|
Article Excerpt
This Article examines whether U.S. district court judges improve their skills at patent claim construction with experience, including the experience of having their own cases reviewed by the Court of Appeals for the Federal Circuit. In theory, higher courts teach doctrine to lower courts via judicial decisions, and lower courts learn from these decisions. This Article tests the teaching-and-learning premise on the issue of claim construction in the realities of patent litigation. While others have shown that the Federal Circuit reverses a large percentage of lower court claim constructions, no one has analyzed whether judges with more claim construction appeal experience fare better on subsequent appeals. Surprisingly, the data do not reveal any evidence that district court judges learn from prior appeals of their rulings. There is no suggestion of a significant relationship between experience and performance. The lack of evidence that Federal Circuit review aids district court judges is disconcerting. The Article explores three possible explanations for the lack of evidence: (1) that the nature of claim construction is indeterminate; (2) that district court judges are incapable of or not interested in learning how to perform claim construction; and (3) that Federal Circuit decisions do a poor job of teaching district court judges how to construe claims. These results shed critical light on the functioning of the patent system. Moreover; the results are relevant to a broader understanding of the relationship between higher and lower courts in general.
TABLE OF CONTENTS INTRODUCTION I. CLAIM CONSTRUCTION AND PREVIOUS EMPIRICAL SCHOLARSHIP A. The Basics of Claim Construction B. Canons of Claim Construction C. Empirical Literature on the Difficulties of Claim Construction II. STUDY DESIGN AND METHODOLOGY A. The Appellate Decisions B. Reliability and Validity of the Appellate Decision Database C. The District Court Judge Lawsuit Database D. Limitations of the Databases III. RESULTS AND DISCUSSION A. Basic Results on Judicial Districts and District Court Judges B. The Impact of Previous Appeals 1. Performance Based upon Number of Previous Appeals 2. Performance After First Reversal C. The District Court Judges with Experience 1. Judges with the Most Patent Experience 2. Judges with the Most Overall Judicial Experience 3. The Age of Judges IV. ANALYSIS AND EXISTING SOLUTIONS A. Claim Construction Indeterminacy B. Possible Solutions at the Trial-Court Level C. Possible Solutions at the Appellate-Court Level D. Solutions at the Patent Office CONCLUSION APPENDICIES
Experience is the name everyone gives to their mistakes.
--Oscar Wilde
INTRODUCTION
Practice makes perfect--the old adage promises that the more you practice, the better you will become. No doubt this logic holds for many endeavors, from playing the trumpet to shooting basketballs to delivering a rousing speech. But does the truism carry over into judicial decisionmaking? Do judges learn to decide cases more accurately through experience?
This Article examines whether U.S. district court judges improve their skills at patent claim construction as a function of experience, specifically as a function of having their own cases reviewed by the Court of Appeals for the Federal Circuit. A core premise of the U.S. legal system is that legal doctrine is taught by higher courts and applied by lower courts. (1) Whether by rules or standards, high courts provide guidance that lower courts must consider. Implicit in this process is the notion that lower court judges are capable of learning the doctrine conveyed by higher courts. This Article analyzes how well this core premise of teaching and learning operates in the realities of patent litigation.
According to the teaching-learning theory, the Federal Circuit teaches district court judges how to perform patent claim construction. (2) Claim construction is the process of interpreting the specific terms or phrases used by the patentee to define the technology covered by the patent. Patent claim construction requires judges to understand the nuances of a particular technology at issue. After mastering the underlying technology in the case, district court judges must apply the claim construction doctrine as instructed by the Federal Circuit. Because learning the technology is essential to construing the claims, reading a body of case law alone is of limited value. Rather, district court judges must learn claim construction through hands-on practice. Previous studies have shown that the Federal Circuit reverses decisions on the issue of claim construction at an alarming rate. (3) However, to date, no one has analyzed whether judges with more claim construction experience fare better on subsequent appeals.
This Article explores the issue for the first time using a novel database containing all Federal Circuit claim construction appeals of decisions from district court judges. The Article provides an extensive analysis into the reversal rates of district court judges with varying levels of patent experience as measured by instances of appellate review. Surprisingly, the data do not reveal any evidence that district court judges learn from appellate review of their rulings. The lack of evidence that Federal Circuit review aids district court judges is disconcerting. It suggests either that district court judges are incapable of or not interested in learning, or that Federal Circuit decisions do a poor job of teaching. The latter explanation is consistent with criticism by others that the Federal Circuit does not provide clear guidance on claim construction. (4)
There may be profound repercussions in our judicial system if judges are not learning how to decide patent cases from Federal Circuit review. Specialized courts are not needed for patent cases if the specialization does not translate into accuracy. Also, Congress is currently debating the creation of quasi-specialized patent trial judges. (5) If experience does not lead to more expert decisions, much of the impetus behind the proposal dissipates.
There also might be serious repercussions if the Federal Circuit fails in its teaching role. Failure means the patent system is inconsistent or unpredictable and that the Federal Circuit has not succeeded in enhancing it. The Federal Circuit acknowledged its high reversal rate on district court judges' claim construction and noted that "a lack of predictability about appellate outcomes ... may confound trial judges and discourage settlements." (6) Unpredictability decreases the possibility of settlement, which raises legal costs and leaves companies, investors, and inventors uncertain whether their technologies will infringe the rights of others. Consequently, the Federal Circuit's failure to foster a predictable patent system may impede innovation.
A teaching failure may also have profound effects on the "students of the class"--the district court judges. Those judges may lose interest in trying to construe claims properly. They may become demoralized and believe that their work on patent claim construction is of no matter. (7) This fatalism may tempt district court judges to resolve issues of claim construction quickly via either a preliminary injunction or a dispositive summary judgment for the sole purpose of obtaining prompt appellate court review. Alternatively, district courts may be wary of enforcing their judgments due to the high reversal rate. (8) Although prompt appellate review may have some benefits, they are likely outweighed by the costs to the litigants and to judicial efficiency.
This Article has four parts. Part I expounds the law of patent claim construction and the various criticisms of the Federal Circuit's claim construction methodology. It then describes other empirical scholarship on claim construction.
Part II explains the study design and methodology. This explanation describes the process of locating and selecting the population of cases and lawsuits, and how the relevant information was coded. Part II also discusses limitations of the data and provides a measure of the reliability and validity of the data.
Part III sets forth the empirical results of the study. These results are broken down into three Sections. The first Section sets forth summary statistics of the information regarding the appealed cases, including the distribution of district court judges and the judicial districts. The second Section reports the claim construction success rate of previously appealed district court judges. The third Section compares other experience, such as overall judicial experience or total number of patent cases handled--as opposed to patent appeals experience--to claim construction reversal rates.
Part IV discusses various potential solutions to the problems with claim construction.
The Appendices to this Article outline details on the methodology used to create the databases, analyze the data using additional metrics of experience, and investigate the potential for selection bias.
I. CLAIM CONSTRUCTION AND PREVIOUS EMPIRICAL SCHOLARSHIP
Patent law is difficult. Not only is the law intricate and ever-changing, but the patents themselves describe complex and often cutting-edge technologies. (9) District court judges are typically legal generalists, with no training in patent law. (10) They also commonly lack scientific training or a technical background. (11)
As part of almost all patent infringement lawsuits, the judge eventually must determine the scope of the patent's reach--the limits of the patentee's right to exclude. (12) This determination is known as claim construction. It is often vigorously contested by the litigants, because claim construction is often the make-or-break determination in patent litigation. This Part sets forth a brief explanation of the law of claim construction and some problems associated with it that have affected district court judges. (13) It then describes other empirical scholarship on claim construction.
A. The Basics of Claim Construction
Claim construction is often the centerpiece of patent litigation. At some point in nearly all patent infringement lawsuits, the patent claims (14) or portions thereof must be interpreted to determine how infringement, validity, and other issues under the patent will be measured. (15) Patents typically have multiple claims, and each claim is considered separately for issues of infringement and validity. (16)
In Markman v. Westview Instruments, Inc., the Supreme Court ruled that judges, not juries, must construe patent claims. (17) Since the Markman ruling, federal district court judges have been exclusively assigned the difficult task of interpreting all controverted patent claims. (18) Shortly after Markman, the Federal Circuit decided that the district court's claim construction analyses were subject to de novo review by the Federal Circuit. (19) This combination of shifting the responsibility for claim construction from juries to judges and raising the standard of appellate review resulted in a substantial increase in the Federal Circuit's discretion in reviewing claim constructions.
B. Canons of Claim Construction
To aid district court judges in construing claims, the Federal Circuit has articulated canons of construction. Decisions of both the district courts and the Federal Circuit utilize these canons in their analysis. On some (if not most) occasions, the various canons are inconsistent and point toward contradictory claim constructions. Below is a general background on the canons of claim construction, with an emphasis on some common claim construction problems.
Sitting en banc in 2005, the Federal Circuit attempted to clarify how to construe patent claims in Phillips v. AWH Corp. (20) There, the Federal Circuit made clear that the preferred way to construe a claim was to study the "intrinsic evidence." (21) This "intrinsic evidence" includes the claim language at issue, (22) other claims in the patent, the remainder of the specification portion of the patent, (23) and the record of correspondence between the patent applicant and the U.S. Patent Office (known as the "prosecution history"). (24) According to Phillips, the court must construe the claims based upon a review of this evidence. (25) Of the intrinsic evidence, the Federal Circuit stated that the specification of the patent document is "the single best guide" for determining the meaning of a claim. (26)
Over the years, the Federal Circuit articulated various canons of construction concerning patent terms. Phillips did not materially change these canons or discourage their use, and it failed to establish any hard and fast rules on claim construction. At first blush, these canons provide a clear explanation as to how district court judges should construe any given claim term. In practice, however, they are more like standards than rules, and they leave gray areas with respect to claim construction. In many cases, at least one of the general canons will support each party's respective position. In these cases, the district court judge must use his or her judgment, guided by Federal Circuit law, to balance the competing canons and effectively construe the patent claim.
Two canons of construction appear particularly contradictory. One canon of claim construction says that the court should not "read in a limitation" from the specification. (27) Said another way, if the claim language is broad and the examples in the detailed description in the patent document are narrow, the claim should be construed broadly. A second canon says that a claim must be read in view of the specification. (28) Arguably, this means that if the embodiments described in the detailed description of the patent are all narrow, then the language in the claim must also be narrowly construed.
In Phillips, the Federal Circuit noted the difficulties raised by these seemingly contradictory canons. However, it asserted that the line between the two canons could be "discerned with reasonable certainty and predictability." (29) It further contended that:
[U]pon reading the specification in ... context, it will become clear whether the patentee is setting out specific examples of the invention to accomplish [the goals of the invention], or whether the patentee instead intends for the claims and the embodiments in the specification to be strictly coextensive. (30)
In other words, the Federal Circuit believed that reading the patent would clarify the applicable canon in each case. Yet, notwithstanding its remarks, the line between these canons is often murky. (31)
In addition to these two frequently contradictory canons, the Federal Circuit has approved numerous others. For example, one canon urges the court to consider "claim differentiation." (32) Essentially, this means that claims should be construed so that each claim has a different scope. It works as follows: suppose a patent has two claims, one of which uses a broader term to describe an aspect of the invention, and the other of which uses a narrower term. In construing the broader term, the doctrine of claim differentiation urges that the broader term not be construed to mean the same as the narrower term. To construe the claims otherwise would render the narrower term superfluous. Other canons instruct that a claim should be construed to preserve its validity whenever possible and that a claim should be interpreted to include its preferred embodiment. (33)
A case which exemplifies district court judges' difficulty in determining which canon of construction to apply is Nystrom v. Trex Co. (34) Nystrom presented the federal district court with a seemingly simple task: to construe the meaning of the term "board." (35) The term appeared relatively straightforward. The patent-in-suit related to a construction material for use in flooring surfaces, and more specifically to boards for use in constructing an exterior surface such as a deck. (36) However, construing the term "board" under the applicable canons of claim construction proved to be anything but straight-forward. (37)
The patentee, Nystrom, argued that the term "board" should be construed according to its ordinary meaning to those of skill in the art: any elongated piece of material for use in building. (38) The accused infringer, Trex Co., asserted that the term "board" must be construed more narrowly to mean a "piece of elongated construction material made of wood and cut from a log." (39) The construction of the term was case dispositive: Trex Co.'s product did not use wood cut from a log. (40) Thus, if the term was construed as requiring that the product use wood from a log, Trex Co. would be entitled to summary judgment. (41)
As support for his claim construction, Nystrom argued that the specification and prosecution history did not disclaim the broad general definition of a "board." (42) Nystrom argued that the claim language did not include any express limitation as to the material of the "board." (43) Trex Co. countered that the specification of Nystrom's patent only disclosed a board as made of wood and cut from a log. (44) The specification stated that the board of the present invention was "a superior product when cut from a log." (45) Trex Co. also argued that arguments made by Nystrom during prosecution (46) of the application that became the patent-in-suit supported a narrow construction of the term "board." (47)
Nystrom urged the court to rely upon the canon that limitations should not be imported from the specifications into the claims. Trex Co. urged the court to rely upon the canon that the claim must be interpreted in light of the specification. After considering the issue, the district court adopted the defendant's proposed construction. (48) The district court found that the term "board" meant "a piece of elongated construction material made from wood cut from a log," and therefore it granted summary judgment of non infringement against Nystrom. (49)
A divided Federal Circuit panel reversed the district court's grant of summary judgment and construed the term "board" to mean "an elongated, flat piece of wood or other rigid material." (50) The Federal Circuit considered the specification of the patent-in-suit. (51) The specification, in relevant part, stated that the "board ... yields a superior product when cut from a log." (52) The Federal Circuit found that this statement, which the district court heavily relied upon as support for a narrow construction, actually supported a broad construction. (53) The Federal Circuit noted that this language "implicitly recognize[d]" that the board may be made of materials other than wood. (54) To the Federal Circuit, it was of primary importance that the claim simply stated "'board,' without restricting the term to a particular material or describing characteristics of wooden boards cut from logs." (55)
Because of the Phillips decision, (56) the Federal Circuit withdrew its Nystrom opinion and substituted a new unanimous opinion by the same panel. (57) The substituted opinion again acknowledged that the claim language did not describe "board" as being cut from a log or made of wood. (58) However, the Federal Circuit reversed its previous holding by stating that "board" must be construed as limited to boards made of wood. (59) According to the Federal Circuit's substituted opinion, Nystrom had consistently used the term "board" to describe wooden decking material cut from a log. (60) Thus, construing "board" as made of wood was necessary to conform to the canon that the claim language must be read in light of the specification. (61) In rehearing the case, the Federal Circuit affirmed the district court's grant of summary judgment of non-infringement. (62)
As Nystrom exemplifies, there are clear problems with claim construction. Despite the fact that the canons of claim construction seem to be useful and practical tools for district court judges, in many cases at least one claim term will implicate contradictory canons of construction. At this point, the district court judge may need to use some reason and discretion in interpreting the claim. It is an open question whether district court judges with proper training can correctly determine which canon of construction to use. If the line between conflicting canons of construction can be objectively ascertained, then district court judges should be able to learn how to properly construe claims. On the other hand, if the nature of claim construction is inherently indeterminate, then experience with claim construction will not likely help.
C. Empirical Literature on the Difficulties of Claim Construction
Others have conducted impressive empirical studies on the Federal Circuit's handling of claim construction cases. (63) These studies reveal a serious problem with claim construction in U.S. courts. (64) None of the studies tracks the performance of the district court judges as individuals, however. Instead, they all treat the district court judges as a group, and only rarely has the information about district court performance been segregated by judicial district.
The most comprehensive study was performed by Judge Kimberly A. Moore, then a professor at George Mason University. (65) Judge Moore studied all precedential, non-precedential, and summary affirmance decisions of the Federal Circuit from April 23, 1996 (the date of the Supreme Court's Markman decision (66)) through 2003. (67)
Judge Moore found that the district courts wrongly construed 34.5% of claim terms. (68) She defined "wrongly construed" to mean that the "Federal Circuit determined that the district court claim construction was wrong (even if it did not actually result in reversal of the judgment) on a term-by- term basis." (69) According to her data, in 37.5% of cases, the district courts wrongly construed at least a single term. (70) Judge Moore also found that 29.7% of the judgments entered in patent cases had to be reversed or vacated because of an erroneous claim construction. (71) Judge Moore's article reports the outcomes of cases by each appellate judge. (72) In doing so, it does not disaggregate or otherwise analyze the claim construction appeals by district court judge.
In an earlier empirical study, Judge Moore had speculated that the cause of the errors in claim construction lay at the feet of the district court judges. She asserted that her data "suggest[ed] that judges [were] not, at present, capable of resolving these issues with sufficient accuracy." (73) In her later study, Judge Moore questioned whether this "high reversal rate could be due to the fact that district court judges lack technical training and repeat exposure sure to claim construction." (74) She rejected this hypothesis as "unlikely" because "construing claim terms in a given patent does not make construing claim terms in a different patent any easier." (75) Eventually, Judge Moore speculated that the Federal Circuit was to blame for the high reversal rate that she had identified. (76) She put forward that the Federal Circuit had not "evolved" the canons of construction sufficiently to guide district court judges. (77)
Christian Chu also performed an empirical analysis of the Federal Circuit's claim construction cases. (78) Chu found that between January 1, 1998 and April 30, 2000, the Federal Circuit modified claim constructions in forty-four percent of the 179 appealed cases. (79) Chu was clear that he included only cases in which the Federal Circuit had "expressly" reviewed a district court's claim construction. (80) But by so defining his study, he excluded all Rule 36 cases, in which the Federal Circuit summarily affirmed the district court's claim construction decisions. (81) Judge Moore, in discussing Chu's study, notes that omitting Rule 36 cases has a "profound ... impact on the results" and on the study of claim construction reversal rates. (82) Because he excluded Rule 36 cases, Chu's results are biased toward showing a higher claim construction reversal rate because all of the excluded Rule 36 cases affirm the district court's claim construction. (83)
Chu investigated whether the more active judicial districts were less likely to be reversed on claim construction. (84) He divided the appellate cases into two groups based upon whether they originated from "more active" or "less active" tribunals. Chu designated a tribunal as "more active" if more than ten of its cases were reviewed by the Federal Circuit during the period of his study. (85) Chu's study could not predict with confidence that the district court from which a case originates affects the likelihood that the Federal Circuit will change the lower court's claim construction. (86) Despite the lack of statistical significance in his results, Chu nonetheless speculated that his study demonstrated a pattern: "[T]he lower reversal rate [among more active tribunals] seemingly supports the theory that claim construction by a trial judge with substantial patent experience may decrease the prospect of reversal on appeal. Perhaps there is some truth in the notion that experience does matter." (87)
Thus, Chu speculated that experience reduces the claim construction reversal rate. (88) Perhaps because of the difficulties involved, Chu did not obtain the identities of the district court judges involved in his dataset. Accordingly, his speculation regarding judicial experience in patent cases was left untested.
R. Polk Wagner and Lee Petherbridge performed a slightly different empirical assessment of the Federal Circuit's claim construction cases. (89) Rather than focusing on reversal rates, Wagner and Petherbridge focused on the methodology used by Federal Circuit judges. They found that the judges on the Federal Circuit were divided between two methodological approaches to claim construction--a procedural methodology and a holistic methodology. (90) Judges either fit into one of these two methodologies or are "swing" judges. (91) Because of this split in approaches, according to Wagner and Petherbridge, claim construction at the Federal...
|
