[00:00:02] Speaker 00:
Case number 18-5326, United Health Care Insurance Company et al.

[00:00:07] Speaker 00:
versus Alex Michael Lazar II in his official capacity as Secretary of Health and Human Services et al.

[00:00:14] Speaker 00:
a balance.

[00:00:15] Speaker 00:
Mr. Shaw for the a balance, Mr. Murin for the appellees.

[00:00:22] Speaker 01:
Good morning, Your Honors.

[00:00:24] Speaker 01:
May it please the Court, Waley Shaw for the United States.

[00:00:29] Speaker 01:
Let me begin by being clear on what the overpayment rule actually requires.

[00:00:34] Speaker 01:
The rule provides that if a Medicare Advantage insurer gets paid for a diagnosis code and it later finds out that that code is invalid, the insurer can't just keep the money and pretend the error doesn't exist.

[00:00:48] Speaker 01:
It has to tell CMS and CMS will take the money back.

[00:00:52] Speaker 01:
The rule does not require insurers to audit all of their diagnosis codes

[00:00:56] Speaker 01:
or report every error, nor would that even be feasible.

[00:01:00] Speaker 01:
Now, UnitedHealth argues that the straightforward requirement is unlawful because the statute bars CMS from recovering for even the most egregious errors unless it first determines the relative error rates in traditional Medicare and in each Medicare Advantage plan.

[00:01:15] Speaker 01:
That would represent an extraordinary change to how the risk adjustment process works.

[00:01:19] Speaker 01:
Suppose CMS finds out that a beneficiary that an insurer reported is having cancer

[00:01:25] Speaker 01:
doesn't actually have cancer.

[00:01:27] Speaker 01:
Under UnitedHealth's theory, CMS can't recover the money it paid for covering a cancer patient unless it first audits both traditional Medicare and the Medicare Advantage plan to determine their overall error rates.

[00:01:39] Speaker 01:
And CMS would have to repeat that process for every single insurer

[00:01:43] Speaker 01:
if it wants to be able to collect for any errors from that insurer.

[00:01:47] Speaker 03:
Well, it has to do it once for itself and once for each of its contracting insurers.

[00:01:53] Speaker 01:
Right.

[00:01:54] Speaker 01:
But still, that would be an endeavor that has never before been attempted in Medicare Advantage.

[00:02:02] Speaker 01:
It would be a degree of auditing that's well beyond what CMS has ever done.

[00:02:06] Speaker 01:
I just want to point out that UnitedHealth doesn't apply any sort of similar limitation on insurers' ability to add new diagnoses.

[00:02:14] Speaker 01:
If an insurer discovers that this purported cancer patient actually has diabetes, the insurer is free to submit that diabetes code and collect payment for the additional code.

[00:02:24] Speaker 01:
And there's no requirement that the insurer demonstrate that its error rate is better than that of traditional Medicare.

[00:02:30] Speaker 01:
Practically, this means that in terms of medical record reviews, the sky is the limit.

[00:02:35] Speaker 01:
When an insurer finds an additional code, it can report it.

[00:02:39] Speaker 01:
But if CMS discovers an invalid code, it has to go through

[00:02:43] Speaker 01:
a huge process in order to collect payment.

[00:02:48] Speaker 02:
Mr. Shaw, there's a number of parts of your opposing party's brief that either make some kind of a factual empirical assertion about how things are done in this industry or says that

[00:03:08] Speaker 02:
an implication of your argument is that some fact exists.

[00:03:13] Speaker 02:
And I made a list of five or six of them.

[00:03:17] Speaker 02:
And I'm wondering if you can kind of maybe confirm or deny whether what they say is true or in some cases what they say you say is actually what you say.

[00:03:30] Speaker 02:
I'll do my best.

[00:03:32] Speaker 02:
The first is the claims adjudication audits

[00:03:38] Speaker 02:
in traditional Medicare remove only a minuscule portion of unsupported codes from the data?

[00:03:46] Speaker 02:
Is that true or is that false?

[00:03:49] Speaker 01:
I think it's a small fraction of codes, and that is an inherent nature of the volume of claims that are submitted to traditional Medicare each year.

[00:03:56] Speaker 01:
And I would point out that the same is also true of Medicare Advantage insurers.

[00:04:00] Speaker 01:
In fact, we have evidence in the record showing that

[00:04:02] Speaker 01:
when HHS's Office of the Inspector General audited one of United House member plans, that in fact it wasn't conducting any auditing except for when it received a request for documentation from an external auditor.

[00:04:17] Speaker 02:
So that leads me to the next question.

[00:04:23] Speaker 02:
They say that, well, you just tell me if this is right or not right.

[00:04:28] Speaker 02:
Medicare Advantage plans,

[00:04:30] Speaker 02:
do not need to reduce the error rate below the rate that CMS achieves through its claims adjudication process.

[00:04:45] Speaker 01:
Yeah, CMS has never imposed any particular requirement that insurers achieve a specific error rate, whether that's an absolute rate or relative

[00:04:56] Speaker 02:
It could be as low as CMS's error rate and that would be okay by you.

[00:05:01] Speaker 01:
No, I mean, that's not, I don't want to say it's okay.

[00:05:03] Speaker 01:
I mean, CMS has never required insurers to achieve a specific error rate.

[00:05:07] Speaker 01:
That does not mean that insurers don't have obligations.

[00:05:09] Speaker 01:
They do have an obligation to have an effective compliance requirement.

[00:05:13] Speaker 02:
Would it be accurate to say that you require plans to do more than the claims review processes that are comparable to CMS's activities?

[00:05:25] Speaker 01:
I can't, I can't, I don't think I can answer that directly because I don't know how to quantify more or less.

[00:05:30] Speaker 01:
There are specific compliance obligations that are spelled out in detail in the Medicare regulations.

[00:05:36] Speaker 01:
And that is what insurers are bound by.

[00:05:39] Speaker 01:
They're not being judged by, you know, any sort of specific error.

[00:05:44] Speaker 02:
Is, I guess, is it conceivable that what you require of Medicare Advantage insurers

[00:05:52] Speaker 02:
is a standard of error correction or error discovery that is more effective than CMSs?

[00:06:05] Speaker 01:
Again, I don't see how I can answer that question because the regulations require specific things like insurers have to have processes for dealing with specific kinds of issues.

[00:06:21] Speaker 01:
These are sort of qualitative obligations that insurers have.

[00:06:24] Speaker 01:
But that's not, it's impossible to compare that to say like what a Medicare administrative contractor does, which is an entirely different program.

[00:06:34] Speaker 01:
The payment model is set up differently.

[00:06:36] Speaker 01:
And so I can't say whether it's more or less.

[00:06:38] Speaker 02:
I don't want to.

[00:06:40] Speaker 03:
Oh, go ahead.

[00:06:40] Speaker 03:
I don't want to interrupt your line of questioning if you had more.

[00:06:43] Speaker 02:
I do, but I'm happy to go for it.

[00:06:46] Speaker 02:
Go for it.

[00:06:47] Speaker 02:
I'm going to switch a little bit.

[00:06:49] Speaker 02:
Imagine a rule that requires United to delete all unsupported codes.

[00:06:56] Speaker 02:
That rule would cause United to be underpaid.

[00:07:00] Speaker 02:
Is that true?

[00:07:00] Speaker 01:
No, it certainly is not necessarily true, which I think is what United Health is arguing.

[00:07:06] Speaker 01:
They are arguing that the result is inevitable.

[00:07:09] Speaker 01:
And I think.

[00:07:10] Speaker 02:
Okay, now that CMS imposes a 5.9% industry wide coding intensity adjuster to offset the financial impact of plan coding activities such as chart reviews.

[00:07:25] Speaker 02:
Plans have no choice but to engage in chart review.

[00:07:29] Speaker 02:
Do you think that's true?

[00:07:31] Speaker 01:
No, I don't think that's true.

[00:07:32] Speaker 01:
I mean, I think we

[00:07:34] Speaker 01:
Sorry.

[00:07:37] Speaker 02:
I have two more and then I'll get out of your way and out of my colleagues way.

[00:07:42] Speaker 02:
Coding errors occur in both directions.

[00:07:45] Speaker 02:
So as a result, the effect of overcoating and undercoating is a wash.

[00:07:54] Speaker 01:
I think it is true that the effects of undercoating and overcoating tend to offset each other.

[00:08:02] Speaker 01:
So yeah, I don't want to be careful to frame it that way rather than to say it's a wash.

[00:08:08] Speaker 02:
They offset each other in a way that's roughly equal.

[00:08:13] Speaker 02:
Okay.

[00:08:14] Speaker 02:
Yes.

[00:08:15] Speaker 01:
I don't want to say that they are precisely the same, but yes, in general, the tendency would be for those effects to offset each other and there's no reason to predict

[00:08:28] Speaker 01:
A bias of one direction or another.

[00:08:30] Speaker 02:
And then this is my last of these clarification questions.

[00:08:34] Speaker 02:
Although some unsupported codes will raise the value of other factors.

[00:08:40] Speaker 02:
The overall impact of unsupported codes in CMS is calibrating data set will lower the overall average value of the factors.

[00:08:52] Speaker 01:
I don't know that that's true.

[00:08:54] Speaker 01:
And the reason is that it's very difficult to determine how changing the inputs to this regression model will affect the outputs.

[00:09:05] Speaker 01:
There are just a lot of effects that can occur depending on the particular mix of patients and how things work.

[00:09:12] Speaker 01:
So I can't represent sort of on a categorical basis what would happen.

[00:09:16] Speaker 03:
But I think what we can say is- Go ahead, finish your answer.

[00:09:19] Speaker 01:
Yeah, I think one thing that's important to remember is that if there are lower codes and higher codes, that the effect on any particular insurer will depend on that patient's mix of beneficiaries.

[00:09:34] Speaker 01:
And in fact, different insurers can have different outcomes, basically, based on sort of the increases and decreases in relative factors, based on each insurer's particular mix of beneficiaries.

[00:09:50] Speaker 03:
I had a question about that, just really trying to understand this program and how it works.

[00:09:56] Speaker 03:
I think the proposition was, or the question was whether the overall effect of unsupported codes in the fee-for-service traditional Medicare is to lower the overall average value of the risk factor.

[00:10:12] Speaker 03:
And I guess the question is whether an unsupported

[00:10:20] Speaker 03:
If FFS Medicare fails to code someone as diabetic, that could have a lowering effect.

[00:10:35] Speaker 03:
at the margin on the value the risk factor is set at.

[00:10:40] Speaker 03:
If that diabetic is more than average expensive to treat, it could also have the effect of raising, I mean, at the very, very margin, we're talking statistical infinitesimal, but it could also have the effect of raising the, or artificially raising the risk factor because the atypically cheap diabetic

[00:11:03] Speaker 03:
is not being figured in, right?

[00:11:06] Speaker 03:
So one would have to have a theory that the FFS undercoding is skipping the expensive people and not skipping the cheap people in any diagnostic category, right?

[00:11:22] Speaker 01:
Unfortunately, again, it's very hard for me to say precisely

[00:11:28] Speaker 01:
how these effects will all play out.

[00:11:30] Speaker 01:
I think maybe the easier way to look at it is to think, I think everyone agrees that there is some over-coding and there is some under-coding.

[00:11:41] Speaker 01:
Now, I think that the aspect of this problem that United's hypotheticals leave out is that when Medicare Advantage insurers delete diagnosis codes by conducting a medical review,

[00:11:54] Speaker 01:
medical record review, they are also likely to add diagnosis codes.

[00:11:58] Speaker 01:
And when they do that, there's no reason in particular to think that they are going to be underpaid as a result of both adding and deleting codes.

[00:12:06] Speaker 01:
And in fact, we have lots of reasons to think that they will be overpaid because insurers structure these programs to generate revenue.

[00:12:15] Speaker 01:
And there are lots of ways that they can go about them, whether it be chart review, whether it be

[00:12:19] Speaker 01:
in-home risk assessments or something else.

[00:12:22] Speaker 01:
So they are likely to structure these arrangements such that the overall effect will be to increase revenue more by adding diagnosis codes than insurers lose by deleting them.

[00:12:32] Speaker 03:
I appreciate that.

[00:12:32] Speaker 03:
Can I take you back to the statute?

[00:12:36] Speaker 03:
Is your primary position that the actuarial equivalence requirement simply doesn't apply to the overpayment

[00:12:48] Speaker 03:
rule?

[00:12:51] Speaker 01:
Yes, I think that's exactly right.

[00:12:53] Speaker 01:
The actuarial equivalence is an instruction to the secretary as to how to design the risk adjustment model.

[00:13:01] Speaker 01:
And the requirements that are reflected in the overpayment rule are an aspect of that model.

[00:13:06] Speaker 01:
But what an insurer is entitled to- Are they an aspect of that model?

[00:13:09] Speaker 03:
Because this is where I'm a little bit confused by your briefing.

[00:13:11] Speaker 03:
It seems like the risk adjustments are set annually based on a data set.

[00:13:19] Speaker 03:
Right?

[00:13:20] Speaker 01:
Yes, the relative factors change.

[00:13:23] Speaker 01:
Well, I want to be careful.

[00:13:24] Speaker 01:
Yes, they change regularly based on recalibrations of the model.

[00:13:29] Speaker 03:
Right.

[00:13:30] Speaker 03:
And the data that CMS has at the particular time that goes into that model is what it uses.

[00:13:36] Speaker 03:
And UHC has not challenged the calibration of the risk assessment factors.

[00:13:44] Speaker 01:
Yes, I think that's exactly right.

[00:13:46] Speaker 01:
And I think that is an odd aspect of this challenge, which is that

[00:13:50] Speaker 01:
sort of the features that we're talking about are long standing features of the risk adjustment model that have existed since its inception, the requirement of medical record documentation, the requirement to return payments due to invalid diagnosis codes.

[00:14:05] Speaker 01:
And CMS, I'm sorry, United has brought up all these issues in the context of the overpayment rule, which is largely a procedural provision that simply

[00:14:15] Speaker 01:
specifies how these substantive requirements interact with the requirements, the procedural requirements of the Affordable Care Act.

[00:14:23] Speaker 01:
And so, yes, I think the far better posture for this case would have been to challenge those rates directly to say that they are too low because of the existence of errors or to sort of challenge the design of the risk adjustment model as a whole.

[00:14:36] Speaker 01:
And that's something that then sure has an opportunity to do every year.

[00:14:40] Speaker 01:
But they haven't done that.

[00:14:41] Speaker 01:
They've sort of smuggled that challenge into the context

[00:14:45] Speaker 01:
a challenge to this very particular overpayment rule, which really just reflects broader substantive requirements.

[00:14:55] Speaker 01:
And I think that makes the court's consideration of the issue much more difficult, because if this had come up in the context of a challenge to the rates, then there would be more opportunity for the agency to examine the data and offer its expert analysis instead.

[00:15:14] Speaker 01:
in the form of a challenge to the omission of an FFS adjuster from this procedural provision.

[00:15:20] Speaker 03:
You, in your brief, you say if we were to affirm the district court, we should do it on arbitrary and capricious grounds rather than statutory grounds.

[00:15:26] Speaker 03:
I guess I wonder whether we can do that without first deciding whether and how the overpayment rule might implicate the statutory requirement of actuarial equivalence.

[00:15:40] Speaker 03:
And I guess by the same token, the requirement of

[00:15:43] Speaker 03:
same methodology.

[00:15:45] Speaker 03:
So can you elaborate a little bit on how it is that you think we could reach an arbitrary and capricious claim without first somehow embracing their notion that the overpayment rule implicates actuarial equivalence?

[00:16:01] Speaker 01:
Yes.

[00:16:01] Speaker 01:
First, we obviously think that the rule is valid.

[00:16:03] Speaker 01:
So I want to make that clear, both against all of the arguments that have been raised.

[00:16:10] Speaker 01:
But to answer your question directly,

[00:16:13] Speaker 01:
I think the way that this court could go about it is to say that the specific challenge that's been brought is to the failure to include an FFS adjuster in the overpayment rule.

[00:16:26] Speaker 01:
And one of the arguments is that CMS failed to adequately explain its decision not to include such an adjuster in light of its decision to implement an adjuster in the context of the separate

[00:16:43] Speaker 01:
RAD-V audit methodology.

[00:16:45] Speaker 01:
And so I think it would be straightforward.

[00:16:47] Speaker 03:
Which it decided at the end of the day not to do.

[00:16:49] Speaker 01:
Right.

[00:16:49] Speaker 01:
It decided and then, but it has it preliminarily decided not to do.

[00:16:54] Speaker 01:
But that rulemaking and the study are still ongoing.

[00:16:57] Speaker 01:
But I think it would be open for this court to say that the agency simply failed to adequately explain its decision to implement the adjuster in the context of that audit methodology, but not implement it in the context of the overpayment rule.

[00:17:11] Speaker 01:
And that would be a straightforward basis for vacating the role and remanding to the agency for further proceedings.

[00:17:16] Speaker 01:
And I would point out that this would give the agency an opportunity to complete, for one thing, the study that is trying to address the precise issues that are brought up by UnitedHealth and to lend its expertise to the question.

[00:17:30] Speaker 03:
Is it trying to address the precise issues?

[00:17:31] Speaker 03:
I thought it was trying to address these RADV issues, which are different issues.

[00:17:34] Speaker 01:
You're right.

[00:17:35] Speaker 01:
I mean, so.

[00:17:35] Speaker 03:
This is where your briefing really, I have to say, really confuses me, because when you say,

[00:17:40] Speaker 03:
that the overpayment rule doesn't violate actuarial equivalence.

[00:17:47] Speaker 03:
It's consistent with it.

[00:17:48] Speaker 03:
It seems to embrace the notion that that requirement is applicable to the overpayment rule.

[00:17:54] Speaker 03:
And yet, it's not entirely clear to me how it could be applicable.

[00:17:59] Speaker 03:
And then in response to my prior question, you said, well, actually, your primary position is that it doesn't apply.

[00:18:04] Speaker 01:
Right.

[00:18:05] Speaker 01:
So I'm sorry for the confusion.

[00:18:08] Speaker 01:
Let me try to be as clear as I can.

[00:18:11] Speaker 01:
The specific FFS adjuster that was proposed and then CMS later preliminarily decided not to implement was in the context of this specific audit methodology.

[00:18:21] Speaker 01:
So it would not apply of its own force to the overpayment rule.

[00:18:25] Speaker 01:
That was a sort of comprehensive audit methodology that would achieve or perform a level of auditing that has not to date been implemented in

[00:18:35] Speaker 03:
in Medicare Part C. And that's contract-wide audit.

[00:18:40] Speaker 01:
Right.

[00:18:41] Speaker 01:
It's trying to, exactly.

[00:18:42] Speaker 01:
It's trying to correct.

[00:18:43] Speaker 03:
Use a sample, audit it, and then extrapolate contract-wide.

[00:18:47] Speaker 01:
It's trying to correct every single error, which is not what the overpayment rule does.

[00:18:51] Speaker 03:
So how would that study clarify the overpayment rule issue that UnitedHealthcare is raising here?

[00:19:00] Speaker 03:
That's what I'm not

[00:19:02] Speaker 03:
following the connection you're making.

[00:19:04] Speaker 01:
Right.

[00:19:04] Speaker 01:
The argument that insurers have made in that case is that the reason a FFS adjuster is required is that errors in traditional Medicare data cause the relative factors for Medicare Advantage to be improperly deflated or to be too low, and therefore an adjustment is needed.

[00:19:26] Speaker 01:
So although the specific adjuster that is discussed there is specific to that particular audit methodology, the sort of underlying effect that the insurers are claiming is the same as the one that UnitedHealth is relying on in this case.

[00:19:44] Speaker 03:
But that makes sense, because there they're challenging the risk adjusters.

[00:19:48] Speaker 03:
They're challenging the data that is the input for the risk adjusters, right?

[00:19:53] Speaker 03:
Which is, I mean, that was sort of the prior question I had for you.

[00:19:57] Speaker 03:
And maybe it's more a question for Mr. Maron that they haven't here challenged the calculation of the risk adjusters.

[00:20:06] Speaker 01:
I think in that case, as in, well, I think it is at least one of the arguments that's being raised with respect to the new audit methodology that an FFS adjuster is required for the reason

[00:20:21] Speaker 01:
that I specified that the relative factors are too low and therefore an adjuster is needed.

[00:20:27] Speaker 03:
And so- On the risk assessment, the calibration of the risk assessment.

[00:20:34] Speaker 01:
Right.

[00:20:34] Speaker 01:
But I think that, right, there are obviously- This adjustment, I'm sorry.

[00:20:38] Speaker 01:
Yeah.

[00:20:38] Speaker 01:
So I think you're right.

[00:20:39] Speaker 01:
So they are alleging that the risk factors are too low.

[00:20:42] Speaker 01:
I'm sorry.

[00:20:42] Speaker 01:
The relative factors are too low and therefore that CMS must implement an FFS adjuster to correct for those factors that are too low.

[00:20:51] Speaker 01:
And so the empirical question that CMS could answer that bears on both whether an FFS adjuster is required in the context of the RADV audit methodology and also on whether an FFS adjuster is required here is whether that sort of decrease in, you know, that sort of deflation of relative factors actually exists.

[00:21:13] Speaker 01:
If it does, then obviously there would be, you know, there would be, that would tend to support,

[00:21:21] Speaker 01:
the arguments on the insurer side.

[00:21:22] Speaker 01:
But if it doesn't, then I think that also undercuts the argument.

[00:21:25] Speaker 01:
So I think that study has the potential to shed substantial light on the issues before this court.

[00:21:35] Speaker 01:
And I think the best thing would be to not short circuit that analysis, but to remand, to vacate the rule and remand so the agency has an opportunity to consider the issue further.

[00:21:45] Speaker 03:
You've kind of confused me because

[00:21:48] Speaker 03:
If the actuarial equivalence statutory standard does not apply to the overpayment rule, then whatever actuarial equivalence is being done on a contract-wide basis,

[00:22:06] Speaker 03:
which that I do understand that they may say, oh, these adjustments are too low, too high, you're skipping the most expensive diabetics or you're rolling them into the average and we don't, the standard, the base rate and none of our people are base rate and so whatever.

[00:22:24] Speaker 03:
But this tail wagging dog theory where they're saying, and you can't collect from us in a documented case of overpayment until you rewrite your risk adjusters.

[00:22:42] Speaker 03:
Like I thought your initial position is just no, the statute doesn't require that.

[00:22:48] Speaker 03:
And then, but then when you say, well, we'll see whether they have it right.

[00:22:52] Speaker 03:
I guess one way of understanding it is they certainly haven't met their burden to show that overpayments are illegitimate because of some skew in the direction of the errors in traditional FFS.

[00:23:12] Speaker 03:
the government as a matter of accuracy might be investigating that under the RAD-V and the FFS adjuster umbrella.

[00:23:27] Speaker 01:
Yes, I think that's a very good way of understanding the situation.

[00:23:34] Speaker 01:
I think what UnitedHealth has not come forward with any evidence that the effects that

[00:23:39] Speaker 01:
it's proposing actually exists.

[00:23:42] Speaker 03:
But to be fair, the evidence is in the hands of CMS.

[00:23:45] Speaker 01:
Right.

[00:23:45] Speaker 01:
And it's currently trying to study that question.

[00:23:47] Speaker 01:
These are not easy issues to answer.

[00:23:49] Speaker 01:
They require extensive study.

[00:23:52] Speaker 01:
And there's actually a tremendous number of comments that were received in response to the study.

[00:23:59] Speaker 01:
And they deal with extremely esoteric topics about econometric analysis and regression modeling.

[00:24:07] Speaker 01:
These are difficult issues to understand.

[00:24:11] Speaker 01:
And I think it really would benefit both the program and the court to have the agency be able to study that question and provide its expertise.

[00:24:21] Speaker 03:
And in the meantime, the overpayment rule would be not in effect?

[00:24:29] Speaker 01:
Right.

[00:24:30] Speaker 03:
Where is it now?

[00:24:31] Speaker 01:
We are not conceding that the overpayment rule is invalid.

[00:24:34] Speaker 01:
And I'm sorry if I gave that impression.

[00:24:35] Speaker 03:
Is it now because of the district court's ruling is the overpayment rule now suspended?

[00:24:42] Speaker 01:
Right.

[00:24:42] Speaker 01:
The agency is not attempting to enforce the overpayment rule at this point.

[00:24:49] Speaker 01:
But again, we are not conceding that it is invalid.

[00:24:53] Speaker 01:
But we are merely saying that if this court decides that it is invalid, that it should rely on

[00:25:00] Speaker 01:
reasoning that does not forever preclude the agency from ever recovering for any errors unless it first conducts a study of the relative error rates in traditional Medicare and each Medicare Advantage plan.

[00:25:13] Speaker 01:
We think that would be a huge change to how the program works, and it could not be what Congress intended when it enacted the actual equivalence requirement.

[00:25:24] Speaker 01:
I see that I'm well over my time.

[00:25:26] Speaker 01:
I have a couple more.

[00:25:30] Speaker 01:
I'm sorry, Mr. Shaw.

[00:25:32] Speaker 02:
Judge Taylor, were you finished?

[00:25:35] Speaker 03:
I was just waiting for Judge Rogers.

[00:25:38] Speaker 04:
All right.

[00:25:38] Speaker 04:
Go ahead, Judge Walker.

[00:25:40] Speaker 02:
No, please, Judge Rogers.

[00:25:41] Speaker 02:
No.

[00:25:45] Speaker 04:
I'm not asking questions, but you know, if other people have covered them, I don't need to just hear my voice repeat them.

[00:25:52] Speaker 02:
I wanted to talk about a different adjuster, not the FSS adjuster, but the code intensity adjuster.

[00:25:59] Speaker 02:
And at one point in the brief, you say that it corrects not only for the unreported diagnoses that United discovers, but instead, and this is not a direct quote, but instead it corrects for the net difference between additions and deletions.

[00:26:19] Speaker 02:
So I just wonder where in the record I could go to confirm that.

[00:26:26] Speaker 02:
And I don't need an exact page, but just

[00:26:28] Speaker 02:
A hint for where to look for it.

[00:26:33] Speaker 01:
I think that's sort of discussed in some of the notices about changes to the risk adjustment methodology and in CMS's responses to that.

[00:26:44] Speaker 01:
So they discussed, for example, the study that they used to initially calculate the amount of the adjuster.

[00:26:49] Speaker 01:
And what they did is they looked to the overall difference in risk scores between the two programs

[00:26:58] Speaker 01:
rather than trying to account for any specific effect of, let's say, you know, insurers conducting medical record reviews or insurers having to delete diagnoses.

[00:27:08] Speaker 01:
So it's a net difference.

[00:27:09] Speaker 01:
And I think that's a very important point to remember because it does not purport to account for all of the overcoding, or I'm sorry, all of the sort of adding of diagnoses that insurers do, rather it

[00:27:24] Speaker 01:
takes into account the net effect of both adding and deleting diagnoses.

[00:27:27] Speaker 01:
So now if insurers all of a sudden are no longer obligated to delete invalid diagnoses, then the coding intensity adjustment would not be enough to offset for the inflated risk scores.

[00:27:38] Speaker 03:
And in fact, it's industry wide, right?

[00:27:40] Speaker 03:
It doesn't it doesn't account for different practices by different insurers.

[00:27:44] Speaker 01:
Yeah, I think and that that's also a very important point.

[00:27:46] Speaker 01:
And there are substantial differences in how different Medicare Advantage insurers conduct error correction.

[00:27:53] Speaker 01:
And so

[00:27:53] Speaker 01:
There are certainly some on the most aggressive end of the spectrum where they are conducting extremely aggressive medical record reviews.

[00:28:01] Speaker 01:
And the effect of those would not be taken into account.

[00:28:06] Speaker 01:
It wouldn't be possible to compensate for those with an industry-wide adjuster.

[00:28:10] Speaker 01:
But that is what is now specified in the statute.

[00:28:15] Speaker 02:
And I suspect the answer to this is no.

[00:28:19] Speaker 02:
Just to be sure, has the government taken any position in this litigation that's inconsistent with any position it's taken in False Claims Act cases?

[00:28:31] Speaker 01:
No, I think it's entirely consistent.

[00:28:35] Speaker 01:
And I think the False Claims litigation, I think, really illustrates some of the things that I've been talking about in terms of medical record reviews and how insurers

[00:28:46] Speaker 01:
try to generate both add and delete diagnoses when they conduct medical record reviews.

[00:28:53] Speaker 02:
Thank you.

[00:28:56] Speaker 01:
If I could, I would like to reserve some time for rebuttal, if possible.

[00:29:02] Speaker 04:
Sure.

[00:29:02] Speaker 04:
No, we'll give you some minutes on rebuttal.

[00:29:06] Speaker 04:
Anything further, though, from my colleagues at this point?

[00:29:09] Speaker 03:
Not from me, thanks.

[00:29:11] Speaker 04:
OK.

[00:29:12] Speaker 04:
All right, so let us hear from Appellees.

[00:29:32] Speaker 06:
Dr. Rogers, good morning and may please the court.

[00:29:34] Speaker 04:
Good morning.

[00:29:35] Speaker 06:
On any reasonable reading of the terms actuarial equivalence and same methodology, they require CMS to engage in a comparison of the two populations.

[00:29:48] Speaker 06:
In the overpayment rule, CMS refused to do that.

[00:29:52] Speaker 06:
Instead, it took the approach that in deciding whether a plan has been overpaid, all you need to do is look at the plan's data

[00:30:01] Speaker 06:
see if there are unsupported codes in there.

[00:30:04] Speaker 06:
And if so, every single one of them is an overpayment, regardless of the rate in CMS's own data.

[00:30:13] Speaker 06:
In fact, without comparing the plans on any metric.

[00:30:17] Speaker 06:
And so there was a discussion, Your Honors, with my colleague, where I think he's got this with respect backwards.

[00:30:26] Speaker 06:
He says, we're saying will inevitably always be underpaid.

[00:30:31] Speaker 06:
if this rule is legal.

[00:30:35] Speaker 06:
It's actually the opposite.

[00:30:36] Speaker 06:
The government is saying no matter what, regardless of all the various factors that it outlined in its argument, even if the coding intensity adjuster fully offsets for a particular plan's coding activities, even if that plan has an error rate that's substantially below CMS's,

[00:30:56] Speaker 06:
Regardless of any of this, it does not matter.

[00:31:00] Speaker 06:
I don't, as an agency, even have to ask those questions.

[00:31:03] Speaker 06:
I simply look at the plan's data.

[00:31:07] Speaker 06:
And what Judge Collier correctly understood, what CMS's own experts in the record said, what the American Academy of Actuaries told CMS, what its own post-judgment study found was, the presence of unsupported codes in their data

[00:31:25] Speaker 06:
on average, reduce the average value of the payments they set.

[00:31:29] Speaker 06:
That's mathematically inevitable.

[00:31:31] Speaker 03:
But Mr. Marin, why is that not just an argument about the setting of the risk adjustment factors, and indeed that's how the agency has looked at it, and that's what the RADV audit methodology is trying to get at, and that's what the FFS adjuster is trying to get at, if all that were

[00:31:49] Speaker 03:
clean and shiny and there was a sense that the adjustments were made and the auditing was done on these samples and extrapolations were, you know, all the actuaries and the statisticians were happy that the extrapolation of error was done, then you would still be arguing that one cannot identify individual overpayments because it all has to be done wholesale?

[00:32:20] Speaker 03:
I just don't even really understand your argument about how the agency would work under the regime that you're talking about.

[00:32:27] Speaker 03:
You come across an error in a claim.

[00:32:31] Speaker 03:
Somebody is identified as diabetic, and they're not.

[00:32:35] Speaker 03:
Their physician asked them, are you diabetic?

[00:32:38] Speaker 03:
And noted in the record, and the patient actually said no.

[00:32:42] Speaker 03:
So it's a mistake.

[00:32:44] Speaker 03:
You come across that.

[00:32:45] Speaker 03:
And because there's been a general auditing of data, it's your position that there's no place for an overpayment rule?

[00:32:56] Speaker 06:
So Judge, thank you for letting me clarify that, because I think my colleague in the government has completely caricatured our position.

[00:33:03] Speaker 06:
So I'd like to clarify.

[00:33:05] Speaker 06:
All we are asking is that the court rule

[00:33:09] Speaker 06:
in the same way Judge Collier ruled, which is this.

[00:33:13] Speaker 06:
The government, consistent with actuarial equivalence, cannot simultaneously both distribute its costs, set payments, over a set of data that contains lots of unsupported codes, and for that purpose, distribute costs to unsupported codes, and then at the same time, say,

[00:33:34] Speaker 06:
Every unsupported code in a plan's data is an overpayment.

[00:33:38] Speaker 03:
When you discover an unsupported code, you're equating something that's done in general as a statistical matter with something that's done in an individual case, which is why my question was, is it your view that only statistical plan level

[00:33:55] Speaker 03:
corrections can be made, because I want to know how you see these two coexisting, because you don't like the way they've done it.

[00:34:02] Speaker 03:
But in my view, what you're raising are arguments that really go to whether they have correctly calibrated the risk adjusters.

[00:34:13] Speaker 03:
And if they have,

[00:34:15] Speaker 03:
then they can look at application and whether they've been accurately applied.

[00:34:21] Speaker 03:
And you're really, I mean, the way I'm seeing it, and you can correct me, this is why I'm engaging with you on this, is that you're saying, well, when CMS asks us to identify an overpayment and give back for that,

[00:34:39] Speaker 03:
We look at that as a business matter in the aggregate and on average, and we think that maybe they have errors that are systemically skewed also, and so they can't ask for that.

[00:34:58] Speaker 06:
No, so Judge, no, I don't think that's my position.

[00:35:00] Speaker 06:
So if I might, and I recognize this is complicated.

[00:35:03] Speaker 06:
So there are two, there has to be, how you define an overpayment can't be approached without asking, how are you paid in the first place?

[00:35:12] Speaker 06:
They go hand in glove.

[00:35:14] Speaker 06:
So there has to be parallelism.

[00:35:19] Speaker 06:
We don't object the current rates.

[00:35:22] Speaker 06:
You're saying, well, why don't we sue about the rates.

[00:35:24] Speaker 06:
The current rates were set by correlating costs and CMS system to the presence of codes and claims data.

[00:35:34] Speaker 06:
Those payment rates are perfectly adequate, so long as they're applied to a similar

[00:35:42] Speaker 06:
criteria set of data in the plans data.

[00:35:46] Speaker 06:
If you're going to correlate costs to the presence of codes and claims,

[00:35:51] Speaker 06:
That's fine, but then we're paid based on the presence of code and claims.

[00:35:55] Speaker 06:
But there's another way to do this that's much simpler, Your Honor.

[00:35:58] Speaker 06:
What Judge Collier said is you can't simultaneously take one approach on your side of one another, which means that on remand, there are two ways CMS can fix it, one of which avoids all the things they're concerned about, and I think that undergird your question, Judge Pillard.

[00:36:13] Speaker 06:
CMS, and it would be not hard at all for CMS to do this.

[00:36:18] Speaker 06:
On remand, they could say, look, the way we're going to achieve equivalence is we're going to recalibrate our rates and figure out how much higher they would have been had we deleted unsupported codes first.

[00:36:28] Speaker 06:
And on an average basis, there's no question they would be higher.

[00:36:31] Speaker 06:
That's what their own post-judgment study found.

[00:36:33] Speaker 06:
It found that 79% when they removed unsupported codes, they read the data, 79% of the codes went up.

[00:36:40] Speaker 06:
The average value went up.

[00:36:42] Speaker 03:
I read the reply brief.

[00:36:43] Speaker 03:
The difference is it's very much smaller than your brief makes out.

[00:36:48] Speaker 06:
So no, Your Honor, it was actually quite material.

[00:36:50] Speaker 06:
So they committed, let me explain.

[00:36:52] Speaker 06:
They found that 34% of their codes were unsupported.

[00:36:58] Speaker 06:
They ran a calculation trying to figure what the beneficiary impact of that was.

[00:37:02] Speaker 06:
And they committed a very basic mathematical error when they did that.

[00:37:05] Speaker 06:
They came up with 3%.

[00:37:07] Speaker 06:
It's in fact 14%.

[00:37:08] Speaker 06:
If Your Honor wants to look at the comments on this, their footnote five in AHIP's brief points you to the rulemaking.

[00:37:15] Speaker 06:
or multiple experts pointed that out.

[00:37:17] Speaker 06:
But I would say, Your Honor, 3% in this program.

[00:37:19] Speaker 06:
But that's not this rulemaking, is it?

[00:37:21] Speaker 06:
No, it's not.

[00:37:22] Speaker 03:
No, it's not.

[00:37:23] Speaker 03:
This is the basic threshold question.

[00:37:25] Speaker 03:
Anyway, why don't you explain to me, on remand, your two different scenarios in which we got a little bit sidetracked, in which you think they could fix, how they could fix this.

[00:37:34] Speaker 06:
OK.

[00:37:35] Speaker 06:
So I think Judge Collier stated, OK, number one, and this is the one that Judge Collier in footnote nine said was the preferable approach, was to

[00:37:44] Speaker 06:
Figure out its rates using only supported codes.

[00:37:48] Speaker 06:
It's not hard.

[00:37:49] Speaker 06:
It runs this model once every few years.

[00:37:51] Speaker 06:
All it has to do is take a robust sample of its data.

[00:37:55] Speaker 06:
And at that point, figure out what an appropriate payment amount is for data that doesn't contain unsupported codes.

[00:38:02] Speaker 06:
If it does that, Your Honor, then all this goes away.

[00:38:06] Speaker 06:
Every single, at that point, every single code that the plan identifies is absolutely an overpayment.

[00:38:13] Speaker 06:
Why?

[00:38:13] Speaker 06:
Because they were paid on the system, but computed costs based on supported codes.

[00:38:18] Speaker 06:
And if you've got an unsupported code, of course it's an overpayment.

[00:38:21] Speaker 06:
More than that, CMS at that point could affirmatively make you look for the codes.

[00:38:26] Speaker 06:
It would, it would be able to bring false claims that cases without doing audits plan wide audits, it would do audits of no plan.

[00:38:34] Speaker 06:
All they would have to do is one audit of itself every few years to set the rates.

[00:38:39] Speaker 06:
And your honor, I think it's important to say this would also be retrospective that the

[00:38:45] Speaker 06:
Overpayment rule in subsection F has a six year look back provision.

[00:38:50] Speaker 06:
Nobody challenged that.

[00:38:51] Speaker 06:
So if on remand they take this route and this is how they decide to fix to achieve equivalence that will have con they can make you go back six years.

[00:38:59] Speaker 06:
So this is Solves all their problems that they seem concerned about.

[00:39:05] Speaker 06:
The only response Judge Pillard on this proposal of ours in their brief with respect makes no sense.

[00:39:11] Speaker 06:
It's a carryover sentence or two between pages 23 to 24 of the reply brief and they say that if they were to audit their data

[00:39:20] Speaker 06:
it would then introduce an inequivalence or mismatch because their data would be at that point audited and ours would be not.

[00:39:27] Speaker 06:
Not at all.

[00:39:28] Speaker 06:
We're saying if you do that route, if you go that route, that option one, you can make us audit our entire data.

[00:39:35] Speaker 06:
There wouldn't be a mismatch and there wouldn't be all these issues.

[00:39:39] Speaker 06:
The only issues that are occurring here, Your Honor, is that for now CMS has not done that option.

[00:39:47] Speaker 06:
The second option that CMS could choose to do

[00:39:50] Speaker 06:
which would be entirely its judgment as a policy matter.

[00:39:54] Speaker 06:
It might decide, Judge Pillard, that it actually prefers, I actually think there's a chance of this.

[00:39:59] Speaker 06:
They might prefer as a policy matter a world in which they continue to set payments the way they do now, in which they use both because it's simpler to do it and because it results in lower upfront payments.

[00:40:15] Speaker 06:
And that the way instead they will achieve equivalence is through some kind of adjustment mechanism.

[00:40:21] Speaker 06:
Now I agree, your honor.

[00:40:23] Speaker 06:
If that's the route CMS chooses at that point in order for a plan to know it's been overpaid, there would have to be auditing done to determine the rate of unsupported codes in its data.

[00:40:35] Speaker 06:
But that would only happen if CMS made the policy determination that of the two options, it defers that method.

[00:40:43] Speaker 03:
But that's where it is now.

[00:40:45] Speaker 03:
And that's what you're saying.

[00:40:46] Speaker 03:
That is the world today, right?

[00:40:48] Speaker 03:
So basically, what you're saying is under the world today, where they continue with setting their payments as they do now,

[00:41:00] Speaker 03:
without doing a super comprehensive audit, and then they have, oh, they don't have the adjustment that you're calling for.

[00:41:08] Speaker 03:
That's what the RAD-V and the FFS was about.

[00:41:12] Speaker 03:
So my question is, that's a different case and a different rule.

[00:41:16] Speaker 03:
It's a whole RAD-V thing.

[00:41:19] Speaker 03:
So it is your view that the only way they can have an overpayment rule is if they have

[00:41:28] Speaker 03:
let's just call it perfect data.

[00:41:32] Speaker 03:
That basically is your position.

[00:41:34] Speaker 06:
No.

[00:41:34] Speaker 06:
So one option would be, if they recalibrate, that's option one.

[00:41:38] Speaker 06:
Option two would be, if they create an adjustment mechanism, tell plans what their rate of error is that are built in, and create a mechanism for a plan to know at that point, OK, I see an unsupported code.

[00:41:50] Speaker 06:
Does that mean I've actually been overpaid?

[00:41:51] Speaker 06:
And the math here, Your Honor.

[00:41:53] Speaker 03:
So wait, wait.

[00:41:53] Speaker 03:
But that's the step that I'm missing.

[00:41:55] Speaker 03:
So they have what they currently have.

[00:41:58] Speaker 03:
And there's an adjuster.

[00:41:59] Speaker 03:
And the adjuster is where?

[00:42:03] Speaker 03:
Because then how would you know when you have an overpayment whether it's already taken account of in the adjuster or not?

[00:42:10] Speaker 03:
There's no such thing in your worldview in that scenario for an individual overpayment.

[00:42:16] Speaker 03:
problem, right?

[00:42:17] Speaker 03:
Because it's all dealt with on average.

[00:42:19] Speaker 06:
No, not at all, Your Honor.

[00:42:21] Speaker 06:
It could be absolutely identifying an overpayment.

[00:42:23] Speaker 06:
So let's say, Your Honor, that CMS, I mean, there's multiple ways that plans are proposed how this adjustment could work.

[00:42:29] Speaker 06:
This is something that the agency would decide on remand.

[00:42:31] Speaker 06:
And as I said, it would still be able to fix it for six years going back.

[00:42:35] Speaker 06:
So this is not in perpetuity.

[00:42:38] Speaker 06:
But let's say, Your Honor, a plan has looked at its data and says 25% of my code is reported.

[00:42:45] Speaker 06:
and it turns out that 20% of CMS's codes are unsupported, then that means the plan has been overpaid by that difference.

[00:42:53] Speaker 03:
Whoa, whoa, whoa, whoa, whoa.

[00:42:54] Speaker 03:
You're making an assumption about unsupported codes and which direction they go.

[00:42:58] Speaker 03:
We need a lot more data to figure that one out, right?

[00:43:01] Speaker 06:
No, I don't think so, Judge, with respect.

[00:43:04] Speaker 06:
If you hold everything else constant, with respect to unsupported code, mathematically, there really is no

[00:43:12] Speaker 06:
disputing what the average impact is.

[00:43:16] Speaker 06:
If I've given a fixed amount of costs, that's what CMS has, if I distribute them across more codes, the value of an average condition has to go down.

[00:43:27] Speaker 03:
And the way you know that is... I just don't even follow that because you don't double count individuals.

[00:43:32] Speaker 03:
Either I'm an individual with diabetes or I'm not.

[00:43:35] Speaker 03:
Right?

[00:43:36] Speaker 03:
And if I'm not, then I'm at the base rate.

[00:43:39] Speaker 03:
And if I'm an individual with diabetes, and I'm in traditional fee-for-service, and I'm an individual with diabetes, then I'm at the base rate plus.

[00:43:46] Speaker 03:
Right?

[00:43:47] Speaker 03:
I mean, it's not like you're creating more people.

[00:43:50] Speaker 06:
No, but that's exactly the point.

[00:43:51] Speaker 06:
So your fixed number of people, your fixed number of costs.

[00:43:55] Speaker 06:
What this model is asking is,

[00:43:58] Speaker 06:
Mr. Marin, what is the average incremental impact of having a particular condition on the expected cost of insuring you?

[00:44:07] Speaker 06:
The overwhelming driver of that question in a universe with a number of people are fixed and the per capita cost is fixed and known, say 10,000 per person across the population.

[00:44:17] Speaker 06:
How many conditions do I have?

[00:44:18] Speaker 06:
If I have four conditions in my data per person, the average cost of a condition across the whole system is $2,500.

[00:44:26] Speaker 06:
If I have... No, no.

[00:44:27] Speaker 06:
That seems totally wrong to me.

[00:44:32] Speaker 06:
Well, Judge, I mean, it's basic division and it's what their own study found.

[00:44:37] Speaker 06:
The 79% go up when you remove unsupported codes.

[00:44:41] Speaker 06:
You're dividing by constant.

[00:44:42] Speaker 06:
You're asking... And the other way you know it is because the way they're too... Wait, wait, wait.

[00:44:46] Speaker 03:
But let me just do this in common sense terms.

[00:44:48] Speaker 03:
I'm a CMS, and I'm doing fee for service.

[00:44:54] Speaker 03:
And a bunch of the doctors, because they're just doing services, they're not necessarily scrupulous about whether they say the person's diabetic or not.

[00:45:01] Speaker 03:
And so let's say 25% of the people that are getting fee for service under traditional Medicare aren't identified as diabetic.

[00:45:12] Speaker 03:
25% of the diabetics aren't identified as diabetic.

[00:45:15] Speaker 03:
So the adjuster is set based on the 75% that are coded as diabetic.

[00:45:22] Speaker 03:
We don't know whether in that 25% that are not coded as diabetic, whether there are a lot of people who are cheaper than average to treat.

[00:45:32] Speaker 03:
or a lot of people more expensive than average to treat.

[00:45:34] Speaker 03:
And I would think, I mean, just again, by hypothesis, that they're probably the people that are cheaper than average to treat because their diabetes is not front and center when they're going to the doctor.

[00:45:45] Speaker 03:
And so actually the under coding or the failure to code there is having an upward effect on the rate adjuster, no?

[00:46:02] Speaker 06:
No, I think Your Honor is conflating two types of errors there.

[00:46:07] Speaker 06:
And if I may try to unpack that, undercoding and overcoding.

[00:46:10] Speaker 06:
And there's two very different issues.

[00:46:13] Speaker 06:
Unsupported codes, which is the issue dealt with by the overpayment regulation, deals with the situation where the code was submitted.

[00:46:22] Speaker 06:
So there's a diagnostic code for diabetes.

[00:46:25] Speaker 06:
But the code is not, in fact, in the medical chart.

[00:46:30] Speaker 06:
Namely, so the person doesn't have diabetes.

[00:46:33] Speaker 06:
Exactly.

[00:46:34] Speaker 06:
So exactly for all the reasons you said, Judge, the effect of adding these, of the presence of these codes of people whose code says diabetes, they don't actually have diabetes, is to dilute the average cost of diabetes when CMS calculates it.

[00:46:52] Speaker 06:
Because it's distributing the actual cost it incurs

[00:46:56] Speaker 06:
over a set of people, some of whom do not have the condition.

[00:47:00] Speaker 03:
But this is the code being submitted.

[00:47:03] Speaker 03:
That only happens on the MAO side, because that's how the MAOs get paid.

[00:47:09] Speaker 03:
On the FFS side, you're getting paid for service.

[00:47:13] Speaker 03:
And what you're saying is, yeah, but the services that this expensive pillard is getting

[00:47:21] Speaker 03:
are not being understood as services for a diabetic because the doctor failed to put that in the code or conversely called her diabetic because they talked about her parents' diabetes in a session and that she doesn't have it.

[00:47:45] Speaker 03:
Could be either way that there's a ton of error where there's unsupported codes or I guess this is your point that the unsupported codes is a one way problem.

[00:47:55] Speaker 03:
And I'm saying, well, in the data, there's going to be Failure to code has both unsupported codes and

[00:48:08] Speaker 03:
Failure to code, like those are both errors.

[00:48:12] Speaker 03:
And what I haven't seen, and it's your burden, what I haven't seen is any evidence or statistical analysis that says given error in traditional FFS, it's much more likely that we're gonna have unsupported codes, no, that the one you don't like is failure to code.

[00:48:36] Speaker 03:
Right?

[00:48:37] Speaker 03:
And when FFS does it, what you're worried about is failure to code because, well, again, we don't know whether the person you fail to code is above average expensive or below.

[00:48:47] Speaker 06:
So there are these two issues.

[00:48:50] Speaker 06:
With respect to, you asked, Your Honor, if there are these codes in Medicare.

[00:48:53] Speaker 06:
Absolutely.

[00:48:54] Speaker 06:
Even though Medicare providers on traditional Medicare are generally paid for services, they do submit the codes, and that's how CMS runs its data.

[00:49:02] Speaker 06:
Now, the reason that, with respect to these two types of errors,

[00:49:08] Speaker 06:
Is the issue of over coding relevant or I'm sorry of under reporting.

[00:49:14] Speaker 06:
So you say, well, maybe in CMS aside, there's also there's some under reporting going on that's offsetting But there's two things there.

[00:49:20] Speaker 06:
Number one is CMS already accounts for that.

[00:49:24] Speaker 06:
That's what exactly the coding intensity adjuster does it says we on CMS aside have more

[00:49:32] Speaker 06:
unreported conditions.

[00:49:34] Speaker 06:
Our data is less complete than yours, and we account for that.

[00:49:39] Speaker 06:
And Your Honor, when you consider, and so it's already accounted for.

[00:49:42] Speaker 06:
Now, the government seems to say two different things about that.

[00:49:46] Speaker 06:
One it says is, well, maybe the coding intensity adjuster is too low on an average industry-wide basis.

[00:49:51] Speaker 06:
That's an extraordinary claim to make, because they set it at a minimum, the Congress sets a minimum, they can go as high as they want.

[00:49:59] Speaker 06:
If the coding intensity adjuster is too low, they can account for it.

[00:50:02] Speaker 06:
So what they really are saying in a more sophisticated way is that it's set on an average basis.

[00:50:07] Speaker 06:
So for some plans who are more than average, it's under compensating.

[00:50:12] Speaker 06:
But of course, by the same token necessarily half of the plans.

[00:50:16] Speaker 06:
It's they're over compensating because half of the plans will be by definition.

[00:50:20] Speaker 06:
If you're setting an average amount

[00:50:21] Speaker 06:
You're over adjusting for half and you're under adjusting.

[00:50:24] Speaker 03:
Or they're setting it low because they don't want to do that.

[00:50:27] Speaker 03:
And then they have to deal with the practices of plans that are more aggressive in finding codes.

[00:50:34] Speaker 06:
Well, so CMS is under our obligation to set it correctly in a JA 212.

[00:50:38] Speaker 03:
Oh, you just explained why there is no one correct.

[00:50:41] Speaker 06:
No, no, no, no, no.

[00:50:43] Speaker 06:
On an average basis, there is one correct.

[00:50:45] Speaker 06:
Yeah.

[00:50:46] Speaker 06:
On an average basis.

[00:50:47] Speaker 06:
So if the issue is particular plans might be to do a more aggressive and average, and that's an offsetting factor, fine.

[00:50:54] Speaker 06:
That's something they can account for.

[00:50:56] Speaker 06:
We have no objection to that.

[00:50:58] Speaker 06:
At page 49 of their brief in the opening brief, they basically said and came almost to admission.

[00:51:03] Speaker 06:
But in order to know in a particular case where there was an overpayment, you have to compare the relative rates of under and overreporting.

[00:51:10] Speaker 06:
fine.

[00:51:11] Speaker 06:
Do that.

[00:51:12] Speaker 06:
That's not inconsistent whatsoever with our position.

[00:51:14] Speaker 06:
Now, but they don't do that.

[00:51:17] Speaker 06:
Now, a couple of things, Your Honor.

[00:51:20] Speaker 06:
The government said that its FFS adjuster for RADV is just a proposal.

[00:51:28] Speaker 06:
That's actually a complete misdescription.

[00:51:31] Speaker 06:
It's a 394 of a joint appendix.

[00:51:33] Speaker 06:
It was a final methodology.

[00:51:35] Speaker 06:
It was proposed at 292 where they put it out for comment.

[00:51:38] Speaker 06:
Their current position is you have to, that in order to know, CMS' own position currently as we speak is in order to know if a plan has been overpaid, you have to compare the relative rates of error.

[00:51:48] Speaker 06:
There's a proposal to change their view on that and that's fine, but the current position of the agency is that way.

[00:51:54] Speaker 06:
Now, they also said,

[00:51:56] Speaker 03:
with respect to... You're talking about RAD-V and the FFS adjuster?

[00:52:00] Speaker 06:
Correct.

[00:52:00] Speaker 06:
Now, in RAD-V.

[00:52:01] Speaker 06:
Now, in the coding intensity adjuster, Your Honor, this is a JA 177 178, because we're saying, look, there's two types of errors, and I agree with that.

[00:52:09] Speaker 06:
Consider how drastically different CMS's treatment is of one type of error versus the other.

[00:52:15] Speaker 06:
When it comes to an error that works against its favor, it would have plans paid more

[00:52:22] Speaker 06:
compared to CMS, which is the issue of underreporting.

[00:52:27] Speaker 06:
CMS says at 177, 178, and it quotes all the congressional committees, that in order for the payments to be accurate, it has to account for that differential.

[00:52:36] Speaker 06:
That even though the codes and plans are adding are accurate in an absolute sense, because they're supported by the medical charts, if CMS doesn't account for the practice of plans to add codes, when CMS does not, it will lead to inaccurate payment.

[00:52:52] Speaker 06:
because you have to have consistent risk scores across both programs.

[00:52:56] Speaker 06:
If Mary Jane is a 1.0 when she's on CMS, she needs to be a 1.0 when she's on Medicare Advantage for the payment to be accurate.

[00:53:03] Speaker 06:
When it comes to this other type of error, where an adjustment would be to our benefit, suddenly it adopts an absolutist stance.

[00:53:10] Speaker 03:
Which type of error are you talking about now?

[00:53:12] Speaker 06:
Unsupported codes.

[00:53:13] Speaker 03:
The RADV?

[00:53:13] Speaker 06:
No.

[00:53:14] Speaker 06:
The issue of unsupported codes that are identified over payment rule.

[00:53:18] Speaker 06:
Yes.

[00:53:19] Speaker 06:
When it comes to that kind of issue.

[00:53:22] Speaker 06:
This is where you're not going to make an adjustment.

[00:53:24] Speaker 06:
We're not going to even compare.

[00:53:26] Speaker 03:
This is where I just don't understand how the statute what your theory is.

[00:53:31] Speaker 03:
of how actuarially equivalence applies.

[00:53:34] Speaker 03:
Actuarial equivalence, as I read the text of the statute, is about the setting of the risk adjustment factor.

[00:53:43] Speaker 03:
And it seems like you're really doing this tail wagging the dog where you're saying, you come after me for an overpayment.

[00:53:50] Speaker 03:
I acknowledge that actually this person doesn't have diabetes.

[00:53:55] Speaker 03:
But you can't come after me for this individual overpayment until your records are all clean, because maybe if my payments are rendered more accurate,

[00:54:13] Speaker 03:
by my recovery from the agency is rendered more accurate by relinquishing overpayments, I will actually be underpaid because of a systemic inaccuracy in the risk factor, right?

[00:54:30] Speaker 03:
And the question is, why isn't that just a case about the accuracy of the risk factor?

[00:54:35] Speaker 03:
Why is this coming in through the tail of

[00:54:40] Speaker 03:
of the overpayment.

[00:54:42] Speaker 03:
I mean, you've never objected to the overpayments until the Affordable Care Act says that you have 60 days, and if you find it, you have to tell us.

[00:54:51] Speaker 03:
That's the procedural part that Mr. Shah is talking about.

[00:54:54] Speaker 03:
I just don't see a world in which, and you didn't actually get to finish your second example, which is the situation where there is imperfect data on both sides,

[00:55:09] Speaker 03:
There is an adjuster.

[00:55:10] Speaker 03:
And then how could there also be an overpayment rule?

[00:55:15] Speaker 03:
There couldn't, could there?

[00:55:17] Speaker 06:
Absolutely, there could be.

[00:55:18] Speaker 03:
I don't understand that.

[00:55:19] Speaker 06:
Because if my errors exceed those that were assumed in the model,

[00:55:25] Speaker 06:
Then, well, in the payment model and if it exceeds that adjustment amount and absolutely and that's the concern here that if a plan is jacking up its its codes it's it's creating errors.

[00:55:36] Speaker 06:
On that are that are different than those and CMS is data, then the problem is at that point.

[00:55:43] Speaker 06:
Its data overstates the health of its population compared to CMS because CMS spread its costs over a certain number of codes and now you've got more errors on yours.

[00:55:52] Speaker 06:
They can absolutely address that by doing it.

[00:55:55] Speaker 06:
Again, I think the simpler and better approach is to adjust the rates and you can do it going back.

[00:56:04] Speaker 06:
In terms of a statutory question, the way they're linked, Your Honor, is this.

[00:56:08] Speaker 06:
The overpayment statute says an overpayment is funds to which you're not entitled under the Medicare program.

[00:56:16] Speaker 06:
So then that asks the question, it begs the question, what are the funds to which you're entitled under this program?

[00:56:22] Speaker 06:
And that's what the actual equivalence provision says.

[00:56:24] Speaker 03:
I understand your policy argument that that would be perhaps permissible, although that would be perhaps permissible if CMS said we want to do it that way.

[00:56:32] Speaker 03:
No, I think it's a statute really require it.

[00:56:36] Speaker 06:
I think so, your honor.

[00:56:37] Speaker 06:
So, and so first of the statute says the secretary shall adjust the payment amount the payment amount shall adjust the payment amount to achieve actual equivalent

[00:56:47] Speaker 06:
CMS consistently in all the briefing below, and I don't see it disputing it here, says what actual equivalence means is that the payment has to be equal to, they have to pay plans the same as they would expect to incur to ensure an identical person.

[00:57:01] Speaker 03:
Right, and we're talking in gross.

[00:57:03] Speaker 03:
They adjust the payment amount by the risk factors in the general system.

[00:57:10] Speaker 06:
But the payment amount is always distributed per person, just so you'll understand.

[00:57:14] Speaker 06:
And so what they're saying is, if it costs them on average $1,000 for diabetes, as long as they set that payment amount at $1,000, they've satisfied actuarial equivalence, and now you give back $400 on the back end, so you're left with a net $600, no inequivalence.

[00:57:33] Speaker 06:
That cannot be right.

[00:57:34] Speaker 03:
No, but there's no an equivalence because, A, you've never shown that there's anything statistically significant about the errors that you're assuming are in their database.

[00:57:46] Speaker 03:
So we don't have a reason not to think that this isn't your scenario one as opposed to your scenario two.

[00:57:52] Speaker 03:
It's your burden.

[00:57:53] Speaker 03:
So if there's some statistical significance in the direction of the errors in the FFS database and that it's actually relevant to the

[00:58:03] Speaker 03:
the condition that you're talking about.

[00:58:07] Speaker 03:
No showing on that.

[00:58:07] Speaker 03:
And I understand you don't have the data.

[00:58:09] Speaker 03:
It's a CMS data.

[00:58:10] Speaker 03:
But it is your burden to establish that.

[00:58:13] Speaker 03:
And I see no statisticians in the record saying anything about that.

[00:58:17] Speaker 03:
I mean, it does seem like, and that's why I keep coming back to it, this seems like a complaint about the Radview system and the FFS adjuster.

[00:58:25] Speaker 03:
That's where all your argument.

[00:58:27] Speaker 06:
Your Honor, when you say there's nothing in the record, the American Academy of Actuaries at 392 told CMS the presence of unsupported codes in its data

[00:58:35] Speaker 06:
Under have the effect of lowering the payment amounts at JA 601 the the head of CMS is internal entity that is in charge of payment and an overpayment recoveries.

[00:58:49] Speaker 06:
Jennifer Harlow analyzed and agreed that plans would be underpaid

[00:58:55] Speaker 06:
if you recovered every single- This is all about RADV.

[00:58:59] Speaker 03:
This is not about overpayment rule.

[00:59:01] Speaker 03:
We don't have anything challenging the overpayment rule on these grounds.

[00:59:04] Speaker 03:
We have no evidence.

[00:59:05] Speaker 06:
But Your Honor, the math is the same.

[00:59:07] Speaker 06:
In other words, the only difference between an overpayment and this is that in RADV, they say we're extrapolating all the errors, and here it's just identified.

[00:59:15] Speaker 06:
So let me address that, because I think the government's being very- Go ahead.

[00:59:20] Speaker 06:
Your Honor, the government is suggesting that because

[00:59:24] Speaker 06:
Judge Collier has vacated the definition of identify that now my colleague said that there's no requirement that you affirmatively look for errors.

[00:59:34] Speaker 06:
All we're saying is if you already have discovered them.

[00:59:37] Speaker 06:
That is absolutely disingenuous and contrary, directly contrary once again, to representations the government made below.

[00:59:44] Speaker 06:
So the government moved to dismiss when we filed our complaint for lack of standing.

[00:59:48] Speaker 06:
And what it argued was that even if the definition of identify were completely eliminated, to eliminate any requirement to exercise reasonable diligence and affirmative compliance,

[01:00:01] Speaker 06:
That would do nothing.

[01:00:02] Speaker 06:
It would do nothing to alter a plan's affirmative obligation to look for overpayments, because that would independently exist by virtue of 42 CFR 422.503 B4VI.

[01:00:15] Speaker 06:
All right.

[01:00:19] Speaker 04:
We're way over.

[01:00:23] Speaker 04:
Judge Walker, do you have any further questions?

[01:00:27] Speaker 04:
You want to ask Mr. Marin?

[01:00:29] Speaker 02:
I do.

[01:00:32] Speaker 02:
How about two?

[01:00:33] Speaker 02:
Because I know we're way over.

[01:00:35] Speaker 02:
But the government says that the coding intensity adjuster factors in over-reporting and under-reporting.

[01:00:45] Speaker 02:
And Mr. Shaw gave me a page to go look at to confirm that.

[01:00:49] Speaker 02:
Do you disagree with that?

[01:00:52] Speaker 06:
I do.

[01:00:53] Speaker 06:
First of all, there's no evidence for that.

[01:00:56] Speaker 02:
Where do I go to look for confirmation that you're right?

[01:01:00] Speaker 06:
Well, so first of all, in page 11 of their opening brief, they said the exact opposite.

[01:01:04] Speaker 06:
They said that the coding intensity adjuster does not account for plans treatment of unsupported code.

[01:01:11] Speaker 06:
Flat out, that's what they said in page 11 of their opening brief.

[01:01:14] Speaker 06:
But Judge Walker, I think the more, I'm not asking you to make a factual finding.

[01:01:19] Speaker 06:
Our point legally, however, is that

[01:01:22] Speaker 06:
All they're really saying there is another version of the coding intensity adjuster might be too low if you are a plan that is removing a lower percentage of codes than the industry average that's built into the adjuster.

[01:01:37] Speaker 06:
And that simply means there's an additional thing to take into account in determining whether the unsupported code in fact is an overpayment.

[01:01:45] Speaker 06:
So fine, if they're right, they can take that into account.

[01:01:48] Speaker 02:
Okay, and then the last question is a follow up on some of your dialogue with Judge Pillard.

[01:01:57] Speaker 02:
You gave two scenarios where you think things would be fair.

[01:02:01] Speaker 02:
I'm gonna ask about the second one where the government continues to use in terms of setting the factors, what you say, the inaccurate data.

[01:02:12] Speaker 02:
I think, tell me if this is your theory.

[01:02:16] Speaker 02:
Imagine there are 20 patients

[01:02:19] Speaker 02:
and with 20 diabetes codes.

[01:02:22] Speaker 02:
And we know that 16 of those are real and four of those are over reported.

[01:02:32] Speaker 02:
Now imagine that traditional Medicare only catches a quarter of false codes.

[01:02:39] Speaker 02:
That means that 19 diabetes codes make it through the traditional Medicare data.

[01:02:47] Speaker 02:
Now imagine that United

[01:02:48] Speaker 02:
catches three quarters of false codes.

[01:02:53] Speaker 02:
I think what you're saying is that United should be allowed to bill for all of the 16 true codes, of course.

[01:03:01] Speaker 02:
They should also be able to bill for the one false code that it didn't catch because of course they can't know that it's false.

[01:03:16] Speaker 02:
And of the three codes they did catch, they should be able to bill for two of those three.

[01:03:21] Speaker 02:
Now, that would come up to 19, which is the same number that made it into the traditional Medicare data of false codes.

[01:03:32] Speaker 02:
Now, my question is, is that your theory?

[01:03:37] Speaker 02:
And explain to me, if it is your theory, how that would be practical.

[01:03:42] Speaker 02:
That's it.

[01:03:45] Speaker 06:
Yeah, this is very complicated.

[01:03:48] Speaker 06:
But essentially, yes, that there has to be some accounting for the presence of codes that are unsupported at CMS's data.

[01:03:54] Speaker 06:
Again, in terms of practicality, our position- Let me make sure I have it right.

[01:03:59] Speaker 02:
So if you catch three false codes, you would get to bill for two of them, and you wouldn't bill for the one because

[01:04:06] Speaker 02:
traditional Medicare would have caught that one.

[01:04:08] Speaker 02:
Am I right about, that's what you want the rule to be.

[01:04:10] Speaker 02:
That's the overpayment rule that you want to exist, assuming that the first option you gave Judge Pillard isn't the option Medicare chooses.

[01:04:19] Speaker 06:
Okay, assuming it doesn't go that route, then it has to figure out some way to do this.

[01:04:22] Speaker 06:
And the reason why, because I understand you said false code.

[01:04:27] Speaker 06:
It's easy to think of that.

[01:04:28] Speaker 02:
Well, overreported code.

[01:04:31] Speaker 02:
And we're so late in the time.

[01:04:34] Speaker 02:
In the hypothetical where the government catches one out of four false codes, you catch three out of four false codes, you would have to give back the money for one, but you would get to keep the money for two of the codes that you know are false.

[01:04:51] Speaker 02:
Is that your theory?

[01:04:53] Speaker 06:
That is one possibility.

[01:04:55] Speaker 06:
There are multiple ways an adjustment mechanism could work.

[01:04:57] Speaker 06:
It could work as a credit system on dollars.

[01:05:00] Speaker 06:
There's multiple ways CMS could solve this problem.

[01:05:03] Speaker 06:
It's really not that complicated for it to solve, but there are multiple ways it could solve it.

[01:05:07] Speaker 02:
And I think you said- Maybe not too complicated for you, but it's complicated for poor people like me who went to law school because there would be no math.

[01:05:18] Speaker 06:
Well, and I appreciate that, Your Honor, but here what the agency did is

[01:05:22] Speaker 06:
We're not even going to try.

[01:05:24] Speaker 06:
We're not even going to consider to what extent our payments already account for and build in for this level of error.

[01:05:31] Speaker 06:
And even if you're right, United, that they build in fully for the errors in your data, it does not matter.

[01:05:38] Speaker 06:
Actual equivalence doesn't come into play, and we get to make you return every one of those unsupported codes.

[01:05:43] Speaker 06:
And I think we gave an example of a 72-year-old female with multiple sclerosis and the unsupported diabetes code.

[01:05:49] Speaker 06:
which I think in one instance, that's a specific person.

[01:05:52] Speaker 06:
It's not population-wide, Judge Pillard.

[01:05:54] Speaker 06:
And it shows that even as to that one person, if you're not paid for that unsupported diabetes code, given the way the payments were set, you're going to be underpaid.

[01:06:03] Speaker 04:
All right.

[01:06:03] Speaker 04:
Let us conclude.

[01:06:06] Speaker 04:
Mr. Marin, is there any essential point you need to make that's not in your brief or that you have not responded to?

[01:06:14] Speaker 04:
in the question.

[01:06:14] Speaker 04:
No, thank you, Your Honor.

[01:06:15] Speaker 06:
You've been very generous with your time.

[01:06:16] Speaker 06:
Thank you.

[01:06:17] Speaker 04:
It's complicated, as you say.

[01:06:20] Speaker 04:
OK, counsel for appellants, why don't you take less than five minutes?

[01:06:31] Speaker 01:
OK, thank you very much.

[01:06:32] Speaker 01:
I do appreciate that.

[01:06:33] Speaker 01:
I'd like to begin with Judge Walker's hypothetical.

[01:06:36] Speaker 01:
I think that hypothetical accurately characterizes what

[01:06:42] Speaker 01:
United is asking for, which is sort of a free pass on errors that it discovers so that it can make up what it thinks it lost relative to what it should have had.

[01:06:53] Speaker 01:
But there's one missing aspect of that hypothetical, which is that in discovering those two or three invalid codes, United Health has to look at the medical record.

[01:07:05] Speaker 01:
And when they look at the medical record, they are also going to look for

[01:07:09] Speaker 01:
under-reported codes, codes that for diabetes diagnoses that are actually valid but were not previously reported.

[01:07:16] Speaker 02:
I get that.

[01:07:16] Speaker 02:
They say the coding intensity adjuster controls for that.

[01:07:20] Speaker 02:
You say that it doesn't.

[01:07:23] Speaker 02:
I'm not sure who's right yet.

[01:07:25] Speaker 01:
OK, yeah.

[01:07:28] Speaker 01:
So I think it is true that the coding intensity adjustment accounts for the net effect, and it doesn't capture all of the increase that insurers are able to achieve.

[01:07:37] Speaker 01:
Something else I want to discuss is

[01:07:40] Speaker 03:
average effect, not the net effect.

[01:07:43] Speaker 01:
Right, the average net effect of having insurers both report additional diagnoses and delete invalid ones.

[01:07:50] Speaker 01:
So it's not just accounting for under-reporting or over-reporting, it's addressing both.

[01:07:55] Speaker 01:
I think in the end, what this comes down to is the secretary has substantial discretion to implement the risk adjustment model and decide what factors need to go in, what needs to be adjusted for, and what does not.

[01:08:10] Speaker 01:
The reason that UnitedHealth says that the secretary needs to do something in particular to take into account this effect that it's talking about is that it says that there is a fundamental asymmetry in treatment between traditional Medicare and Medicare Advantage, and that's just not true.

[01:08:28] Speaker 01:
In both cases, in both programs, all diagnoses codes need to be supported by medical record documentation.

[01:08:36] Speaker 01:
And there's actually even an overpayment rule that also applies to parts A and B. So for example, if a physician is treating in traditional Medicare, let's say part B, if a physician is treating a patient for cancer or seeks reimbursement for cancer treatment, it has to report a cancer diagnosis code on that claim or else the claim will be rejected as not medically necessary.

[01:09:00] Speaker 01:
Now, if the physician collects payment for that treatment, and then it later turns out that the patient doesn't have cancer, then that payment is an overpayment, and the physician has to give the money back.

[01:09:11] Speaker 01:
So that is the same rule that applies in traditional Medicare Advantage.

[01:09:16] Speaker 01:
So the idea that there's fundamentally different treatment is just incorrect.

[01:09:22] Speaker 01:
And it certainly is not the case that CMS requires Medicare Advantage insurers to catch all errors.

[01:09:28] Speaker 01:
I think what UnitedHealth wants to have is a system where they have a method of, they want to be able to say, even after application of the risk adjustment model, according to all the rules that haven't previously been specified, according to the contracts that these insurers have signed, they want to be able to say that, well, there's this separate, actuarially equivalent amount that exists independent of the risk adjustment model.

[01:09:55] Speaker 01:
And we want to be able to,

[01:09:57] Speaker 01:
claim that amount.

[01:09:59] Speaker 01:
So we want to be able to say that the risk adjustment model produced a result that was too low and now we want more money.

[01:10:04] Speaker 01:
And that's not how it works.

[01:10:05] Speaker 01:
The actuarial equivalence requirement is an instruction to the secretary.

[01:10:09] Speaker 01:
It is not an entitlement of any particular insurer to a particular amount of money.

[01:10:15] Speaker 01:
So finally, I think that the idea that CMS can't define something as an overpayment in the event that

[01:10:26] Speaker 01:
all those overpayments recovered that insurers would be underpaid.

[01:10:31] Speaker 01:
I think that's just wrong because the secretary obviously has to design the risk adjustment model to take into account the real world behavior of insurance companies, of beneficiaries and so forth.

[01:10:42] Speaker 01:
And it is not the case that insurers are reporting all erroneous or all the invalid codes that they're finding them and reporting them.

[01:10:50] Speaker 01:
And they're not even coming close.

[01:10:51] Speaker 01:
And we know that from the record,

[01:10:54] Speaker 01:
with respect to some of United's own plans.

[01:10:58] Speaker 01:
And finally, I just want to emphasize that there are always ways that the secretary can improve the risk adjustment model.

[01:11:10] Speaker 01:
Someone could say, well, medical costs are correlated with income.

[01:11:14] Speaker 01:
So you should have an income adjuster to take into account that maybe

[01:11:20] Speaker 01:
Traditional Medicare patients have lower income than Medicare Advantage patients, or some Medicare Advantage plans have higher average income than others.

[01:11:29] Speaker 01:
And so the secretary has to have discretion to be able to decide whether these adjustments make sense in any particular case.

[01:11:37] Speaker 01:
And that's in fact what the statute says.

[01:11:38] Speaker 01:
The statute says that the secretary may add to, modify, or substitute for such adjustment factors if such changes will improve the determination of actuarial equivalence.

[01:11:47] Speaker 01:
Now in this case, I think the secretary reasonably

[01:11:50] Speaker 01:
exercised his discretion not to include the FFS adjuster.

[01:11:54] Speaker 01:
And there's no, there's neither empirical data, which it was United Health's burden to produce that proves that this was impermissible, nor is there any logical or mathematical reason why that the decision not to include an FFS adjuster would necessarily produce underpayment.

[01:12:14] Speaker 01:
And so I think absent either of those things, there's no way United Health can sustain its claim.

[01:12:19] Speaker 03:
So I should have asked this of Mr. Meron.

[01:12:21] Speaker 03:
And actually, with the indulgence of Judge Rogers, he also has it.

[01:12:27] Speaker 03:
But let me ask you, Mr. Shaw, is the elephant in the room here the False Claims Act liability?

[01:12:33] Speaker 03:
I'm trying to make sense of what this case is about.

[01:12:36] Speaker 03:
And is it your understanding that a ruling on the statutory ground would bear on False Claims Act liability or not?

[01:12:48] Speaker 01:
I wouldn't characterize it as the elephant in the room.

[01:12:51] Speaker 01:
Certainly the government has alleged that UnitedHealth has, while mining medical records for additional diagnosis codes, in fact found many of them and collected billions of dollars of payment for those codes, while at the same time ignoring evidence that other codes were invalid.

[01:13:12] Speaker 01:
So that's definitely the government stands by those allegations.

[01:13:16] Speaker 01:
But I don't think that's

[01:13:18] Speaker 01:
sort of the motivation for the government defending the overpayment rule here.

[01:13:24] Speaker 01:
I think the real elephant in the room is that the overpayment rule reflects substantive requirements that have existed since the inception of the risk adjustment model.

[01:13:35] Speaker 01:
The requirement that insurers report diagnosis codes that are supported by medical record documentation, and that CMS can collect payments

[01:13:46] Speaker 01:
or recover payments when those codes turn out to be invalid.

[01:13:49] Speaker 01:
So I think that's the real elephant in the room, which is that if this court were to rule that this longstanding approach is impermissible, that would require really a drastic change to how CMS has operated the risk adjustment process for a long time.

[01:14:03] Speaker 04:
All right.

[01:14:04] Speaker 04:
Do either of my colleagues have any questions of counsel for CMS?

[01:14:11] Speaker 04:
Thank you, Mr. Shaw.

[01:14:12] Speaker 01:
Thank you.

[01:14:13] Speaker 04:
Judge Pillar, did you want to ask Mr. Maron this question about the elephant in the room?

[01:14:19] Speaker 03:
If you haven't answered either of the points that Mr. Shah made about whether this is really about False Claims Act liability and what about the fact that this rule has been longstanding and it's really only the procedures that were added in the Affordable Care Act.

[01:14:39] Speaker 06:
Okay, so let me let me address the second one, because that's a myth with respect to Mr. Shaw.

[01:14:45] Speaker 06:
The industry and CMS have had a disagreement on this issue since 2009, the very and towards the very beginning of the system.

[01:14:52] Speaker 06:
So let me and and there's been no subtle understanding on the relevant question.

[01:14:57] Speaker 06:
The relevant question being, does the existence of an unsupported code in itself mean you've been overpaid?

[01:15:02] Speaker 04:
So the answer to Judge Pillard's question is no.

[01:15:07] Speaker 06:
Correct.

[01:15:08] Speaker 06:
It's not a long standing.

[01:15:08] Speaker 06:
I'd like to, if I could.

[01:15:10] Speaker 04:
The only question she asked, not reopening the entire argument is, is the.

[01:15:20] Speaker 06:
I do not believe it's been a lot.

[01:15:21] Speaker 06:
The answer is no, it's not a long standing interpretation.

[01:15:24] Speaker 06:
I would look, can I cite.

[01:15:25] Speaker 04:
Can I just finish?

[01:15:26] Speaker 06:
Oh, of course, judge Rogers.

[01:15:28] Speaker 04:
The question she asked, and she can correct me if I'm misstating it is.

[01:15:33] Speaker 04:
is false claims liability the elephant in the room?

[01:15:39] Speaker 04:
And why this case is before us in this posture?

[01:15:45] Speaker 06:
OK, so I think there were two questions.

[01:15:46] Speaker 06:
I was answering her second question about whether this was a longstanding interpretation.

[01:15:51] Speaker 03:
I'd love to hear your answer to the false claims act question.

[01:15:54] Speaker 06:
OK, so the parties, there's lots of things the parties disagree on.

[01:15:58] Speaker 06:
This is one of them.

[01:16:00] Speaker 06:
The government's position is that a ruling in our favor on the statutory grounds would have absolutely no impact on the False Claims Act case.

[01:16:06] Speaker 06:
That's the government's position.

[01:16:08] Speaker 06:
We, depending on the scope of it, may disagree on that.

[01:16:11] Speaker 06:
What we don't disagree on is that there are certainly many ways the government, depending on its theory, can pursue False Claims Act claims against the industry

[01:16:20] Speaker 06:
or against participants in the industry, depending on the particular theory and the particular activity.

[01:16:24] Speaker 06:
I mean, it's a very broad question.

[01:16:27] Speaker 06:
There's lots of issues being litigated.

[01:16:28] Speaker 06:
I can't just say yes or no, but I know the government's position is it will have no effect.

[01:16:33] Speaker 06:
With respect to your second question, Judge, just very quickly, the current risk adjustment system was phased in from 2004 to 2007.

[01:16:41] Speaker 04:
That's your point, all right, that there's been this disagreement, all right?

[01:16:46] Speaker 04:
But we're going on and on and on here.

[01:16:49] Speaker 04:
And I don't think much clarity is being added by repetition.

[01:16:53] Speaker 03:
Thank you.

[01:16:54] Speaker 03:
Thank you.

[01:16:54] Speaker 04:
In my opinion, if my colleagues disagree, let's question on.

[01:16:59] Speaker 03:
All right.

[01:17:00] Speaker 03:
Thank you very much.

[01:17:01] Speaker 03:
And thank you, Judge Rogers, for the extra time and the willingness of counsel to answer so many questions today.

[01:17:10] Speaker 06:
Of course.

[01:17:11] Speaker 06:
Thank you.

[01:17:11] Speaker 06:
Thank you very much.

[01:17:12] Speaker 06:
Thank you, Judge Rogers.

[01:17:13] Speaker 02:
Thank you.

[01:17:13] Speaker 04:
Let me just ask, Judge Walker, did you have anything?

[01:17:16] Speaker 02:
I don't.

[01:17:17] Speaker 02:
Thank you, Judge Rogers.

[01:17:18] Speaker 04:
All right.

[01:17:19] Speaker 04:
Thank you, counsel.

[01:17:20] Speaker 04:
We will take the case under advisement.