Latest / Becker’s Healthcare Digital Health + Health IT / Fathom & ApolloMD Partnership: Achieving 90%+ Coding Automation and Why It Matters

Transcript

Fathom & ApolloMD Partnership: Achieving 90%+ Coding Automation and Why It Matters · Becker’s Healthcare Digital Health + Health IT

• 0:00
  Hello, everyone. This is Jacob Emerson with the
• 0:02
  Becker's Healthcare Podcast. Thrilled today to be joined
• 0:05
  by 2 special guests. Andrew Lockhart is the
• 0:08
  CEO of Fathom, and doctor Yogan Patel is
• 0:12
  the CEO of Apollo MD. Doctor Patel, Andrew,
• 0:15
  thank you so much for taking the time
• 0:17
  to be with us on the podcast today.
• 0:19
  Thank you for having us.
• 0:20
  So, gentlemen, before we dive into everything we
• 0:23
  wanna talk with you about, can we hear
• 0:25
  a little bit more about yourselves? What do
• 0:26
  you do right now with your respective organizations?
• 0:29
  And what led you to this point in
• 0:31
  terms of health care career backgrounds? Doctor Patel?
• 0:35
  Thanks, Jacob. I serve as the chief executive
• 0:37
  officer for Apollo MD. I'm a practicing emergency
• 0:40
  physician.
• 0:41
  I've been in the space now,
• 0:44
  17 years with Apollo MD. So
• 0:47
  thrilled to be part of the organization and
• 0:49
  to serve.
• 0:51
  We have the distinct honor of serving over
• 0:53
  4,000,000 patients a year,
• 0:55
  mostly in emergency medicine. We also do some
• 0:57
  hospital medicine,
• 0:58
  anesthesia, and radiology.
• 1:00
  And we are, you know, a private practice
• 1:03
  model, private independent group. I've been in the
• 1:05
  space for over 40 years. How about you,
• 1:08
  Andrew?
• 1:09
  My name's Andrew Lockhart. I'm the cofounder and
• 1:11
  CEO of Fathom.
• 1:13
  My background in health care so I really
• 1:15
  got the bug,
• 1:17
  from Toronto originally
• 1:18
  and was working for a design firm there
• 1:20
  where are a lot of our
• 1:23
  our clients were US based health care organizations,
• 1:26
  Mayo Clinic, UHG,
• 1:28
  and was spending a lot of time,
• 1:30
  you know, exploring the US health care system.
• 1:32
  And
• 1:33
  the infinite complexity
• 1:35
  of the system was something that just was
• 1:38
  like a mind trap for me that got
• 1:39
  me very excited, interested in health care.
• 1:42
  Went on to do my MBA at Stanford,
• 1:44
  and then coming out of that, really had
• 1:46
  a distinct desire to do something in health
• 1:48
  care.
• 1:49
  My first swing there was in an on
• 1:51
  demand pharmacy,
• 1:53
  which, you know, I can give you lots
• 1:55
  of color on what's problem problematic about the
• 1:59
  way the pharmaceutical industry works. But, ultimately,
• 2:02
  you know, that journey eventually led us towards
• 2:05
  finding Fathom, a solution that we felt was
• 2:08
  much needed,
• 2:09
  timely,
• 2:10
  and, you know, ultimately, something that, you know,
• 2:13
  is very scalable. Sure. Absolutely. And, you know,
• 2:16
  for our listeners who might not be familiar,
• 2:17
  let's take a step back and talk about
• 2:19
  how the partnership between your two companies
• 2:21
  came about. How long you've been working together,
• 2:24
  and what were you trying to achieve when
• 2:25
  you when you first began working together?
• 2:28
  Yeah. I'll I'll start. Thanks, Jacob. I was
• 2:30
  actually just looking back at our emails. So
• 2:32
  we actually
• 2:33
  first,
• 2:35
  started talking to Fathom back in December of
• 2:37
  2019.
• 2:39
  So, we were,
• 2:41
  fortunately now an early adopter.
• 2:43
  And for us, you know, as a private
• 2:45
  group that's been in this space for a
• 2:47
  long time,
• 2:48
  we had always
• 2:49
  relied on kind of a human
• 2:51
  centered RCM
• 2:53
  model. So rep all of our revenue cycle
• 2:55
  required human coders.
• 2:58
  We actually do the entire revenue cycle in
• 3:00
  house. But the one component that we had
• 3:02
  historically outsourced was our coding.
• 3:04
  And what we found was over the years,
• 3:07
  that human workforce
• 3:08
  just didn't have the same resiliency.
• 3:11
  And the quality started to become a little
• 3:13
  bit more variable, especially as coding guidelines were
• 3:16
  changing.
• 3:17
  And so we were looking for a higher
• 3:18
  fidelity solution.
• 3:20
  And and ultimately, as we started to explore
• 3:22
  the idea of automation,
• 3:23
  we
• 3:24
  stumbled upon Fathom. And, you know, at the
• 3:27
  time, we probably looked at several
• 3:29
  different coding solutions,
• 3:32
  you know, computer assisted coding,
• 3:34
  you know, other other groups that were early
• 3:36
  in this space.
• 3:37
  But really we're looking for somebody that could
• 3:39
  deliver
• 3:40
  top quality and accuracy,
• 3:42
  sustainability,
• 3:44
  and and great value.
• 3:45
  And I think, even from the very first
• 3:48
  conversations, Fathom
• 3:49
  clearly kind of met those benchmarks.
• 3:52
  And then really as a organization that was
• 3:54
  naive
• 3:55
  to artificial intelligence and automation,
• 3:58
  they were great about kind of holding our
• 3:59
  hand as we kind of journey together. And
• 4:02
  it was,
• 4:03
  you know, it's been a,
• 4:04
  it's been a good collaboration. Sure. Absolutely. And
• 4:07
  I wanna get into what some of those
• 4:08
  outcomes are in just a second. But, Andrew,
• 4:10
  do you remember the initial goals you you
• 4:12
  set out to achieve with your partnership?
• 4:15
  Yeah. I mean, like, when we started Fathom,
• 4:18
  I think,
• 4:19
  we would have thought that
• 4:21
  50% automation for emergency department coding, which, you
• 4:25
  know, I'd put in the, like, moderate high
• 4:27
  complexity on the scale Mhmm. Was something that
• 4:30
  was maybe within reach.
• 4:32
  And then by the time we were able
• 4:33
  to connect with Apollo MD, we felt pretty
• 4:36
  confident we could get to 70.
• 4:38
  And
• 4:39
  the fact that we're able to jump to
• 4:41
  where we are today is really,
• 4:44
  you know, just a testament to the to
• 4:46
  the to the engineering prowess of the team,
• 4:48
  but also partners
• 4:49
  who were able to work with us and,
• 4:51
  you know, iron out edge cases. Sure. And
• 4:53
  you're at 90% today. Correct?
• 4:55
  Yeah. I think 91. Wow. So even even
• 4:58
  higher than that. I mean, can you discuss
• 5:00
  that journey of of improving automation rates to
• 5:02
  90 over 90%?
• 5:05
  And, also, you know, doctor Patel, why is
• 5:08
  such high automation? Why is that an important,
• 5:11
  achievement for the organization?
• 5:13
  Yeah. I mean, as I mentioned, you know,
• 5:15
  this has been a 4 year journey for
• 5:16
  us. And at the start of the journey,
• 5:18
  I think we would have been, you know,
• 5:20
  we would have been thrilled at 70%. So
• 5:22
  it really is a testament to how well
• 5:25
  the platform
• 5:26
  has learned and developed over the years.
• 5:29
  It also allowed us
• 5:31
  for the more complex cases
• 5:34
  to really focus the human coders
• 5:37
  on those cases.
• 5:38
  And so we're really deploying the right resource
• 5:42
  to the to the right work workflow.
• 5:45
  And so for us, that has been kind
• 5:47
  of a game changer. I mean, we have
• 5:48
  a what I'd like to think of as
• 5:50
  a kind of 6 Sigma process almost where
• 5:52
  the auditing process is best of breed now.
• 5:55
  And really the partnership with Fathom allowed for
• 5:58
  that. Mhmm. You know, in addition to kind
• 6:00
  of automation accuracy, I would say is even
• 6:02
  more important.
• 6:04
  And for us,
• 6:05
  it's been a pleasure to see that that
• 6:07
  has been a priority for Fathom.
• 6:09
  And every time we've had kind of a
• 6:11
  question,
• 6:12
  a query,
• 6:13
  they have been right there to kind of
• 6:14
  investigate,
• 6:15
  walk that path with us, and improve the
• 6:18
  technology. Yeah. Any other outcomes you'd you'd point
• 6:20
  out, Andrew? Well, maybe to answer your question
• 6:22
  about the journey to 90%.
• 6:24
  Sure.
• 6:25
  One
• 6:26
  extraordinarily challenging to get there with any organization.
• 6:29
  Yeah. But Appol MD is a physician group
• 6:31
  servicing, what, how many emergency departments? Like, over
• 6:33
  a 100? Yeah. You've got disparate coding guidelines,
• 6:37
  disparate EHR formats, disparate documentation
• 6:40
  practices
• 6:41
  that you need to cover go all the
• 6:43
  way across.
• 6:44
  And that's where, you know, really the benefit
• 6:47
  of a technology platform that's continuously snowballing data.
• 6:51
  So we're able to get going with with
• 6:53
  clients like ApolloMD. As we've got more and
• 6:55
  more clients, the AI became more and more
• 6:56
  aware. And I'd say at this point, you
• 6:58
  know, our system has seen north of 3
• 7:00
  quarters of a 1000000000 encounters. Wow.
• 7:03
  And that's enabled it to get this very
• 7:05
  nuanced understanding of the ins and outs of
• 7:07
  all these different data formats
• 7:09
  and let it work across that so effectively.
• 7:14
  And then, also, like, I I wanna give
• 7:15
  credit to the team. Like, we've been very
• 7:18
  fortunate that we've been able to pull in.
• 7:20
  We've got engineers on our team who have
• 7:21
  joined us from
• 7:23
  AI teams at meta, AI teams at, like,
• 7:25
  teams at Neuralink, literally building brain computer interfaces,
• 7:29
  Amazon,
• 7:31
  Microsoft. And
• 7:32
  these folks, you know, have been really attracted
• 7:35
  to the idea. Interestingly, like, basically, probably half
• 7:37
  our engineering team is pre med.
• 7:41
  And
• 7:42
  so they're interesting in this idea that, one,
• 7:44
  enormous amount of data, 2, something that has
• 7:46
  a really scalable impact,
• 7:48
  and then 3, like, a really hard problem.
• 7:51
  Like,
• 7:51
  you know, our CTO has a lot of,
• 7:53
  like, vision and really pushes the team to
• 7:56
  continually work with the latest, greatest. Sure. Sure.
• 7:58
  Absolutely.
• 7:59
  So, doctor Patel, you're in over a 100
• 8:01
  emergency
• 8:02
  departments.
• 8:03
  What would you say have been the direct
• 8:05
  and indirect impacts of autonomous coding on your
• 8:07
  physician's
• 8:09
  day to day experiences
• 8:11
  and work life satisfaction?
• 8:13
  Yeah. We are unique in the industry, I
• 8:16
  think, in that we actually
• 8:18
  compensate
• 8:19
  our physicians,
• 8:21
  on a almost a purely performance driven model.
• 8:24
  And what that requires is our ability to
• 8:26
  understand their productivity in a given month, and
• 8:29
  within 2 weeks, translate that into payroll.
• 8:33
  And so what we're talking about is their,
• 8:35
  you know, RVU
• 8:37
  productivity,
• 8:38
  their collaboration
• 8:39
  with,
• 8:41
  advanced practice clinicians, so PAs and nurse practitioners.
• 8:44
  And at some practices, their collaboration with resident
• 8:47
  physicians.
• 8:48
  And so it's a fairly complex box. Right?
• 8:51
  And it does require this kind of precision
• 8:54
  and execution
• 8:55
  on a very tight timeline.
• 8:57
  And with human coders that had always historically
• 8:59
  been a challenge,
• 9:00
  we have some unique workflows that that complicated
• 9:03
  that further.
• 9:04
  And what we found is as we partnered
• 9:06
  with Fathom, you know, it took a lag
• 9:08
  of several days,
• 9:11
  that human coders required down till, like, less
• 9:13
  than 2 hours. And so really, it just
• 9:16
  closed the feedback loop to our team, you
• 9:19
  know, dramatically.
• 9:20
  And that allowed us to, you know, improve
• 9:22
  our chart delinquency rates. That allowed us to
• 9:25
  improve the accuracy and capture of the procedures
• 9:26
  and
• 9:29
  and documentation,
• 9:30
  and really ultimately translated, I think, into a
• 9:33
  more
• 9:34
  accurate compensation model for our clinicians. So
• 9:38
  it has been, you know, I think we're
• 9:39
  we're thrilled with where we are today. The
• 9:41
  system continues to improve.
• 9:43
  Andrew, other things that kinda stood out to
• 9:45
  you in that journey?
• 9:47
  No. I mean, I like, I think the
• 9:49
  one thing that gets maybe understated
• 9:52
  with AI solutions is everyone's like, oh, this
• 9:54
  is, like, a thing that's getting rid of
• 9:56
  people.
• 9:58
  But,
• 9:59
  I I think, really, it just empowers people.
• 10:01
  Sure. You know, the coders
• 10:03
  on, you know, both our teams and
• 10:06
  Paul and Dee's team, they're able to rather
• 10:08
  than focus on things that are pretty rote
• 10:10
  and monotonous
• 10:11
  and
• 10:12
  really dive into the details and, like, build
• 10:15
  better systems rather than just, like, basically plugging
• 10:18
  holes in the dam. Yeah. And I think
• 10:20
  that that is something that is, you know,
• 10:22
  really exciting. And, you know, ultimately, the promise
• 10:24
  of this is, like, you
• 10:26
  with these tools, you can basically
• 10:29
  maximize your executive function rather than just, like,
• 10:32
  getting tasks done. Yeah. And to hit home
• 10:34
  on that where you use empowerment of these
• 10:37
  physicians, are you finding that that really fits
• 10:39
  into how Apollo MD prefers to operate?
• 10:43
  Absolutely.
• 10:44
  You know, we are we're kind of a
• 10:45
  clinician cooperative. So we have,
• 10:48
  an ownership model that allows our physicians, our
• 10:50
  PAs, our nurse practitioners, our company employees to
• 10:53
  all own in the organization.
• 10:55
  So that need for empowerment
• 10:57
  and autonomy of work,
• 11:00
  is kind of ingrained in us. Sure. And
• 11:02
  frankly, the ability to have better transparency
• 11:05
  and reliability on the coding
• 11:08
  really kind of facilitates that. It also, as
• 11:11
  Andrew pointed out, rather than displace people, we
• 11:14
  were a lot we were able to focus
• 11:16
  our coding folks into an audit program that
• 11:19
  really has taken,
• 11:20
  I think, the fidelity of the entire system
• 11:22
  to another level. So, you know, and when
• 11:24
  we onboard new clinicians,
• 11:26
  we're able to focus those auditors on their
• 11:29
  work and give them early feedback, which, you
• 11:31
  know, is kind of a it's it's a
• 11:33
  lifelong journey for them. Right? So,
• 11:37
  yeah, very excited about the future with Fathom.
• 11:40
  Sure. So, Andrew,
• 11:41
  let's talk about another
• 11:44
  administrative favorite for for physicians,
• 11:47
  and I use favorite lately. The issue with
• 11:49
  payers administrative processes,
• 11:51
  processes, prior authorization,
• 11:53
  denials,
• 11:55
  submitting claims,
• 11:57
  all considering all of that, how does Fathom
• 11:59
  help Apollo MD navigate what we know are
• 12:02
  increasing challenges with payer denials and audits right
• 12:05
  now? So this touches nicely on your earlier
• 12:08
  question around why automation rates matter. Mhmm. The
• 12:11
  biggest benefit to AI is your ability to
• 12:14
  programmatically
• 12:15
  perform certain tasks. And a lot of the
• 12:16
  denials, they come about because,
• 12:19
  you know, what a lot of payers work
• 12:21
  on is they have these sort of disparate
• 12:22
  guidelines that, you know, this will never reimburse
• 12:25
  this procedure with this diagnosis code, etcetera. And
• 12:28
  these can be insanely hard to track and
• 12:30
  enforce against a large, you know, manual
• 12:33
  pool of workers, but, like, basically, like, the
• 12:36
  Paul and d MD team comes to us
• 12:38
  and says, like, hey. For this payer, we
• 12:39
  can't submit something that looks like this anymore.
• 12:42
  You know, that's a that's a code change.
• 12:44
  And then it's enforced across
• 12:46
  not 50% of the volume, 90% of the
• 12:49
  volume. And so you're inherently
• 12:51
  your your coding related denials go down.
• 12:54
  And, also, you know, the other benefit is
• 12:57
  with that high automation rate, your compliance processes
• 13:00
  go up because you've you've created these very
• 13:02
  robust coding guidelines and documentation on how you're
• 13:05
  gonna approach things, and it's enforced consistently. So
• 13:08
  that if you are in an audit and
• 13:09
  they're like, hey. Why did something happen this
• 13:11
  way? There's not a ton of inconsistencies in
• 13:13
  the system. There basically is like, hey. This
• 13:15
  is very consistent.
• 13:16
  The same way we wrote this down, the
• 13:18
  same thing that we agreed to, and this
• 13:19
  is our logic.
• 13:21
  Yeah. No. That makes sense. And, doctor Patel,
• 13:23
  what's your experience? And what are you hearing
• 13:24
  from your employees?
• 13:26
  Yeah. I mean, I for us, you know,
• 13:28
  it has taken a lot of the variability
• 13:30
  of documentation
• 13:31
  out of the work. And, you know, an
• 13:33
  earlier point that Andrew made is I'd say
• 13:34
  half of our practices are epic sites.
• 13:37
  The other half are are spread across a
• 13:40
  smattering of EHR. So, you know, Cerner, Meditech,
• 13:43
  Med Host,
• 13:44
  and and each has its own documentation
• 13:47
  kind of idiosyncrasies.
• 13:49
  So the ability for us to message to
• 13:51
  our teams to use
• 13:53
  standard language
• 13:54
  in the way they talk about
• 13:57
  certain procedures.
• 13:58
  FATHOM, you know, picks that up with such
• 14:00
  a high degree of accuracy
• 14:02
  that it has minimized, you know, lost procedures,
• 14:06
  missing,
• 14:07
  you know, electrocardiograms
• 14:08
  or CPTs.
• 14:09
  So for our clinicians, it's, you know, I
• 14:12
  I think net net has been a huge
• 14:14
  validation of their work
• 14:16
  because they're doing the work at the bedside.
• 14:19
  But, you know, documentation
• 14:21
  documentation challenges,
• 14:23
  sometimes leave them without getting full credit for
• 14:25
  the work they do. Sure. So fair to
• 14:27
  say you've seen payer denials decrease then? Yes.
• 14:31
  And and, you know, some of that is
• 14:32
  we've we've layered on other technologies as well,
• 14:34
  but Fathom has been a big part of
• 14:36
  that for us. Absolutely.
• 14:37
  It's
• 14:38
  attribution can be difficult because pair denials, generically,
• 14:41
  are increasing.
• 14:42
  And so to, you know, know
• 14:45
  exactly what's what's what and what's because Fathom's
• 14:48
  doing a better job versus, you know, what
• 14:50
  is some unnamed plan tightening screws? Hard to
• 14:53
  know. Yeah. That is a great point. So,
• 14:55
  Andrew, for for providers considering autonomous coding that
• 14:59
  really haven't dipped their toes into this space
• 15:01
  yet,
• 15:02
  advice that you'd give for all the leaders
• 15:05
  listening in in terms of selecting the right
• 15:07
  partner,
• 15:07
  managing the implementation,
• 15:09
  what what goes into this?
• 15:11
  I mean, I'd say the biggest thing in
• 15:13
  selecting a partner
• 15:15
  is
• 15:17
  make the contract reflect the promises made.
• 15:20
  You know, the industry has evolved a lot.
• 15:22
  It's much more mature. The technology is much
• 15:25
  more mature. Sure. You know, serve vendors like
• 15:28
  Fathom know what we can deliver.
• 15:30
  And, yeah, if we say x, like, put
• 15:32
  in the contract. We get punitive if we
• 15:34
  didn't. Like, when we got started with Apollo
• 15:36
  MD, it was like, hey. Like,
• 15:38
  this is new. Like, we think this, but
• 15:39
  the range was pretty wide.
• 15:41
  But, you know, we're now in a place
• 15:43
  where, you know, you should hold your vendors
• 15:46
  to SLAs on accuracy, SLAs on
• 15:49
  automation. Because automation, it, you know, drives all
• 15:51
  these compliance benefits. It drives your ROI. And
• 15:54
  if you're not getting that, then, you know,
• 15:56
  there should be a penalty. The other thing
• 15:58
  is make them demonstrate their wares. Like, see
• 16:01
  the AI work across a large volume,
• 16:04
  before you commit to anything
• 16:06
  or make it contingent on seeing that because,
• 16:09
  you know, the nice thing is is, like,
• 16:11
  most of these systems can run at relatively
• 16:12
  low marginal cost. And so,
• 16:15
  you know, push on that. Yeah. Absolutely. And
• 16:18
  doctor Patel, let me reframe the question for
• 16:19
  you. What do you wish you had known
• 16:21
  before
• 16:22
  you had entered,
• 16:23
  looking for a partner within autonomous coding?
• 16:27
  Yeah. I mean, 5 years ago for us,
• 16:30
  we were
• 16:31
  relatively novice with the whole idea of automation
• 16:34
  and artificial intelligence.
• 16:35
  So I think there's been a huge leap
• 16:38
  in just familiarity
• 16:40
  and comfort with with this type of engagement.
• 16:44
  What we did do well, I think, was
• 16:46
  that quality was our North Star.
• 16:48
  And we kind of had this relentless pursuit
• 16:50
  of accuracy
• 16:52
  and sustainability in this process.
• 16:54
  And I think we continually return to that
• 16:57
  to say as we're measuring milestones,
• 17:00
  are we are we more accurate? Are we
• 17:03
  are we higher quality than we were? And
• 17:05
  I think that has informed
• 17:07
  much of the journey for us.
• 17:09
  You know, I we spent we spent our
• 17:11
  time on this. Right? And and again, as
• 17:13
  Andrew pointed out in the earlier days of
• 17:15
  this, we were very methodical about a phase
• 17:19
  rollout across our sites,
• 17:21
  about
• 17:22
  validation,
• 17:22
  validation of the model, parallel coding, you know,
• 17:25
  third party audits,
• 17:27
  and then really developing an auditing system
• 17:31
  that could be continued going forward.
• 17:33
  I think they have that playbook now. And
• 17:36
  and I think for their newer clients,
• 17:38
  can expedite a lot of those things, including
• 17:40
  the onboarding of new clients. So we're proud
• 17:42
  to have been part of that journey. That's
• 17:44
  great advice. What else are we missing? Anything
• 17:46
  we didn't get to today that our listeners
• 17:48
  should know?
• 17:49
  I would say if you're a a a
• 17:51
  health system out there that still is relying
• 17:54
  on
• 17:55
  humans
• 17:56
  or in particular, physicians to do their own
• 17:58
  coding,
• 17:59
  you you need to run, not walk to
• 18:02
  a
• 18:03
  more sustainable,
• 18:05
  more compliant solution.
• 18:06
  And I think the promise of automating revenue
• 18:10
  cycle is already here.
• 18:12
  Yeah. I dovetail on that. I think
• 18:14
  one of the things that has me most
• 18:16
  excited about Fathom is the recent movement in
• 18:18
  clients that
• 18:20
  are in
• 18:21
  practices health systems where clinician coding is being
• 18:24
  performed. And what we're seeing in terms of
• 18:26
  impact of clinician experience and bandwidth
• 18:29
  and ability to better deliver care because an
• 18:33
  AI is sitting there performing the coding on
• 18:35
  their behalf. Mhmm. And further, what we are
• 18:37
  discovering is, generally speaking,
• 18:40
  most health care organizations are undercoding.
• 18:43
  And so not only do you get a
• 18:45
  step up in compliance, but you used to
• 18:46
  get a step up in acuity
• 18:48
  and ensure that you're being fully reimbursed or
• 18:50
  accurately reimbursed for all the work that your
• 18:53
  team is performing.
• 18:55
  Absolutely. Well, I think that's a great place
• 18:56
  to leave it. Gentlemen, thank you so much
• 18:59
  for taking the time to be with us
• 19:00
  on the podcast,
• 19:01
  for sharing your insights with our listeners. We
• 19:03
  really appreciate it. Thank you. Thank you. If
• 19:06
  you'd like to listen
• 19:07
  more podcasts from Becker's HealthCare, you can visit
• 19:10
  our podcast page at beckershospitalreview.com.
• 19:13
  I'd also like to thank Fathom as well
• 19:15
  for sponsoring this episode.