SUPREME COURT ADOPTS DATA IN-DATA OUT PATENT TEST

In these times of super-partisanship, all major decisions of the U.S. Supreme Court are along the lines of five-to-four, right? Conservatives on one side, liberals on the other, a swing vote or two determining the outcome?

Well, it is certainly easy to have this impression. But even in this period of extreme dogmatism, all nine justices occasionally find themselves in agreement on a case, and not merely a case of little import. One such instance occurred this past spring. In a unanimous decision, the Court just might have taken a positive step toward easing a significant economic conflict. That conflict is between an Industrial Age patent system on one hand and the realities of the Information Age on the other.

On March 20, 2012, the Court decided the case of Mayo Collaborative Services v. Prometheus Laboratories, Inc. Justice Breyer delivered the Opinion of the Court. There were no dissenting opinions. There were not even any concurring opinions, those ramblings in which a Justice or Justices say in effect, “We sorta agree with you, but. . . .” In short, everybody was on board.

In the decision, the Court held that a process of applying researchers’ findings about the proper dosage of specific drugs was not entitled to a patent. The drugs were thiopurine drugs to treat autoimmune diseases. The process involved three steps: (1) an “administering” step in which a doctor administered a drug to a patient, (2) a “determining” step in which the doctor measured the resulting metabolite levels in the patient’s blood, and (3) a “wherein” step in which the doctor was to increase or decrease the drug dosage if the metabolite level was outside a specified range.

This three-step process, in the Court’s view, was nothing more than stating an unpatentable law of nature—the researchers’ findings—and adding the words “apply it.” Something more was needed before the application of a law of nature qualified as a patentable invention.

An observer’s first reaction might be, “So what’s the big deal?” The big deal is that the process patent in the Mayo case is an example of what a sizeable portion of the patent system has become in the Information Age: words, sometimes precise but more often vague, describing how to do something that involves a substantial degree of intangibility. Business methods patents are one variant of the group. Another common term is software patent, which at an unpatentable level is just a computer program. But add a little hardware, and U.S. Patent Office Examiners, striving to meet performance goals, are often persuaded.

Included in Information Age patents are financial arrangements and manipulations that result in exotic financial products, business and managerial analysis procedures that purport to optimize operations, and scoring and ranking exercises that lend an aura of certainty to murky decision-making.

U.S. patent law authorizes patents for processes, machines, manufactures, and compositions of matter. In an outburst of exuberance, a Congressional Committee report in 1980 stated that “anything under the sun” was patentable. Realizing this was not quite the extent of the patent concept, the Supreme Court over the years has narrowed patentability to exclude laws of nature, natural phenomena, and abstract ideas. But the boundary between what is and what is not patentable remains ill-defined. Indeed, as the predominantly tangible Industrial Age has given way to the intangibles of the Information Age, the boundary has become more and more opaque.

In 2010, the Supreme Court, in Bilski v. Kappos, 561 U.S. __, 2010, grappled with the boundary’s location. The Court affirmed a lower court decision that a process involving a form of financial derivative, specifically a hedging transaction in the energy markets, was not patentable. All nine Justices agreed on the ultimate result, but they took several paths to that result.

Four Justices—Kennedy, who wrote the opinion of the Court, Roberts, Thomas, and Alito—said as little as possible. They defined the proposed patent as an abstract concept and rejected it for abstractness. But they did not hold that business methods in general were unpatentable. Business methods presented special Information Age challenges—the opinion specifically recognized the existence of the Information Age—but the four Justices found no general prohibition on the granting of patents for business methods.

Four other Justices—Stevens, who wrote a concurring opinion, Breyer, who joined in that opinion and also wrote another concurring opinion, Ginsburg, and Sotomayor—took a more expansive view. They would have prohibited business methods patents outright. Justice Scalia joined a portion of Kennedy’s opinion and a portion of Breyer’s opinion. The portions of Kennedy’s opinion and Breyer’s opinion that Scalia joined were portions that did not raise questions or concerns about patents for business methods. Interestingly, the portions of Kennedy’s opinion Scalia did not join were also the only portions of the opinion that referred to the Information Age. Perhaps a Constitutional Originalist and the Information Age don’t mix well.

Thus, although everyone agreed on the ultimate outcome, the varying rationales made for a rather messy situation. The one point of agreement in the rationales seemed to be that a machine-or-transformation test provided a useful and important clue for determining whether some claimed processes were patentable, but it was not the sole deciding factor. The machine-or-transformation test for an invention that involves a process is if (1) the process is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing. But to emphasize, the machine-or-transformation test only provided a clue regarding patentability.

So Bilski left much uncertainty in its wake.

One approach to bringing a little more clarity to the question of what is and what is not patentable might be the application of a data in-data out test. Data, incidentally, is a fundamental pillar of the Information Age. In a data in-data out test, if the subject matter of a patent application consists primarily of introducing data into a system—system being broadly defined—applying or otherwise manipulating the data, and then extracting a largely intangible result of predominantly data, then the threshold of patentability has not been achieved.

And without saying expressly that was what it did, that indeed was what the Supreme Court appears to have done in Mayo. The Court’s explicit analysis was centered on whether the process in Mayo was a patently permissible application of a law of nature, or merely a statement of the law of nature with the directive to “apply it.” But the Court’s analysis also lends itself to a data in-data out interpretation. For example, on pages 10-11 of the slip opinion, the Court says: “The upshot is that the three steps simply tell doctors to gather data from which they may draw an inference in light of the correlations.” The process’ data in? “gather evidence.” The process’ data out? “an inference.”

The Court discussed two precedents in depth: Diamond v. Diehr, 450 U.S. 175 (1981), and Parker v. Flook, 437 U.S. 584 (1978). Each precedent involved a process of three steps. And each three-step process involved a basic mathematical equation that, like the law of nature in Mayo, was itself not patentable. In Diehr, the process was determined to be patentable. In Flook, the process was determined to be not patentable. The Court’s analysis in each case can be viewed in data in-data out terms, with the difference in the results being attributed to a greater tangibility component in Diehr.

Diehr’s process was a method for molding raw, uncured rubber into various cured, molded products. The three-step process used a known mathematical equation, called the Arrhenius equation, to determine when to open the mold’s press. The three steps were: (1) continuously monitoring the temperature on the inside of the mold; (2) feeding the resulting numbers into a computer, which would use the Arrhenius equation to continuously recalculate the mold-opening time; and (3) configuring the computer so that at the appropriate moment it would signal a device to open the press. The Mayo Court appeared to attach great significance to the output in Diehr being not just data but the automatic opening of the press at the proper time.

Flook’s process was a method for adjusting “alarm limits” in the catalytic conversion of hydrocarbons. Operating conditions such as temperature, pressure, and flow rates were continuously monitored in the conversion process. When the operating conditions exceeded certain alarm limits, inefficiency or dangers were indicated. The three-step process amounted to an improved system for updating the alarm limits. The three steps were: (1) measuring the current level of the variable, temperature for example; (2) using an apparently novel mathematical algorithm to calculate the current alarm limits; and (3) adjusting the system to reflect the new alarm-limit values. Unlike in Diehr, the output of the process in Flook was somewhat removed from anything tangible. The output was simply adjusted alarm limits. The connection to hardware, to a means for setting off an alarm, was unclear.

So one can argue that another way to describe the state of the law regarding patentability is that a pure data in-data out process does not appear to be patent-eligible. But if a bit of hardware is added for the data out to interact with, then perhaps a patent is warranted. The caveat to this perhaps idealistic view concerns the field of business methods patents. Here, the data out may be an intangible such as a financial instrument, a customer rating, a fraud score, or an optimization selection. One can argue from Flook, Diehr, Bilski, and Mayo that a predominately intangible output in a process should not result in a patent. Unfortunately, each week the Patent Office awards a number of patents for data in-data out processes that have little tangible output and involve not much more than general purpose computers running particular software.

A saying one sometimes hears around the Patent Office is that a patent is an invention and words describing the invention. But in many patent applications in the Information Age, the words predominate. The invention is an intangible, an idea, a series of steps, an abstraction. Maybe some legal protections are desirable for these types of inventions. Perhaps a copyright. Perhaps an entirely new concept. But ramming the intangible innovations of the Information Age into the largely tangible-oriented patent system of the Industrial Age is not a particularly pretty sight.