Oral argument — Maxmed Healthcare, Inc. v. Thomas Price

I have not had a chance to, mostly, generally speaking, to read the whole record before we come in here. So if you can give us record sites, we'll appreciate it. And we will also appreciate your taking note of the time schedule on our clock. The yellow light means you have two minutes left in your argument. When the red light comes on, please conclude unless, quickly, unless you're answering one of the court's questions. So, with that, I will call the first case, number 16-50398, Maxmed Healthcare v. Price. Ms. Terrell. May I please go? Yes, ma'am. My name is Lourise Terrell. I represent Maxmed Healthcare, Inc., a home health provider in the Medicare system. Are you in San Antonio? Is it located in San Antonio? It is located in San Antonio. The principal is sitting in the gallery, Olu Oye. Who does it, um, does it work out of a particular hospital or, uh, you know? It is a standalone home health agency. Okay. My office address, I am not, I don't have in front of me. Okay. And they send nurses out from there. I'm from, I'm from San Antonio, so I was just interested. The issues presented in our brief are whether the district court erred in affirming the decision of the Medicare Appeals Council, which I will refer to as the council, to uphold the statistical sampling and extrapolation methodology used to calculate an overpayment assessed against the healthcare provider. Or whether CMS gets to rewrite the rules of statistical mathematics. Our second issue is whether the district court erred in concluding the U.S. Department of Health and Human Services violated their own rule of thumb, which is a prohibition against using screening tools for home health beneficiaries because the beneficiary must qualify for the services under two, um, individualized reasons. They must be homebound, which is a natural inability to leave the home, um, without assistance. They must require assistance to do that. And it's a taxing effort for them to do that. Otherwise, they could go get medical services instead of receive them at their home. They also must require the medical services. And CMS has a rule of thumb, which requires that the home health beneficiary be regarded only for its own, his or her own situation. Because those two requirements make a matrix. It could be an amputation allows them, makes them homebound. It could be, um, other ailments that allows them to be homebound. Can you elucidate a few things for me? I don't, I didn't, how big is MaxMed? How significant is this? This ended up as a $700,000 repayment order, right? Yes. Okay. Plus at 10.35% interest, I believe. My heavens. Okay. And, um, what percentage of the patients of the company does that involve? I do not know at the moment what percentage of the patients that involves. I can find out for you if that is something that you would like to know. Well, I'm sure this background is in the record somewhere. I was just sort of trying to. It is. The sampling was done with just a certain time period, a two-year time period. So they took those two years of clients. 130 claims or clients? 130 claims is how they couched it. That would, that is, I don't know exactly how many clients that is, that constitutes. And I don't know what his current census is. He is, three-quarters of a million dollars is a whole lot of money, no matter who you are. To finish point two, if I may, the rule of thumb, whether CMS gets to write in a rule of thumb and then completely ignore it. Well, they've got a statute that says that requires statistical, or at least permits statistical sampling. There is a statute and the regulations permit statistical sampling. That is true. Whether it's permitted for home health beneficiaries, it could be permitted for durable medical equipment suppliers. It could be permitted for hospitals. It could be permitted for lots of other providers. However, CMS put in their guidelines, which are their interpretation of the regulations and statutes that govern them, that Medicare recognizes that determinations of whether home health services are reasonable and necessary must be based on an assessment of each beneficiary's individual care needs. Therefore, the denial of services based on numerical utilization screens, which would be statistical sampling, diagnostic screens, diagnosis, or specific treatment norms is not appropriate because each beneficiary is homebound and needs medical necessities on their own basis, not on anyone's comparable basis. Are there any cases where that provision that you just read has been held to be over and above the statistical sampling? In other words, when there is a conflict, or even if there is no conflict, then the rule of thumb should be used over the statistical sampling as a matter of priority, or is it simply are you saying that it's better to have done it with the rule of thumb instead of the statistical sampling? I know of no case law that addresses this particularly. There may be some cases in our brief that touch on this or are analogous to this, but we're allowed to sample, but you can't sample these folks because of the uniqueness of their qualifying for the home health services benefit. And this is only about the home health services benefit. That provision is in their Medicare policy benefits manual. Who certified these people? Their physician certified them as... Was there one particular physician who certified them? No, a myriad of physicians. Various ones. And so when you're talking about 130 claims, is that a claim for each visit or a series of visits after the person is certified? It will be a series of visits after this person is certified. Home health works on a 60-day episode. An episode can have a few visits. Home health could be required after a person leaves an institution, a hospital, or a nursing home for training and education on new medication they must take or wound care or things like that, which would be a few visits for education typically, or someone who needs more care such as a wound care patient who needs a dressing change every day in a wound that they can't get to or they can't manage. That would be the episode would have many more visits in that instance, but they're all based on the individual's needs. Could you move back to the statistical problems? Because I had some questions about the sampling that was done. Certainly. The statistical sampling, and why I say the secretary gets to rewrite the rules of math, the statistical sampling parameters are written in the Program Integrity Manual, Section 8. It used to be Section 3. You'll see that reference is now Section 8. They track pretty much the same as Section 3. They are contained in approximately 22 pages of the Program Integrity Manual, and the sampling is just math. You take a small number. Say you want to know the length and weight of 1,000 footballs. NFL footballs are all about the same. How many do you have to measure to figure out what the length and weight will be? For footballs, they're fairly homogenous. It wouldn't take a whole lot, so you may have a small sample size there. For other things like fish, you net a bunch of fish from the bay. That weight and length of 1,000 fish would be much more difficult to calculate, and statistics in math has certain formulas and programs that you follow to determine what you're looking at, see how you need to measure it, measure it for the sample, and then extrapolate. I think I followed what was written. Did Judge Ezra not even have an oral hearing on this case? No. He canceled it just before. The only oral hearing you ever had was a telephonic two-day thing before the ALJ? Yes, Your Honor. So all the rest of this is totally on paper? Yes, Your Honor. Unbelievable. Anyway, as I understood it, the board, was that the MAC, said, oh, yes, well, Dr. Holler, the plaintiff's doctor, which is not accurate, relied too much on textbooks and treatises, and under our guidelines, you don't really have to do that. But did their opinion, which I haven't read, explain how the methodology that was used was sufficiently accurate? No, they did not. Dr. Holler was the administrative law judge's expert. Yes, well, the brief pointed that out, but I think Judge Ezra may have been a little confused about that, too. The counsel doesn't want to be bothered with the textbook, academic texts, and other rules for statistics, which statistics is just an estimate of the end result anyway, and if you follow the proper parameters, you get a reasonableness of the accuracy of the estimate. But the program integrity manual is what the counsel stands on, and they were quoted in Judge Ezra's ruling as, these are simply not flaws. They're referring to the comments that Dr. Holler has made about the inefficiencies and the inaccuracies of the sampling that was done. These are simply not flaws in the sampling cognizable by the guidelines. The PIM doesn't require the academic accuracy that the texts and the tomes do require in mathematics. They have written a different set of requirements, thus negating the accuracy foundations for the end result. Now, are you challenging the program integrity manual as being not accurate or the way in which the sampling was done in this case? It was challenged on the way the sampling was done in this case. The initial sample that was taken did not have enough independent variables for the individual claim lines. They took 40 claims out of 130. And the experts, the two experts, the hearing officer's expert and the petitioner's expert, both said that there were flaws in that sample. If you have a sample that's not representative of the universe or is in some way adversely weighted one way or the other, for or against, then the end result is not within the parameters required for an accuracy confidence level. And that's basically what the experts said. When you say experts, you mean Dr. Howler? Howler and Dr. Whitmer. Now, there was an expert that the health integrity used, I believe, the underlying contractor that calculated this statistical sampling. And that expert, that statistician, said we follow the program integrity manual guidelines. And so what CMS is saying is that as long as you follow our guidelines, even if we don't do it by the academic textbooks, by the rules of math, you can't contest the validity of our extrapolation number. And so they've rewritten the rules of math to allow them to win every time. And there's a due process issue with that, if nothing else. I was about to ask you about that. When was this first raised, the questioning of the methodology first raised? At the initial evaluation or at the ALJ? Initially it was raised. It was raised at the two prior levels of appeal. We have a request for redetermination and then a request for reconsideration, and then we go to the ALJ level. We complained about their accuracy of their statistical extrapolation at each level. We received their documentation of how they did the statistics just before the ALJ hearing, which is my third issue, is if CMS has to show their work, when do they have to show it to us? And even so, we were fairly strapped to get an expert to review it and be ready for testimony. But we prevailed at the ALJ level, and we prevailed on the math side of the extrapolation. The provider still has an overpayment. The actual claims that were reviewed, that is still the overpayment that's provided. My understanding is you're not challenging 258,000 overcharge. It's just the extrapolation? In this case, that is true. So the government still has a large overpayment that they've recovered. But they have recovered it. You're just seeking it back. They've recovered the 258,000. They've also recovered the 775,000. The client is still paying the interest off, 10.35% interest. I think every seven years that doubles. So he's got a ways to go to pay that. However, and I may have misspoken, he may have paid almost all of it off. I don't know how close he is to that payment. I didn't calculate that before I got here. I apologize. Would you say that the record was complete at the time the ALJ got the matter? I did not say that. We were able to get an expert, Dr. Whitmer, to testify. We did not know until we showed up at the hearing that the ALJ had hired an expert of her own. Well, I'm just interested that the way this happens, because if a judge is going to hire an expert, they tell the parties, right? Well, there are no parties. The provider is the only party. Oh, is that right? CMS, unless they announce that they are going to be a party, CMS just remains a participant. They can give clarification, or their contractors can give clarification, present issues, and explain to the ALJ, but they don't actually have the rights of a party. And if CMS does not participate as a party, we are deprived of discovery. There's no discovery. If they do participate as a party, there is discovery. Oh, so I take it they did not participate as a party here. I've been doing this 10 years. They've never participated as a party in any case I've been involved in. Do you have a medical background? No, I don't. Probably have a little bit by now. By now, I've read a lot of medical records. I used to do workers' compensation. Okay, well. So the premise is, does CMS get to do the math wrong and win every time? Do they get to withhold the math? If you wanted to summarize for us, what are your strongest points about the inaccuracy of the sampling here? The inaccuracy, it would be the points of the statistical experts that testified for the ALJ and the petitioner in that the sampling was not performed properly to give a reasonably certain accurate estimate at the end. And therefore, the ‑‑ Well, let me say, I think this, I put this down from the district court's opinion. He listed four things that the ALJ found deficient. And one of them was a failure to keep the actual random numbers so that you didn't have that data. But you haven't mentioned that in your oral argument. Failure to define the sample units, failure to demonstrate that the sample units were independent of each other, which I personally think is pretty important, and a failure to demonstrate that there was a normal distribution of average overpayment. And I take it he got those from you or got those from the ALJ? He got those from the ALJ's decision. Okay. That's what the ALJ found on the merits of the statistics, is that they didn't ‑‑ the Program Integrity Manual required that the estimation be replicable, and you have to give us more documentation than the Program Integrity Manual says that you have to give for us to test the accuracy of the overpayment. And that was what we complained about, is we didn't get any documentation on it until just before the ALJ hearing. This is the odd system, because they say that you don't give any deference to the ALJ's findings. So why do you have an ALJ? It is an odd system, and when, if I may continue. Yes. Thank you. When a contractor requests review, and the contractor requested that the council look at the statistical sampling issue on its own motion, because they weren't a party in the ALJ hearing, they can't make an appeal. But they can request that the council look at this on their own motion. When the council takes something on their own motion, they are limited to a glaring error of law that had a bearing on the decision, or public policy. And so what they do is they say, the ALJ didn't apply Rule 86.1, which says we can sample. It's lengthy. It says a lot of things, but it's about sampling. He didn't apply that, or she didn't apply that correctly, so we get to review this. When they would otherwise, if you're going to allow that the sample can be destroyed by bad selection of numbers at the beginning, by a nonrandom sample, by a sample that's not representative of the universe that it's representing, then you say, well, they didn't apply 86.1. You make all issues of fact an issue of law, and therefore the council can review it and overturn his facts, even though they wouldn't normally be able to do that without CMS having been a party. Amazing. Well, you have time for rebuttal. Thank you. Ms. Krueger? Now you're with the Western District of Texas. Yes, Your Honor. How many of these cases have you done? This is my first Medicare overpayment case. Have you done other Medicare cases? My area of practice, Your Honor, is False Claims Act work primarily, so it does entail a good deal of Medicare, Medicaid, any sort of fraud against government. But this obviously is in a different category. This wasn't pled as a fraud, though. This is an overpayment. Not at all, Your Honor. All righty. May it please the Court and the Council, my name is Mary Krueger. I am an assistant United States attorney from the Western District of Texas. I am here today on behalf of the Secretary. The Secretary, as you know, is charged with administering the Medicare program, and a large part of that program is a system of payments that is done before review of claims. As an aspect of signing up for reimbursement, as a government health care provider who received government reimbursement, people are made aware that there is often a post-payment review of claims that have been submitted and paid by the system. And the Secretary... When you say often, what triggers this post-payment review? Is there some type of suspicion that there's an inaccuracy or an overbilling? Correct as to this case, Your Honor. In this case, what happened is the contractor who is hired to ensure program integrity did a sampling of payments that were going out in home health care in this region and found the MaxMed's payments to be an outlier. And what's meant by that could be many things, but it basically means it showed up as outside the norm or what would be expected. That is what then generated the review of the claims that had been submitted. So they then selected a time period, about a year and a half, and in that time period, it was a universe of 130 claims. Of those 130 claims, they had a nurse practitioner review the medical documentation pertaining to 40 individual claims. Once those 40 individual claims were reviewed, they found an error rate that was very high, practically 98%, 97.5. That error rate is when 39 of 40 claims were found to have been inappropriate. Now, when you say error rate, is that charging for services not rendered, charging for services not certified, charging for services not necessary? What? It is not the first. There's no allegation that the services were not rendered. The allegation is that the services were not appropriate and the claims should not have been paid. There was issues of homebound status, the need for skilled nursing care, and then a combination of those two factors in this case pertaining to homebound status. Does the nurse practitioner go to the doctor who issued the certifications? I don't think that they do as a matter of course. Is it in the record who the doctor was? It would be in the medical records that are before the court, yes. So basically the nurse practitioner is second-guessing the doctor's orders? The system for home health care payments is designed so that it would be more than simply a doctor certification of the need for services, but also certification of homebound status, for example, in this case. And that is to be the records as to that and the judgment as to that are to be kept by the home health care provider. Let me go back. You mentioned 130 in the universe. You've selected this time period, and there's 130 as opposed to 500 or 600 or 40. Correct. But we have now chosen 40, and are those chosen randomly out of the 130? Those are selected randomly? Yes, they are, Your Honor, and actually there's no allegation in the record that there was an attempt to bias the sample of the 40 that was selected. I understand. And the process for reviewing these is to have a nurse practitioner, I think you said, to go through and look at the file and make certain calls about either necessity or appropriate documentation or appropriate charging. At what point is it decided to use a statistical sample, in other words just the 40 that are randomly chosen, as opposed to doing the whole 130? That doesn't sound like, to me, and I've never done it before, I'm learning a lot about this, but 130 doesn't sound like a lot more than 40. I mean, if we were talking about several thousand and you chose 40, that might have more reason to the rhyme, but I'm not quite sure I understand why we would do 40 instead of 130. That's about, my math real quickly, maybe a third, 33%? Yes, Your Honor. In this case, it was about a third that were actually reviewed. There are cases, particularly with regards to hospitals or very large providers, where we would be talking about thousands or potentially tens of thousands, 100,000 of claims at issue. That is not the case here. And admittedly, 130 claims is a sizable universe, but the secretary being charged with the administration of this enormous government program has made a clear determination for policy reasons that sampling and extrapolation are necessary. So while it would only be one-third of 130 that could theoretically in this specific case be reviewed, that would be more than double the resources required, more than double the time required in order to effectively go through all of those files and make a determination. And in other cases, it would be... Entirely up to the secretary whether sampling is going to be used as opposed to an actual accounting? Well, I would not say entirely in that there is a statute that says that statistical extrapolation and sampling are permissible for the secretary to engage in where you have one of two conditions. One is high outlier, and the other is where education has been attempted and has failed. So I think that Congress has put some parameters here. Education like communist re-education? Does that mean you... Is that a euphemism? I apologize for not knowing what re-education means in this context, Your Honor, but I know that the reason in this specific... That doesn't sound good. It sounds like a warning. Yes, Your Honor. In this case, it was not re-education that was at issue. It was the high outlier status. That's what generated the review, and that's why we would say statutorily the secretary was doing more than simply her mere discretion in applying the sampling and the extrapolation. And after the 39 claims were found to be paid in air, and the air rate of almost 98%, it then did go to the ALJ, who individually did go through each of the 40 claims again and found a different outcome as to one. So the number that we are dealing with for purposes of appeal is 38 claims to be found. Well, for the sample, the whole purpose of the appeal here is not the sample because they're not challenging that at this point. They're just challenging the extrapolation. So why don't you move on to that? Yes, Your Honor. The extrapolation... The statistical methodology that was used was a valid method to reach an extrapolation. Who says it's a valid method? Well, I would say at the administrative record level, Mason definitely testifies to that. Ms. Mason is the expert who was testifying on behalf of the contractor who is doing this work on behalf of the secretary. And so her testimony or her report says that it is a valid methodology. Was she an independent expert hired by the Or was she in-house for the contractor? She is an employee of the contractor. Okay. And the issue that I think is raised by the independent expert, in this case Howler, is the dependency of the claims. And what he argues is that because the claims could be dependent on one another versus independent from one another, it would then skew the extrapolation that is done. So you reach a valid estimate, so you reach a 95% error rate. You then can properly extrapolate that to what it would be on 130 claims. But the real debate here between, I think, the experts and what was raised by Howler below in his expert report is how now you frame up the confidence interval and how correct you are about your estimate. So in this case, Howler. And where is that going to be clear? Because I didn't read that as being, I mean, I'll take your word for it, but is that going to appear in the MAC opinion or in Judge Ezra's? If you wanted to see Howler's actual report, it is a CD, which my understanding is the court has access to because it was something the district court judge reviewed, but it would have to be requested. Page 6 of our appellate record references the fact that these CDs are with our district court. If you then went to the administrative record, you would find Dr. Howler's report primarily at 824 to 842. And I would also mention, if you wanted to look at those expert reports, Whitmer's is at 817 to 822, and then Mason, who we think is relevant, is 874 to 892. And also, in rebuttal to something that was said earlier, there was a specific report issued by Mason that's at 510 to 518 for the court to the opinions that were provided by Dr. Howler. So there was rebuttal to the points that he was making. But the argument in this case is not so much about where the number comes out if you were to estimate it to the extrapolation, but it's how wide of a margin you should put on either side in your confidence in that number. So in this case... Because if it's too wide, you have no confidence. Correct. At some point, you would have no confidence, or you would get to a point where if you could not say that I have 130 claims that... It's not statistically significant. Correct. But you would get to a point where you would say 90 of those claims we know could not have been payments because we've already looked at 40, and we know that 38 of those we are claiming are in error. So that would be the point at which the margins stop because we now know that this is no longer possible. But in this case, what the government does, what the secretary does through this program, is it places a very generous one-sided confidence interval on the estimate that it calculates. Generous means pro-repayment. In favor of the provider, actually. I didn't read it that way, so you have to explain it. The 90% confidence interval rate is suggesting that it is 90% likely that the true amount due would have been a higher amount. Well, I thought the difference was something about 2.5% and 8.5%. I think that is the margins on either side of the estimate that you have. Should it be a 2.5%, should it be an 8.5%. But in this case, where the government applies that 90% confidence interval rate, it errs in favor of the provider. So it says that we're going to go to the outer limit on one side, the provider's side, and then make that calculation from there. And that is done to favor the provider because of the nature of these extrapolations. And so my argument to this court would be that the question is whether or not the Secretary's final decision, and that is the MAC decision, which is the decision that is upon review, is supported by substantial evidence on this issue of extrapolation. And I think that Judge Ezra correctly did that analysis, which was to go in a very detailed fashion through the record and determine if there was substantial evidence in favor of what the Secretary determined. Well, I mean, with all due respect, Judge Ezra is a very, very smart fellow. But it looked to me as if he basically did what courts routinely do in the Social Security cases, which say, well, the doctor says, you know, the claimant's doctor says this, the Social Security doctor says that. We take the Social Security doctor at his word, and anyway, whatever the SSA credited is substantial evidence. In other words, he ticked off all the points, but then he said, well, they said something here, so it's substantial evidence. In other words, it seems sort of conclusory to me, and that's why I was asking all these questions about the statistics. Your Honor, I think if you look to the record and to the administrative record that he was reviewing, one can find that there is substantial evidence to support the Secretary's conclusion. And I would argue to this Court that that's more than a mere scintilla of evidence. Unfortunately, you're not able to tell us whether ALJs normally hire their own statistician in these cases. I do not practice before the lower administrative process, so I would not know the frequency of how often they hire their independent expert. I do know in this case we would point to the record that has evidence that refutes what Dr. Haler said as a basis to show that there is substantial evidence on the record to affirm the Secretary's decision as to the extrapolation and the overpayment calculation. I'd like to touch on rule of thumb, as that was an argument that was also presented today. The rule of thumb argument, in the government's view, is misconstrued in this case. The rule of thumb provision is part of a manual. There's two manuals that are at issue here, the MPIM, which is the Medicare Program Integrity Manual, and the MPBM, the Medicare Benefit Policy Manual. And so there is a whole chapter on the Integrity Manual that deals with statistical sampling and extrapolation, which is what was done in this case and has the procedures for that. It's a chapter of that manual. There is then this Medicare Benefit Policy Manual, which applies in a wide range of circumstances. And Maxman is arguing that one provision of the policy manual should basically trump the provisions of the Integrity Manual, the chapter of the Integrity Manual, that would allow for extrapolation and sampling. I would also argue that because, statutorily, we know that Congress has now added a statute which says that extrapolation and sampling can be done in the two instances that I discussed before, that of high outlier, or if there is failure to heed education on the point, that what Maxman is trying to do is argue that one provision of the policy manual would also null that statutory authority for doing this. In this case, what was done on the 40 files that were reviewed individually was not a rule of thumb analysis at all. A rule of thumb analysis would say, every time I see claims for injection of insulin, I'm going to say that those were improper. What was done in this case was a review of each of the 40 files and a determination based on each individual beneficiary's status and need as to whether or not the claim was appropriate. And I think that that distinction is important when reading those provisions. Also, I'd like to point the Court to one of the cases that was cited in our brief, which is Stavis County Home Health Care. That is a D.C. Circuit opinion from 1991. It was before the statute was implemented that I've discussed here today with the two situations in which extrapolation and sampling were permitted. But in that case, I would argue that the Court did confront something very similar, an argument about whether or not you can have a rule of thumb provision in one place and statistically sampling in another. And the analysis that they provided highlights to the Court why that is possible. Finally, as to any due process violation, it's alleged that materials were not provided to MaxMed during the course of the administrative process. I think the administrative record is silent as to exactly what was requested at the first two levels of review. But certainly I think there's no dispute that by the time of the third review by the ALJ, all the materials had been provided that were requested, and each of the experts who testified, the independent expert from the Court, the expert for MaxMed, and the expert from Health Integrity, all were able to recreate the sample, talk about the sample, and knew which claims were and were not at issue. Therefore, I would suggest that MaxMed do notice of what was proceeding and the opportunity to be heard on those claims and arguments. Do you agree with Ms. Terrell that the MAC has plenary review of the findings and conclusions of the ALJ? I think that the MAC, the MAC Council, has a de novo standard of review of the ALJ and can review and make its own credibility determinations irregardless of the ALJ's outcome. I think that both by statute and by regulation, that review is to be de novo and there's no limitation on that. And I would suggest to the Court that in some sense the process of the administrative proceeding is to reach a singular decision. It's to reach the singular decision of the Secretary. So in this case, the administrative process occurred, the MAC made its ruling, and that is the final singular decision of the Secretary that's upon review. I don't think anybody's disputing that. Okay. Thank you, Your Honor. Yep. Ms. Terrell? Briefly, I would like to direct you to 42 CFR 405.1110, the 1100s are for the Council. And it, at small c, gives the standard of review when a referral by a CMS after participation is made and CMS or its contractor participated in the appeal and the Council exercises its own motion of authority. That is where the limitation is its review, either broad public policy, broad policy, or procedural issue that may affect the general population or an error of law material to the outcome. And did Judge Ezra cite that provision? I don't recall Judge Ezra did. Okay. Counsel, where did the 258,000 number come from that apparently is undisputed? Was that a calculation? Obviously, on your side, your client's side, there was a calculation made, and you believe that to be the correct number as opposed to the 700-and-something thousand. How did that number come about? That number came about with the nurse practitioner's review, reviewed the 40 claims. When they take a sample, they do the hard work and review the claims. Now, those services were all certified by the physician, so often what you'll find is a paperwork issue. Often CMS looks for physician's notes in home health records and for home health notes in physician's records, and so it makes it very difficult for a provider to know what to do. The rules have changed as of late for the home health agency to obtain the physician's records and to get attestations from physicians to help in those reviews to support the certification more. But, again, for purposes of your client's purposes, the 258,000 is a number that is generated by your client's examination of the 40 files. Did I understand you correctly, what you just said? No. We brought those 40 claims through the administrative process as well. We prevailed on one. We did not appeal those 40 claims to this court for re-adjudication. Judge Ezra quotes the MAC as saying that the provider could attack the statistical validity of the sample or it could challenge the correctness of determination in specific cases identified by the sample for an including waiver of liability where medical necessity or custodial care is at issue. Did you take the latter tack at all? That is what we appealed on the way up. If we prevail on those, which we did prevail on one. Okay, but with the statistical sampling, you don't even know. If you were actually going to challenge that, you could bring in the 90 additional case files, I suppose, and dump them in front of the ALJ and say, here's what disproves the sampling. The ALJ has, at that level, no jurisdiction to hear those other 90. They have to be brought through the process. Those are the only ones reviewed. In order to reduce the overpayment on the sample claim side, we can win on those individual claims and on each claim that is reversed, such as the ALJ reversed one. That would actually make a difference, change the sample, and the extrapolation should be done anew, as a matter of fact. In the Program Integrity Manual, and I left it back at the table, there's a provision for a re-determination of the initial claim, which that would be. Okay. Thank you very much. Thank you.