Supreme Court Alters Standard of Review for Patent Appeals
Posted Thursday, January 22, 2015.
Before I went to law school, I had only a simplified understanding of how an appeal from an adverse trial decision worked. I knew that, if you lost the trial, you could appeal the verdict. And, if you lost your appeal, you could appeal again all the way to the Supreme Court. There is, of course, much more to it than that. I did not know, for example, that only certain aspects of the trial can be reviewed during an appeal.
During trial deliberations, a jury will consider evidence, including testimony, to decide upon a version of the facts that were presented during the trial. Those findings of fact, when combined with the law, lead to a verdict. Rules called “standards of review” dictate which trial content – facts or law – may be reviewed during an appeal. The general idea is that the appellate court ordinarily reviews only those aspects which can be determined directly from the written record of the trial that the appellate court received. Questions of law, often embodied in written texts like statutes or contracts, are normally reviewable on appeal. Questions of fact arising from physical evidence or oral testimony, however, are usually not.
To understand why questions of fact aren’t normally reviewable, consider that a question of fact is resolved by a jury in consideration of the evidence presented. The evidence might include, for example, in-person witness testimony. Because the judges of the appellate court were not present at trial to personally view the in-person witness testimony or any trial exhibits (physical evidence, e.g.), they are not in any position to override the findings of the jury members who experienced the evidence first-hand. A question of fact may be reversed on appeal if and only if there was clear error, as in no reasonable jury could have possibly reached the factual conclusion that it did. Accordingly, on appeal the appellate panel will give a high degree of deference to the factual findings at trial and not tinker with them.
On the other hand, legal issues arising from the trial may be reviewed at will by the appellate court. The appellate court has the discretion to determine that at trial, a statute (and therefore its application to the facts at hand) was misinterpreted. Written instruments such as contracts may also implicate legal findings by a judge, and such contract interpretation may be reviewed on appeal. This standard of review is called “de novo.” Questions of law are considered anew by an appellate court. On appeal, if the appellate judges find that the law was misapplied at trial, it may reverse or vacate all or part of the trial rulings.
In a patent case, where a plaintiff accuses a defendant of infringing one or more claims of a patent owned by the plaintiff, the outcome is likely to turn on the exact meaning of the patent claims. Recall that the patent claims are the legal description of the intellectual property protected by the patent. Claims are akin to the legal description of a piece of real property, and describe the boundaries inside of which the property is owned. The words used in a legal description of a piece of property don’t vary much, and for those who need to understand such legal descriptions (realtors, county recording offices, mortgage companies, assessors, etc.) the meaning of the words is clearly understood. The words used in a patent claim vary widely from one patent to the next, however, and the same word or phrase may have different meanings in different patents. That’s why the outcome of a patent case can hinge almost entirely on “claim construction,” which is the determination of the precise meaning of the words in a patent claim.
Claim construction occurs during a phase of a patent trial called a “Markman hearing,” which is a proceeding unique to patent trials. At a Markman hearing, each side submits a proposed definition for patent claim terms (their “claim constructions”). Expert testimony may be heard as to the meaning of claim terms during a Markman hearing. Afterwards, the trial judge will issue a ruling on how the claims will be understood during the remainder of the trial based on the party’s proposed constructions and other evidence taken at the hearing. At that point, a jury may be seated and the remainder of the trial can begin.
When a verdict is rendered in a patent case, an appeal may be taken to the United States Court of Appeals for the Federal Circuit (CAFC). Federal cases are normally appealed to an appellate circuit covering the geographic area of the trial court. In the Seattle area, a federal case would normally be appealed to the Court of Appeals for the Ninth Circuit. However, some time ago Congress decided that a single appellate court for patent appeals from all over the country would be a good idea, so all appeals of patent cases are heard by the CAFC.
That brings us back to the standard of review. The CAFC is like any other appellate court, in that findings of fact are usually undisturbed unless they are clearly erroneous, but findings of law can be considered anew. And in 1995 the CAFC decided, in the case of Markman v. Westview Instruments (the case that gave the name to the Markman hearing), that claim construction was a matter of law, which meant that it could be reviewed in its entirety during an appeal. That CAFC decision was subsequently unanimously upheld by the Supreme Court in 1996, which was itself later noted by the CAFC in Cybor v. FAS reaffirming their own right to review claim construction “de novo” (i.e. to determine claim construction again from scratch).
The Markman holding was interesting. Claim construction involved the taking of evidence, which could include in-person expert testimony (which is commonly used to understand highly scientific or technical language of patent claims), and evidentiary findings are normally questions of fact undisturbed on appeal. The CAFC’s ruling that claim construction was actually a matter of law meant that it had the power to reverse a claim construction from the trial court even without the benefit of the evidence on a first-hand basis. The Court noted that the Supreme Court had previously considered claim construction to be a matter of law, but also noted that a patent is a written instrument. It was almost as if a patent claim is a statute, and since the application of statutes is fully reviewable on appeal, so should patent claims be.
The effect was that the CAFC became almost a second trial court for patent cases. Since the outcome of the patent case is normally highly dependent on the meaning of the claims, and since the meaning of the claims was reviewable by the CAFC, there was almost no downside (other than expense) for the loser of a patent case to ask the CAFC to take another look. I have heard that some U.S. District Judges privately complained that they didn’t know why they should even bother with patent cases since the majority of it would be re-heard on appeal anyway. And of course, the possibility of virtually every patent case being appealed meant that patent litigation itself could be very expensive. From a practical standpoint, patent litigants could almost expect to have to fund two trials.
Which brings us to this past Tuesday, when the Supreme Court altered the standard of review for patent cases in its holding in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc. Justice Breyer, writing for a 7-2 majority, found that while claim construction is a matter of law reviewable de novo on appeal, factual findings taken from extrinsic evidence heard at trial which lead to the claim construction are not normally reviewable on appeal, except in cases of clear error.
“Extrinsic evidence” refers to evidence such as expert testimony that is outside the patent document itself. “Intrinsic evidence” means the patent itself, where the detailed description offers its own documentation of the meaning of the claims, as well as the record of the proceedings at the patent office which led to issuance of the patent. So, intrinsic evidence in a patent, which is all in writing and easily able to be placed in the record at trial, and its subsequent impact on claim construction, may still be considered de novo. However, the aforementioned extrinsic evidence is no longer set aside on appeal (except in cases of clear error, as mentioned above).
For example, in Teva v. Sandoz, the trial judge listened to expert testimony about what the claim terms meant. Each side presented their own expert, who provided an account of what the claim terms would mean to an ordinary person with skill in pharmaceuticals who was reading the claim. The trial judge chose a claim construction consistent with Teva’s expert witness testimony, rejecting the testimony of the expert witness from Sandoz. On appeal, the CAFC re-construed the claims, and this time rejected the testimony of the Teva expert witness, offering no deference to the trial judge’s finding in regards to the Teva expert’s testimony.
The Supreme Court now says this method of review is wrong, and that the CAFC must give deference to factual findings stemming from extrinsic evidence (for example, which expert witness testimony taken during a Markman hearing is correct) even when they inform the claim construction, which is still a question of law. The CAFC may still reinterpret the intrinsic evidence (the patent’s own detailed description, for example) leading to the claim construction and/or reverse the trial judge’s ultimate construction of the claim. However, going forward the CAFC must give full credit to the trial judge’s findings of fact regarding the extrinsic evidence such as resolution of conflicting expert testimony presented about the meaning of the claims. The only way for the CAFC to ignore such evidentiary findings informing a claim construction is if they are in clear error.
Pulling some issues for review on appeal off the table, as the Supreme Court did in Teva v. Sandoz on Tuesday, will lessen the possibility of a patent trial being heard once at the trial level, and then again in its entirety at the appellate level. Patent litigation costs may drop over time as a result.