Opinion | Eli Lilly’s odyssey to use a fake rule and fake news to protect bad patents

This column originally appeared on STAT on August 16, 2017.

Someone high up in Eli Lilly and Company must be a “South Park” fan. When the company failed to ensure that the patents it filed on two blockbuster drugs, Zyprexa and Strattera, complied with Canadian law, it decided to “Blame Canada” rather than take responsibility for its error.

The company sued the Canadian government under the North American Free Trade Agreement for $500 million based on a classic trope: invent a fictional character to take the blame for one’s own failings. It lost the case, but confused everyone so much that the Supreme Court of Canada recently had to kill off the character.

In Oscar Wilde’s play “The Importance of Being Earnest,” one of the main characters created a fictional friend, called Bunbury, to obfuscate his own departures from proper conduct.

Eli Lilly’s Bunbury was more obscure than its literary predecessor. In essence, the company miscombined and mischaracterized statements taken out of context to invent a non-existent legal rule it called the promise doctrine and then blamed its application by Canadian judges for the loss of two valuable patents. It then spread fake news that Canada possessed this fictional rule, and that it specifically hurt pharmaceutical patent holders.

A team of McGill University students and I went through a vast collection of publicly available documents — court and tribunal filings, testimony, expert reports, decisions, news reports, and letters from politicians — to track how Eli Lilly invented the promise doctrine and spread “fake news” about it. We also collected and analyzed the results of all Canadian patent cases since 2000 to investigate the company’s assertions about Canadian patent law. We found that Eli Lilly — whether by mistake, misunderstanding, or otherwise — created a legal rule that had never before been invoked by a court. Worse, its claim that this legal rule had harmed pharmaceutical patent holders was unsupported, and its statistical methods misleading at best.

The story begins in 1991, when Eli Lilly applied for a Canadian patent for its antipsychotic drug, Zyprexa. Five years later, it applied for a Canadian patent for its attention deficit hyperactivity disorder drug, Strattera. In both cases, the company copied its U.S. patent applications without making changes to address the specificities of Canadian patent law, presumably as a cost-saving measure.

Patents are government-granted rights to exclude others from making, using, or selling an invention for a period of 20 years. Governments grant these rights to provide inventors with a way to make money to offset the costs of developing and selling a product. At the same time, each patent represents a hurdle in an obstacle course that must be covered by those wishing to develop and sell their own products. Patent law attempts to balance these contradictory effects. Each country achieves this balance in its own way, though there are many similarities.

A central principle of patent law is that you cannot patent the same thing twice; otherwise people could continue to patent the wheel. Because the drugs making up both Zyprexa and Strattera had already been patented, Eli Lilly could not simply patent them again. To get around this problem, it had to show it had discovered a previously unknown use for or advantage of each drug.

Zyprexa belonged to a group of compounds that had already been patented as being useful in the treatment of mild anxiety and psychotic conditions such as schizophrenia. Given that, Eli Lilly claimed that Zyprexa distinguished itself by showing marked superiority and a better side effect profile than other antipsychotic agents, and had a highly advantageous activity level. Strattera had also been previously patented, as an antidepressant, so the company claimed it had discovered that the drug could also treat ADHD.

In 2007 and 2008, the Canadian generic firm Novopharm (subsequently acquired by Teva) and Eli Lilly started litigation about the two patents. Novopharm argued that Eli Lilly simply did not have a basis for making its claims of new uses for or advantages of the drugs. Without them, there were no inventions and thus could be no patents.

Experts from both Eli Lilly and Novopharm explained the drugs and the patents to Canadian judges. In both cases, the judges determined that Lilly did not prove its assertion that the two drugs had newly discovered advantages, and that experts in the field would not conclude that Eli Lilly had a logical basis for those assertions. After several trial and appeal decisions, the Strattera patent was finally thrown out in 2011; the Zyprexa patent was tossed in 2013.

Rather than accept that it had bet and lost, Eli Lilly began its Bunburying. It created the fictional legal rule that, after several failed formulations, came to be known as the promise utility doctrine or just the promise doctrine. It was loosely based on the requirement in both U.S. and Canadian law that an invention be “useful” in order to gain a patent.

For more than 60 years, Canadian courts had been applying the same set of rules to determine if an invention is useful. Eli Lilly claimed that the courts had thrown out these rules and replaced them with something so new and radical that it was not found in any other patent system. The firm claimed that Canadian judges went out of their way to find statements about advantages that were impossible to prove and then struck down the patents because the drug firms had not proven the impossible.

Eli Lilly said that it should be good enough that a drug did something, even if it did not do what was stated in the patent application. Never mind that no other country would grant a patent on such a trivial function.

While artfully articulated, the promise doctrine never existed. Our comprehensive analyses showed that no Canadian court had ever acknowledged these rules, had never used the term “promise doctrine,” and had never agreed with Eli Lilly’s description of the legal rules. Courts did try to figure out what the inventions related to these two drugs were supposed to do, though our analysis determined that the courts showed no particular eagerness to find unreasonable advantages.

What Canadian courts have always required is simply that experts evaluating an invention would think that it did, or could be predicted to do, what the inventor said it would do. The same rule exists in Australia and Israel.

U.S. law accomplishes much the same thing as the “usefulness” rule, though under a different name: the enablement requirement. As Eli Lilly’s paid expert on U.S. patent law, Robert Merges, stated: “a patent that claims a range of uses must enable all or virtually all of the claimed uses in order to satisfy the enablement standard.” In other words, a patent is only valid in the U.S. if the patent provides a sufficient basis to believe that the invention does what the inventor says it does.

Beyond making up the promise doctrine, Eli Lilly had no solid evidence that it actually changed the outcome of pharmaceutical patents. At the time Eli Lilly submitted its official complaint in November 2012, only three other patents beside those for Zyprexa and Strattera could arguably be considered to have failed solely because of the invented doctrine. Interestingly, two of the other patents belonged to Eli Lilly, suggesting that the problem lay with the firm, rather than with the Canadian patent system.

To dance around those issues, Eli Lilly created “fake news” by artificially inflating the number of drugs affected by the nonexistent promise doctrine. It did this by including with the five real cases patents that were so bad they would have been thrown out no matter what; by adding cases that had nothing to do with the doctrine; by including cases from before the company claimed the promise doctrine began; and by counting the same invalid patent multiple times. Eli Lilly also provided its paid statistics expert, Bruce Levin, with prefabricated data rather than letting him conduct his own dataset. He stated that if some of the assumptions that went into the collection of data were changed, his conclusions would not hold up.

On March 16, 2017, the NAFTA tribunal unanimously ruled that Eli Lilly had failed to make its case. The tribunal determined that there had been no radical change in Canadian law and no evidence of discrimination against pharmaceutical patents. The tribunal specifically criticized Eli Lilly’s statistics for “not attempt[ing] to isolate ‘promise cases’” from other cases and for moving back the date from when it started collecting cases to before even it claimed the promise doctrine began.

Our own review of the data confirmed the NAFTA’s tribunal’s findings. Even supposing that the promise doctrine existed, we could not find any statistical effect of it. Further, there was no basis to think that the “rule” affected the pharmaceutical sector more than any other.

While Eli Lilly did not deceive the NAFTA tribunal, it did spread fake news about the promise doctrine and its alleged effects directly and through experts and institutes affiliated with the company. The company’s campaign was quite effective.

Eli Lilly asked members of the U.S. Congress — Daniel Coats (now director of national intelligence), Erik Paulson, Robert Menendez, Aaron Schock, David Nunes, Ron Kind, and Roy Blunt — to draft letters in support of the firm’s claims, repeating its false statistics. (Interestingly, Eli Lilly financially contributed directly to each of their campaigns, most in the same year in which they signed the letter.)

Think tanks that Eli Lilly funded, including the Fraser Institute and the Macdonald-Laurier Institute in Canada, and lobbying groups to which it belonged repeated the false statistics in reports and in comments to the media. Lawyers and industry insiders tied to the firm or its associates also published comments about the promise doctrine in the media and in trade journals in both Canada and the United States. Law firms representing pharmaceutical companies commonly repeated the false claims on blogs.

In 2014, turning fiction into reality, a Canadian court accepted the existence of the promise doctrine. While it did not accept the premise that Eli Lilly put forward about the rule, it acquiesced to the general — but false — belief that the doctrine existed.

It took the Supreme Court of Canada to kill off Eli Lilly’s Bunbury. On June 30, 2017, it threw out the promise doctrine although without being clear on what happens next. One thing seems certain: with the fictional character gone, Canada will now either return to what it had always been doing about patents or find another way to keep inventors to what they promise.

Drug patents can be worth billions of dollars. This case shows how far a pharmaceutical company might go to protect them. And long before the term “fake news” entered our lexicon, firms such as Eli Lilly generated and traded in it to distract shareholders and politicians from patent cliffs and leaky drug pipelines. The company went to great lengths to create its Bunbury rather than accept responsibility for its own faults. It almost succeeded, but pushed the argument too far.

Richard Gold is professor of law at McGill University and a senior fellow at the Center for International Governance Innovation.