Friday, December 14, 2012

Comity Cannot Solve the Problem of Inconsistent NOC Decisions

Apotex Inc v Allergan Inc / COMBIGAN (NOC) 2012 FCA 308 Noël JA: Stratas, Webb JJA aff’g for different reasons 2012 FC 767 Hughes J
            2,440,764 – COMBIGAN – brimonidine / timolol

While the FCA has affirmed Hughes J’s decision (blogged here) granting an order of prohibition preventing Apotex from launching a generic version of COMBIGAN, the FCA decision is a thoroughgoing repudiation of Hughes J’s reasoning, in respect of both the main issues, comity and obviousness. The FCA decision is important on both these points. Today’s post will deal with comity and related issues, while my next post will focus on obviousness. 

The ‘764 patent covers a composition of two active ingredients, brimonidine and timolol, for the treatment of glaucoma. Such a composition is sold by Allergan as COMBIGAN. The ‘ 764 patent had previously been the subject of another NOC proceeding between Allergan and Sandoz, in which Crampton J had granted the order of prohibition sought by Allergan: Allergan v Sandoz / COMBIGAN (NOC) 2011 FC 1316. As discussed in my posts here and here, the sole issue respecting the ‘764 patent in the Sandoz / COMBIGAN litigation was whether it was invalid for obviousness. Crampton J held it was not. In the decision under appeal in Apotex / COMBIGAN (NOC), Hughes J found that the ‘764 patent was invalid for obviousness [FC 189], expressly disagreeing with Crampton J. This raised the question of the extent to which a FC judge should be influenced by a prior inconsistent decision. Hughes J therefore went on to review the case law and proposed a five-part approach for a court to adopt when faced with an earlier NOC decision involving failed invalidity allegations [FC 82]. Despite his conclusion that the claimed invention was obvious, Hughes J then went on to issue the prohibition order sought by Allergan, though it does not appear that Hughes J actually applied the five-part test which he had developed; rather he granted the order on the basis that the law regarding comity, and hence the precedential effect of Crampton J’s conclusion, required clarification by the FCA, and “[t]he only practical way to get the matter before the Court of Appeal is for me to grant the Order for prohibition in the likely expectation that Apotex will appeal” [FC 194].

The question of the extent to which a FC judge should be influenced by a prior decision is an important one. Recently, in Merck / alendronate (NOC) 2012 FC 1235 Hughes J similarly invoked comity in following a prior a decision on a point of law with which he disagreed, while in Janssen-Ortho / levofloxacin (NOC) 2009 FCA 212 the FCA reprimanded Shore J 2008 FC 744 for having failed to make his determination of facts independently from the prior decision of Hughes J. Nonetheless, on appeal, all were agreed that Hughes J was wrong to grant the order for the purpose of provoking an appeal and having the law clarified. “It is common ground [between the parties], and I agree,” wrote Noël JA, “that prohibition could only be granted if the conditions set out in the PM(NOC) Regulations were met” [33]. 

While Hughes J was wrong to issue the order simply to have the law clarified, he nonetheless achieved his objective, in that the FCA went on to clarify the law, albeit not in the way Hughes J had suggested. Hughes J had pointed out that it will be considered an abuse of process for a patentee to bring an application for prohibition against a second generic in respect of a patent which had earlier been held to be invalid in proceedings against a first generic, while a second generic will not be estopped from challenging a patent that had previously been found valid [FC 69-75], and he remarked that this was “somewhat contradictory” [193]. The FCA responded at [40-41] that

contrary to the Federal Court judge’s view, the decisions of this Court on point are not inconsistent (reasons, para. 193).

As noted by Sexton J.A. in Novopharm [2007 FCA 163], each party must put forward its entire case, complete with all relevant evidence, at first instance. He emphasized that this applies both ways: “Generics likewise must put forward their full case at the first opportunity” (Novopharm, para. 50).

I entirely agree with this conclusion: see my post discussing Hughes J’s decision. The reason for the apparent inconsistency is simply that in the former situation the patentee had already had an opportunity to put its case forward, while in the latter situation, the second generic will not yet have had a chance to be heard. 

Moreover, as Noël JA pointed out, the cases analyzed by Hughes J were really directed at the doctrine of abuse of process [40]. Nonetheless, as well as reaffirming the law respecting abuse of process, the FCA went on to discuss of the doctrine of comity at [43]-[48]. The FCA rejected the five-step approach proposed by Hughes J. The Court summarized the law by noting first that “the principle of judicial comity only applies to determinations of law. It has no application to factual findings” [44]. When a question of law is at issue

the general view appears to be that the conclusions of law of a Federal Court judge will not be departed from by another judge unless he or she is convinced that the departure is necessary and can articulate cogent reasons for doing so. On this test, departures should be rare [48].

This is conceptually logical, and is consistent with the Court’s decision in Janssen-Ortho / levofloxacin (NOC), in which questions of fact were in issue. In my view it also provides a reasonable balance between consistency in law while avoiding the problem, discussed here, of having the law locked in by a single decision at first instance. However, in practice it is unlikely to provide consistency of outcome, because validity turns on questions of fact as much as questions of law. This is a pervasive problem in the law generally. It would be naive to suppose that two different judges would necessarily come to the same conclusion on the same facts when applying the same legal principles to construe a contract or determine negligence. The problem is simply less embarrassing in other areas, because it almost never happens that different judges will decide exactly the same issues on the same facts, while this is common under the NOC regime. (Note the similar embarrassment stemming from inconsistent national decisions regarding the same patent that is partially driving the Unified Patent Court in Europe.) Moreover, as discussed below, I am not sure that departures on a point of law will be as rare as supposed by the FCA. With all that said, I can’t think of a better rule respecting comity than that set out by Noël JA. 

Hughes J had held that comity did not apply to the same extent in the interpretation of a patent because interpretation of a patent is more like a interpretation of a contract than like interpretation of a statute or regulation [FC 149]. Against this, Allergan had relied on Binnie J’s statement in Whirlpool 2000 SCC 67, [49e] that “When the patent issues, it is an enactment within the definition of "regulation" in s. 2(1) of the Interpretation ActR.S.C., 1985, c. I-21.” Hughes J, relying on an article by Hayhurst, on “The Distinction between ‘Letters Patent’ and ‘Patent Specification’” (2007), 57 CPR (4th) 161, had countered that “the regulation spoken of is the one-page document attached to the patent specification, which is the page granting the patent, not the patent itself. The specification is a document drafted by the patentee, not Parliament or the Governor in Council” [FC 149]. The FCA rejected Hughes J’s view that comity does not apply to interpretation of a patent for this reason: “Construing a patent in order to identify the inventive concept is no less an exercise that leads to a determination of law because the document being construed is drafted by the patentee. . . The analogy which the Federal Court judge drew with a contract (ibidem) is no more helpful since all else being equal a contract should not be subjected to contradictory constructions any more than a patent, or a statute” [52]. In my view the FCA is entirely correct on this point. Lord Hoffmann explicitly pointed out in Kirin-Amgen [2004] UKHL 46 at [30] that a contract and a patent specification are governed by the same principles of interpretation, on the basis of the general principle that both are communications for a practical purpose and so should be construed as it would be by the people to whom they are addressed. This is entirely apart from the Canadian Interpretation Act. (I must add that my view is that Lord Hoffmann’s interpretation of the particular patent at issue in Kirin-Amgen was a disaster, but that does not affect the broader principle.)

While that is enough to dispose of the point, I can’t help but point out that so far as the Interpretation Act goes, there is nothing to the point made by Hughes J and Hayhurst at 167, that the specification is not drafted by the Parliament or the Governor in Council. Legislation is drafted by legislative drafters, not by Parliament. These are often government employees, but a document drafted by a government bureaucrat does not derive any legitimacy from that fact, and in any event it is not uncommon for legislation to be initially drafted by outside contractors (especially at the provincial level). Legislation gets its authority, not from the person who actually holds the pen, but from the Crown’s approval of the final product. Moreover, while Hayhurst’s detailed historical analysis of the origin of the provision is interesting, he fails to undertake a purposive analysis of the definition of “regulation” in the Interpretation Act, which is of course required by modern principles of statutory construction. Consequently, his conclusion that only the cover page of a patent is a defined to be “regulation” verges on the absurd (in the statutory interpretation sense), as it is difficult to see what purpose there would be in legislatively providing that the title and abstract, but not the rest of the patent, are defined to be legislation.

At the more specific level of the application of comity to determinations of obviousness, the FCA held that “construing a patent in order to identify the inventive concept . . .gives rise to a question of law,” [50] and the doctrine of comity therefore applies, while“[a] finding that an invention is obvious because the solution proposed was plain to see is one of fact,” and comity therefore does not apply [50]. As will be discussed in more detail in my next post, Hughes J disagreed with Crampton J as to the construction of the inventive concept. (Hughes J also gave less weight to the evidence of Allergan’s expert witness than did Crampton J, but this was not relevant to his construction of the inventive concept, and so did not affect the application of comity to that construction.) The FCA held that because construction of the inventive concept is a question of law, “unless [Hughes J] could demonstrate that Crampton J.’s construction of the patent in order to determine the inventive concept was wrong or that distinct evidence adduced before him compelled him to reach a different conclusion, it would have been preferable for him to adhere to it” [50].

The Court went on to say that “[Hughes J] did not identify any error nor did he rely on distinct evidence to explain his diverging view. He simply chose to construe the patent differently and held that the inventive concept did not extend to the improved safety profile which Crampton J. had included” [51]. This is not entirely fair to Hughes J. Crampton J was of the view that it is “both necessary and permissible” to look to the patent as a whole to determine the inventive concept [51], while Hughes J took the narrower view that the “inventive concept” is merely “a statement of what the claim, properly construed, says ‘stripped of unnecessary verbiage’” [137], and that the patent as a whole could be considered only in relatively narrow circumstances [141]. This amounts to a disagreement on a point of law. Consistently with this, Hughes J’s interpretation of the inventive concept at [145] was based directly on the claim itself, and was consequently narrower than that accepted by Crampton J. Hughes J also explained why he believed that Crampton J’s interpretation was wrong, even in light of the specification as a whole [146-147]. Thus in my view Hughes J did provide reasons based both in law and on the evidence as to why Crampton J’s construction was wrong, and therefore, even on the test articulated by the FCA, he would have been justified in departing from Crampton J’s construction.

Whether or not Hughes J should have followed Crampton J’s interpretation as a matter of comity, he did not. Consequently, the FCA noted, “there are now conflicting and equally authoritative decisions as to how the ’764 patent is to be construed and it falls upon this Court to determine which is the correct one” [53]. As will be discussed in more detail in the next post, the Court went on to hold that Hughes J’s preferred approach to construing the inventive concept was wrong as a matter of law [FCA 72]. Because of this error, his conclusion that the claims were obvious could not stand [75], and after de novo review, the FCA held that the inventive concept was not obvious and the claims were therefore valid [92].

If I am right that comity would not have prevented Hughes J from departing from Crampton J’s construction of the inventive concept, this case illustrates that while Hughes J’s solution may have been off the mark, the problem of inconsistent decisions that he was concerned about remains very much alive. If Hughes J had applied the test for comity as stated by the FCA and then departed from Crampton J’s conclusions, he would have refused to grant the order of prohibition. While we now know that he would have been wrong to do so, as his analysis of the law was wrong, the question would have become moot as soon as the NOC issued to Apotex, and the FCA would have refused to hear an appeal. Other generics would then also seek NOCs, which could not be contested by Allergan, as to do so would be an abuse of process. (Presumably, however, Sandoz would not be able to obtain an NOC as it would be an abuse of process for it to relitigate the same issue in respect of the same patent.) The consequence would have been that Hughes J’s erroneous legal analysis of obviousness would have been determinative in practice of the right of generics to enter the market under the NOC proceedings, though of course his error would have been open to correction in a subsequent infringement action. Pharma patentees will view this as one more illustration of why innovators need a right of appeal in NOC proceedings, while generics will no doubt respond that the infringement action provides a sufficient remedy. Another response would be to amend the regulations to provide for joinder of interested parties, as Hughes J pointed out is possible in US law [FC 76]. Whatever the ultimate solution to the problem of inconsistent NOC decisions, I am inclined to think that the FCA approach to comity is sound, and while it will not solve the problem, this is because the problem cannot be resolved by an appeal to comity.

No comments:

Post a Comment