Oral argument — In Re Niaspan Antittrust Litigation

Good morning. I'm ready to call our first case in Re Niaspan Antitrust Litigation. You may proceed. I'm Hunt for Endpayer Plaintiffs. I'd like to reserve five minutes for rebuttal. Done. Fifteen million times a day, consumers go to pharmacies and pick up prescriptions that someone else pays for. This system works only because each prescription links to the entity that's ultimately responsible for footing the bill. You know, we've got a lot to talk about. We've read your briefs. We know the background and the seriousness with which you take the position that you've got. The policy implications are important. We understand all that. But let's let's do dive into the issues because there is a lot on the table. One of the most important things I think for us to get the grip on is whether your suspenders can be considered by the court, so to speak. You major your overalls, belts and suspenders analogy and suspenders, as I understand it, are affidavits. The other side is pressed the point that that showed up no place until a reply brief footnote in passing. So why don't you help us understand why that's properly in front of us when under Jaluti, it arguably was never in front of the district court. Yes, John, I'm happy to say the first time the parties really joined issue on the administrator issue, which is what the affidavits respond to. It's indeed this expert report, which was filed after the opening brief in the district court. The district court allowed us to respond to that brief, which we did in a reply to that report, which we didn't reply and in craft expert reply. And in that reply, we raised the issue of affidavits as a way to resolve any lingering confusion. Well, you've got the burden of proof, right? Yes, sir. OK, if it's your burden of proof, how is it incumbent on them to raise an issue that you're responding to instead of you in the first instance explaining to the district court? This is how we this is how we intend to resolve this problem. We've got these data sets and and we can bolster it with affidavits. You had city select. You had birds. You had other precedent before you. Why is it why is it on them to bring something up and you to respond in a reply brief? Well, I think because your honor are going into this issue is that overall would work. You know, there would be no need for these affidavits because the data could be resolved. And everyone agrees, in fact, that when end payers are clients, that data can be resolved on its face. The administrator issue, which has been raised that, hey, in this small subset of transactions, there's going to be this problem because the PBMs contract with administrators is the only thing that triggered the need for this individual verification. Right. I'm I'm I'm trying to get to this point, though. What you say in your briefing, what you appear to be saying here is we couldn't have known. How are we to know that there was this gap in our proof until they pointed it out? Do you agree or do you not agree that this is a problem, that there's a gap in your proof without the affidavits? Maybe we should start there. So we disagree that there's a gap in the proof without the affidavits because. PBMs can identify end payers even when the contract with an administrator and the record reflects that they can do that. I thought there was an agreement that there's some percentage that that is not the case for where there and that that is what just the boys found, that there's a set of cases where even if we accept that your PBM data is going to bring us to a narrower point, it still has a branch there where it could be one of two kinds of entities at the end of it. And the only way to know that one way or another is through an affidavit. Is that wrong? So our overall method is that, in fact, that is wrong and that PBMs will always be able to identify the end payer, even when there's an administrator. And the record reflects that they can do that. Our named plaintiff was able to get their PBM data AFL, even though they work with an administrator. Now, the only thing in the record there's no there's no example in the record where they are unable. The only the only thing in the record on that is when you say there's no instance in the record where they're unable. I thought I thought as a finding of fact, maybe I'm wrong about this. Correct me if I'm wrong. I thought it was a finding of fact by Judge Du Bois that that was not the case. So you're telling us that that was clearly erroneous. Your Honor, I, Judge Du Bois did not make a finding of fact on the issue of whether or not PBMs can identify end payers when there's an administrator. If you look at a. Well, let's look at five. Because now I now I am genuinely confused. I thought he said you haven't done it. You haven't done it because of what I said just a minute ago. You haven't done it because there's some subset of this data which you can't tell whether you're going to be dealing with an ASO or with an end party. And you can't know that based on this data. But did I misread his ruling? I think if you look at 85, he made specific findings on particular parts of craft methodology. But that does not include the overall methodology. But even if you disagree with us about that and you think that it's not clear on the face. At the end of the day, everyone agrees that there are objective records that allow the end payer to be identified. And we explain this in reply 14 to 15. Everyone agrees it can be done when it's PBM clients. It's agreed it can be sometimes a lot of things can be done, but they're not really feasibly done. Isn't that the problem you have here? You've got 20 million transactions, right? And you've got, I think, if I remember, 3400 or 4400 different contracting parties with the PBMs. Is that right? So, your honor, there are 1300 insurance companies that could conceivably act as either administrators or intermediaries. So we think in the worst case scenario, you would only need to send inquiries to those 1300. And that is within the scope of... You'd only need to send what to those 1300? Inquiries or requests for records. So, let me step back for a second. That's what I'm wrestling with here. Craft seems very facile when she just says, oh, this can be done, this can be done. Just throwing some search terms, look at the numbers. And the more I dug into her deposition, the more she didn't have answers to why there were gaps in the record. And one of the big ones to me is that just because you see a name, pick one, Humana or whatever, that doesn't answer the question whether that Humana is acting as a TPA, an ASO, or as an insurer. You need to drill down to another layer to ascertain whether that would be a member of the class, correct? Yes, your honor, but I think it's important to understand when that layer comes up. We don't need to know which entity paid to do notice here, or liability, or damages in the aggregate. The only point at which we need to know which entity paid is at the claims administration stage. And at that stage, it's an individual verification process that this court explained in Byrd is a normal part of class litigation. But I thought the PBM authorizes payment in the claims administration phase. Isn't that why PBMs exist? I mean the claims administration phase of the class litigation. Oh, I'm sorry. Yes, I'm sorry. You don't need to know which entity paid. At the time that the pharmacy transaction happened, there is a record that links each transaction to an end payer. That sounds like you're asking us to green light a certification based on some sort of assumption that we'll all work out. When the dust settles during the claims administration process, that inverts the analysis, doesn't it? No, your honor, not an assumption, an agreement. So if you look at Dietz's testimony at 981 to 982, he agrees that in the PBM data, there is information that links to the end payer. Every transaction. He also said some of that information could well be proprietary. If it's proprietary, you're never going to be able to drill down to figure out whether the person's a member of the class or not. He said that's why the PBMs may not know how to read that code, but the point is that the administrators do. This is remarkable, right? If the PBMs can't read the code, how is this so administratively easy that you guys are going to be able to read it? When Ms. Kraft dealt with it, you know, I don't mean any disrespect to her at all, but she left the district court judge here distinctly unimpressed because when she picked four examples to dig into, according to the judge, she got two of them wrong. She blew it. I mean, if that's accurate, you know, a 50% failure rate when you're dealing with four is a little disturbing when the assertion is this is easy and we're going to be able to nail it. So, your honor, I want to explain why those two examples are not accurate. When I said there's a lot to talk about, you're on our time now. Go ahead. Okay. So, I want to come back to why those two examples are actually accurate and the district court clearly erred. But let's step back and say why it actually doesn't matter, even if you disagree with me. First, it was two out of 22. So, there's no dispute on the other 20 that they were accurate. And you can see those are 1010 to 1011 in the JA. But stepping back from that, the task the district court put before the plaintiffs here to generate a list or prove that they can generate a list from the PBM data alone is not what this court's ascertainability standard requires. And that's what I think why it's important in Byrd that the court played the class forward. We're not disagreeing with that. I think you'll understand that this conversation started with saying, why should we hear you talk about affidavits now? You're actually making the point I was trying to ask you about earlier. There is an opportunity for a plaintiff to come forward and say, here's the data and where there's problems, we have affidavits that we believe will deal with that. But you never did that until you got to a pride brief responding to an expert report. So, now we're kind of full circle. Why should we be hearing about that now if the burden proof is on you? And if you don't have that benefit of those affidavits, that's an argument point. What does that do to your case? Okay, so I'll start, I'll come back to waiver. Again, we presented it in our third brief of five. Defendants had an opportunity to respond to that and in fact responded to other points from our reply brief and in response to filing, the district court allowed them. We then reiterated again in a fifth brief. So, it was aired before the district court with defendants having an opportunity. All right, but even if it wasn't forfeited, then we look at the quality of the presentation. And if I understand Judge Jordan's concern, it's a concern I share, which is instead of coming forward and saying, we can do it with affidavits, what you should do is you produce affidavits. And the affidavits show the district judge, this is feasible, look at this, here's the evidence. Instead, you didn't do that. This is like reading the Kraft testimony was like watching a whack-a-mole game. She throws one thing up, it gets batted down. She throws another thing up, it gets batted down. The whole issue of the 5500s, do I remember that number correctly? Oh, the form 5500? Yeah. She confessed error on that, right? Yes, Your Honor, but that wasn't the… And then we get to the end and it's like, oh, no, no. Okay, now it's the affidavits. Okay, we'll come forward with a raft of affidavits that show that this works. Instead, what we get is we can produce affidavits. That's a really thin read for the district judge to go on, isn't it? Your Honor, we came forward with better than affidavits. Every named plaintiff came forward with records of their knife ban transactions, whether they worked with the PBM directly or not. And that's the A989 in the appendix. And in fact… Well, when you say every… Maybe we should ask it this way. It appears that the district court judge's concern was that there was not going to be a fair way to test the program here, that there was going to be a gap in proof. And you say, look at these affidavits. Now, if we accepted that there wasn't a forfeiture, I even have a preliminary issue, too. You should have come forward with affidavits, which is, did you really put it in front of the district court in a way that the district court understood and could grapple with the issue? That's not an insignificant point. But let's say we got past that. We're operating under an abuse of discretion standard here, right? Is it an abuse of discretion for the district court not to pick up on your footnote in your third brief or your passing reference to it in your fifth? Your Honor, I think in the fifth brief, it was not a passing reference. It's on multiple pages, and really it's 1074 to 1076. And it was in that brief, it was directly responsive to the district court's concern about the gaps that the district court raised through those two examples. So the answer is, it's an abuse of discretion. Your Honor, it's an abuse of discretion, but here, the district court ruled out, by requiring a programmatic list, ruled out the possibility of individual verification at all. And that's a legal error that infects the district court's analysis. Where do you see that in the district court's opinion? I think it's clearest at A87, Your Honor, where the district court says, EPPs must prove that identifying class members will not require individualized fact-finding. And then at 89... Wait. And how did you characterize that? What is the standard, the legal standard you think is in that, and that it's a legal error? I think the legal standard that's in that is saying there can be no individual verification process along the lines of what was accepted in Byrd and Hargrove and City Select. Yeah, I think you did a good job pointing out that that could be legal error or it could be an overstatement that's inconsistent with some of the other things the district court did, so we have to read the opinion as a whole, right? Yes, Your Honor. And the one that you latched on, I thought, I think the one you just identified, at least to me, was rather persuasive. The other one that I thought was less persuasive was when you said, you cited that the district court said, this is more than de minimis, right? And I took that to mean the district court saying, this isn't some sort of side issue, this is a meaningful issue. That's not really a... Taking it as more than de minimis, that's not really a legal error, is it? I think taking it as not a legal error, it still reflects the district court's confusion about the scope of the problem. Because if you think, if you take even just the... If you thought that you had a gap in your proof, that's not a small problem, is it? Well, I have two responses to that. It's not a... A gap in proof that can be resolved by individual verification where the entity does not need to be identified until the claims administration stage is not a problem. It's not a problem at all because, as Bert explained, individual verification where a class member comes forward and says, I'm an end payer, and you cross-check that with records is a normal part of class litigation. That doesn't make it unmanageable. And that's what all the named plaintiffs were able to do here. Now, then respond... When you say that's what all the named plaintiffs were able to do here. If I understand the arc of your argument, it is that we had some named plaintiffs, and they were okay. Therefore, you should not have concern with over-inclusiveness and an inability to test the accuracy of affidavits overall. If I understood you right? Yes, Your Honor. And I think that's a lesson from Hargrove in the sense that in Hargrove, they did the piecing together of records, which was much more complicated than what we have here. And they did it for named plaintiffs. And that was sufficient to show it could be done for everyone. And I can even play out how the target example would happen in real life. In real life, what would happen... Let's say the craft is wrong. Again, we dispute that, and we don't think that they've shown that. But let's say the craft is wrong in the data that Target down as a fully insured plan. When Target actually comes forward to claims administration and says, hey, no, I'm actually self-funded, which is what defendants say it is, haven't proved it, then the claims administrator would say, oh, well, our records show fully insured. Can you clear this up for us? And Target can go get its claims just like every single named plaintiff was able to do. In addition, we have the PBMs. So is that all there is to it? But they make the pitch, of course, that the plaintiffs are waving their hands over a pretty darn big problem, which is we would have to be able to test the veracity of all the affidavits that would come in within this group. And this is really like trying to figure out which of the two $20 bills is counterfeit, right? Why don't you respond to the question to the counterfeit bill argument, which was pretty catchy. Yes, sir. So first, as everyone agrees, including Dietz, when the PBM directly has a client, the PBM can call out that counterfeit bill. So the PBM can flag a transaction as an administrator, at least sometimes. Dietz won't agree to all ways, but he has no examples when it never happens. But if you look at 981 to 982, Dietz agrees that even if the PBM itself cannot flag the counterfeit bill, there is a code within the PBM data that allows the administrator to flag the counterfeit bill. Everyone agrees that administrators have to segregate their insured and administrator business. They're required by law to do so. Let me make sure I understand what you're saying there. So does that mean that... Let's pick a PBM. I'll try to use real parties instead of fictional ones. So let's say it's Caremark. That's a big one, right? So Caremark's the PBM, and let's use an HMO. Kaiser, yeah, okay. So Kaiser's the HMO. The record makes pretty clear, I think, correct me if I'm wrong, that there's a possibility that in that... Let's use two transactions, both involving Caremark and Kaiser. There's a possibility that in one of those transactions, Kaiser is acting as an insurer, and in another transaction, Kaiser is acting as an ASO or a TPA. Is it your argument that it can be readily ascertained which hat Kaiser is wearing in each of those transactions just by looking at a code? Or is that where the affidavits are required? And you have to ask Kaiser. In claim one and claim two, which, of course, we have to replicate over tens of thousands or hundreds of thousands of claims involving Kaiser, are you wearing the insurer hat or the administrator hat? So the record reflects that Caremark would know what role Kaiser was playing. There's nothing in the record. There's no example where Caremark doesn't know, except, I agree, deeds of kinds that Caremark may not know and will not always know. So I'll accept... It might be proprietary, too. It might be proprietary. So let's accept... When you say there's nothing in the record, there's expert testimony that was accepted by the district court. Is that clearly erroneous? Well, I think where the district court went wrong is the district court agrees that Caremark will know, to stick with this example, at least some of the time. And there's no assessment of how often Caremark doesn't know. But there is agreement... Whose problem is that? Whose problem is it that there's no assessment of that? Is that the defendant's problem or is that your problem? I think it's the district court's problem. I think we put in front of the district court the... It's not the district court's job to sift through 20 million claims or to sift through 1,300 parties to figure out who's on first and who's not. Absolutely, Your Honor. But let's accept that 50% of the time, to give everything to deeds, that Caremark won't know who the end payer is on a transaction. And so we have to ask Kaiser. That's only 1,300 asks. It's not the district court sorting through it. It's records that exist that link to the PBM records that Kaiser can come forward with. And beyond that, we know from our named plaintiffs... I'm sorry to interrupt, but I understand the other side's argument to be, if I've got it right, this isn't sorting through just 1,300. You won't know... This is back to their counterfeit argument. You won't know which ones are problematic unless you look at all of them. You know, you got to... Or you stumble... You'll be lucky if you stumble across the couple of counterfeits, but you just don't know without looking at all of them. Now, that's... Tell me where they're overstating or that they've got this idea wrong. So I think where that goes wrong is that even if you take Dietz's number, that, let's say, 56% of employers use an intermediary. And that's the highest N I can give it, okay? So that's, if I'm doing my math right, 44% that don't. Those employers show up as PBM clients. There's no question about them. There's agreement about them. You couldn't have any issue where you think they could be acting as a Kaiser or a Humana. For the rest, Dietz agrees. Sometimes we have... First we have PBMs saying that their clients include third-party administrators and that they can identify their clients. So that supports that they can identify when their clients are acting as administrators. Dietz agrees that they at least sometimes can do that, and we have AFL coming forward with its records even though it works with Blue Cross. So we know that they sometimes can. So now we're down to maybe half the time and only 13... It doesn't matter if it's 20 million transactions. It's 1,300 entities for which there's a code, and this again is at 981, in the PBM data that identifies that end payer to that administrator. So... Okay, well then, if that's true, why would you ever need an affidavit? Well, Your Honor, I think... I don't know that we would actually need an affidavit if people just bring forward their records. Then why are we fighting about affidavits? Affidavits is a shorthand for the individual verification process that occurs at claims administration. Yeah, we're just... Okay. I'm just maybe seeming obtuse here, but I hear you arguing it was wrong for the district court to say we need objectively verifiable stuff here without affidavits, but also hear you saying we can do this without affidavits. I hear you saying it was an abuse of discretion for the district court not to pay attention to your references to affidavits, but then I'm hearing you say that's never going to be an issue because it's all in the codes and that can be figured out. I feel like I'm getting a little bit of whiplash. Do you need affidavits or don't you need affidavits? The record establishes, in our view, that you don't need affidavits because the record establishes that PBMs can do this and we don't think the district court made an express finding on that. But if you disagree, and this is with that, if you disagree, then we can do it with affidavits. I'm still troubled with the PBMs can do this. That seems to be the whole theme of the case. So then... PBMs can do the heavy lifting. They can do this. They're not third parties to the case. They're not... They're just third parties over there. And those affidavits from the PBMs were, gosh, they were as thin as thin can be. They were boilerplate language and there's so much could, can, would. Where's the... And maybe your argument is that's for later. This is... We were held to too high a standard at the certification phase. Is that kind of the crux of your argument? That is the crux of your argument, Your Honor. And I would say those PBM affidavits might have been general, but they're third parties. They don't have a stake in this and they said that they can identify end payers. And even if you accept that all they said is they could identify client end payers, that's a big portion. But if you disagree, if every single one of you disagrees and every single one of you has a disability, what the disability asks is when you need to know which entity paid, can you objectively answer that question? And that is where I think DITSA's agreement that there's a group number, a code in the PBM data that conclusively ties that transaction to an end payer. It's only a matter of how much work we have to do to read that. And it's very little work. In our view, it was none, but you disagree, okay. It's very manageable. And that kind of individual verification is all that affidavit attainability requires. Well, when you say it's very little work, is that a question of fact? Is that a question of law? Or is it a mixed question of law and fact? Because the district court's conclusion was this would require substantial work, many trials. Is that a finding? Or is that somehow a legal conclusion? How would you characterize that? I would characterize that as a mixed... I mean, I think that wraps the wrong legal standard onto a clearly erroneous fact finding. You have to make it a legal standard, don't you? I mean, you can't say, well, I disagree. I mean, that standard review is not going to get you what you want. I think what you want to say is as a legal standard, ascertainability cannot require, that cannot be applied here to say this is too difficult. But if that's the position, then what's the holding we're writing that's going to allow district courts to manage the next case that comes along? I think that's a fairly important question. I think the holding is that ascertainability does not and has never required plaintiffs to prove they can programmatically generate a list of class members. If that serves no practical purpose for litigation, as Byrd made clear, it's a forward-looking inquiry when you're looking to see when you need to know who's in and who's out. And as Byrd also made clear, if you're doing that at claims administration, that can be an individual process. It's individual, and that's not an issue so long as it's not mini-trials. And it's not mini-trials here because we have records at the end of the day to verify those, whoever comes forward and what they say about their status. Let me just probe something you said a minute ago because it didn't sit right with me based on my recollection. As I look at the affidavit on 344, this is from the Caremark rep, Steven Shaper. What I heard you say was these affidavits said, the PDM said they could identify the end payers. And the reason that didn't sit right with me was because if you're correct about that, I think you've got a much stronger argument than what I understood your argument to be. And when I look at this affidavit, what I see here in paragraph six is Caremark maintains records by which clients and clients' members can be identified. That doesn't say end payers. And there's the rub because the PBMs can identify a client. And what we don't know is whether that client is acting as an end payer or an administrator. Is not that the real problem with your case? And if it's not a problem, tell me why. If it is a problem, tell me why you overcome it such that the district court erred. Yes, Your Honor. So first, for this particular affidavit that you highlighted, the reference to a client's members indicates that this is talking about end payers. Wait a minute. Wait a minute. Who procured these affidavits? Didn't Kraft or counsel? I mean, yes, they were in response. I mean, this just didn't come out of thin air, right? Somebody said, can you give me an affidavit about how you handled this, right? How difficult would it have been for those affiants to say on an issue of incredible importance to this certification, we can identify end payers. You're now telling us that they didn't want to say that or they were too careless to say that. Instead, they hit the ball and said, we're going to talk about clients and clients' members, and the court is supposed to, through clairvoyance, understand clients' members to refer to end payers? That sounds really far-fetched. No, Your Honor, but that is just one affidavit, and I would point you to 349, where Prime says, Prime has readily accessible records by which third-party payers can be identified on every purchase of NYIS BAN that Prime adjudicates on behalf of its third-party payer clients, and that's similarly at 351 for Optum, and then at 357, there is express scripts saying that their clients include third-party administrators, so they're recognizing their clients, but nonetheless, again, if this court says the district court rightly found or we won't disturb a finding that PBMs cannot identify end payers in every case in which they contract with an administrator, that goes to this mythical requirement to produce a programmatic list or prove you can. Well, that, when you say the mythical requirement, that depends on our reading the district court's opinion the way you're reading it. You're saying that's the standard that the court held you to. When I read the district court's opinion, what comes out to me is, well, I'll just, I'll read it to you. It says, EPPs have not persuaded the court that distinguishing between class members and mere intermediaries which are excluded from the class will not require consideration of the individual contractual relationships underlying each transaction. That's the linchpin holding here. And then the court goes through and says, in essence, I don't believe they're expert. I mean, he's a lot nicer about it than that, but he says, I don't believe they're expert. First, she's talked about the form 5500s, and then she collapsed on that. Second, in the renewed motion, the data collected by the PBM to the standards mandates the use of fields that identify TPP associated. Ms. Kraft admitted the data merely contains code numbers, not names or descriptions. So it doesn't look like it's as easy as she was making it out to be. I don't believe her about that. Third, he mentions her deposition testimony and how she had to back off her statement about categorical this or that with HMOs. Fourth, he says she selected four to give particular attention to. I know there were 22, but there were four given specific emphasis. And he said she blew it on half of them. So why isn't this straight up old school determination of credibility, where there's experts saying things on both sides, there's evidence and a lot of it in front of the judge, and the judge has the unhappy but necessary job of saying, you know, I believe this side more than I believe that side. And I just don't believe what this person is saying because they've hurt their credibility in front of me. When it comes down to it, why isn't it as simple as that? Your Honor, I'll respond to the credibility. I want to go back to the sentence you quoted right at the beginning of your question. Embodied within that sentence was the wrong legal standard because the district court was saying that we failed ascertainability if we could not prove that individual contract review was not required. There were a lot of negatives in that. But at the end of the day, this is a programmatic method to go straight from the data. Or is it saying it's too much work, it's not determinative enough? It's not saying, I don't read him to be saying it's got to have a programmatic. I read him saying they have not persuaded me that we can deal with these exclusions without consideration of individual contractual relationships underlying each transaction. Why isn't the fairest way to read that to be this is not supposed to be mini trials? That's the whole point of ascertainability, right? Right. Well, there's no dispute on commonality. Again, we don't need to know which, we don't need to know, there's a predominance dispute, but we don't need to know which entity paid to determine liability or aggregate damages. I think I understand your position, which is you read that to mean him saying there's no room for affidavits in this. It's all got to be data-driven programmatic. Even though, arguably, and the other side will argue, I'm sure, you never put the issue of affidavits in front of him, but accepting that you did, is it not equally plausible to read this as saying I can't have mini trials on each of these transactions, and I think that's what they're going to be leading me to do? I don't think it's equally plausible, but if the district court did say this was going to require mini trials, that would be clear error because, again, and to go back to, and this is explained in reply 1415 and DITSA statements 981 to 982, there's no dispute here that the administrators can look at the data and say exactly who that end-payer is. It's encoded in each transaction. So, at the end of the day, your statement is, even if he didn't believe crap, if he thought she's just not credible anymore, there's something in DITSA's that saves the day and makes it clearly erroneous. Yes, Your Honor. And, of course, I also don't think there's a dispute in addition. I understand. Yes, Your Honor. Yeah, I gotcha. Anything else? Okay, thanks very much, Ms. Hunt. We'll have you back for rebuttal, and don't worry, I'm sure it'll be longer than five minutes, and that'll be okay. So, please, the court, Elaine Goldenberg for AbbVie, Teva, and the other defendants, this is a case about factual findings. Can I interrupt you, because I don't want to forget what I'm thinking about right now. I'm sorry to do that to you, but I think Ms. Hunt raises an important correction, perhaps, in my thinking, and I'd like you to respond to it, which is I've been looking at this ascertainability issue and reviewing the district court's opinion with 20 million claims in mind, and she tells me that's the wrong way to look at it. I should be looking at 1,300 parties, contracting parties, instead. When you're dealing with ascertainability, 1,300 is a much smaller number so do you concede the point that 1,300 is the right frame of reference and if not, why? No, I definitely do not concede that point. 1,300 is the number of insurance companies that are members of a particular professional organization. That doesn't even tell us how many insurance companies there are, but you're not just looking at insurance companies to ask this question if you had to send out affidavits, because part of the whole point is you don't know in looking at the PBM data who's who. You don't know what role the different entities are playing in advance that tells you whether you're looking at a fully insured plan or a self-funded plan, even. So you would have to send affidavits, as I understand it, to all of the payers in the carrier and account fields with respect to all of these transactions and even if you were just looking and that method is 20 million transactions, even if you were just looking at the number of entities that could be in the class, the number of potential class members here, we think, is in the hundreds of thousands. Are you saying that? Yes, certainly. So the people who could be in this class would be insurers, when they're the end payer for a fully insured plan, or self-funded plans that have the ultimate payment responsibility. How did they get to 1,300 and why are they so wrong? So they say that 1,300 is the number of insurance companies operating in the United States. As I said before, I don't think that's an accurate description of the number of insurance companies operating in the United States. So the question is, are you the end payer? And if you were to look up in these pairs of fields, even on plaintiff's own view, are you the end payer? Because you can't tell in advance. And if you were to decide... And is that when you say, are you the end payer? You're talking about claim by claim. That's right, because they would need to look at the number of insurance companies. Even if you were just going to add together the hypothesized number of insurance companies, 1,300, and the number of self-funded plans in the United States, there are 2.2 million group health plans in the United States during this cost period. But their response to that is, you can take a gigantic number of those off the table immediately, because everybody will agree that they're not end payers. So when they tell you it's millions and millions, you're like, oh my, it's just not realistic. They're throwing scare tactics at you. There's a confined universe here. Don't listen to their millions, like Carl Sagan, billions. Just focus on the real stuff here, which is we've got like 1,300 people, maybe, or entities, that you'd have to deal with. How do you get from that to hundreds of thousands of members? Assuming we thought, okay, maybe that's 1,300 pieces, right. How do we get from that to hundreds of thousands of potential class members? Yeah, and the 1,300 assumes, this is the counterfeit gold example, that you can somehow tell at the beginning where the confusion is going to be. There's going to be confusion here under Plano's own theory that you have these pairs of entities in the carrier and account fields. There's going to be confusion with respect to a self-funded plan and an administrator. There's going to be confusion with respect to a fully insured plan and an insurer. There may be yet other entities that show up in those fields that aren't any of those. Every single time there's going to be confusion. I hear Plano's today saying that there is some magical code somewhere in the PBM data that's going to fix this for you and that Dietz admitted that's right. We certainly do not agree with that in any way, shape, or form. What Dietz said is, yes, of course, as you go along the chain, each person in the chain, someone pays the PBM and that person then is maybe being reimbursed by someone else. Those people can figure that out. Why don't you take us to the Dietz language that they're pointing to that you think they're misinterpreting. Well, as I understand what Dietz is saying on the pages that they have cited to, what he's saying, and this is just a snippet. I don't have it in front of me, I'm afraid, Ron. I believe it was 980 to 981, but I'm not positive about that. This is just a snippet of Dietz's deposition testimony, and so we also have to look, of course, at all the other evidence in the record and Dietz's reports and all that. What Dietz is suggesting is, is there something in the PBM data that tells you what the plan is, that identifies the plan? Sure, there is. That may mean that as you go back along this chain of payers, which could be three or four or five payers all the way back to the end payer in various scenarios, then can each person in that chain look back and say, here's who owes me money, here's who owes me money, here's who owes me money. Yes, maybe that is true. Although, again, I think it's just a passing reference by Dietz, but that is not at all the same thing as suggesting that you can look at the PBM data and somehow embedded in the PBM data is the identity of the end payer. That's just not right. There's nothing in the record that supports that, and if it were right, I'm not sure why we'd be having all this debate over the expert testimony telling you what the PBM data can and cannot show. There's no evidence that PBMs have ever come forward with identities of end payers or that anyone, including Kraft, has ever used PBM data to figure out who the end payer is. Is that really an issue here? Because I thought Dietz acknowledged, maybe I'm wrong about this, I thought Dietz acknowledged that the PBM data would, in fact, take a gigantic percentage of this off the table, that it would be clear that there's a better than half that you would be able to look at. Not at all, definitely not, and I think your Honor is referring to the 6% argument that plaintiffs have made in their briefs that there will only be confusion 6% of the time. We think that is irrelevant, we think it's wrong, and it's contradicted by evidence in this record, and we think it's forfeited as well. So if I could talk about each of those in turn. Since you're of a mind to speak about forfeiture, why don't you deal with the affidavit issue? Absolutely, affidavits. If it were the case that the problem here could be solved by affidavits, is this a case where the public policy implications are important enough that it would be incumbent on us to say, okay, well it may not have been in front of the district court in quite the way it should have been, but we're gonna consider this argument because, you know, under the law pertaining to reverse payments and overpayments dependent on that, there's a significant cause of action here that should be allowed to go forward. No, I don't think so, Your Honor, because it is not a legal issue. It is a factual issue. So they have the burden of proof to put in facts about affidavits into this record to show who the affidavits would go to and how they'd be verified and how they would work, and there is not a single scrap of evidence in this record about affidavits, not one. Their expert, Ms. Craft, never talks about affidavits. It only shows up in this passing footnote in their reply brief to which we did not have an opportunity to respond. Well, they say it's in their fifth round brief. Right, and maybe I should walk through that briefing to make clear what happened there, to make clear that we actually had no opportunity to respond, nor would we have thought that we would have had to respond to this passing footnote supported by no evidence at all. It shows up in the footnote in Plaintiff's reply brief, and that's the last brief for the class certification motion that you're considering. After that, plaintiffs come to the court and say, we want to put in a reply report from Ms. Craft, and we say, well, if you're going to do that, we have to be able to respond to it. They put in the reply report from Ms. Craft. The judge orders that we can respond to the craft reply report, and the craft reply report says absolutely nothing about affidavits, so we cannot respond, or again, we thought we needed to given that there's no evidence and there's been no argument on affidavits up to this point. The so-called fifth brief is the brief that they filed when the district court said to them, look, it seems like Ms. Craft has gotten half of her examples wrong here. What do you have to say about that? And they filed a three-page brief that said, we aren't going to identify the end payers and these payers. We don't have to, and they again mention affidavits in passing in the text of this brief a couple of times with no evidence, so it's not simply, I think when this court for public policy reasons, and I don't think the public policy reasons exist here, and I like to talk about that, but when this court for public policy reasons overcomes a forfeiture, in my reading of the cases and my experience, that is a legal issue. It's not an absence of proof in the record, and here you have both the forfeiture and the absence of proof in the record. Hold on just a second. That's true. We've said there's a legal issue, pure question of law, and this is important. Is the question of whether affidavits are appropriate and ought to be considered, Because of course their assertion is that the district court's opinion was that you had to have a purely objective, no inquiry beyond data way of ascertaining, and that that is a legal issue, wrong as a matter of law. So now they didn't put it in quite these words, but you would be addressing it as a legal issue because it was the wrong legal standard, so if you sent it back, you'd send it back so that you could consider affidavits, which you should have done in the first place, or whether affidavits were appropriate here. I still think the failure of proof would be a problem for plaintiffs in that circumstance, but the predicate of Your Honor's question is that there was legal error by the district court here, and we think it's extremely clear on any fair reading of the district court's opinion that there was not that legal error here at all. Well, let me quote the district court. This is at page 168, 555F sub 3rd at 168. The district court said that the plaintiffs had not shown that they could identify class members, quote, without individualized inquiry, end quote. That's not our standard. Understood, Your Honor, but what the district court is saying here, and I'm happy to quote you some other places in its opinion that make this clear, is that there has to be individualized fact-finding every single time here. Every time you meet with the district court. He didn't really say that. That would have been helpful if he said that. He did say that. I appreciate your candor. You've, I think, conceded that what I just read to you is inconsistent with our standard. No. I didn't mean to concede that, Your Honor, I'm sorry. Well, it's okay. I didn't mean to lose the case. I understand, but I don't concede that. I see you lose that point, but what I heard you say was what I just read from page 168 is inconsistent with our standard, but I heard you say that if you read the opinion in context as a whole, he didn't apply that invalid standard. That's one way to look at it. Is that not what you're saying? Well, the last part of that is definitely right. All I meant to say with respect to the first part is that this court, too, has used that same location in Carrera that you shouldn't have to have individualized fact-finding and has used it in the same way that the district court used it here, meaning you shouldn't have to have individualized fact-finding or you're not allowed to have individualized fact-finding every single time. Let's see if we can all, including your able adversary, let's see if we can all agree on what our standard is. The standard is there can be some individualized inquiry, but individualized inquiry can't dominate. If individualized inquiry dominates, that's sort of incompatible with the idea of an ascertainable class, right? Yes, absolutely. And the district court in a number of places... So explain to us why what I read isn't the driving force behind what the district court did. Happy to. So at pages A79, 87, and 89, that's where the district court is quoting and applying the standard from this court that there shouldn't be much individual fact-finding. At A84, the court says that the plaintiffs have not persuaded the court that there wouldn't need to be, quote, consideration of individual contractual relationships underlying each transaction. That's individualized fact-finding every time. At page 87 of the appendix, the court says that the target and MITRE disputes are emblematic of this. At page 89, the court says again, plaintiffs have not shown that they can identify without individualized inquiry to each individual, class members and cross-examples, let alone the millions of transactions at issue in this case. I think the only fair way to read the district court's opinion here is what it is saying is you would leave literally individualized fact-finding every single time you tried to ascertain a class member. That is a huge amount of individualized fact-finding. Which takes us back to where we started. Is it 1,300 or 20 million? You're saying because knowing the identity of the contracting party with the PBM does not tell you whether that contracting party is in fact an end payer, we must drill down to every one of those 20 million transactions. Yes, that's right. And that's a huge disagreement between you and Ms. Hunt. And if you're right, it seems obvious you should win. And if she's right, maybe they should win. Well, I don't agree with that, Your Honor, but maybe I could just give an example of where I think Ms. Hunt's gone wrong here. So one of the things that plaintiffs say is that when you have a fully insured plan and an insurer, the insurer will be the end payer, and the insurer will be the class member every single time and will also always be the PBM's client every single time. And therefore, you'll always know that the PBM's client in that instance is the end payer. We disagree with that. It's not an argument that plaintiffs have made before. There's evidence in the record that goes against that. What is the evidence? Well, at page A470, for instance, in a footnote of Dietz's opinion, he explains that there are times where there can be a separate administrator. So you can have a fully insured plan, an insurer, and an administrator. And in that instance, you could expect the administrator to be the PBM's client. You've got the whole government issue too, because as Caltech, you know, difference between county versus state, you know, is county part of the state or is it different from municipalities? That's of a similar ilk, right? I think that there are other problematic premises embedded in this 6% argument. That's just one of them. With respect to the Medicare... Well, wouldn't the district court say, I think the plaintiffs win on the government stuff. They can sweep that out. Yeah, that's true. How's that consistent? If it's like and like between the government and these other plans, and he says, government stuff, they can do that on the data. It's completely consistent and it shows that Judge Du Bois was calling balls and strikes here, I think. With respect to the government plans, plaintiffs were not relying purely on PDBM data. They identified a number of other data sources, including lists of government plans that they would look to and the district court recites that in his opinion. There was also a statement by a defendant's expert that the district court relied on in that section of the opinion. So I don't think it's fair to say that the district court accepted what they were saying about government plans. It means somehow that what Ms. Kraft was saying about this problem with respect to identifying the end payer in this context is somehow solved. What the district court... When we're talking about data, answer the American Antitrust Institute, not sure I got their name right, Antitrust Institute's amicus brief argument that there really is superb data here because all the players have the highest incentive to keep good and accurate data. And so the concern that AbbVie and others have skillfully created in the mind of Judge DuBois shouldn't be in our minds because this data is... This is the classic example of non-litigation driven data which you can rely on and which will lead you to a correct result because everybody outside of litigation who's involved in this has an incentive to be able to provide the data so you can trace and get to the correct end payer. So the fact that data may exist out there in the world does not mean there's a reliable and administratively feasible way of ascertaining the class. And so, yes, it may be true that... So what you're saying, when you just say, well... I suppose that's the end of the argument if you just say they're wrong. But what I hear them saying is, wait, outside the context of hired guns, no offense, including experts, when in the real world where people are doing business, they need to be able to identify whose wallet is going to get a hand put into. And they'll keep that data because it's valuable. And that's something that should be paid attention to. And it's not being adequately paid attention to. And by not adequately paying attention to it, there's an entire class of potential wrongdoing here that's being ignored because you will always, in a pharmaceutical end payer case, be able to say, what is being said here, which is it just can't be done. It just can't be done. And what the amicus seems to be saying is it can be done. Give it a chance to go forward. What's wrong with that argument? Well, we don't think that the record here shows that it can be done. And each case, of course, needs to be decided on its own record. We point to cases where end payer classes have been certified or end payer classes have not been certified in our briefs. It depends on the record in any particular case. Does it depend on the circuit that you're in? No. Does it depend on whether you went to the second or the third and that there's a problem with the circuit case law here? No, I don't think so at all. A number of circuits have agreed with this court's ascertainability precedence and even some courts that haven't, like the Second Circuit, actually conduct what's the equivalent of an ascertainability inquiry under the rubric of predominance, which this court has been very careful to separate out from ascertainability. So, no, I don't think it does depend. We are not suggesting that the court should rule that you can never certify a class of end payers in any pharmaceutical case. All that the court needs to do here is a very narrow ruling based on the standard of review that says the district court made factual findings about crafts, made factual findings about the PBM data not being enough, made factual findings about individualized fact finding being required every single time. Those are not clearly erroneous findings and the court didn't otherwise abuse its discretion. And I'd like to pick up also, if I could, on something that Your Honor said earlier on in the question, which is, is wrongdoing going to go unpunished here? The answer to that is, first of all, of course, we don't think there was wrongdoing, but if there was, the answer is no because there is a class of direct purchasers that has been certified that's proceeding in the district court and because the class members here are not individuals, they are entities including some very large insurers and when classes aren't certified in end payer cases, those entities are perfectly capable of bringing their own claims. So it's not as if we're going to walk away with a bag of money if this class doesn't get certified. That's not how it's going to work here. And so, I think the idea that somehow the fact that there's this data that is out there in the world somewhere will satisfy acceptability isn't enough and I'd like to give an example of that. So for instance, in this court's decision in Marcus, that was the case about the BMW purchasers and the deflated tires, right? So there's an ascertainability problem in that case. Now, if you had infinite time and infinite money and infinite investigative resources, could you go out, could you find the addresses of all the BMW purchasers and canvas the gas stations and the service stations around their area and look at their credit card bill and see if they bought replacement tires? Yes, you could. That data is out there in the world, but that doesn't mean that the class is ascertainable in some way. The problem that plaintiffs have here is that they're trying to twist the PBM data to do something that it is not designed to do because the PBMs have no special reason to want to know who the end payer is. The PBM has a client. That client, often an administrator or some other party, pays the PBM and that's what the PBM needs to know and Dietz testified that in his experience, PBMs may ask who the end payer is and maybe sometimes they're going to get told that, but often they're not because it's proprietary information that's going to be particularly true where you have a smaller PBM contracting with a larger PBM, which does happen and that's part of the multiple entity chains of payment that I was referring to before. So it's not the case that this PBM data just magically gives you the answer in some way and the affidavits argument is problematic. It's problematic not only for the reasons that we've been talking about so far, which are the lack of proof and the waiver, but it's problematic on other grounds as well because it flies in the face of these factual findings that the district court made that you would need individualized consideration of the underlying contractual relationships every single time. That essentially makes affidavits futile and because we know that named plaintiffs here sometimes got wrong things about the status of their own plans or didn't have the underlying contractual documents at hand because this cost period was a long time ago. And so I think if I could, let me just circle back to one other thing that came up. Sorry, I got lost during the argument here. All right, I'll give you a minute to wrap up. Okay, thank you. So another thing that was alluded to here is that the PBMs themselves can just identify the impayer because their affidavits say that they can. That is not a fair reading of those affidavits. They say that the PBMs can identify their clients. Their clients may or may not be impayers and PBMs probably don't know whether their clients are the impayers or whether they aren't. What I read said clients, but then Ms. Hunt pointed to other affidavits where it did say and it said third-party payers. Interestingly, it didn't say end-party payers. It says end-payers. I didn't know what to make of that. You keep reading that sentence, Your Honor. It says third-party payer clients. So I think third-party payer is a term that's sometimes used loosely. Well, if they're administrators, they're not payers, right? They're pass-through entities. Right, third-party payer is sometimes... It's ambiguous to some extent. Well, I think it's sometimes used loosely to mean somebody who paid in the chain rather than an impayer and it's not clear to me actually how the PBM affidavits are using that term, but I think at the least what they're saying is that they can identify their own clients and as we discussed, that's not necessarily good enough and as I think also came up earlier in the argument, the district court at 834 discounted those PBM declarations, which it was entitled to do as a fact finder and said that they weren't you know, they were very bare-bones, they were boilerplate, they didn't have enough facts to satisfy the district court, that there was an actual way that PBMs could do this. So again, I think this is a narrow case. It's a case that's driven by the standard of review. It is not some kind of broad referendum on class actions in pharmaceutical cases by any means. The district court made factual findings, accredited defendant's expert, it disbelieved plaintiff's expert and plaintiff's expert's testimony is the basis of plaintiff's arguments here and so that's enough to affirm the district court. Okay. Understood. Thanks, Ms. Goldenberg. We'll hear Ms. Hunt on rebuttal. Senator, I'd like to start with the PBM data and the fact that as AAI pointed out and others have represented, this is a very well-documented industry. No one disputes that for practical purposes, every time someone picks a prescription, the payment chain has to be recorded. Well, when you say it has to be recorded, when you say everybody agrees, I just, Ms. Goldenberg, say that's not true. The PBM's interest is in them getting paid. They don't give a hoot or holler whether somebody else gets paid so they don't care who the unpayer is. They care whether they get paid. Maybe I misunderstood her but I took that to be the import of her statement. Is she mistaken in what the incentives of the PBMs are? Even Dietz agrees that PBMs want to know and it's important for their financial security and this is in the 1080-1081 of his deposition testimony but setting that aside, all I was trying to say actually is that the PBMs have to track for the administrators who the unpayer is. So, the data is there. The data linking every prescription to an unpayer is in that PBM data. Is it possible for us to agree 100% with what you just said and still affirm the district court because the fact that as a matter of reality, it's there is a different question than whether you created a record in the district court to satisfy the district court that not only is it there but it's feasible to gather it all together and not have this extraordinarily complex process of chasing down 20 million claims. Yes, Your Honor. I think you have to... If you agree with me, it's there and I'll point you to Dietz in a bit as to why I think that's undisputed. Actually, while you're at it, tell us exactly what you're looking at when you say it. Dietz admits it. 981, I think is the best one. I'll read it for you. He was asked, do PBMs typically collect information on the ultimate payer associated with a given transaction? It's page 109 of his deposition, the bottom right-hand box. The PBM would collect information that would be passed to the ASO or the TPA so the ASO or TPA can conduct their business processes but I'm not aware the PBM's going to know in all situations the funding. Then he's asked, but does the PBM know in some situations? He says yes. He agrees... Wait a second. Maybe I'm just missing this. I'm looking at the exact same thing you are and you read it accurately. If the question is, does the PBM know in some situations the funding relationship and he says yes, how does that advance your argument? He's saying, yeah, you know it in some situations. The part that advances my argument is actually the part before that where he says they have to collect the information that's passed on to the ASO and you can see it on that same page at the top where they said the PBM's going to know the number. The PBM's going to know the end payer by a number at the very least. So the data is there and the only question is how manageable it is to interpret it. But when you say the data is there, okay, at least I now pinned to exactly what you're looking at but when you say the data is there, we do come back to this assertion that this is what the record shows and the district court was clearly erroneous in thinking something else was needed. So your honor, I think the district court got the law wrong and this is what I'm trying to get to. Please don't move off my question. My question to you is, is your position that the district court was clearly erroneous in thinking that the PBM data doesn't have it all because you're saying it's got it all. You've just quoted something from Dietz that you said it's all in the PBM data, period. And he didn't think that was so. Was he just clearly in error in that? Is that the legal position you're taking? My legal position would be he misapplied the law when he said that individualized fact finding would be required to interpret that data. Okay. I don't think he made a finding as to whether or not the PBM data included the end payers identity by number. I don't think he made a finding on that. How can we read his opinion as not concluding that the PBM data is insufficient in itself? I mean, it's the whole conclusion of his opinion. The PBM data is not sufficient in itself. I would have to do a whole bunch of other stuff. How is there any way to read that opinion other than that way? Your Honor, I agree. His opinion says the PBM data is insufficient, but I don't think it's factual. Embedded in that is there not necessarily a factual assertion? No, Your Honor, because the reason why I understand him to say the PBM data was insufficient is it would take some work to figure out what it means. And this is where I think this is different from Carrera or Martin. Wait, wait, wait. Let's pay attention to what he said. At the end of the day, he said, this will bring you down to a branch. Now we really are back to where we started. This will bring you down to a decision branch. They'll come, even with all their work and their PBM stuff, if we accept it, you'll hit a point where you won't know because it could be one thing or it could be the other, and at that point, you would have to do something else. It's not in the data. Is that not in his opinion? I think that's in his opinion. But the something else is taking a code and putting a name to it. So the point I'm making is the data has it at least in code, and all we're talking about is reading it. So again, I don't think the district court found the answer was unknowable or couldn't reliably be determined. I think he found it would take time. You're right. The opinion is rife with suggestion or commentary that it would be too unwieldy, too hard, too onerous. And that's why comments like Kraft's comment that, oh, this would just cost $250,000 were kind of laughable. Right? So I want to, yes, I want to step back and consider again what ascertainability is for. Carrera, Marcus, and Hayes all go off on the idea that the end of the day when someone comes forward with an affidavit and I said, I have a tire that ran flat or I bought a weight loss supplement, you can't verify that. Here, it's not going to be 20 million transactions because there's no universe in which this is done by transaction. And 24,000 class members is actually defendant's estimate and that's at A55. You're going to have 24,000... There's no universe where it's claim by claim. How... How do you know? How do you identify the class members, Ms. Hunt, if you don't know the character of the transaction? If you don't know... If the same parties and we've talked about this here, Judge Hardman asked you about it, it's all over the briefing, it's in Judge DuBois' opinion, the same players wear multiple hats. Because... So, how do you know without getting into transactional data? Well, because they can do what the named plaintiffs have done. So, let's say I'm Humana or United Healthcare. I act as an ASO. I also act as an insurer. I get notice of the class as no one's disputing or sending notice through a different database. I get notice and I go, oh, great, I'm going to print out the ones I paid for and I'm going to bring them to the claims administrator. Those records are there and they link to the PBM data because as Dietz agrees, that PBM data has to link to an NPO. Do that, then, to know whether they are in the class or not. You have to wait for them to tell you. You have to wait for them to say, my transactions are such that some of them are end-payer transactions. So, in the world in which we're setting Kraft entirely aside and we're just saying we discredit Kraft, which, again, I don't think the district court did because the district court accepted Kraft on government plans. But if we do that, then everyone agrees at the end of the day that there's an objective record that they're going to get noticed. There's no dispute. This has nothing to do with notice. They're going to get noticed. They're going to be able to do what the main plaintiffs did here, all ten of them, I'm sorry, all eight entities and bring forward records of their claims. You know, maybe this is just not to any person, but I'm going to try this one last time and it's probably my fault, so I apologize to you. Here's what I understand the district court to be saying. I understand, at the end of the day, the district court to be saying they can wear more than one hat. So whether they're really in that class or not, because he accepted everything from you folks except the final exclusion, right? He accepted that you could, that this was objectively verifiable as a class in theory. He accepted that you could manage five of your six exclusions and that that wouldn't be a problem. It all came down to this one thing. Can you separate out final payers from others, right? And I understand him to be saying, I don't think you can do that without getting into transactional data because the data that you got based on the record in front of me don't tell you end payer or not. Now, what I've just heard you say is that's not a problem because they'll get notice and they'll come forward and tell you. They'll come forward and tell you who they are. Right? Yes, Your Honor, and I think that's exactly what was the case in city select. Notice is going to go out to the whole database and they'll come forward. I got you. I think I understand your argument. Good? Okay. Thank you very much. It's well argued on both sides and helpful to the court. We got the matter under advisement.