Wednesday, February 22, 2012

Guest Post - Pay For Delay, There For The Taking?

This is the first gues post that DDL has had by a law student in its five+ year history.  We weren't sure at first, when the offer to post was made, but once we read it, our concerns vanished.  Heck, it's probably better written than half the stuff we throw out there.

So all of what follows - and all credit or blame therefor - belongs to Brenna Jenny, a Harvard Law 3L whom we now know is a dedicated fan of the blog.


The most topical constitutional issue implicated by the pharmaceutical industry has been the intersection between off-label promotion and the First Amendment.  While we continue to await the Second Circuit's decision in United States v. Caronia, a new constitutional consideration has been receiving increased attention: the Takings Clause.  In a recent article in Food and Drug Law Journal, Professor Richard Epstein argues that the Biologics Price Competition and Innovation Act of 2009 ("Biosimilars Act") raises Fifth Amendment concerns.  66 Food & Drug L.J. 285 (2011).  Professor Epstein's argument may have force against other legislative fixes Congress would seek to apply to the pharmaceutical industry, such as Representative Bobby Rush's (D-IL) recent proposal (HR 3995) to ban all reverse payment settlements between brand and generic drug manufacturers.

First some background on the Biosimilars Act.  In order to facilitate FDA approval of "biosimilar" biological products (the analogue to generic drugs in the Hatch-Waxman context), the Biosimilars Act allows the FDA to rely on the pioneer's biologics license application (:BLA") when determining whether a new entrant's product is "highly similar" to the existing version.  (The FDA earlier this month released some much-anticipated draft guidance on this, and other, topics).  As under Hatch-Waxman, the second-comer is allowed to introduce far less clinical data than the innovator, and this shortcut allows the copycat product to make it to market sooner, with lower cost.  The Biosimilars Act mirrors the quid pro quo created in the Hatch-Waxman Act:  although innovators lose on one hand (the trade secrets disclosed in their applications are used by the FDA in approving a competitor's products) they gain on the other (the innovator not only is granted a twelve-year period of exclusivity, but the filing of a biosimilar application is considered an artificial act of infringement, allowing the innovator to file suit and litigate any patent claims before the biosimilar reaches the market.)

Companies choosing to file BLAs after the passage of the Biosimilars Act can be seen as "opting in" to this quid pro quo, agreeing to allow the FDA to indirectly use their intellectual property in exchange for the benefits provided by the Act.  But Professor Epstein maintains that applicants who filed BLAs before passage of the Act:
had investment-backed expectations - based in statute, FDA regulations, and longstanding FDA practice - that their data would not be used or relied on by the agency, directly or indirectly, for the purpose of approving competitors.  Taking these trade secrets for the benefit of a competitor thus requires just compensation in order to avoid a constitutional violation.

The term "investment-backed expectations" alludes to the Supreme Court case Penn Central Transportation Co. v. New York, in which the Court held that regulatory interference with the investment-backed expectations of property owners is a critical factor weighing in favor of compensation.  438 U.S. 104 (1978).

Despite universal agreement that patents are property, whether patents fall within the protection of the Takings Clause is actually an unsettled issue.  Although a 2006 Federal Circuit decision held that patent infringement by the government is not a cognizable Fifth Amendment violation, Zoltek Corp. v. United States, 442 F.3d 1345, 1352 (Fed. Cir. 2006), Congress should not feel liberated from the constraints of the Takings Clause when regulating the intellectual property of the pharmaceutical industry.  Just four years prior to Zoltek, the Supreme Court insisted that "courts must be cautious before adopting changes that disrupt the settled expectations of the inventing community," because "[f]undamental alterations ... risk destroying the legitimate expectations of inventors in their property."  Festo v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 739 (2002).  This language directly reflects the Court's discussion of investment-backed expectations in Penn Central.

Therefore, when Congress acts in ways that significantly alter the "legitimate expectations of inventors in their" patents, it does so at least under the shadow of the Fifth Amendment.  A recent legislative proposal regarding reverse payment settlements modifies the ability of patent holders to craft settlements, threatening to trigger the Takings Clause.  Reverse payment settlements have been a hot topic recently, as the FTC and DOJ have found themselves stalled at an utter impasse with the courts over how to treat them.  The FTC in particular has begged Congress to step in and act.

What are these settlements and why do they create such an uproar?  Reverse payment settlements occur when a generic drug company seeks to enter the market before the expiration of a patent.  Citing either patent invalidity or non-infringement, the generic manufacturer, pursuant to the Hatch-Waxman Act, must notify the branded company of its intention to enter.  The branded company then has an opportunity to invoke a thirty-month stay, during which it can litigate the patent claims.  Many of these cases end in settlement, which may involve a payment from the branded company to the generic, with an agreement that the generic will not enter the market until a pre-determined date (prior to the patent's expiration).  The FTC and DOJ have argued that these settlements are nothing but a smoke screen for a branded company to buy off generic competition by sharing a portion of its monopoly profits.

Although the FTC and the DOJ have passionately insisted these settlements violate antitrust laws, and should be deemed presumptively illegal, courts have consistently been unpersuaded.  Both the Federal Circuit and the Second Circuit (and arguably the Eleventh Circuit as well, although its case law is subject to differing interpretations) have granted reverse payment settlements significant room to thrive:  as long as the branded company's suit against the generic is not a "sham," and the resulting settlement meets a few low hurdles, such as allowing other generic companies to subsequently challenge the patent and not restricting the marketing of non-infringing products, the court settlement does not violate antitrust laws.  Arkansas Carpenters Health & Welfare Fund v. Bayer AG, 604 F.3d 98, 106 (2d Cir. 2010). Citing a patent's presumption of validity, courts have tended to view these settlements as falling within the bounds of a patent holder's property rights.

Previous congressional proposals, including one introduced last year by Senator Kohl (S. 27), would have granted the FTC and DOJ's requests and made reverse payment settlements presumptively illegal.  Rep. Bobby Rush's bill goes a step further: any settlement involving a generic drug filer receiving something of value and agreeing not to research, develop, or sell a drug that is the subject of an infringement claim would be an "unfair method of competition," in violation of section 5 of the FTC Act.

This bill takes an arrow out of the proverbial patent holder's quiver of property rights, and a segment of settlement options is now off the table.  When considering reverse payment settlements, several courts have approvingly cited Judge Posner's insistence that a patent holder "is entitled to defend the patent's validity in court, to sue alleged infringers ... whatever its private doubts ... and to settle the suit to avoid risking the loss of the rights.  No one can be certain that he will prevail in a patent suit."  Asahi Glass Co., Ltd. v. Pentech Pharmaceuticals, Inc., 289 F. Supp. 2d 986, 993 (N.D. Ill. 2003).  Some opinions have even gone so far as to say that reverse payment settlements are a "natural consequence" of the Hatch-Waxman Act.  King Drug Co. of Florence, Inc. v. Cephalon, Inc., 702 F. Supp. 2d 514, 529 (E.D. Pa. 2010).  When patent holders filed under the previous regulatory regime, their disclosure provided them with a right to exclude, by settlement if necessary, and courts have recognized this right.  Rep. Rush's bill would alter those expectations.  Given the near infamous expense and uncertainty associated with patent litigation, it is no small change to hamstring a patent holder's terms of settlement when his property is challenged.

To be sure, patent holders certainly cannot have an investment-backed expectation of engaging in monopolization.  But although antitrust law prohibits patent misuse, HR 3995 raises concerns that the government is leveraging antitrust law to take away property rights which patent holders possess under Hatch-Waxman.  Courts that have addressed reverse payment settlements have drawn a line around activities taken within the patent's zone of the right to exclude and segregated them - as valid exercises of a property right - from the reach of antitrust law.  As the Federal Circuit has pointed out, patents are inherently anticompetitive, and there is a certain degree of anticompetitive behavior which antitrust law cannot touch.  In re Ciprofloxacin Hydrochloride Antitrust Litig., 544 F.3d 1323, 1333 (Fed. Cir. 2008).  By redefining "unfair method of competition," HR 3995 seeks to expand the boundary line of what antitrust law can reach, and correspondingly what will no longer fall within the property rights of a patent holder's right to exclude.  If Congress simply changed the scope of the property right, without invoking antitrust, patent holders may have a real claim to a taking under Penn Central.  The question posed by Rep. Rush's bill is not whether it is good policy to redefine an "unfair method of competition," in order to ban reverse payment settlements, but rather whether the government can use antitrust law as an end-run around a takings inquiry.  As Congress continues to contemplate if, and how, it wishes to step in and change courts' treatment of reverse payment settlements, it should be aware that the pharmaceutical industry may have valid constitutional claims to raise about its methods.


Max Kennerly said...

Nice post, but it raises two issues:

First, if, as you argue, it's a "taking" within the meaning of the Takings Clause when Congress alters the rights available to patent infringement plaintiffs (by regulating reverse settlement payments), then wouldn't you agree that, if Congress alters the rights available to personal injury plaintiffs with accrued claims, then that, too, constitutes a "taking" worthy of compensation? Would you agree that, for example, the Medical Device Amendments of 1976 (MDA) to the Food, Drug, and Cosmetic Act, which killed off thousands of accrued state law tort claims against medical device manufacturers, constituted a "taking?"

Second, why should any party be granted a property right in information they voluntarily shared with the government to obtain a license to sell the product and the many protections that come with that license? Clinical data on medical devices, standing alone, is worthless: the value comes in applying that data either to obtain governmental approval or in applying that data to develop something else that will obtain government approval. Similarly, what property was actually "taken" here when, as you note, the company retains the right to file a patent infringement suit, and can file the same even before real infringement begins?

As a general matter of IP law, inventors get the choice of withholding their inventions as trade secrets or publishing them and earning patent protection. Under your argument, medical device companies are entitled to the full benefits of trade secret and patent protection. That would be anomalous and contrary to the purposes underlying the distinction between trade secrets and patents, and so some real justification for that needs to be given.

Brenna said...

Hi Max,
Let me address your second point first. You seem to be asking why, as normative matter, patents should be property rights. That normative question is beyond the scope of the issues I address: the fact is, patents ARE property rights (see, e.g., Festo, 535 U.S. 722, 730 "The patent laws 'promote the Progress of Science and useful Arts' by rewarding innovation with a temporary monopoly. U.S. Const., Art. I, § 8, cl. 8. The monopoly is a property right."). What is unclear is whether they are "constitutional property," in other words whether they are protected by the Takings Clause. (for an excellent argument about why they are, see Adam Mossoff, Patents as Constitutional Private Property: The Historical Protection of Patents Under the Takings Clause, 87 B.U. L. Rev. 689). The only question then, given that patents are property, is how does that limit the ways in which Congress can regulate them? Any answer will necessarily remain speculation until the Supreme Court rules on the issue, but the unsettled nature of the inquiry is something Congress should keep in mind. There are ways to address reverse payment settlements that don't involve such blunt instruments, and would definitively steer clear of the Takings Clause - so why not use them?
This property right in the use of the underlying molecule or process is generally granted far in advance of a company actually obtaining FDA approval to sell a drug or device. So in that sense, an applicant's disclosure is not really performed in order "to obtain a license," but rather to obtain a patent because if you don't, someone else might, and then you'll have to pay license fees to use your competitor's patented substance or process. (The biologics context is different, and more closely aligned with your comment, in that an application involves private disclosure to the government of trade secrets, in return for a sort of government license). I disagree with the characterization of clinical data as "worthless" - it is valuable precisely because one can use it to get a patent. You're right that in terms of value to society, raw clinical data is relatively worthless, as compared to a final, usable product that is safe and effective. And one can certainly imagine a quite reasonable world in which patents are unavailable until someone crosses the finish line, so to speak, and obtains FDA approval on a drug or device. But that is not the regulatory world we live in.
As for what property right would be taken: Property, or so the metaphor goes, is like a bundle of sticks. It's true that patent holders would retain many of the sticks in their bundle, such as filing for infringement, but some sticks are taken away by HR 3995, namely the ability to settle on a certain broad category of terms. This is but one facet of an intangible right, to be sure, but seems implicit in a property right involving a temporary right to exclude, at least in the view of many courts that have considered the issue. As Judge Posner explained, the right to settle with challengers is part of the property right. Additionally, courts seem to view settlement as a particularly important aspect of the Hatch-Waxman regulatory scheme which patent filers opted into. The reason why courts have viewed reverse payment settlements as a "natural consequence" of Hatch-Waxman is because the Act places a disproportionate amount of risk on the patent holder in litigation, which increases his desire to settle. If you're interested, I've included an excellent summary, by the Second Circuit, at the bottom of this post explaining why this is the case.*

Max Kennerly said...

I don't doubt that a patent is property and I assume for purposes of this discussion that if the government literally "took" a patent — as in, suddenly wiped it away, allowing all the public to use the invention — that such would be a "taking." (Put aside, for the moment, the United States itself infringing on a patent; crafting a remedy raises complicated issues of sovereignty.)

My focus is a more narrow because the bills themselves are more narrow. None of the proposed bills do anything remotely like completely "taking" a patent in the same way that, say, eminent domain "takes" real estate from its owner. The pending bills (1) allow the government to use in a different context information that was voluntarily given to it in exchange for the FDA's approval or (2) restrict a particular form of settlement that bears no relationship to true enforcement / exclusivity of a patent but which damages consumers & insurers & the government by inhibiting the proper functioning of the market for drugs and medical devices.

It's not enough to say that patents create a property right and so anything that affects the value of a patent is a taking. That's like saying real estate is property, and ergo every last zoning restriction or building code amounts to a "taking." I have an "investment-backed expectation" that none of the homes on my street have home businesses operating out of them, but if the township grants a variance on one of them, I can't claim that's a taking. So it goes with the two bills you've discussed: there is at best a vague claim of damage, one that doesn't seem at best speculative, much less so concrete and certain — like ownership of land — as to justify mandated compensation.

Brenna said...

As to the sovereignty issues, I'll just note up front that the Supreme Court in Jacobs v. United States (290 U.S. 13 (1933)) found that the Taking Clause is self-executing, so the United States need not waive sovereign immunity in a particular statute in order to be a defendant in a takings claim.

Eminent domain is a per se taking, but regulatory takings, as here, need not be as extreme as complete loss of the physical property in order to be a compensable taking. For example, the Supreme Court in Kaiser Aetna was faced with a dispute between the United States and a marina owner (444 U.S. 164 (1979)). The government wanted to require the marina to provide public access to its waters, in order to allow the public to access a government-owned pond. The Court held that Congress could force the marina to do so, but it had to compensate for the loss of exclusive use.
You're right that "It's not enough to say that patents create a property right and so anything that affects the value of a patent is a taking." Under Penn Central, this is clearly true, and certainly government in the modern regulatory state would be unable to function if otherwise. I noted that the bill may implicate the Takings Clause, thereby necessitating a Penn Central analysis, but ultimately the inquiry is whether the regulation which infringes on the patent holder's exercise of his property rights is significant enough to be compensable. Your arguments go to the belief that they are not. But the focus of my note was less on whether I think patent holders would ultimately prevail under a Penn Central analysis, and more that Rep. Rush's bill is veering into territory that could require the application of this balancing test. Given the uncertainty of the law in this area, I was not arguing that patent holders would necessarily win, but more that they might have a cognizable claim - so why not just craft a better piece of legislation that avoids the problem?

On that note, I also disagree with your implication that every reverse payment settlement damages consumers. There is actually significant disagreement in the literature; some certainly share your views, but others do not. What everyone agrees on is that these settlements are bad if they cause generics to enter the market later than they would have without a settlement (in other words, later than they would have if the parties litigated to an end result). But if the patent holder won the underlying case, and could continue to exclude until the patent expiration, then a reverse payment settlement is procompetitive, because it led to generic entry well before the end of the patent. Which outcome would have resulted from the underlying litigation is purely but-for speculation. What we want then is a test which bars only the anticompetitive settlements. There really has not been sufficient empirical evidence to show that these settlements warrant "per se illegal" treatment, and consistent with antitrust doctrine, they should therefore be analyzed under the "rule of reason." Along this vein, several commentators have suggested multi-factor tests courts could use in making such an inquiry. One that I particularly like, set forth by Butler & Jarosch (96 Iowa L. Rev. 57, 115-19) examines factors such as: the size of the reverse payment in comparison to saved litigation costs and the value of the patent, length of time between the expiration of the patent and when the generic is allowed to enter, ability of the ANDA filer to market the drug without the financial assistance provided by the payment, etc. Courts may reach a point in time when reverse payment settlements consistently fail a rule of reason inquiry, and at that point per se treatment would be appropriate - but not yet.

Bexis said...

I'll leave the takings related issues to Brenna, since that's more up her alley than mine. The first point about "taking" and preemption (or indeed any other statute that eliminates causes of action) is interesting. Due process being what it is, the same question of whether there's an established property right in an inchoate cause of action would seem to govern (1) whether there's a taking, or (2) whether a particular statute is unconstitutional for eliminating "accrued" claims. The plaintiffs' bar has preferred to attack allegedly retroactive tort reform statutes under #2 rather than under #1. Your side has had decent success with that. I admit I haven't followed it closely, but I don't know of any resorts to #2 where #1 has failed.

As far as the MDA. That statute was passed in 1976. I don't think it's ever been used against a claim that was "accrued" at the time the statue went into effect. Any post-1976 claims would not have a takings right because congress did not change an "accrued claim." Preemption didn't really come to tort law until after Cippolone, which was late 1980s at the latest. It didn't come to the MDA until those first collagen cases in something like 1993 or so. Thus I don't think that MDA preemption can be considered a taking, because no preempted claim had "accrued" when the statute passed, which has been the relevant date for Due Process purposes.

Bexis said...

I'll leave the takings related issues to Brenna, since that's more up her alley than mine. The first point about "taking" and preemption (or indeed any other statute that eliminates causes of action) is interesting. Due process being what it is, the same question of whether there's an established property right in an inchoate cause of action would seem to govern (1) whether there's a taking, or (2) whether a particular statute is unconstitutional for eliminating "accrued" claims. The plaintiffs' bar has preferred to attack allegedly retroactive tort reform statutes under #2 rather than under #1. Your side has had decent success with that. I admit I haven't followed it closely, but I don't know of any resorts to #2 where #1 has failed.

As far as the MDA. That statute was passed in 1976. I don't think it's ever been used against a claim that was "accrued" at the time the statue went into effect. Any post-1976 claims would not have a takings right because congress did not change an "accrued claim." Preemption didn't really come to tort law until after Cippolone, which was late 1980s at the latest. It didn't come to the MDA until those first collagen cases in something like 1993 or so. Thus I don't think that MDA preemption can be considered a taking, because no preempted claim had "accrued" when the statute passed, which has been the relevant date for Due Process purposes.

Soronel Haetir said...

Brenna ,

The entire point of patent protection is that the applicant, even before gaining their exclusive license, has to disclose what it is they have. And that disclosure, after the license lapses, is then available to everyone. While I fully realize that the FDA process is not handled through the PTO I do believe the courts would find that that this activity falls under the same power of government to grant exclusivity in exchange for disclosure.

Actually, come to think of it, has there been litigation in the past over whether FDA -- or PTO for that matter -- even has the power to grant such a license in the absence of public disclosure?

Unless there has been such litigation (which obviously would have had to come out in FDA's favor for them to have continued operating in the manner they have) I don't don't think you have a winning argument that just because FDA has been withholding information that rightly belongs to the public that they must continue doing so (or at least keep doing so for information acquired before a certain date).

Brenna said...

Soronel - I'm not claiming that the terms of what a patent holder gets in exchange for his disclosure can't be changed - they can, the Hatch-Waxman Act itself is certainly evidence of that. But if the change to the patent property rights is significant and if it is applied mid-stream to patent holders who already disclosed under one quid pro quo regime but now, after they have disclosed, receive a set of rights other than the rights available to them when they filed - this is a scenario in which it is at least plausible to consider whether the Takings Clause is implicated.
Also, the thrust of my arguments are based on Hatch-Waxman and reverse payment settlements. In that context (unlike with biologics), you're not dealing with a situation of the FDA using an applicant's (non-publicly disclosed) trade secrets. Instead, the issue is a narrowing of the patent holder's property rights (by limiting their ability to settle with alleged infringers) following their earlier disclosure. This has nothing to do with the FDA continuing to withhold information. (For more on that issue, see Professor Epstein's article; the rules governing biologics are a different animal. Pre-Biologics Act, the FDA wasn't "withholding information that rightly belonged to the public" because the system of obtaining approval for a biologic generally involves use of a trade secret in the approved product, not a patent, or at least the disclosure of the trade secret after twelve years is the point of contention for pre-Act filers. Trade secrets rightly belong to their owner, not the public, because the owner has chosen to forego the property rights attendant to a patent in exchange for being able to keep his innovation a secret.)
As for your question - I'm a little unclear about what you mean by "license"; do you mean FDA approval to sell a drug for a certain use that has been proven safe and effective? The PTO grants patents, not licenses. An applicant certainly must make a public disclosure in order to get a patent from the PTO, but an applicant for FDA approval of a drug need not publicly disclose any heretofore undisclosed information.

Soronel Haetir said...

Perhaps I am unclear on the specifics here, I thought that an successful new drug application entitled the filer to exclusive rights for a time regardless of whether the filer also received patent protection from PTO. Certainly I see no constitutional reason that such grants need to be limited to PTO, the power belongs to Congress and Congress could choose to have any number of agencies wield part of it.

In the case of biologics, I was further under the impression that irregardless of the processes being guarded as trade secrets that FDA does in fact disallow duplicate products during a period of exclusive license. That company B can't just come along with a duplicate product a year after company A brought some product to market (assuming that B in fact duplicated all the work to create the product, not relying on disclosure). If my understanding is correct that FDA does in fact grant monopolistic rights I don't see a company getting very far with a claim that the FDA filings are meant to remain secret. Now if B actually is free to duplicate the work and bring an equivalent product to market without relying on A's secrets I believe A would have a pretty good case for the disclosure remaining their property.

But if the grant is monopolistic I don't see the fact that FDA was lax on requirements of possibly constitutional scope as being particularly relevant. If that is the situation then A would be relying on a promise the government had no power to make. And I do see the requirement of public disclosure before monopolistic rights are granted as being of likely constitutional significance. I think there would be an excellent argument for unconstitutionality, for example, if Congress were to pass a statute purporting to remove the requirement of public disclosure before a patent is granted. I just don't think the word 'patent' or going through the PTO has any magic meaning that doesn't attach to any market monopoly based on invention (other forms of monopoly grant could get entirely different answers because they would flow from other powers of Congress).

Brenna said...

As an initial point, let's be sure we're on the same page as far as "monopoly" terminology. A patent is not a monopoly in the way we normally think of monopolies, in that one can command economic monopoly profits. Instead, it's a legal monopoly: the holder has exclusive right to make and use his invention. But someone could come up with a competing product that doesn't infringe the patent, and consumers would not be facing a monopoly in terms of product choices, or paying monopoly prices. The Supreme Court is perhaps at its most lucid on this point in Illinois Toolworks.
Your statement in the first paragraph is correct to a certain extent. There is a "new drug product exclusivity" which provides some holders of NDAs with certain benefits, such as that the FDA won't approve a competing ANDA for a certain period of time (three or five years, versus a patent period of twenty years). Note that the FDA isn't granting a patent though, in the sense that no 'exclusive right to make or use' has been awarded. Additionally, this benefit is only available for some NDAs: "new chemical entities" and drugs which represent "significant changes" to already approved drugs. My impression is that new drug product exclusivity tends to come into play in situations where the innovation is often otherwise unpatentable (maybe it is viewed as obvious, or there is a public use bar), yet the FDA still wants to incentivize research into beneficial improvements.
As to your second paragraph, let me preface my comments by saying that the intention of my post was not to defend Professor Epstein's thesis regarding the biologics market, and I am far less familiar with the Biosimilars Act than Hatch-Waxman. You're right that the Biosimilars Act creates the period of exclusivity, but my impression is that this period of exclusivity was completely new - it was one of the new carrots to balance out one of the new sticks (such as that the FDA can eventually use the trade secrets in a BLA to approve a competitor's biosimilar). It's incorrect to say that the company can't "get[] very far with a claim that the FDA filings are meant to remain secret" - the FDA has in fact promised exactly that. Furthermore, this period of exclusivity is not a "grant [of ]monopolistic rights," it just means that the FDA won't approve a copycat product (using the innovator's data) during that time period. But if another innovator, say C, comes along and submits an application for a new biologic, even if C's biologic could compete on the market with A's earlier product (they have some overlapping treatment uses), the FDA would still approve it. Professor Epstein's idea is that there is a partial taking of A's property when the FDA uses his confidential, trade secret information (which wasn't really "disclosed" as we use that term in the patent context) for the purpose of approving a competitor's drug. A invested a lot to get the know-how encompassed in those filings, and the knowledge is no longer exclusively A's, because B is able to short-circuit all of that expensive R&D and get a biosimilar approved on the basis of A's know-how - in other words, B can indirectly use A's know-how/intellectual property.

I'm not entirely clear on your criticism of the arguments in the Hatch-Waxman/reverse payment settlement context.