Experimental Use Exception to Anticipation

Bayer Inc v Apotex Inc / drospirenone (NOC) 2014 FC 436 Hughes J
            2,382,426 / drospirenone & ethinylestradiol / YAZ

The anticipation argument in Apotex / YAZ turned on clinical trials conducted by Bayer in the US and Europe more than one year before the filing date, in which samples which embodied the invention were delivered to trial participants and their doctors. The doctors and participants did not know the precise ingredients (they knew of the active ingredients and dosage, but apparently not the micronized form) [100]. While Bayer took reasonable precautions to require participants to return unused samples [112], it appears that neither the doctors nor participants had signed confidentiality agreements [99]. While there was no evidence that anyone had ever actually analyzed the tablets [111], there was evidence that it was inevitable that some tablets would not be returned, given the large scale nature of the trials [113]. On these facts, Hughes J held

there has been established a ‘theoretical’ possibility that a tablet could have been kept and analyzed, [and] therefore the requirements of subsection 28.2(1) (a) of the Patent Act have been met” [118].

While Hughes J drew on European case law, and did not cite the recent FCA decision in Wenzel Downhole Tools 2012 FCA 333, it seems to me that this holding is consistent with the test set out by the majority in Wenzel [74, original emphasis] that it is sufficient that “there was an opportunity to access the relevant information.”

However, this did not end the matter. Hughes J held that “The law in Canada has long been established that experimental use in order to bring the invention to perfection, does not constitute public use” [119]:

[121] In the present case clinical studies were necessary to prove that the drug was safe and effective and, thereby, gain government approval for sale. Until this had been demonstrated, no commercial sale of the drug could have been made. Bayer took reasonable steps to ensure the confidentiality of the relevant documents and to ensure that unused tablets were returned. The theoretical possibility that some tablets were retained and analyzed is just that, theoretical. This theoretical possibility does not preclude the fact that the studies were experimental, and of necessity, conducted by the provision of tablets to members of the public. Thus these clinical studies are exempted from public use.

This seemingly establishes a broad experimental use exception to what would otherwise be anticipating disclosure, which applies to any clinical trial, so long as reasonable steps are taken to ensure that the unused tablets are returned.

As authority, Hughes J relied on Conway v Ottawa Electric Railway Co., (1904), 8 ExCR 432, 442; Gibney v Ford Motor Co. of Canada, [1967] 2 ExCR 279 [49], citing Elias v Grovesend Tinplate Co. (1890), 7 RPC 455, 466; and Hi-Qual Mfg Ltd v Rea’s Welding & Steel Supplies Ltd (1994), 55 CPR(3rd) 224, aff’d 61 CPR(4th) 270 (FCA). Of these cases (and I will not go into all of the other case law cited in those cases), Conway does stand for a broad experimental use exception to anticipation, but the others are weaker authority. Gibney and Elias both insisted that the experiments must be kept secret if anticipation was to be avoided, and in both cases the patents were held to be invalid as anticipated. In Hi-Qual the experimental nature of the use was an alternative ground for holding that patent not anticipated, the first ground being that the invention was not publically available. (The invention, relating to farm equipment, was outdoors on the inventor’s property, but the property was not open to the public, and the invention could not be seen from public property.)

It seems to me that one decision more than a century old is not strong enough authority to say this point is well-established. With that said, while old, Conway raises an interesting point which is directly applicable to clinical trials. The invention in question was a snow-plow for clearing street car tracks. An important feature of the plow as compared with the prior art was that it was adapted so as to accommodate the irregularity of real street surfaces. Consequently, it was only possible to develop the invention by experimenting in public, and this fact was relied on by the court in holding the experiments did not destroy novelty. Evidently, with a mechanical snow-plow, even more than with pharmaceuticals, if the device is operated in public it will be disclosed to any knowledgeable observer, and it is impossible to exclude the pubic from the public streets. This suggests that if it is impossible to develop the invention in secret, an experimental use exception may be required. With that said, it is not clear that Bayer could not have filed a patent prior to conducting its clinical trials, as clinical trials are not necessary to establish patentable utility. But in any event, the problem of inventions which must be developed in public is an interesting one in principle.

A different consideration is that it does appear that if the inventors had signed all the doctors and participants to a confidentiality agreement, then the invention would not have been available to the public in the first place. But is this mere legal formalism? It is not clear that signing a confidentiality agreement would have been any more effective than the (unspecified) reasonable steps Bayer actually took to ensure the unused product was returned. Perhaps the law should recognize that taking reasonable steps to ensure product is returned is tantamount to a confidentiality agreement, at least when dealing with unsophisticated patients in a clinical trial. As the Fed Cir put it in Dey v Sunovion

The fact that a tiny fraction of the thousands of vials were lost without penalizing the responsible test subject(s), or that the practicalities of the study required self-administration at home rather than physician administration in a closed facility, does not preclude a reasonable jury from concluding that the use of Batch 3501A was sufficiently controlled and restricted, rather than unfettered and public.

As noted, strictly, this goes to the question of whether there is disclosure to the public in the first place, rather than whether there is an experimental use exception, but this logic does reach the same result as Hughes J on the same factual basis. A response to this argument, as to the point above, may be that the pharmaceutical company should get its patent before starting clinical trials. (Note that in Dey prior trials by the accused infringer were at issue.)

In summary, while there is some case law supporting Hughes J’s position, I would not say that the experimental use exception to anticipation is well-established, and the policy issues are difficult. This aspect of his decision raises important legal and policy questions for the FCA, which might be dealt with either in terms of an experimental use exception, or by a refinement of the Wenzel approach to disclosure. Patently-O posts here and here discussing the same issue indicate that there has been a shift in US case law which has narrowed the experimental use exception, while at the same time adopting a more restrictive test for what constitutes public use.

Addendum: When I wrote this post, I was focused on the interesting legal issue of whether there is an experimental use exception to anticipation, and I glossed over the prior issue of whether there was disclosure to the public in the first place. On further reflection, I do not find the analysis of that point to be entirely satisfactory (though of course this may be a consequence of the way the case was argued). A disclosure to a party under an obligation of confidence is not a disclosure to the public, and while it appears that neither the doctors nor participants had signed confidentiality agreements, it is clear law that an explicit confidentiality agreement is not required to establish an obligation of confidence. The question is whether the product was conveyed in circumstances giving rise to an obligation of confidence: see generally Corlac v Weatherford 2011 FCA 228 [36]- [65], which is now the leading FCA case on the point, and which was not cited by Hughes J. The question of whether the circumstances were such as to give rise to an obligation of confidence is ultimately one of fact. From this decision, all we really know of the relevant facts is that Bayer “took reasonable precautions so as to keep relevant information confidential and to require participants to return any unused tablets” [112]. That at least raises a serious issue as to whether the circumstances were such as to give rise to an obligation of confidence. The simple fact that there was no explicit written or oral confidentiality agreement is “ significant, but not dispositive” (Corlac [53]). On the whole, I am inclined to think that this case should really turn on whether the circumstances were such as to give rise to an obligation of confidence. If such an obligation did arise, then there is no need to invoke an experimental use exception to anticipation; if it did not, then it is difficult to see why an exception should be invoked to protect a party which had disclosed an invention to the public without adequate regard to its confidentiality. Furthermore, whether or not such an exception is desirable is a matter for the legislature. Canadian law does have a grace period for disclosure by the inventor, but many jurisdictions do not. Whether such a grace period is desirable is a debatable matter of policy, and one of the most contentious issues is the uncertainty which it can create (see Part I of the recent Tegernsee Group report). If a further exception is to be made for experimental disclosure, this should be done by the legislature, rather than through an open-ended and poorly defined judicial exception.