Patent Litigation Weekly: Stanford Researcher Loses at PTO, Still Fights On

In the April/May issue of IP Law & Business, we reported on the plight of Bob Shafer, a Stanford University scientist embroiled in a dispute with Luxembourg-based Advanced Biological Laboratories, S.A. (ABL). ABL claimed it owned two patents that were being infringed by the world-renowned HIV-drug resistance database Shafer created, HIVdb.

To recap quickly: After ABL sent Stanford claim charts and a letter asserting its patent rights, Stanford sued to invalidate the patents. That case settled, with Stanford agreeing to post a notice on the database stating that users of HIVdb might infringe ABL's patents.

ABL subsequently sued Stanford, saying the university had violated the settlement terms; the company also sued Shafer for defamation, based largely on comments he posted on his website, harmfulpatents.org. Shafer, tapping his own savings, hired attorneys at Day Casebeer to file reexams on the two ABL patents. (Day Casebeer has since been purchased by Howrey LLP.)

ABL's counterclaims against both Stanford and Shafer individually have now been settled. And it certainly appears that Chalom Sayada, the French doctor who is ABL’s CEO, has concerns about how the litigation makes him look to HIV researchers. The joint settlement statement reads, in part:

"According to ABL, it never intended to shut down or adversely impact the HIVdb. To further demonstrate its commitment to the HIV community at large, ABL clarifies that it will not assert its U.S. Patent Nos. 6,081,786 and 6,188,988 against any non-profit institution, or any hospital, researcher, or individual doctor that utilizes in the treatment of HIV."

Sayada, who called The Prior Art to express his unhappiness with the IPLB story, insists that his suggestion in 2007 that Stanford University and its affiliated hospital needed to license his patents didn't "adversely impact" HIVdb. But Stanford lawyers certainly saw ABL’s actions as an implied threat, and filed a declaratory judgment suit seeking to invalidate the patents.

The settlement agreement between ABL and Shafer also calls for Shafer to provide a link on harmfulpatents.org to ABL's response to his arguments. As of late August, the ABL response page noted that "patented methods and products form the backbone of the global healthcare industry," and said a direct response to Shafer’s statements "will be posted soon."

The full settlement document is public, and available on Shafer's website.

As for Shafer's reexaminations at the USPTO, they also concluded over the summer—and not in his favor.

Patent examiner Deandra Hughes decided that all 66 claims of the 6,188,988 patent are, indeed, patentable, despite more than 200 pages of evidence submitted by Shafer and his lawyers. Even though doctors had used databases to help choose therapies to treat various ailments for decades before the first relevant patent application at issue was filed in 1998, Hughes said the '988 patent should be allowed.  Her reasoning: the prior art references didn't distinguish a system with exactly three "knowledge bases." And that distinction alone—having three "knowledge bases"—is a patentable advance, Hughes decided. See Notice of Intent to Uphold the Claims of the '786 patent [PDF].  

Not surprisingly, Shafer isn't pleased with the decision. "I expected that examiner Hughes would choose the course requiring the least amount of work and uphold the patents," Shafer writes on his website. He submitted a second set of reexamination petitions in July, together with a sworn declaration from Ted Shortliffe, president of the American Medical Informatics Association. Shortliffe's work on Stanford's MYCIN system, a computer system that used expert knowledge rules and data structures to give therapeutic advice related to disesases as early as the 1970s, provides the basis for Shafer’s primary arguments for why ABL’s patents are invalid.

Shafer realized in January the reexam was unlikely to succeed. "I was extremely dismayed, and it took me a number of weeks to recover," he says. "But Jonathan Loeb [of Howrey] did an outstanding job on the reexams, and Ted Shortliffe deserves the gratitude of the whole biomedical informatics community for his declaration." Between filing two sets of reexams and defending the defamation suit, Shafer's legal costs have topped $200,000; he has raised some $25,000 through harmfulpatents.org.

In his declaration [PDF], Shortliffe explains that "subsequent work by myself and others built upon that early research (known as the MYCIN System) and invalidate the claims for novelty and invention that were evidently the basis for the patent applications in question."

He further notes:

I should add that, although we published the work extensively in the 1970s, 1980s, and 1990s, we viewed it as public-domain knowledge and methodology.  We never sought to patent the work or the underlying methods and viewed it as contributing to a public good for the benefit of patients and health care.  Our “reward” for our efforts was the acknowledgement from colleagues that the work was important and that the knowledge-based methods that we devised were worthy of adoption and expansion by others.

Shortliffe was "appalled" when he learned that ABL aimed to enforce these patents "by demanding licensing royalties for work that, in my opinion, had long been in the public domain." His declaration goes on to describe several examples of medical-expert systems that have separate and distinct "knowledge bases." In any case, Shortliffe argues, the idea that having three distinct "knowledge bases" should constitute a patentable advance is silly: 

This is a trivial distinction without a practical difference since multiple knowledge bases could be merged into a single entity and have long been separated into multiple representations largely for computational convenience and clarity.

See Shortliffe Declaration [PDF].

And so the lesson of this story appears to be: a patent granted for a description of a system that differs very slightly—if at all—from earlier technology is exceedingly difficult to overcome, even after considerable time, expense, and the donated expertise of a top practitioner to the effort.

Shafer notes that while the reexam has narrowed the patent—because now only expert systems with exactly three "knowledge bases" can be said to infringe—it probably won't make that much of a difference to any potential defendant. "Unfortunately, this knowledge does not prevent continued patent infringement lawsuits because a careful review of a patent and its prosecution record...does not take place in a typical patent infringement lawsuit until several hundred thousand dollars have been spent by the defendant."

More links:

• Shafer's website: Front page, a page on the ABL patents saga, and a list of documents related to the reexams and other litigation. 
• IPLB story on Shafer, "Dr. No"

 Photo: Jason Doiy / The Recorder

 

Comments

[Eric Delwart]

Dr Shafer is fighting for all researchers at the mercy of patent trolls. The mere threat of litigation is enough for many to fold but Dr Shafer showed that standing up to them publicly can bring important concessions. The quality of USPTO reviews is also frightening. Hopefully that particular decision will be reversed.

[Gio Wiederhold]

Thanks for the information,
I'll get a patent for a system requiring exactly 17 knowledge bases and start suing. I am sure the cost of the patent will be offset by settling, say. at $10,000 per defendant, since most companies can't afford to a have a lawsuit on the books that they are not challenging, and they will not be able to hire a lawyer that cheaply. I bet I'll get $100,000 easy in settlements. The patent that I do have only cost me about$23,000. That a better RoI than any other `health care' investment I think of.
And I'll donate half to Bob Shafer, unless he decides to file a patent for 13 knowledge bases.

Gio

[bigpicture]

Makes you wonder about the extent of systemic corruption versus the public good and impartiallity. What is the price to buy a politician, a judge or a Patent Examiner?

[neil]

So to work around this patent you need to extend your 3 knowledge bases into 4, either by splitting one or adding another of marginal relevance? In some circumstances merging them to reduce 3 to 2 might be better?