The Supreme Court’s decision in Bilski v. Kappos was greeted by the patent bar as a partial victory for clarity and a partial defeat for predictability. The opinion rejected the Federal Circuit’s attempt to make the machine-or-transformation test the exclusive test for patentable subject matter under 35 U.S.C. § 101, and it confirmed that the hedging method at issue was an unpatentable abstract idea. What it did not do was give the lower courts and the Patent and Trademark Office a workable framework for the next case. This Article argues that the absence of such a framework has been particularly costly for diagnostic and personalized medicine patents, and that the patent system risks throwing out a category of inventions worth protecting along with the business method patents that Bilski was designed to discipline.
The Bilski Decision and Its Reach
Bernard Bilski and Rand Warsaw filed a patent application claiming a method of hedging risk in commodities trading. The Federal Circuit, sitting en banc, rejected the application and used the case as the occasion to formalize the machine-or-transformation test, under which a process claim is patent-eligible only if it is tied to a particular machine or transforms an article into a different state. The Supreme Court affirmed the rejection but rejected the test as the exclusive test for § 101 eligibility. Justice Kennedy’s opinion for the Court emphasized that the patent statute is to be read with a measure of flexibility, that abstract ideas are not patentable, and that the concept of hedging is itself an abstract idea.
The opinion is short on machinery and long on principle. It does not tell a future patent applicant how to know whether her process claim crosses the line from a permissible application of an abstract idea to an impermissible claim on the abstract idea itself. The Federal Circuit and the district courts were left to develop those rules case by case, and the cases that followed have not been kind to claims at the boundary.
The Diagnostic Patent Problem
A diagnostic patent typically claims a method of measuring a biological marker, correlating that marker with a clinical condition, and informing a clinician of the correlation. Before Bilski, claims of this form were routinely allowed. The Patent and Trademark Office issued thousands of such patents, the biotechnology industry built business models around them, and the academic literature treated them as a settled category of patentable subject matter. After Bilski, every element of the claim became contestable. A correlation between a marker and a condition is, on one reading, a natural phenomenon. The act of measuring is, on another reading, a routine and conventional step. The act of informing the clinician is, on yet another reading, a mental step that adds nothing of substance to the natural phenomenon.
The Supreme Court extended Bilski to exactly this kind of claim two years later in Mayo Collaborative Services v. Prometheus Laboratories, holding that a method of optimizing thiopurine drug dosage by measuring metabolite levels was unpatentable as a claim on a natural law. The same analysis was applied to gene patents in Association for Molecular Pathology v. Myriad Genetics. By the time the dust settled, a substantial fraction of the diagnostic patents that had been issued in the previous two decades were of doubtful validity.
The Babies and the Bathwater
The Bilski Court was concerned, rightly, that overbroad business method patents were imposing costs on the economy without producing offsetting innovation. The hedging claim at issue was a paradigm case: a description of a familiar financial practice, dressed up in the vocabulary of process claims, asserted against entities that had been hedging long before Bilski applied for a patent. The Court was right to reject that claim and right to send a signal that the patent system would not absorb every clever recharacterization of a known practice.
The difficulty is that the framework the Court adopted does not distinguish well between business method overreaching and diagnostic method legitimacy. Both kinds of claims involve a process. Both can be characterized, at a sufficient level of generality, as claims on an abstract idea or a natural law. The features that make a diagnostic patent valuable — the years of research that produced the correlation, the analytical work that translated the correlation into a clinically useful test, the investment in clinical validation that made the test reliable — are not visible at the level of generality at which Bilski’s analysis operates.
The result is that the same legal move that disposes of Bilski’s hedging claim also disposes of the Mayo correlation between thiopurine metabolite levels and dosing. One is the bathwater and the other is the baby, but the doctrine treats them identically. As we have argued in our analysis of pharmaceutical patent disputes, the patent system is at its weakest when it cannot distinguish between the kind of monopoly the system was designed to support and the kind it was designed to prevent.
The Industry Response
The biotechnology industry has responded in three ways. The first is to draft around the doctrine. New diagnostic patents add language about specific machines, particular reagents, and concrete physical steps in an effort to satisfy the residual machine-or-transformation analysis without claiming what the inventor really wants to claim. The drafting tax is real and the resulting claims are narrower and less useful than the underlying invention would justify.
The second response is to retreat to trade secrecy. A diagnostic test that cannot be protected by patent can sometimes be protected by keeping the algorithm confidential and offering the test only as a centralized service. The retreat to secrecy is itself a cost: it concentrates expertise in a few firms, limits academic verification, and produces fewer of the spillover benefits that the patent system was designed to encourage.
The third response is to lobby for legislative correction. Bills introduced in Congress in the years after Mayo and Myriad would have clarified that diagnostic and personalized medicine claims are categorically eligible for patenting, with the validity questions left to the more conventional doctrines of novelty, nonobviousness, and enablement. Those bills have not become law, and the courts continue to apply the § 101 framework to the very kinds of inventions the bills would have rescued.
A Path Forward
The problem Bilski identified is real. Business method patents that claim known practices in process-claim language are exactly the kind of patent the system should resist. But the doctrine the Court used to resist them is too coarse to spare the diagnostic and personalized medicine patents that the system has historically supported and that the industry continues to need. Three corrections would help.
First, courts applying § 101 to diagnostic claims should pay closer attention to the inventive contribution rather than asking only whether the claim recites a natural correlation. The test for § 101 has become a test for the wrong question; the right question is whether the claim adds something to the natural law it recites, and that question should be answered with a serious look at the prior art rather than a categorical move at the eligibility stage.
Second, the Patent and Trademark Office should issue guidance distinguishing between claims that recite a natural correlation as the inventive contribution and claims that recite a natural correlation as the basis for a clinically useful method. The distinction is workable in practice and tracks the distinction the Supreme Court was trying to make in Bilski between the abstract idea itself and applications of the abstract idea.
Third, Congress should consider the legislative correction that has been on the table since Mayo. The doctrinal repair the courts have failed to perform is exactly the kind of thing legislatures are designed to do. Whether one believes diagnostic patents are valuable or not, the current framework is a blunt instrument that produces results few observers would defend on first principles.
Conclusion
Bilski v. Kappos was a necessary correction to a Federal Circuit that had begun to read § 101 out of the Patent Act. It became a problem when the framework it introduced was applied to cases that the Court had not contemplated and to which it does not fit well. The pattern is familiar in patent law: a doctrine developed for one technology proves too broad when applied to another. The babies-and-bathwater language is a metaphor, but the underlying point is not. A patent system that cannot tell the difference between Bilski’s hedging method and Mayo’s drug-dosing correlation is a patent system that needs better tools.
Frequently Asked Questions
What did Bilski v. Kappos decide?
In Bilski v. Kappos, 561 U.S. 593 (2010), the Supreme Court rejected the Federal Circuit’s machine-or-transformation test as the exclusive test for patent eligibility under 35 U.S.C. § 101 but agreed that Bilski’s hedging method was an unpatentable abstract idea. The decision left the boundaries of patentable subject matter unsettled and applied broadly to any patent claiming a process, including diagnostic methods.
Why did Bilski affect life science patents?
Although Bilski involved a financial hedging method, its analysis of what constitutes an unpatentable abstract idea applies to any process claim. Diagnostic and personalized medicine patents typically claim methods of correlating biological data with clinical outcomes, and these correlations can be characterized as abstract ideas or as natural phenomena. The Supreme Court applied Bilski’s framework in Mayo v. Prometheus and AMP v. Myriad to invalidate diagnostic and gene patents.
What is the babies-and-bathwater concern?
The babies-and-bathwater concern is that Bilski’s broad rejection of overreaching business method patents also threatens diagnostic and personalized medicine patents that the patent system has historically protected and that the biotechnology industry depends on.
Related articles: Hatch-Waxman Pharmaceutical Patents · Provigil and Anticompetitive Behavior · IP Privateering · The Medicines Patent Pool · Software Patents and § 271(f)