving medications coming to market that, again,
• 14:13
  have the potential to change lives and save
• 14:15
  lives.
• 14:16
  But at the same time,
• 14:18
  that means an ever
• 14:21
  increasing pressure to ensure that all systems have
• 14:25
  a
• 14:26
  high degree of precision
• 14:28
  around navigating ever evolving payer requirements, especially as
• 14:32
  we look ahead to 2026.
• 14:34
  Yep.
• 14:35
  A really important note to leave our listeners
• 14:38
  with. So thank you, Will. And I know
• 14:40
  this isn't our last conversation together. Wanted to
• 14:43
  flag for our listeners that Will will be
• 14:45
  joining us,
• 14:46
  for a second episode of part two, where
• 14:48
  we'll talk about driving revenue and access and
• 14:50
  specialty pharmacy.
• 14:52
  We'll take a deeper dive there and,
• 14:54
  really looking forward to that, Will. And of
• 14:56
  course, we'd like to also give a big
• 14:58
  thank you to Mandolin for sponsoring today's episode.
• 15:01
  So listeners, be sure to have a great
• 15:03
  day and join us again on the Becker's
• 15:04
  Healthcare Podcast Series soon.