Tuesday, November 13, 2018

Bill C-86 and the Use of Prosecution History in Claim Construction

Bill C-86 has introduced several proposed amendments to the Patent Act. The most problematic, in my view, is the proposed s 53.1 (introduced by s 191 of the Bill), which makes prosecution history relevant in claim construction. More specifically, the proposed s 53.1 provides a statutory definition of communications constituting the prosecution history, and provides that such communications “may be admitted into evidence to rebut any representation made by the patentee in the action . . . as to the construction of a claim in the patent.” (Section 187 of the Bill has an apparently related amendment to s 10(1) which would make the file history available to the public when the patent is laid open; and s 197 of the Bill introduces a new s 123.1 which provides that the same principle applies in respect of a certificate of supplementary protection.)

The proposed s 53.1 raises two questions: What is the intended effect? Is the change desirable? One might have thought that the effect of an amendment would be clear, and the difficult question is whether it is sound as a matter of policy. But s 53.1 is not entirely transparent, in part because “representation” is neither a term of art, nor even a commonly used term, in patent law. In Canadian law, at least at present, claim construction is an inquiry as to how a person of ordinary skill in the relevant art (the POSITA) would understand the claim, in light of the relevant context. The relevant context includes things such as the common general knowledge of a person skilled in the art, but not, under current law, the prosecution history. The relevant context is established on the basis of evidence, such as expert witnesses or textbooks. With the context established, the parties make submissions to the court (aka arguments) based on this evidence, as to how the POSITA would understand the claim. Does “any representation” in the proposed amendment refer to the evidence, or the arguments? And what would it mean to “rebut” a representation? Consider the parallel with the question of using legislative history in statutory interpretation – a patent is deemed to be a regulation by s 2(1) of the Interpretation Act, claim construction is like interpreting the statute (Whirlpool 2000 SCC 67 [49(e)]), and the prosecution history corresponds to the legislative history. Suppose that legislative history could not be used in statutory interpretation, and an amendment were made to the Interpretation Act saying that legislative history may be admitted into evidence to “rebut representations” made by a plaintiff in litigation as to the interpretation of the statute. What effect would that have?

Rather than trying to further parse the proposed s 53.1 directly, I will consider the various ways in which prosecution history is used in other legal systems which permit its use; whether s 53.1 might be understood as amending Canadian law analogously; and whether such a change would be desirable. (To emphasize the main points, I am overstating the distinction between these various approaches, which may run together at the margins.)

First, statements made by the patentee during prosecution might be admitted as evidence of what the POSITA would understand the claim to mean. This is apparently how it is used in German law: “it is permissible … to use statements made by the applicant [and the examiner] during the grant procedure as an indication of how the person skilled in the art understands the subject matter of the patent” but “such indications cannot be readily used as the sole basis for construction” (Eli Lilly v Actavis Group PTC, Case No X ZR 29/15 BGH, paras 39-40, as quoted in Actavis UK Ltd v Eli Lilly and Co, [2017] UKSC 48, [84]). This is an evidentiary approach: statements made by the patentee during prosecution are admissible as direct evidence of how the POSITA would have understood the claim, which is weighed along with all the other relevant evidence. That approach might be particularly useful in the context of claims which remain ambiguous in light of all the other evidence: see Actavis ibid noting the Dutch approach.

Is this what is intended by the proposed amendment? Perhaps, if we understand “representation” to refer to evidence, and “rebut” as meaning weigh against. Normally, we refer to a presumption being rebutted, where the presumption arises as a matter of law on certain evidence being established, and the presumption is rebutted by the introduction of evidence to the contrary. This typically has a flavor of a binary switch, rather than merely weighing as a matter of evidence. While the word “rebut” seems to be seldom used in Canadian legislation, I have found two instances of the use of the term, both of which refer to rebutting evidence by way of introducing other evidence: see Military Rules of Evidence — Can. Reg. 1049 (National Defence Act), s 21(1), and Canadian Transportation Agency General Rules [Repealed] — Can. Reg. 2005-35 (Canada Transportation Act), s 60(3). These examples link the concept of “rebutting” with evidence in a manner that is at least broadly similar to the way prosecution history is used in German and Dutch law. On the other hand, if the drafters intended that prosecution history could be used to rebut evidence, it would have been simpler to use the word “evidence” as in these examples. Nonetheless, this is at least a plausible interpretation.

Would an amendment implementing this evidentiary approach be desirable? In general, it would be a relatively modest change to Canadian law, which would simply have the effect of expanding the pool of evidence which is relevant to claim construction, without otherwise altering any basic principles. The advantage of this approach is that the extra evidence might help resolve ambiguous cases. The disadvantage is that it would add additional complexity to every case, in the hope of more easily resolving a few cases. It is important to recognize that the main effect of introducing prosecution history into claim construction will not be felt only, or even primarily, in litigation. In litigation the extra effort of reviewing the file history will be minimal. The problem is that in a freedom to operate search, the extra effort is multiplied over dozens or hundreds of patents that might need to be reviewed. But if the file history rarely has much impact, it may be practical to ignore it in an FOI analysis. On the whole, this strikes me as a change that might be either good or bad on balance, but the effect is not likely to be large either way.

There is a more specific problem. If the prosecution history is to be used as evidence, it is not clear why it should be used only to rebut a position taken by the patentee during litigation. For example, suppose that in litigation the defendant argues for a broad construction of a term, intending to argue that the claim, broadly construed, is invalid as being anticipated. The patentee argues for a narrower interpretation which would avoid the prior art. The prosecution history shows that the examiner had originally interpreted the term in the same broad sense as was being proposed by the defendant, and had wanted a narrowing amendment, but the patentee had responded by arguing that an amendment was unnecessary because the term actually had a narrow meaning – the same narrow meaning that it subsequently advanced in litigation. The examiner, on reconsideration, agreed that the POSITA would understand the term narrowly. In that context, the prosecution history would support the patentee’s litigation argument as to the meaning of the term. If the prosecution history is to be made admissible as evidence of how the term would be understood by a POSITA, it’s not clear to me why it should be admissible only to counter a position taken by the patentee, and not to support it.

Second, prosecution history may be used in the context of the doctrine of equivalents (DOE). More prominently, this is the primary way prosecution history is used in US. Under the doctrine of equivalents, the first step in claim construction is to understand what a POSITA would understand the claim to mean. Then, protection may be extended beyond that normal meaning (under certain conditions which define the scope of the DOE). In US law, the patentee is estopped from relying on the DOE to reclaim ground which was given up during prosecution, particularly by narrowing amendments to the claims, under certain conditions which define the scope of the doctrine of prosecution history estoppel. This approach is not merely evidentiary. Under this approach statements made in prosecution create a substantive limitation on the doctrine of equivalents.

Can the proposed s 53 be interpreted as introducing a prosecution history estoppel of this type? It is not clear whether the operative words “rebut any representation” are intended to effect a prosecution history estoppel, but there is a more basic problem: Canadian law doesn’t really have a two-step doctrine of equivalents. It doesn’t make sense to introduce prosecution history as a limit on the DOE when we don’t have a DOE in the first place. Now, we do have a doctrine of essential elements which is in some ways akin to a DOE. (And it is possible that the law might shift towards a true DOE under the influence of the UKSC Actavis decision.) However, it is difficult to read the proposed s 53.1 as having this effect, or at least as being limited to this effect, given that it makes no mention of the doctrine of essential elements.

Would this type of amendment be desirable? There is an intuitive fairness argument to the effect that the doctrine of equivalents does the patentee a favour by giving the patentee protection beyond what it had claimed, and it would be unfair to allow that favour to extend to subject matter that the patentee had disclaimed during prosecution. But it also matters why the amendment was made. In US law, prosecution history estoppel only applies to amendments made for a "substantial reason related to patentability": Warner-Jenkinson (1997) 520 US 17 at 33; Festo (2002) 535 US 722 at 735. In Warner-Jenkinson the Court indicated that an estoppel will normally arise when an amendment to a claim is made to avoid novelty-destroying prior art, but may not arise if the amendment is made to address an enablement objection, or to impart further precision to the claim: see Warner-Jenkinson at 32-33; Warner-Jenkinson Brief for United States as Amicus Curiae at 22-23, cited in Warner-Jenkinson ibid. In Festo the Court clarified that an amendment made for any reason of patentability may give rise to an estoppel (736), but a complete bar was inappropriate (737). Instead, a narrowing amendment would not normally give up “equivalents unforeseeable at the time of the amendment and beyond a fair interpretation of what was surrendered,” nor would it prevent claims of equivalence “for aspects of the invention that have only a peripheral relation to the reason the amendment was submitted” (738). There are further complications when the reason for the change is not apparent, leading to questions of burden: Warner-Jenkinson at 33-34. This is enough to give a flavor for the issues that arise, and I won’t try to further summarize the law. At this point is suffices to say that even within patent law, which is sometimes said to be a particularly complex area of law, the interaction of prosecution history estoppel and the doctrine of equivalents is notoriously complex. In Free World.2000 SCC 66 [66], the SCC rejected the use of prosecution history, saying:

To allow such extrinsic evidence for the purpose of defining the monopoly would undermine the public notice function of the claims, and increase uncertainty as well as fuelling the already overheated engines of patent litigation.

The US experience amply testifies to the validity of these concerns. As Lord Hoffmann noted in Kirin-Amgen [2004] UKHL 46 [39]:

the difficulties involved in deciding exactly what part of the claim should be taken to have been withdrawn by an amendment drove the Federal Court of Appeals in Festo Corporation v Shoketsu Kinzoku Kogyo Kabushiki Co Ltd 234 F3rd 558 (2000) to declare that the law was arbitrary and unworkable. Lourie J said:

The only settled expectation currently existing is the expectation that clever attorneys can argue infringement outside the scope of the claims all the way through this Court of Appeals.

I think it is fair to say that the current US approach is generally considered to be unsatisfactory: see eg Sarnoff, Abolishing the Doctrine of Equivalents and Claiming the Future After Festo, (2004) 19 Berkeley Tech LJ 1157. No doubt it would be possible to develop a better body of doctrine, but the US experience shows, at the very least, that the problem is a difficult one.

A third approach is that statements made in prosecution history might be binding on the patentee as a substantive rule, on the view that it should not be permissible for the patentee to say one thing to get the patent granted and then turn around and say something else during litigation to win its case. This is related to second approach, just discussed, in that the patentee would be bound by its prior statements as a substantive rule, with the difference that this substantive rule would apply to claim construction generally, and not just in the context of the DOE. US law does have a doctrine of this type, referred to as “prosecution disclaimer”: see eg Omega Eng’g, Inc. v. Raytek Corp., 334 F.3d 1314, 1326 (Fed. Cir. 2003).

Can the proposed s 53.1 be understood as implementing a doctrine of this type? This interpretation avoids the objection to the second alternative, that there is no reference to the doctrine of essential elements. It is possible that “representation” refers to substantive submissions. While “representation” is not a term which is used in patent law, in the law generally it tends to have the meaning of a statement made by a person as to a state of affairs which is relied on by another, as per the tort of negligent misrepresentation, or the Competition Act s 52. That meaning doesn’t seem applicable, as submissions made by counsel to the court isn’t really an assertion as to a state of affairs. However, occasionally it is used to mean submissions to a court or tribunal: see Immigration Division Rules (SOR/2002-229) s 48. So that meaning is a bit obscure, but it is plausible that that is what was intended by the proposed s 53.1. As noted above, other legislation seems to use “rebut” to mean “weigh against, though it could plausibly be taken to mean substantively “limit” or “estop.” Indeed, perhaps “rebut” is used rather than a more standard term like “estop”, precisely to mean that it will not necessarily estop the patentee from making the argument in question. Rather, the patentee’s statement during prosecution will be weighed against the statements (representations) made during litigation. So, it is at least plausible that “to rebut any representation made by the patentee” was intended to mean something like “to weigh against submissions made by the patentee during an action.” This seems like a sensible interpretation, though perhaps a bit strained on the language.

Would such an amendment be desirable? The broad form of prosecution disclaimer does not operate solely as a limitation on the doctrine of equivalents. That means that in principle it would apply even to claim language that is in itself clear to a POSITA. (That is not to say prosecution disclaimer is normally used that way in US law.) This type of amendment would apply not only to ambiguous claims, as with the first approach, nor only to claims where we have extended protection beyond the language, but also to a claim that is perfectly clear on its face.

In US law, prosecution disclaimer is justified in part because “The public has a right to rely on such definitive statements made during prosecution. Notice is an important function of the patent prosecution process”: Digital Biometrics, Inc. v. Identix, Inc., 149 F.3d 1335, 1347 (Fed Cir1998). This cannot be the rationale for the amendment, if it intended to implement the third approach, because under current Canadian law the public does not have the right to rely on statements made in prosecution; notice is not a function of the patent prosecution process. It would be circular to argue that the law should be amended to allow the public to rely on statements made during prosecution because the public is entitled to rely on statements made during prosecution.

Thus this kind of amendment is not based on improving the notice function of the claims. Rather, it appears to be based on the idea that it is not fair to allow the patentee to say one thing to the examiner to get its patent granted, and then another thing to the court when it comes to enforcement. This intuitive fairness argument is much weaker than for prosecution history estoppel as applied as part of the DOE. As noted above, the DOE gives the patentee protection beyond the claims as literally interpreted, and in that context prosecution history prevents that extended protection from encompassing material which was effectively disclaimed in prosecution by a narrowing amendment. In this third approach, prosecution history prevents the patentee from claiming material which was not disclaimed in prosecution, and which, indeed, was actually claimed.

Whatever the merits of this fairness argument, it will, as the SCC noted in Free World, undermine the notice function of the claims. Under present law, whatever the patentee might have said to the examiner, it has said only one thing to the public, namely what is in the claims. The claims, in this scenario, may be entirely clear, and the effect of the proposed amendment would be to change the meaning of otherwise clear claims in the interest of this perceived unfairness. The question is whether the unfairness of allowing the patentee to change its position between prosecution and litigation, outweighs the adverse effect on the notice function. Consequently, in US law, prosecution disclaimer is applicable only when the patentee has “unequivocally disavowed a certain meaning to obtain his patent”: Omega Eng’g, 1324. That is, there is a requirement that the disclaimer be sufficiently clear. Further, as discussed above, the reason the statement or amendment was made matters in the context of prosecution history estoppel. Presumably it would also matter in the context of prosecution disclaimer, thus giving rise to the same problems of complexity.

Thus, in assessing the balance between the potential unfairness in allowing the patentee to say one thing in prosecution and another in litigation, against the notice function, we must recognize that introducing prosecution disclaimer will substantially increase the cost and complexity of litigation. We have seen that is true even of the more limited use of prosecution history estoppel in the context of the DOE. As with the use of prosecution history in the context of the DOE, it is far from clear that the game is worth the candle.

Now, I should acknowledge the foregoing has overstated the distinction between these approaches in order to highlight some of the key issues. In practice, they may run together. So, the distinction between the first and third approaches turns on how much weight is given to the prior statement by the patentee in comparison with the other evidence. The same kind of gradation applies in terms of the policy considerations. At one end of the spectrum, prosecution history is just one more source of evidence as to the meaning of the claims, to be weighed along with all the other evidence. In the second approach, it is not just one source of evidence, but conclusive evidence, amounting to a substantive rule, that is applicable to the DOE. In the third approach, it is a substantive rule which applies broadly. As the power and breadth of the use of prosecution history increases, it becomes more complex, as more safeguards are needed to ensure a just result. As the power and breadth of the use of prosecution history increases, it also requires more careful consideration, as the potential effects, both for good and for ill, are magnified. For a powerful doctrine to be beneficial on balance, it must be carefully limited. The complexity of the doctrine of US prosecution history estoppel reflects the difficulty of crafting suitable limits.

This problem is illustrated by the Actavis decision, [2017] UKSC 48, in which Lord Neuberger recognized the possibility of using prosecution history in claim construction under UK law. Lord Neuberger held that as a general matter, the prosecution history should only be used where “the point at issue is truly unclear if one confines oneself to the specification and claims of the patent, and the contents of the file unambiguously resolve the point” – that is, the first approach discussed above – or where it would be “contrary to the public interest for the contents of the file to be ignored” [88]. The second scenario “would be exemplified by a case where the patentee had made it clear to the EPO that he was not seeking to contend that his patent, if granted, would extend its scope to the sort of variant which he now claims infringes” [89]. The word “variant” implies this refers to the context of the doctrine of equivalents – that is, the second approach described above. The word “exemplified” suggests it is not confined to that context, so the third approach is also possible, but has not yet been directly addressed in UK law.

In Actavis, the question of prosecution history arose in the context of the doctrine of equivalents. After the Court had provisionally concluded that the scope of the claims extended beyond its language, the question was whether the prosecution history should cause the Court to revise that provisional conclusion [81]. On the facts, it was perfectly clear that the original claims would have encompassed the variant in question and the patentee had narrowed its claims in response to an objection by the examiner: Actavis [76]-[80]. Lord Neuberger nonetheless declined to revise the provisional conclusion [89]. He drew a distinction between “the patentee's response to a disclosure or added matter objection,” which he apparently considered would not normally give rise to an estoppel, and an amendment made to overcome an objection based on the prior art, which Lord Neuberger described as “a very different point” [91] (without saying that it would necessarily, or even normally, give rise to an estoppel). The idea, apparently, is that there is a distinction between an objection which goes to scope, such as novelty, and one which does not, such as disclosure [90]. Now, I must admit that the rationale for the distinction is not immediately clear to me; if the amendment narrows the claim, so that the patentee expressly gave up subject matter which it later seeks to reclaim, I’m not sure why the reason for the narrowing amendment is relevant: see Festo 736-37. With that said, there is undoubtedly something to the point, which escapes me due to my lack of familiarity with prosecution history estoppel, as US law does – or at least used to – draw a very similar distinction, as noted above. Whether Lord Neuberger’s distinction is well-drawn or not, the point is that the very first decision of the UKSC to deal with prosecution history estoppel has already raised a difficult distinction, the bounds and implications of which are uncertain.

I have found the proposed s 53.1 somewhat difficult to interpret, but that is a saving virtue; to the extent that it is open-ended, maybe the courts will have more or less free reign to make something sensible out of it. I must say that I don’t feel that the operative phrase “rebut any representations” will help in making good law, and it may hinder to the extent that the courts feel constrained by it to depart from what would otherwise be good policy. If the use of prosecution history is to be allowed, I would rather have seen either an entirely open-ended approach, to the effect that prosecution history is admissible in the context of claim construction, to allow the courts the unconstrained ability to develop sound doctrine; or, a much more specific provision which would have directly addressed the difficulties canvassed above. This provision is certainly much closer to the unconstrained end of the spectrum, and if it is enacted as is, we may hope that the courts interpret “rebut any representations” flexibly.

Even under an open-ended approach the problem is a very difficult one. The introduction of prosecution history into Canadian law will no doubt be good business for lawyers and academics. Whether it is good for the Canadian patent system will depend on exactly how the details of the doctrine develop. There is a very good argument to be made that the US doctrine does more harm than good. Perhaps we will go down a better road. But Actavis suggests that there is no easy road. And in any event, we will be guaranteed a lengthy period of uncertainty – at least a decade, and likely more – as the courts try to develop the doctrine.

1 comment:

  1. Excellent post, Norman.

    From the examination side of things, it may be worth pointing out that CIPO applies very different criteria to the assessment of essential elements than the courts do (you have written about this, of course). Applicants routinely contend with examiner assertions that claim elements are "non-essential" during examination, according to CIPO's prescribed approach. One is sometimes forced to argue "essentiality" under that framework during examination merely to have an examiner read and give full consideration to all claim features.

    It is difficult to envisage how file wrapper estoppel will work with CIPO's approach being so divergent from the claims construction of the courts. What is a "representation" when the courts are interested in backgammon and CIPO has insisted on flipping over the board to play checkers?