The question of whether a “fair expectation of success” is the correct legal standard applicable to an obvious to try analysis was finally laid to rest in the Federal Court of Appeal’s recent important decision of Eli Lilly v Mylan (2015 FCA 286). A unanimous panel held that a “fair expectation of success” was not the correct standard. Instead, the higher “more or less self-evident” test articulated by the Supreme Court of Canada in Sanofi-Synthelabo (2008 SCC 61) was reaffirmed as the correct legal standard. The higher standard for obviousness makes it more difficult to invalidate a patent, thereby strengthening protection for patent-holders.

The Eli Lilly case was a NOC case. Lilly sought to prevent Mylan from obtaining marketing authorization for generic CIALIS®. The asserted ‘948 patent was directed to a pharmaceutical formulation of reduced particle size of the active (tadalifil) with particular excipients for ED treatment.

At first instance, the applications judge found the allegation of obviousness justified. With respect to the obvious to try analysis, the court recognized two different legal standards: (i) the so-called “fair expectation of success” standard; and (ii) the “more or less self-evident” standard articulated by the Supreme Court in Sanofi-Synthelabo.

The applications judge held in 2015 FC 178 the correct legal standard was a “fair expectation of success” and applied it, finding the invention obvious to try:

[150]    I am also of the view that the person skilled in the art would have considered it obvious to try the excipients in the ‘948 Patent to achieve a stable, rapid onset tadalafil tablet. … Moreover, the test is not whether a skilled person would know for certain that a formulation would work or whether there is a guarantee that particular formulations would work … This would set the bar too high. The test, rather, is whether the skilled person had good reason to pursue predictable solutions or solutions that provide a “fair expectation of success”. …

The Federal Court came to the same conclusion, however, after ostensibly applying the higher “more or less self-evident” standard from Sanofi-Synthelabo:

[151]    In light of the prior art and applying this standard, … a person skilled in the art would have had a high expectation of success in using the excipients and the surfactant of the ‘948 Patent in obtaining a tadalafil formulation with the desirable bioavailability, and that it was indeed more or less self-evident that such a strategy ought to work. Therefore, even if the standard is not “fair expectation of success”, I find that the invention was “obvious to try” according to the test in Apotex Inc v Sanofi-Synthelabo Canada Inc, 2008 SCC 61, at para 69, [2008] 3 SCR 265.

This dichotomy reflects a tension that has been growing in the law of obviousness ever since the Supreme Court released its leading Sanofi-Synthelabo decision. Far from being a mere matter of linguistics, many obviousness decisions have since applied the relatively low “fair expectation of success” standard (e.g., 2015 FC 1261, 2015 FC 247, 2013 FC 283, 2013 FC 246, and 2010 FC 612). The underlying concept is that when a skilled person is motivated to try to obtain the invention using routine techniques and has a fair expectation of success, then the invention is “obvious to try”. In such a situation, the invention would inevitably have been discovered by a skilled person as a result of routine testing.

The application of this relatively low standard is, however, an error of law and no longer correct. On appeal, a unanimous panel of the Federal Court of Appeal rejected a “fair expectation of success” as the applicable standard. Instead, it held that the relatively higher “more or less self-evident” standard is the correct standard to be applied. Given that the applications judge found the invention obvious to try under both standards, however, the decision was not disturbed despite the recognized error of law:

[4]        First, Eli Lilly asserts that the Judge erred in law in his obviousness analysis by applying an incorrect test for obviousness when he wrote, at paragraph 150 of his reasons, that the “test, rather, is whether the skilled person had good reason to pursue predictable solutions or solutions that provide a ‘fair expectation of success’”. We agree that the correct test, and the test that ought to be applied by the Federal Court, is that articulated by the Supreme Court of Canada in Apotex Inc. v. Sanofi-Synthelabo Canada Inc., 2008 SCC 61, [2008] 3 S.C.R. 265, at paragraph 66: “For a finding that an invention was ‘obvious to try’, there must be evidence to convince a judge on a balance of probabilities that it was more or less self-evident to try to obtain the invention. Mere possibility that something might turn up is not enough.” This said, the Judge went on, at paragraph 151 of his reasons, to note that it was “more or less self-evident” that the invention ought to work. The Judge then concluded “even if the standard is not ‘fair expectation of success’, I find that the invention was ‘obvious to try’ according to the test in” Sanofi-Synthelabo. It follows that any error by the Judge in his initial articulation of the test for assessing obviousness was not material to the Judge’s decision.

This important appellate decision lays to rest the question of what legal standard is to be applied in an obvious to try analysis. It reaffirms the Supreme Court’s 2008 ruling that an invention is not obvious to try unless the invention is more or less self-evident, or very plain, in advance of routine testing. An invention is not obvious to try even when a motivated skilled person using routine tests would have had a fair expectation of success and would have inevitably discovered the invention. For example, as the Federal Court of Appeal held in 2013 FCA 186, where the inventive concept includes the properties of a compound, the lack of advance knowledge of the properties renders it not obvious to try to obtain the invention, even where the skilled person is motivated to do so and would have inevitably done so using routine tests.