Focus of the Obviousness Inquiry Should be on the Solution to the Objective Problem

Apotex Inc v Shire LLC 2021 FCA 52 Rennie JA: de Montigny, Gleason JJA affg 2018 FC 637 Fothergill J

            2,527,646 / lisdexamfetamine [LDX] / VYVANSE / NOC

My previous post discussed some puzzles that I see as arising from Vyvanse. In this post I’ll try to address those puzzles. In Bristol-Myers 2017 FCA 76 [65] Pelletier JA remarked that

It may be helpful to keep in mind that the obviousness analysis asks whether the distance between two points in the development of the art can be bridged by the Skilled Person using only the common general knowledge available to such a person. If so, it is obvious. The first of those points is the state of the prior art at the relevant date. References in the jurisprudence to “the inventive concept”, “the solution taught by the patent”, “what is claimed” or simply “the invention” are attempts to define the second point.

While Pelletier JA went on to note that these terms are often treated as being synonymous [66], there is arguably a distinction to be drawn between a focus on the information disclosed in the specification, and the claimed subject-matter. This has given rise to a debate as to whether the proper focus of the obviousness inquiry—the “second point”—is the inventive concept or the claims. The debate has centred on the second and third steps of the Windsurfing / Pozzoli approach to the obviousness determination, endorsed in Sanofi 2008 SCC 61 [67], which seems to distinguish the inventive concept from what is claimed while at the same time treating them as alternative means of defining the second point. In Ciba 2017 FCA 225 Pelletier JA indicated that the claims were the proper focus of the inquiry. In Vyvanse Rennie JA held that it is inventive concept that is the ultimate the end point for the obviousness inquiry [65]. At the same time, Rennie JA indicated that the inventive concept is the same as the “solution taught by the patent,” as that term was used in Bristol-Myers. In this post I will argue that Bristol-Meyers holds the key, though it is somewhat buried. The proper focus of the inquiry is not the inventive concept, nor is it the essential elements of the claim; it is whether the subject-matter defined by the claim is the solution to the objective problem that would have faced the skilled person.

There is certainly a powerful and well-founded intuition that the focus of the inquiry should be on the inventive concept. I’ll start off by making the case in favour of the inventive concept, and then I’ll show where that approach breaks down. My next post will look at where this leaves the Windsurfing test.

1.1.1   The quid pro quo is the new information provided by the patent

In this section I’ll make the argument in favour of a focus on the inventive concept. The basic argument is that the quid pro quo for the patent exclusivity is the new information disclosed in the specification and so we should focus on that new information in assessing obviousness.

What is the invention? Viscount Dunedin answered this question in Pope Appliance (1929) 46 RPC 23 (JCPC) 55:

After all, what is invention? It is finding out something which has not been found out by other people. . . [A] patent represents a quid pro quo. The quid to the patentee is the monopoly; the quo is that he presents to the public the knowledge which they have not got.

See also Halocarbon [1979] 2 SCR 929, 944–45 and Pfizer v Ratiopharm 2006 FCA 214 [23], quoting Pope Appliance. The SCC made the same point long ago, in Ball v Crompton Corset Co (1887) 13 SCR 469, 477, quoting with approval Smith v Nicholls 88 US 112 (1874) 118, stating “A patentable invention is a mental result”; Dann's Patent [1971] 88 RPC 425 (HL) 452 (Lord Diplock) “The counterpart of the temporary monopoly granted by the patent was the provision of the information”; Aerotel [2006] EWCA Civ 1371 [32] (Jacob LJ) “Patents are essentially about information as to what to make or do”; Biogen [1996] UKHL 18 [14] (Lord Hoffmann) “Whenever anything inventive is done for the first time it is the result of the addition of a new idea to the existing stock of knowledge”; American Home Products [2000] RPC 547 (Pat) [18] (Laddie J) “The invention is a novel mental concept which is embodied in, or exploited through, new products or processes.”

In short, “An invention is a piece of information”: Merrell Dow [1995] UKHL 14 [28]; Baker Petrolite 2002 FCA 158 [42]. More particularly, it is new information—the knowledge which the public “have not got.” The new information provided by the specification is sometimes referred to as the “invention,” but it is also often referred to as the inventor’s “discovery” or “idea,” depending on which term is more natural in context: see eg Shell Oil [1982] 2 SCR 536; Tye-Sil (1991) 35 CPR(3d) 350 (FCA).

This new information is provided to the public by the disclosure. So in Consolboard [1981] 1 SCR 504, 517, the SCC famously stated that “The description of the invention [required by s 27(3)] is the quid pro quo for which the inventor is given a monopoly for a limited term of years on the invention.”

Thus, it is well established that the quid pro quo for the patent is new information disclosed in the specification. The aim of the anticipation requirement is to ensure that the information provided by the specification is indeed new: Merrell Dow [1995] UKHL 14 [28]; Baker Petrolite 2002 FCA 158 [42]. The aim of the non-obviousness requirement is to ensure that the new information provided by the specification is an advance sufficient to warrant the reward of a patent: “Is there such an addition to the stock of human knowledge as to entitle the Patentee to a monopoly?”¹

The new information may in principle be found anywhere in the specification, but the information is normally found primarily or entirely in the description, not the claims. The function of the description is to disclose the invention, while the function of the claims is to define the scope of the monopoly.

Now, the new information disclosed in the specification—the “mental result” or “new idea” or “novel mental concept”—cannot be patented as such: “A disembodied idea is not per se patentable. But it will be patentable if it has a method of practical application”: Shell Oil [1982] 2 SCR 536, 554; and see similarly Tye-Sil (1991) 35 CPR(3d) 350 (FCA) 364–65; Christiani v Rice [1930] SCR 443, 457. It is the claims that define a practical embodiment of the new information. As the SCC stated in Ball v Crompton Corset Co (1887) 13 SCR 469, 477, quoting with approval Smith v Nicholls 88 US 112 (1874) 118:

A patentable invention is a mental result. . . . The machine, process or product is but its material reflex and embodiment.

Thus, what is claimed is related to the new information disclosed in the specification, but it is not the same as that new information; the difference is between the new information itself and a practical embodiment of that information.

This suggests that the focus of the obviousness inquiry should be the new information that is disclosed in the specification and embodied in the claim. Both conditions must be satisfied. It is not enough to claim something new and inventive if the specification does not actually disclose the information needed to attain it: a perpetual motion machine would undoubtedly be new, useful, and inventive, and many applications have been filed claiming one, but none have been granted, as none so far has disclosed the information necessary to actually construct a perpetual motion machine. At the same time, it is not enough that the specification discloses new and inventive information, if that information is not embodied by the claimed subject matter: so, in BVD v Canadian Celanese [1937] SCR 221, the specification disclosed a method of stiffening shirt collars that was new, but claimed a method that was old, and so the claim at issue was invalid.

To summarize, the quid pro quo for the invention is the new information supplied in the patent. If the new information would have been obvious to a skilled person, then the patentee has not met its end of the bargain. So, if the new information disclosed in the patent and embodied in a claim would have been obvious to a skilled person, the claim will be invalid for obviousness.

My next post will discuss how the term “inventive concept” is used in the leading cases. I will suggest that it is used in two or three ways. Sometimes it is used to refer to the information disclosed in the specification that is actually new. I will refer to this as the “objective inventive concept.” In other cases it is used to mean the information disclosed in the specification that the applicant believed was new. I will refer to that as the “subjective inventive concept.” It is also commonly used to refer to the information that the patentee, during litigation, argues is new. I will call that the “putative inventive concept.” All of these are related in that they refer to different aspects of the information disclosed in the specification. In many cases they collapse into the same thing. When the applicant was right about what information in the specification was actually new, the subjective and objective inventive concept are the same, and that is put forward in litigation as the inventive concept. In those cases, which are very common, the question is not what information should be the focus of the analysis, but whether that information actually required inventive ingenuity. It is only necessary to make the more refined distinctions I have suggested when we are trying to pin down what was intended by the various steps of the Windsurfing test, or when the nature of the inventive concept is contentious.

For now, we can focus on the more common case. I will suggest that the “inventive concept” corresponds broadly to the new information disclosed in the patent. For example, in Vyvanse, the new information disclosed in the patent included the properties of the claimed compounds, and the focus of the analysis was whether those properties were inventive.

1.1.2  Vyvanse

In this section I’ll suggest that an analysis that identifies the inventive concept with the new information provided by the patent is largely consistent with Vyvanse and provides a rationale for many of the points made by Rennie JA. The next section follows up with “yes, but. . . .”

Identifying the inventive concept with the new information provided by the patent explains the appeal of Rennie JA’s statement that “The inventive concept, properly construed and applied, remains the end point for the obviousness inquiry” [65]. The quid pro quo for the invention is the new information supplied in the patent. If the new information would have been obvious to a skilled person, then the patentee has not met its end of the bargain. Since the bargain requires the patentee to provide information and not a thing, it makes sense to focus on the information itself and not on the thing that is claimed.

A focus on the information implies a focus on the specification. The role of the disclosure is to disclose new information which is the quid pro quo for the patent. The role of the claims is to define a practical embodiment of that inventive concept. These roles are distinct. It is possible for elements of the claim to expressly reflect the inventive concept, but it is not necessary. The point is illustrated by the canonical example, emphasized by Rennie JA in Vyvanse, where the invention is based on the discovery that a novel compound is effective in treating a disease. The new information is the discovery that the compound treats the disease; what is claimed is not that information as such but the physical compound. The compound is not the same as the information: a skilled person would not be able to discover, from examination of the compound alone, that it is effective in treating the disease. The compound nonetheless embodies the information in a practical form, because the physical compound will effectively treat the disease if administered to a patient.

So far, this supports Rennie JA’s analysis in Vyvanse. However, the same point about the different roles of the disclosure and the claims suggests that it will never be possible to identify the inventive concept from the claims alone. The subject-matter defined by the claim must embody the inventive concept, but it need not expressly state the inventive concept. Again, this is illustrated by the example of a bare chemical compound claim, where the inventive concept lies in the properties of the compound, and those properties are embodied in the claimed compound [96]. But this is not a special case. It is simply a particularly clear illustration of the general point that the description discloses the information and the claims embody it.

There is no need for the claims to expressly reflect the inventive concept, so long as the claimed subject matter embodies the inventive concept. Conversely, the express elements of the claim often comprise subject matter that is not new or inventive at all. This is a consequence of the requirement that the claims must define a practical embodiment of the new information. For example, in Shell Oil, the new information was that certain previously known compounds were useful as plant growth regulators (538), but to be used practically for that purpose, it was necessary to mix the compounds with adjuvants (547). The claims were drawn accordingly, even though the need to mix the compounds with an adjuvant was not part of the new information contributed by the patentee. A similar example is a pharmaceutical composition with claims to a specified compound combined with a pharmaceutically acceptable carrier. In these examples, we can often make a good guess as to what elements are inventive—but formulation claims are also common, and sometimes the inventive step really does lie in finding an appropriate salt form, or the right adjuvant or other excipient. In general, claims will often combine some elements that directly reflect the inventive concept, and some that are necessary to define a practical embodiment, and we can’t tell which is which simply by looking at the claims. That means we have to identify the inventive concept before we can decide whether it is expressly present in a particular claim.

This also explains why different claims can have different inventive concepts. The information has to be disclosed in the specification, but it has to be embodied in the claims. Different claims may embody different aspects of the information disclosed in the specification.

As discussed in my last post, patents may have different claims to make it easier to establish infringement. A patent based on the discovery that a compound is medically useful might claim the compound in both a tablet form and in an injectable form, where preparation of the salt form and the injectable form are both routine. The inventive concept would be exactly the same in those two claims.

But the second main reason to use multiple claims is that the patentee might be wrong about what the inventive concept really is. Consider the 446 patent, at issue in Teva / Viagra 2012 SCC 60, which disclosed that PDEv inhibitors can be administered orally to treat ED. Accordingly, Claim 27 was to the use of a PDEv inhibitor administered orally to treat ED. The specification also disclosed that sildenafil is a PDEv inhibitor, and Claim 7 was to the use of sildenafil administered orally to treat ED. Claim 6 was to a different compound, also a PDEv inhibitor, administered orally to treat ED. This illustrates Rennie JA’s point about “a single, overarching inventive concept [that] connects every claim of a patent” [86]. The overarching inventive concept, reflected in Claim 27, is that PDEv inhibitors can be administered orally to treat ED. Claim 7 is connected to that concept by the fact that sildenafil is a PDEv inhibitor. Claim 6 is similarly connected to that concept by the fact that the claimed compound is also a PDEv inhibitor. Claims 6 and 7 embody different inventive concepts—that the two different compounds can be administered orally to treat ED—but they are connected by the single overarching inventive concept that PDEv inhibitors can be administered orally to treat ED.

Now consider a twist on the Viagra scenario. Suppose that unbeknownst to the patentee, the use of a PDEv inhibitors to treat ED was already known in the form of the traditional Chinese herbal medicine yin yang huo (aka epimedium, aka horny goat weed), which contains a small amount of icariin, a known PDEv inhibitor. If that were the case, then the overarching (subjective) inventive concept, reflected in Claim 27, that PDEv inhibitors can be administered orally to treat ED, would not be inventive after all. However, the inventive concept of Claim 7, that sildenafil can be administered orally to treat ED, would still be inventive. In that example, the difference between the overarching inventive concept and the specific inventive concept matters; one is inventive and the other isn’t.

It might be said that in the forgoing example there is nonetheless a single overarching inventive concept. Now consider a modified version of this example. Suppose a team of researchers from Company A discovered that PDEv inhibitors could be administered orally to treat ED. However, their lead compound and preferred compounds were structurally entirely different from sildenafil. They obtained a patent — the 111 patent — with Claim 27 being to the use of PDEv inhibitors administered orally to treat ED. Let’s suppose this claim is valid (forget about epimedium). The inventive concept of Claim 27 is that discovery that PDEv inhibitors could be administered orally to treat ED—just as in the 446 patent.

After that patent is laid open, other companies enter the field. Company B surprisingly and unexpectedly discovers that sildenafil is a PDEv inhibitor and obtains a patent — the 222 patent — with Claim 7 to the use of sildenafil administered orally to treat ED. The inventive concept of Claim 7 of the 222 patent is not that PDEv inhibitors can be administered orally to treat ED, since that was already known. It is that sildenafil is a PDEv inhibitor. That inventive concept is not any part of the inventive concept of the 111 patent, as the inventors of the 111 patent did not know that sildenafil was a PDEv inhibitor.

In this scenario, Claim 27 and Claim 7 have completely different inventive concepts. The two patents are related by the fact that sildenafil is in fact a PDEv inhibitor — the 222 is a selection patent over the 111 patent—but the inventive concepts are entirely distinct. It is not possible to assess the inventiveness of Claim 7 of the 222 patent by looking to the overarching inventive concept of the 111 patent. If that is true when Claim 7 and Claim 27 are in different patents, where Claim 27 is prior art over Claim 7, I would suggest that it must also be true if Claim 7 and Claim 27 are in the same patent, but Claim 27 is anticipated, so that the subject-matter of Claim 27 is effectively part of the prior art over Claim 7.

Consequently, I would suggest that it is not possible to assess obviousness in light of a single overarching inventive concept because different claims are likely to have different inventive concepts whenever the inventors are mistaken about what they have actually discovered. But this example also illustrates that this point is entirely consistent with the view that we must assess obviousness in light of the information in the disclosure. It is the information in the disclosure and embodied in the particular claim that must be new and non-obvious if the claim is to be valid. The reason that different claims can have different inventive concepts is that they may reflect different information that is disclosed in the specification. This different information may be joined by an overarching inventive concept, but it need not be.

More generally, it doesn’t matter if a patent with two completely different inventions—say a new kind of paddle wheeler and a new kind of steam engine (see Morgan v Seaward (1836) 1 WPC 187)—somehow slips through the administrative unity of invention requirement. The inventive concept of a particular claim is that new information that is disclosed in the specification and embodied in a particular claim. This is true whether or not there is an overarching inventive concept. This point, that different claims have different inventive concepts, does not imply that the focus on the claim must be what is claimed as opposed to the information in the disclosure. It means that the information in the disclosure that is the focus of the analysis must be that which is embodied in the claim at issue.

1.1.3  Not all new and non-obvious information warrants a patent

Now for the “but.” I have argued above that if the new information disclosed in the patent and embodied in a claim would have been obvious to a skilled person, the claim will be invalid for obviousness. In this section, I will make two points. First, this requirement is necessary but not sufficient for the claim to be non-obvious. There are cases in which new and non-obvious information about the subject-matter of the claim is disclosed in the specification, and yet the claims are nonetheless invalid for obviousness. Second, I will argue that it follows that a focus on the inventive concept is not always adequate.

To the first point, consider the 277 patent at issue in Wellcome / AZT 2002 SCC 77. AZT was a known compound. The inventors discovered and disclosed that it was useful in the treatment of HIV/AIDS. Claim 22, to the use of AZT in treating HIV/AIDS, was valid. Claim 1, to AZT itself, was not valid: [2001] 1 FC 495 (FCA) [27]. This is even though the specification disclosed new and non-obvious information, namely that AZT is useful in treating HIV/AIDS, and Claim 1, to AZT itself, embodied that new information in just the same way as it would if AZT had been a novel compound, so that Claim 1 would have been valid. The principle is that “[m]erely knowing something new about something old is not enough”: Genentech Inc's Patent [1989] RPC 147 (CA).

Of course, Claim 1 of the 277 patent was invalid for anticipation, but the same principle applies in the context of obviousness. To paraphrase, merely knowing something unexpected about something obvious is not enough. The point was made in Hallen v Brabantia [1991] RPC 195 (CA) 216, noting that “an added benefit, however great, will not found a valid patent if the claimed innovation is obvious for another purpose." The same point was made more recently in Actavis v ICOS [2019] UKSC 15, in which the patentee, in the course of routine dose ranging studies, had unexpectedly discovered that tadalafil was effective in treating ED at a very low dose. The claim was invalid even though the discovery was a surprise [88]:

It was obvious to embark on that exercise and carry out tests in a routine way until that appropriate dose was ascertained. . . . The fact that tadalafil at the dose of 5mg, while remaining effective as a treatment of ED, also, and unexpectedly, had the additional benefit of reduced side effects was an added benefit which does not prevent the identification of 5mg as the appropriate dose from being obvious.

Thus, even though the specification disclosed new and non-obvious information about the subject-matter embodied in the claim, the claim was nonetheless invalid.

1.2      Reconciling Bristol-Myers

That is essentially what happened in Bristol-Myers 2017 FCA 76. The 736 patent at issue in Bristol-Myers claimed atazanavir bisulfate, which is used in the treatment of HIV/AIDS. As discussed here, the first class of anti-HIV drugs were known as nucleotide reverse transcriptase inhibitors and, while they were effective against HIV, they had significant side-effects: [14], [15]. A new class of anti-HIV drugs was developed, known as protease inhibitors [16]. These had a number of advantages over NRTIs, but the first-generation protease inhibitors had poor bioavailability, which resulted in poor compliance and a consequent potential for the development of drug resistance [22]. There was therefore a strong motivation to develop second-generation protease inhibitors [23]. One of these was atazanavir, which came on the market in 2003 and which remains one of the best second-generation protease inhibitors. The earlier 840 patent, which was part of the state of the art with respect to the 736 patent, claimed atazanavir and its salts, but disclosed only the free base form. While the free base had good inherent bioavailability, it was not sufficiently orally bioavailable in solid form to be a viable product. The problem facing the inventors of the 736 patent was to develop a form of atazanavir having properties suitable for an oral pharmaceutical dosage form, including improved oral bioavailability. They performed a salt screen and solved the problem by developing the bisulfate salt, which was claimed in the 736 patent.

The question was whether the 736 patent was invalid for obviousness. Mactavish J found that the inventive concept of the 736 patent included: (1) improved oral bioavailability of the bisulfate over the free base; (2) crystallinity; and (3) stability [12] [FC 446]. She held that improved bioavailability was obvious. The crystalline form and stability were advantageous and could not have been predicted in advance [FC 506]. If those properties were part of the inventive concept, then, so the argument went, the inventive concept was non-obvious and so the claimed invention was non-obvious, even though improved bioavailability was obvious. Even though those properties could not have been predicted, they were easily discovered using routine techniques, and “did not add anything inventive” [FC 507], [26]. Mactavish J consequently concluded that atazanavir bisulfate was obvious [FC 509].

But how can the claimed invention be obvious if the inventive concept is non-obvious? Bristol-Myers appealed, and the key question was [32]:

whether the Federal Court erred when it found that the development of Type-I atazanavir bisulfate was obvious in spite of the fact that only one of the three elements of the inventive concept, improved bioavailability over the free base of atazanavir, was predictable and the uncontradicted evidence was that the other two elements, crystallinity and stability, were not. [32]

The FCA agreed that it was obvious, but held that Mactavish J had erred in identifying the inventive concept. The true inventive concept was solely the improved bioavailability [75]. Since that was obvious, the invention was obvious. On this analysis, the problem of the finding that the inventive concept was obvious, even though two parts of it were non-obvious, disappeared.

Bristol-Meyers illustrates that not all new and non-obvious information disclosed in the patent is part of the inventive concept. The key question then, is how to identify the inventive concept. As noted by Rennie JA in Vyvanse [76], the FCA in Bristol-Meyers held that it was equivalent to the “solution taught by the patent,” [65]–[68], “which was often treated as synonymous with ‘what is claimed in the patent’ or ‘the invention’” [66], and this understanding of the inventive concept was not changed by Sanofi [68].

1.2.1  Objective or subjective inventive concept?

While identifying the solution taught by the patent was the central issue in Bristol-Myers, the decision does not provide any explicit definition. I’ll suggest that the key point is not identifying the solution, but identifying the problem. In particular, is the relevant problem that which motivated the actual inventors—the problem identified in the specification—or is it the problem that would have motivated the hypothetical skilled person? The emphasis on the role of the specification in Vyvanse, both in general statements of principle ([67], [70]–[74], [79]), and in determining the inventive concept of the claims at issue ([80]–[84]), suggests the former. As discussed below, Bristol-Myers suggests the latter.

In my view, it is the problem which would be faced by the POSITA—the objective problem—that must be considered, and not the subjective problem disclosed in the specification, which motivated the actual inventors. The basic reason is that it is very well established that the obviousness inquiry is objective, in the sense that the question is whether the claimed subject matter would have been obvious to the skilled person, not the actual inventor. So, the question whether the skilled person “would, in the light of the state of the art and of common general knowledge as at the claimed date of invention, have come directly and without difficulty to the solution taught by the patent”

Beloit (1986) 8 CPR(3d) 289 (FCA) 294; and see Sanofi [67]. It is not much of a leap to suggest that the problem must also be objectively defined, but the point is not always addressed explicitly in the leading cases, probably because in most cases there is no difference between the subjective and objective problem.

So, at the risk of belaboring the point, I’ll argue that the problem must also be defined objectively, as a subjective approach to the problem would depart from the principle that obviousness must be determined objectively.

More specifically, there are two related problems with a focus on the problem disclosed on the specification. First, while Rule 56(1)(d) requires the description to be set out in terms that permit the technical problem and its solution to be understood, the problem need not be set out expressly—there is no equivalent to claims in that respect. This means that identifying the problem taught by the patent would require a parsing of the specification to identify what the patentee considered the problem to be. That may be more or less clear, depending on the particular patent. This calls to mind the exercise in construing the promised utility under the old promise doctrine. So, in a selection patent, the prior art genus patent will typically disclose a class of compounds with utility in treating a disease, and the selection patent will disclose a particular compound from that class that is useful in treating the disease and that has greater efficacy or reduced side-effects. We know from Bristol-Myers that the three properties disclosed in the specification—usefulness in treating the disease, increased efficacy, and reduced side-effects—are not necessarily all part of the inventive concept. But which are? If the court determines that the usefulness in treating the disease is the only property that is part of the inventive concept, then the claim will be obvious, as the genus patent tells us that all members of the class have that property. That is effectively what happened in Alcon v Apotex / travoprost 2014 FC 699. Now, one may well argue that this identification of the inventive concept is wrong—that is what I did argue—but what is the test that allows us to say that in Bristol-Myers improved oral bioavailability was the inventive concept, while crystallinity and stability were not, while in Travoprost reduced increased efficacy and reduced side-effects are the inventive concept and usefulness in treating the disease is not? In the absence of some structured approach, asking the court to simply construe the disclosure is likely to result in the same kind of arbitrariness that plagued the promise doctrine.

The second problem is worse. What if the inventor pursued the wrong problem? Consider a twist on Bristol-Myers. Suppose that all the facts about the state of the art and the common general knowledge are the same. In particular, suppose it was known that atazanavir was a promising HIV drug that was limited in use because of poor bioavailability and that it would have been obvious to carry out a salt screen and thereby discover that atazanavir bisulfate had good bioavailability. However, suppose that the inventors had not been motivated by the search for improved bioavailability, but were instead motivated by an idiosyncratic pursuit of stability. Suppose that it would not have been obvious to try a salt screen in a search for stability, and the inventors had actually used a different approach and had only discovered that atazanavir bisulfate was stable after prolonged and arduous testing. In other words, suppose that the problem taught by the patent was to find a stable form of atazanavir, and atazanavir bisulfate was a truly non-obvious solution to that problem. Let’s suppose that the inventors were so focused on stability that they never even tested bioavailability, and the patent made no mention of it. In that case, if we take the inventive concept to be “the solution taught by the patent to the problem taught by the patent,” it would be non-obvious.

That can’t be right. It is well established that obviousness is objective. If the actual inventor is not very bright, and thinks the invention is a clever solution to a difficult problem, it is nonetheless obvious if the skilled person would have found it to be an easy solution to the same problem. The same must be true if the actual inventor thought the invention was a clever solution to an obscure problem, but the skilled person would have thought that invention was an easy solution to an obvious problem.

The two problems build on each other. If the rule is that the inventive concept is the solution taught by the patent to the problem taught by the patent, an applicant could turn an obvious invention into a patentable one by manipulating the nature of the problem that the inventor had set out to solve; on that approach, the patent in Bristol-Myers would have been valid if the applicant had drafted the disclosure to emphasize the importance of stability rather than bioavailability.

These problems do not arise if the focus is on the objective problem, which looks to the problem that the skilled person would have faced, rather than the problem that the inventors themselves had in mind. The nature of the objective problem is determined by the evidence, including evidence of expert witnesses as to the state of the art, and does not turn on the wording of the specification. It is quite common that the inventors did not have full knowledge of the relevant state of the art (and this will be even more common now that Hospira 2020 FCA 30 has held that obscure prior art is part of the state of the art) and the objective technical problem is often different from the subjective problem for that reason. But focusing on the objective problem also addresses the more extreme scenario of idiosyncratic motivation.

A focus on the objective problem is entirely consistent with Bristol-Myers, though the point is buried a bit. The relevant discussion in Bristol-Meyers is very brief:

[75] For the reasons set out above, I find that the “inventive concept” is not materially different from “the solution taught by the patent”. Had the Federal Court applied that definition to the facts, it would have found that the inventive concept in this case is atazanavir bisulfate, a salt of atazanavir which is pharmaceutically acceptable because it has equal or better bioavailability than the atazanavir free base. Atazanavir’s limited bioavailability was the source of the motivation to pursue the solution. The fact that claim 2 of the ‘736 patent claims a pharmaceutical dosage form of Type-I atazanavir bisulfate confirms its acceptability for pharmaceutical purposes.

The “reasons set out above” do not define the inventive concept or the solution taught by the patent, but simply make the point that Sanofi was not intended as a departure from prior law. The primary authority for the “solution taught by the patent” was the famous passage from Beloit (1986) 8 CPR(3d) 289 (FCA) 294, where the FCA described “The classical touchstone for obviousness” as being the technician skilled in the art “having no scintilla of inventiveness or imagination. . . [would] have come directly and without difficulty to the solution taught by the patent.” The passage was primarily concerned with the threshold for obviousness — hence the description of the characteristics of the skilled person and the references to “directly and without difficulty” — and not the identification of the inventive concept, which was not at issue. So, while Beloit is the main authority for the “solution taught by the patent” approach to the inventive concept, it does not actually help us understand what that phrase means.

Pelletier JA’s statement that the solution taught by the patent is “atazanavir bisulfate, a salt of atazanavir which is pharmaceutically acceptable because it has equal or better bioavailability than the atazanavir free base,” simply identifies the solution taught by the patent, without explaining how to identify that solution, or, more to the present point, how to identify the problem.

The most informative statement is that “Atazanavir’s limited bioavailability was the source of the motivation to pursue the solution” [75], though that statement does not tell us whether the relevant motivation was that of the actual inventors or the skilled person. The only prior relevant mention of the motivation was at [23], where the FCA referred to the finding of Mactavish J at trial that “As for the question of motivation to find the claimed solution, the Federal Court found that the limited bioavailability of atazanavir would have given the skilled person every reason to try to improve its solubility – and therefore its bioavailability – by making its salts: Reasons at paras. 483, 497.” Note the reference here to the “skilled person.” This reflects Mactavish J’s decision at the cited paragraphs, in which she referred to the motivation of the POSITA, and remarked eg that “the skilled person would have had every reason to try making salts of atazanavir in order to improve its solubility and bioavailability” [497]. Thus, it is clear that Mactavish J considered the relevant motivation to be that of the skilled person, not the actual inventors, which supports the view that it is the objective technical problem that must be considered. Further, throughout his analysis Pelletier JA adverted to the objective nature of the obviousness inquiry: see eg [65], [78].

In itself, this is strongly suggestive rather than definitive. While Pelletier JA made it clear that the question was whether the skilled person would arrive at the solution to the problem, he did not directly address whether the problem itself is the objective problem that would be faced by the skilled person or the problem that was set out in the specification. On the facts, it appears that the motivation of the inventors was the same as that of the skilled person.

As discussed above, Vyvanse clearly focused on the problem identified in the patent, which suggests a focus on the subjective problem. However, Vyvanse approved the Bristol-Myers reference to the “solution taught by the patent,” which is consistent with a focus on the objective problem. And as in Bristol-Meyers, it appears that on the facts the problem identified in the specification is the same as the problem that would have faced a POSITA, so that the nature of the problem was not directly at issue. Moreover, Rennie JA explicitly recognized the objective nature of the obviousness inquiry, albeit not in the context of identifying the relevant problem: [103].

Given the well-established principle that obviousness is objective, I would suggest that a clear holding with an equally clear rationale would be required to adopt a subjective approach to the nature of the problem at issue in the “solution taught by the patent.” Bristol-Myers clearly supports an objective approach, though it is not strong authority given that the point was not explicitly addressed. Vyvanse suggests a subjective approach, though weakly. Both are consistent with an objective approach.

I will also note that this explains why Bristol-Meyers treated “the inventive concept,” “the solution taught by the patent,” and “what is claimed” as being essentially equivalent. The solution to the problem is information about how to solve the problem, and what is claimed is the tangible solution which embodies that information.

1.3      The solution to the objective problem

To recap, to this point I have argued that the claim will be invalid unless the specification discloses new and non-obvious information regarding the claimed subject-matter. This is why it is normally appropriate to focus on the new information disclosed in the specification, which I have termed the objective inventive concept. If the objective inventive is obvious, then the patent will be invalid.

But that is not the end of the story. That requirement is necessary, but not sufficient. It is not any new and non-obvious information that will support a patent. The new information must be the answer to the problem that would have faced the skilled person. This implies that the inventive concept cannot be the ultimate end-point of the inquiry. The inventive concept, whether subjective or objective, is a property of the specification. That is, the inventive concept of the patent relates to the information disclosed in the patent itself: either the information that the applicant believed was new, or the information that was in fact new. But the claimed subject-matter may be obvious even if it embodies new and non-obvious information disclosed in the specification, as is illustrated by Bristol-Meyers and Actavis v ICOS [2019] UKSC 15.

Now, in both Bristol-Meyers and Actavis v ICOS the (obvious) information addressing the objective problem was disclosed in the specification, along with other (non-obvious) information. From this it might seem that we can still focus on the inventive concept of the specification. When there is a variety of information in the specification, it is just a matter of picking the right information. After all, the specification will always have obvious information. The specification will normally discuss the state of the art, all of which is by definition known or obvious to a skilled person. For example, take the canonical example of a claim to a new compound where the inventive contribution lies in the discovery of its new properties. In such cases, the compound is often easy to synthesize, but the way to synthesize it must nevertheless be disclosed as a matter of the disclosure requirement. So, even in the most straightforward case, it is necessary to distinguish new information provided in the specification that is properly the part of the inventive concept (eg the properties of the new compound) from that which is not (how to make it). That means that even in the easiest cases we need a test to distinguish new information that is relevant to the obviousness inquiry from that which is not. The view that the focus of the inquiry is the solution taught by the patent provides such a test, whether in the easy case (excluding obvious information about how to synthesize the compound) or in harder cases (identifying which of the various properties of the compound are relevant).

It is therefore tempting to conclude that “the solution to the problem” test does not detract from the principle that the end-point of the obviousness inquiry is the inventive concept—the test simply allows us to identify the inventive concept from among all the information in the specification.

The problem with this view is that the patent must provide a non-obvious solution to the objective problem. That is illustrated by the hypothetical twist on Bristol-Meyers where the inventors were idiosyncratically focused on stability, even though the objective problem was bioavailability. In that example, all of the relevant information provided in the specification was new and non-obvious. The claim was invalid, but not because the solution to the problem was obvious; it was invalid because the inventors focused on the wrong problem. That is, the claim was invalid because the claimed subject-matter embodied an obvious solution to the objective problem. That is not a conclusion that can be arrived at by a focus on the information disclosed in the specification, no matter what information we choose to focus on. This is the essence of the problem-solution approach to obviousness used by the EPO, which focuses on the “objective technical problem,”² which focuses on whether the claimed subject-matter would have been obvious: “Thus the question to consider, in relation to any claim defining the invention, is whether before the filing or priority date valid for that claim, having regard to the art known at the time, it would have been obvious to the person skilled in the art to arrive at something falling within the terms of the claim” EPO Guidelines § 4. The Supreme Court of the United States also endorsed an objective approach in KSR v Teleflex 550 US 398 (2007) 402 (my emphasis):

The first error of the Court of Appeals in this case was to foreclose this reasoning by holding that courts and patent examiners should look only to the problem the patentee was trying to solve. The Court of Appeals failed to recognize that the problem motivating the patentee may be only one of many addressed by the patent’s subject matter. The question is not whether the combination was obvious to the patentee but whether the combination was obvious to a person with ordinary skill in the art. Under the correct analysis, any need or problem known in the field of endeavor at the time of invention and addressed by the patent can provide a reason for combining the elements in the manner claimed.

Asking whether the claimed subject matter is an obvious solution to the objective problem is also consistent with the principle that it is the information provided by the specification is the quid pro quo for the patent: a solution to a problem is information about how to solve the problem. Recall that in Vyvanse, Apotex had argued that for a claim-based end point, and more specifically the focus of the inquiry was “the bare chemical compound, without its features or advantages, as those can only be found in the specification” [61]. That is, on Apotex’s argument, a focus on the claims would exclude all the information in the specification. Asking whether the claimed subject-matter is an obvious solution to the objective problem focuses on the claims, but does not exclude consideration of the properties of the claimed subject-matter, because it is only a solution to the problem of abuse resistance if it has the right properties.

Put another way, Apotex argued that the focus should be on the claims, while using the objective problem and solution approach, the focus is not on the claims, but on the subject-matter defined by a claim. In the case of a bare chemical compound claim, the claim itself does not include any properties of the compound, but the claimed subject-matter—ie the compound itself—has properties, and in asking whether the claimed subject-matter is a solution to the objective problem, we must take the properties of the compound into account.

The key holding on the facts in Vyvanse was that the properties of LDX must be taken into account, and the objective problem and solution approach is entirely consistent with that holding. As discussed above, this approach is supported by Bristol-Myers. The objective technical problem was to improve the bioavailability of atazanavir. Atazanavir bisulfate was an obvious solution to this problem. Since the claimed subject-matter was an obvious solution to the objective technical problem, it is obvious. Because it focuses on the claimed subject-matter, the objective problem and solution approach is also consistent with the text of s 28.3; as I pointed out in my previous post, this textual point is a problem for an approach that focuses on the inventive concept.

Thus, a focus on the inventive concept is a convenient short cut, when the actual inventors were motivated by the objective problem, as is very commonly the case, and the only question is whether the solution disclosed would have been obvious to a skilled person. But the objective nature of the obviousness inquiry implies that the ultimate focus of the inquiry should be on the subject-matter defined by the claim.

Finally, it might be objected that the Canadian courts have not previously explicitly endorsed a problem-and-solution approach to obviousness, though Vyvanse itself nearly does so at [76]. Further, both the Canadian and English courts have routinely referred to the problem addressed by the patent and the solution proposed in the patent in the context of the obviousness inquiry: see eg Vickers v Siddell (1890) 15 App Cas 496 (HL) 500; General Tire [1972] RPC 457 (CA) 475-476; Uhlemann Optical [1952] 1 SCR 143, 152; Northern Electric v Brown's Theatres (1940), 1 CPR 180 (Ex Ct) 185 affirmed [1941] SCR 224; and there are many more cases, both new and old, which could be cited. In Biogen [1996] UKHL 18 [53] Lord Hoffmann went so far as to say that “[a] proper statement of the inventive concept needs to include some express or implied reference to the problem which it required invention to overcome.” To adopt a problem-and-solution approach to obviousness would not be a departure from prior law; it would be in the common law tradition of explicitly recognizing a principle that was long implicit in the prior caselaw.

1) Blakey v Latham (1889) 6 RPC 184 (CA) 189 (Lopes LJ) prefacing the inventive step inquiry. Recall that Blakey was the first English case to use the word “obvious” to assess inventiveness. See also White v Toms (1867) 37 LJ Ch 204, one of the first cases refusing to grant a patent for a mere analogous use, in which Malins VC remarked at 207 that “the protection of those laws shall only be given to those who really invent something that is for the public benefit"; Britain v Hirsch (1888), 5 RPC 226 at 232, Cotton LJ, asking "whether there is sufficient invention to justify a monopoly being granted”; Graham v John Deere 383 US 1 (1966) 11, stating that the inventive step requirement address the problem “of weeding out those inventions which would not be disclosed or devised but for the inducement of a patent”; Alphapharm [2002] HCA 59 [143], quoting Graham; Société Technique de Pulverisation [1993] RPC 513 (CA) 519, stating the question is whether “the patent discloses something sufficiently inventive to deserve the grant of a monopoly.”

2) This is completely unrelated to the problem and solution approach to claim construction and patentable subject matter that is used by CIPO and nothing in this post should be taken as any kind of endorsement of the CIPO approach to patentable subject-matter.