"The Federal Circuit was wrong," wrote Justice Breyer in a majority opinion in which the United States Supreme Court's dissatisfaction with the Federal Circuit was evident. Justice Breyer's opinion in Teva Pharm. USA, Inc. v. Sandoz, Inc., delivered in late January, became the most recent in a long line of decisions in which the Supreme Court has overturned interpretations of patent law by the Federal Circuit. eBay Inc. v. MercExchange, LLC, 547 U.S. 388, 126 S. Ct. 1837 (2006) (injunctions); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 127 S. Ct. 1727 (2007) (obviousness); Bilski v. Kappos, 130 S. Ct. 3218 (2010); Octane Fitness, LLC v. ICON Health & Fitness, Inc., 134 S. Ct. 1749 (U.S. 2014) (attorneys fees). Criticized over the years for giving the decisions of district courts too little deference, the Federal Circuit again did just that in Teva, causing the Supreme Court to vacate the Federal Circuit's judgment with orders to apply the proper standard of review on remand.
Who is affected when the Federal Circuit gets it wrong? And what exactly did it do wrong in Teva?
The Federal Circuit hears all appeals where the original action arose under U.S. patent laws. Unlike other federal Courts of Appeal, the Federal Circuit's jurisdiction in patent cases is nationwide. It could be argued that the Federal Circuit's reversals by the Supreme Court stem from a perception that it does not confine itself to its role as a federal Court of Appeals, instead deciding issues without enough deference to the discretion or fact-finding of the trial court. The Federal Circuit's encroachment into the territory of the trial court is exactly what raised the ire of the Supreme Court in Teva, where the high court held that the Federal Circuit was giving district courts engaging in patent claim construction less deference than they were due by reviewing their factual findings de novo. 574 U.S. ___, 135 S. Ct. 831, No. 13-854, 2015 U.S. LEXIS 628, at *11, 29 (Jan. 20, 2015). Specifically, the Federal Circuit had, wrongly, ignored the trial court's factual determinations regarding how a person of ordinary skill in a certain industry would interpret a term in the patent at issue.
What is "claim construction," and how does it involve questions of fact? And why should the factual findings of a district court be given deference?
The coverage of a patent is dictated by the language it uses to describe the invention it "claims." An invention accused of infringing a patent must include every component described in the patent before it can be deemed an infringing product. If it is missing any component described in the patent, then, generally, the product does not infringe. Sometimes the difference between infringement and non-infringement depends on how to interpret a particular term in a patent. For example, consider a plaintiff who holds a patent for a vehicle used to lay gravel that is steered with a "steering wheel." The plaintiff wishes to sue a defendant who sells an identical gravel-layer, except that it is steered with a joystick. The defendant will claim that its gravel-layer does not infringe, because it has no steering wheel as required by the patent. The plaintiff and defendant likely will engage in a bitter dispute as to how to interpret "steering wheel." The plaintiff will argue that a "steering wheel" should be interpreted as any device used for directing a vehicle, while the defendant will argue that a "steering wheel" should be interpreted as a circular device used for directing a vehicle. Who settles the dispute?
Nearly 20 years before Teva, the Supreme Court ruled that patents, like contracts or deeds, are legal instruments to be interpreted by a judge rather than a jury. Markman v. Westview Instruments, Inc., 517 U.S. 370, 116 S. Ct. 1384 (1996). However, trial judges interpreting patent language – in a phase of litigation commonly called "claim construction" – must sometimes consult outside sources for technical terms and phrases not commonly understood. The trial judge might (for example) consult extrinsic evidence such as a dictionary, or for more complex issues rely on the testimony of experts. Questions that require outside evidence to resolve usually are deemed factual – not legal – questions, and the Federal Rules of Civil Procedure require appellate courts to accept a trial court's decisions on a factual question unless they are "clearly erroneous." Fed. R. Civ. P. 52(a)(6). The policy behind Rule 52(a)(6) is simple: "a district court judge who has presided over, and listened to, the entirety of a proceeding has a comparatively greater opportunity to gain . . . familiarity [with a subject] than an appeals court judge who must read a written transcript or perhaps just those portions to which the parties have referred." Teva, 2015 U.S. LEXIS 628, at *16. Unlike an appellate judge, the trial judge may, if necessary, question experts offered by the parties, examine an invention that is the subject of a patent, rely on a court-appointed expert, etc. Once the trial court resolves factual questions regarding a term in a patent (e.g., how a person of ordinary skill in the relevant industry at the time a patent was issued would interpret a term such as "steering wheel"), the district court must resolve differences in interpretation among the parties.
Was the Federal Circuit wrong to apply the de novo standard to the district court's claim construction determination?
Not entirely. While Markman gave trial judges power to interpret the claims of a patent, it made the ultimate construction of a patent claim a legal question. The traditional rule is that appellate courts are never bound by a district court's interpretation of purely legal questions. Thus, after Markman, the Federal Circuit began reviewing trial courts' ultimate constructions of a patent claims de novo – meaning that the Federal Circuit could ignore the trial judge's ultimate interpretation of a patent claim and start from scratch. Cybor v. FAS Technologies, 138 F.3d 1448, 1451 (Fed. Cir. 1998) (en banc). But, in Teva, the Federal Circuit followed its practice since Cybor of using the de novo standard for reviewing a district court's underlying factual determinations as well. According to the Supreme Court, it should not have. Consistent with Fed. R. Civ. P. 52(a)(6), where the district court needs to consult extrinsic evidence in order to understand, for example, "the background science or the meaning of a term in the relevant art during the relevant time period," the appellate court may not overturn the resulting factual finding unless the trial court made a "clear error." Teva, 2015 U.S. LEXIS 628, at *22-23. Only after the Federal Circuit has accepted the district court's underlying factual findings that are not clearly erroneous may the Federal Circuit embark on its de novo review of the district court's ultimate claim construction. That is precisely what the Federal Circuit did not do in Teva. It rejected the district court's factual findings regarding how persons of ordinary skill in the relevant industry would interpret the term "molecular weight" – without evaluating whether the district court's finding was clearly erroneous – and deemed the patent at issue invalid. Id. at *10. The Supreme Court vacated the Federal Circuit's judgment and remanded the case.
How will Teva affect patent litigation?
There are many predictions about how Teva will affect the way in which patent suits are litigated:
 In fact, the Federal Circuit rejected the notion that claim construction truly involves factual issues at all: "Nothing in the Supreme Court's opinion supports the view that the Court endorsed a silent, third option - that claim construction may involve subsidiary or underlying questions of fact." Cybor, 138 F.3d at 1455.