Tip:
Highlight text to annotate it
X
Thank you, Wendy, and thanks to the sponsors for inviting me to give a talk today. And
as was said before, I think, because CPTU was already addressed and reimbursement has
already been talked about, I think I have nothing more to say about my talk. So I can
just probably gloss over this whole thing. First point of disclosure, while I am at CMS
often, probably once a month talking about new technologies and getting the federal government
to on-board these things - I'm not an employee of CMS or Medicare. I don't represent Health
and Human Services. So if you have any complaints about Obamacare or the healthcare exchanges,
you know, don't direct them my way - not my area. We are in very tumultuous times,
though. There's no question - not only at a macro, federal healthcare reform level;
the healthcare exchanges. But specifically with Next Gen sequencing and advanced sequencing,
the molecular diagnostics. The policy landscape has been changing over the last three years.
And we're - it's coming to a head, and there are some developments that I can share with
you in terms of how Medicare and other payers have decided to reimburse for some of these
tests. But the vast majority of the policies that you are probably interested in have yet
to be developed. And if there's one takeaway from my talk today that I'd like to share
with you, is that you all who are developing these diagnostic tests, and taking them from
the lab and bringing them to the clinic - you all will play a vital role in setting reimbursement
policy for these tests. And if you embrace that role, then you'll have a change to affect
what is going to be. The alternative - and I'll allude to some of the comments made
earlier - the alternative is to wait for the bureaucrats and the policy makers to make
the decisions for you. And then you are left, as Dr. Corless said, to basically hunt and
peck around for codes and reimbursement rates, and make it work in your lab. That is the
fork in the road that you're at, and you can decide which pathway you're going to
take. And I'm going to show you some numbers. And
whenever I give these reimbursement talks, once I show you some numbers, nothing else
is going to matter. So that's why I want to make sure I frame my takeaway to you early
on, which is you working with your local Medicare contractors, your Blue Cross Blue Shield plans,
and the payers that you insure, the patients that you treat - that dialog needs to happen,
and you - again, can remain passive and let them make the decisions, or you can be an
active participant in the process. And I do hope that we get to the point where you feel
comfortable having that dialogue and that engagement, because that will be the key for
success in the future. Just some basic topics that we're going to
cover - we'll talk about the new coding, the new payment rates that have been issued
by CMS that, in many cases are already in effect, but are going to be consolidated here
in 2014 and beyond. There are additional mechanisms to code for next gen tests that are going
to be coming out, and there are avenues for you. We talked about test panels, and we'll
talk about why the payers may not like that concept. But there are opportunities to get
those panels coded if you want to, in the short term and we'll talk about that. And
then you know, we'll talk about some of the other challenges and opportunities that are
on the horizon. When we think about reimbursement, a classic
construct takes reimbursement and breaks it down into three separate but related concepts.
And the first one is coverage - which I'll come back to because it's probably the one
that's least well understood. Coverage is really - what's the value of the test? It's
a bit of a binary question. Should I pay for a service, a procedure, a technology, an intervention?
It's really yes/no. It could be yes in certain situations, limited to certain indications.
But coverage - when Dr. Corless said earlier, we're in an era where payers don't like
the concept of a panel. That's a coverage decision. You're screening patients. You're
doing all these things that may not be clinically relevant. That's why I don't like a panel.
I like targeted tests, like a clinical chemistry, a BRAF test. Okay, I understand that. But
when we talk about a panel, it seems like you're collecting all this data for data
purposes. That seems like a research application, so I don't want to cover it. So denial of
coverage is an independent, again, usually binary decision.
Then of course everyone talks about coding. And the question earlier is, you know, what
codes are available. And that, to me, is the hunt and peck strategy. I go find a bunch
of codes. That's what I'm going to put on my charge master, and that's the revenue
I'll make due with. Coding is just a lexicon of the nomenclature describing things that
are done. The really important thing that's related to coding is payment, obviously. Coding
without adequate payment is, you know, nothing to me. It's only coding with adequate payment
which forms the basis of proper reimbursement. So what happened? This has actually been about
five years in the making. The AMA CPT at some point, who is in charge of writing the lexicon
for medical procedures, including clinical chemistry, molecular, pathology, etc. finally
got around to saying, 'We need to develop codes for this broad proliferation of tests.'
And they moved away from what were the old code stacks, which are methodology based codes
- RTPCR; there was a series of array codes, and they really said, 'We need to be more
specific in the coding to describe the anolyte, or what is being done, as opposed to how it's
being done.' And that was the major change. And you know, we worked through Committee;
we came out with - everyone knows the hundred or so Tier 1 codes; there's Tier 2 codes;
there's multiple anolyte algorithm analyses codes, the MAAA's. We're now working on the
next gen sequencing codes. Again, my takeaway is if you are waiting for the codes to come
out, keep waiting, okay. They are not going to solve the problems that you're facing
the next quarter, the next year, because the codes are always - they're like treatment
guidelines. They are a trailing indicator, not a leading indicator. Okay. And I'm going
to show you where we're at. I mean, we're five years into this, and again, I'm going
to show you some numbers and you'll see where this is at and where this is going. And they're
- I'm concerned. I'm concerned if we all just simply wait around for the codes to come
out, it's not going to be a very pretty picture. So what happened to these codes? These codes
actually came out in 2012. There was a one year moratorium; they had to delay the implementation.
And in 2013, all the local Medicare contractors have been running around trying to price these
codes., because CMS went with what's called a gap fill methodology. And gap fill is really
where the local contractors have to figure it out, submit their data to CMS, CMS aggregates
it and then issues the fee schedule, which just came out a bit ago. What are the methodologies
that the local contractors can use in setting prices? They can use cost based methodology.
They can use what they paid historically as a methodology. They can use what other payers
charge in their regions, so they can do survey type data. They have all these different tools
at their disposal. And the dominant form that came out from most of the codes on the Tier
1 list is they tried to cross-walk it back to what they would have paid previously, had
someone used PCR codes, basically. And that's what's the dominant form - but not the only
form, and I'll show you some examples of that.
So I just picked a series of codes here. There's BRAF, EGFR. And this shows you what was published
by CMS as the national amount, that's the first column, that will be in effect in 2014.
So while there's some variability in different parts around the country, these will be the
benchmark, not only for Medicare, but we know other players will look at these fee schedules
and use them, as well. In the other two columns, what I'm showing you is back in 2012 when
we were still under code stacking, just a couple commercial reference laboratories,
what they were basically getting in reimbursement from Medicare when they used to code stack.
All right. So you can see, there are some numbers that came in between; there are some
numbers that are significantly lower. And in a couple cases, for diseases that are primarily
non-Medicare populations, because no claims were actually paid during this era, CMS didn't
have actual data points with which to set the methodology. So they punted. So of the
hundred or so codes, about 60-some got priced, and the rest were left off and will continue
to be contractor price. Okay. So if you break down the pie chart of what happened - the
vast majority of the codes that got priced were cross walked to old codes. But that's
not the only result. So I'll give you an example of the BRAC test from Myriad. List
price, $3,333. It's priced at $2800 on the national fee schedule. So - that's a sole
source test, been around a long time, been paid a long time. I think everyone - no one
would dispute the clinical value of that test. It's 80 genes, $2800, national fee schedule,
now. There's another company out there, Genomic Health, out in California - they have Oncotype
DX breast cancer 21-gene panel - they are not even on this list. They don't want a
code, because they routinely get $3,000. Their market cap is, I think, 900 million dollars.
Foundation Medicine, which was mentioned earlier, they just went IPO, raised a hundred million
dollars in the public markets. Their market cap's about 500 million dollars. Do you think
the investors put that kind of money in because they're going to get $300 a test? No. Foundation
Medicine is working with the payers and establish pricing at a price point that they are comfortable
with, and that's how they're able to raise that kind of money. So while many of you will
do tests that are on this list - and it is correct, in the short term, you know, if you
perform an EGFR, you should bill this code. That is proper billing, and you will live
with this reimbursement. For many of the panels and new tests that you are developing - think
about Myriad; think about Genomic Health. Think about those models as a path to reimbursement,
as opposed to looking on a fee schedule, finding a code, and living with the reimbursement.
You will live in this hybrid world, where there will be different situations. But if
there is clinical utility for the test, and value for the test, you shouldn't be thinking
about these types of fee schedules. You should be thinking about establishing that value
with the payer, with codes at a price point that you're comfortable with.
So many in the lab industry have said this approach is not sustainable. There's legislative
efforts, political efforts - who knows where those things are going to go. The private
payers are using these data, and they will be starting to negotiate with the labs, potentially
using these as starting points for negotiation. But at the end of the day, that's going to
be a discussion that you are going to have to have with your payer. And individual contracts,
you know, those - when the contracting cycle comes up, for those of you who may not be
familiar, you may need to go to your business department, managed care departments, and
work with them, because many times what will happen is - and I think we've all experienced
it - is someone who doesn't really know what we do negotiates a contract, and then the
department head or Finance comes back and says, 'Oh, look at - you know, this is your
budget for the year.' And you're like, 'Whoa, wait a minute, how - you know, who decided
on these rates.' So again, active engagement and participation is really what I recommend.
The code sets are not perfect. It has been stated by the AMA that those single anolyte
codes, those somatic mutation codes, they're not intended to be stacked or billed when
you do something like NGS. In practice, that's what everyone does anyway, because it's the
only outlet that many people believe are available; that's not exactly true - there are unlisted
codes which we recommend a lot of labs use to distinguish panels and complex testing,
as opposed to just billing what's available to you. But the coding - again, you think
about the number of tests that we run, and the fact that there's 68 codes with a price.
I mean, is the code - you know, again, don't expect the coding system to catch up with
you. It's always going to lag multiple years behind what we're doing in the clinic. And
that's why, again, you're in this driver's seat in determining what the reimbursement
rates should be. What AMA has realized is they can't - they
don't have the bandwidth or the capability of describing all the different tests and
test panels that we're creating in the clinic. And that's why they brought in McKesson.
And McKesson is the outlet, if you will, that if you have a test - and I really want to
be very upfront about this - if you have a test where there are clear therapeutic alternatives
or decisions, the clinical utility is well supported. And we'll talk a little bit about
what that means. Then going to McKesson, they can issue you a code within weeks, and the
payers can have a test specific code that they can use to process and price your claim.
That's what the whole McKesson/AMA partnership is about. It's to create something that's
far more rapid, far more nubile to deal with the innovation, because AMA is too bureaucratic
and too slow to do so. So as I said before, there's unlisted CPT codes that are always
available, and then through McKesson there's these things called Z codes which are available,
that all the big commercial payers are onboarding onto this platform. And if you want a code
for your tests in your lab, that outlet is available to you.
This just describes the process. McKesson is trying to aggregate this, so they're a
part of the Diagnostic Exchange is what it's called; they're trying to be the intermediary
between providers and payers, so they've got a system set up. This is going to be launching
here within the next few months. It's been officially announced, but a lot of the logistics
need to get worked out. The key to participating in this process is making sure that you have
a test that has clinical value. And a common mistake when I'm working with people in the
lab is - they've got their research projects here. You know, we saw the slide where thereís
basic research and clinical research. Take research off the table. Payers don't want
to know about research. Payers want to understand - how is this affecting the patient, patient
care, patient outcomes. And if you don't have tests of that ilk, this is not for you.
So I don't want to be misleading, oh, and have everyone run off and go get codes for
a bunch of things. Because if you - you know, just because you have a code doesn't mean
you get coverage. So there are a lot of people who go rush to get codes and at the end of
the day, they won't pay for it anyway. So again, you want to be selective if you go
down this pathway. Some of the benefits potentially of this system
are that McKesson will link it to your drug treatments. We'll have more information about
the clinical outcomes. So in some cases, people are using this platform as a tool to gather
more data that we need. And payers are very open to that. We've worked with various labs
where the data are good - they're not great, they're not perfect - and the payers have
said, 'Let's give this a try. We'll go ahead and code for your test and we'll pay
for it. And then let's look in six months, nine months, in terms of how we would affect
treatment decisions, use of expensive biological therapies, patient survival. And then we'll
make another decision.' So again, the tools are out there to have that negotiation. But
the payers aren't going to come to you with that proposition. You're going to have to
approach them to do so. The last thing about coding I'll mention
- we came out of Committee just a couple weeks ago in L.A., where AMA, AMP is one of the
leading specialty societies in this effort - where we - there's eight different sub-groups
that are developing a series of codes for Next Gen panels, if you will, for the lack
of a better term. And I am cautiously optimistic that we'll get through this process. But
again, the case studies and the panels that we're creating are for tests that are in
the rear view mirror, not for tests that we're going to be doing in the future. And the process
- I'll show you the timeline here in a second - the process is long, and it's not very
fast moving. So here's an example of one of the codes. This is for a test in the prenatal
market for aneuploidy. One of the interesting things about these tests is that they're
set up as a family, so that the technical component is coded separately from the interpretational
report. And this is something obviously AMP and physicians would be interested in, is
to create separate reimbursement for the interpretative component, which is a good development. So
a lot of these codes that you'll see will come in these pairs for the technical versus
professional, right. And this is just one example. There's a whole group working on
whole exomes, whole genome sequencing. But most of these, again, are well defined, established
tests, as opposed to coming up with a system that would allow us to onboard the next tests.
That doesn't exist. There was thinking of do you simply create an NGS code, like the
old code stacks, and just have any new tests dropped into a code. Oh, but there's really
no interest in that. They want the code descriptors to be explicit as to the clinical scenario
and the rationale for the test. They want to bake that in.
So the reason I'm cautiously optimistic - I may even be pessimistic, as - you think about
the types of patients that we see, and the reason we do tests. And is that really something
that you can put into a 600-page book? You know, this technical appendix could be so
long, I don't see where they're going to house all this information. So that's why
so many different people are working on it. But again, I do not expect easy solutions
to be available to us in the near future. Here's the rough timeline. The earliest any
of the NGS codes will be available will be January 2015. That's the timeline that we're
on. So applications right now go in in November. They get discussed at a public meeting in
February for implementation in 2015. That's just the release of the code. As we've said
before, the real battle is not the release of the code; the real battle is how are they
going to be priced. And that could not - that may not be until 2017. So again, that's the
type of timeline that we're on. So in the meantime, what are your options? Unlisted
code - that is what Genomic Health still bills for oncotype DX - an unlisted CPG code. Thereís
McKesson. There are options for you to go get something on your own, again, if you feel
like it. But ultimately, all those mechanisms don't replace the conversation you need to
have with your payer about why you're doing the test, what is its value, and what's an
appropriate price point. That's all this going to come down to. The coding system is
not going to replace - and do not use it as a crutch to not have that conversation.
I'm just going to talk - I'll just finish up on this slide, and then I'm happy to answer
questions. When I talk about clinical utility, what does it mean? And it does mean different
things. You will see it described differently, in different ways. But this hierarchy shows
the terminology the payers use. Analytic validity - whether it's accurate, precise, reproducible
- that's usually what you're doing as part of your CLIA certification. That's what FDA
wants to see. But that's not what payers are interested in. That's a mandatory, but
insufficient, okay. So we have lots of people going into payers, talking analytical validity,
and the payers say, 'I haven't heard a single thing that I'm interested in yet.' Okay.
The next thing is clinical validity, which is - is it actually valid. If you find a mutation,
what does it really mean? And there are many times we do tests, and we find things, and
we don't know what it means. That's again, don't be going to the payers, talking about
things that don't have clinical validity. Okay. Clinical utility is really what they're
interested in, which is the, 'So what? If I run all these mutations but they're not
actionable,' or they involve a product, and we saw earlier - they involve a product in
clinical development and it's not FDA approved - well, that immediately says to me that's
a research tool. That should be used in the investigational setting. Why am I paying for
that? Okay. The products - the interesting discussion is, let's say it's an off label
product - that you found a BRAF mutation in a glioblastoma patient. Okay, maybe there's
some data on that. Maybe it's in the compendium. Now we're talking. If it's clearly an FDA
approved drug for that indication, I think we can set that aside. But there are gray
areas in between, as far as why my panel includes these mutations for these patients.
And if you have the data for that - that's one of the things Foundation Medicine and
some other people do really well, is they're able to communicate that, 'I'm doing these
things for a reason; that there's a body of literature that can help guide patient
treatment.' And that's how they got to where they are. So think of it from a - does it
change patient diagnosis? Does it change patient management? Is there data on patient outcomes.
That's a lot to imbed into a proposition for a test. But those are the questions that
they payers are asking. And again, if you're going to go down these pathways of McKesson,
and talk to payers, negotiate contracts - make sure that you have that story really well
told through your literature, your evidence, as you engage in these discussions. With that,
I think I'll - I'm out of time. So thank you.