Patent Law & Policy

Keep a Nonobviousness Doctrine That Doesn't Lose Suction

It slices and dices with no messy cleanup, it painlessly removes unwanted hair, it cleans room after room and doesn’t lose suction, it tones all the muscles of your lower body in one effortless movement, and it should be held to the same standard of patentability as every other invention. In his article, Uncertainty and the Standard of Patentability, Robert P. Merges proposes a new model for interpreting nonobviousness to lower the threshold of patentability for very high cost research. Although the economics of his risk management justifications are sensible, this proposal departs from the basic policies underlying patent law, as well as the lessons of real-world examples.

Merges argues that the best patent scheme is one that only rewards those inventors who would not have invented but for the expectation of a patent. This minimizes the social costs associated with the number of patents granted, while maintaining the supply of inventions incentivized by exclusive rights. He also suggests that the patent system has a greater influence on the development of technology, rather than actual invention. Thus, major research and development firms are best influenced by patent schemes, as their efforts typically involve more development work. Among other considerations, the prospect of securing patent rights to an invention encourages hi-risk research by these firms. Research that that involves more uncertainty and higher cost translates to higher risk. Merges concludes that by lowering patentability standards for very high-cost research, we will achieve a socially optimal level of research and development.

There are reasons to believe that this new patent doctrine will not serve its purpose, as well as reasons to believe that it is inconsistent with the motivation behind requiring an inventive leap over the prior art, even if successful. Since the proposal uses the standard of obviousness to lower the patentability threshold, the purpose of 35 U.S.C. § 103 must first be examined. The fundamental principle behind § 103(a) is to reward actual brainwork, rather than opportunism. By separating the quick-dissolve gelcaps from the pencil erasers and porcelain doorknobs, society is ensuring that patentees are only rewarded for developing something that could not have been invented by just anyone (having ordinary skill in the art). It is not, however, and never has been, a recognition of sweat-of-the-brow, capital invested, or time spent. Allowing the standard of patentability to be determined by the cost of research would go against a long history of well-founded legal indifference to the labor behind inventions. The worth of an invention to the public is determined by the size of the inventive leap over the prior art, not the risk the inventor took in pursuing his idea. Uncertainty in the field and the difficulty of research may relate to a product’s market price, but it does not relate to inventive value. This is not to say that uncertainty in the field does not impact a determination of obviousness for a new invention. On the contrary, uncertainty already makes up a large part of §103 analysis via the determination of the person having ordinary skill in the art. Determining the phosita is perhaps the most crucial step in deciding whether an invention is obvious, and the research and development firms that would benefit from Merges proposal already get the benefit of the doubt over ordinary skill in an uncertain art. The level of ordinary skill in the art of pharmaceutical research and biotechnology is considered to be such that the patentability bar is already lower than industries like business models or electronics. Biotechnological processes often yield wildly unforeseeable results, and the complexities of cellular machinery, substrate affinity, and genetic mutation make the success of experimental designs unpredictable. Section 103 (b) reduces the scope of § 103 (a) invalidating prior art for biotechnology (the industry subject to the highest level of uncertainty). By shifting the entire focus of § 103 onto uncertainty, there will be an even greater disparity in the application of nonobviousness to various industries. Removing the consideration of uncertainty from the confines of phosita analysis undermines the intent of § 103 to be case-specific and industry neutral. It also overshoots the socially optimal level of risky research that Merges strives to attain.

For example, BIOGEN-Idec, maker of the formerly successful multiple sclerosis drug Avonex, is a large drug design and development firm that is very patent conscious. The company has derived all of its wealth from one uber-patent, and has been looking for more patents ever since. This is the model firm for Merges’s proposal, in that every risk in research is balanced against the chances for patentability of a product. Even in this ideal example, the Merges proposal does not promote the progress of useful arts.

Granting patents for more obvious inventions in risky fields will not create a greater incentive for BIOGEN to develop useful new drugs, it will only create more pressure to seek patents along the way, and it will make the patents granted less meaningful. Crowding the patent landscape and increasing the burden of obtaining licenses from competitors deters research from over-patented fields, and slows the progress of these developments. Merges tries to head off this argument by assuring that his proposal will not lead to the anti-commons tragedy of the fisherman, since inventions, unlike fish, have to be dreamed up, and are not just swimming around for the taking. This assurance seems to forget that inventions below the obviousness threshold are, by definition, already swimming around in the collective consciousness of the phositas, and do not require substantial brainwork to dream up.

Fields with higher uncertainty also face bigger challenges designing around existing patents. It is difficult enough to identify the cause of a disease, find a way of combating it, design a drug to deliver the treatment, and make that drug safe and practical to administer, without having to design around other patented methods. Increasing the patent presence in BIOGEN’s field raises the costs of high-cost research, while lowering the value of the reward. The drawbacks to the Merges proposal are not confined to the targeted kinds of firms in the targeted kinds of fields. By creating a system that favors the patentability of uncertainty, the law would comparatively underpay inventors in more predictable fields for greater contributions. Disfavoring the man who invents a better vacuum cleaner for the firm that makes a very minor improvement to a method for tagging an unpopular restriction enzyme raises both practical and ideological concerns.

Practically, it raises problems with the relationship among high-cost, high-risk, and uncertain research. The patent-induced behavior model only works if all firms fit the BIOGEN profile, but this is not the case in reality. The risk management decision formula in the Merges proposal is too sophisticated for the small-time inventor, who does not have multiple projects from which to choose based on patentability standard. This inventor does not risk other avenues for success by focusing on vacuum cleaners over, say, hair removal creams. He researches the field of his expertise or idea, he invests what he can afford, and he risks failure. All research is high-cost research for the small inventor, and uncertainty is out of the inventor’s control. Lowering the standard for patentability in uncertain fields has no way of influencing the behavior of the small-time inventor to the benefit of society.

Merges may not be bothered by this problem, because he only seeks to reward the risks of those inventors that would not develop their ideas but for the expectation of patentability. Assuming that large firms are the only entities in this category is incorrect. Small inventors would not bear the cost of seeking a patent if it were not important to them, and they would not invest their limited time and resources if they did not think they could profit from their invention. Small inventors may desire patents even more than large firms, since they do not have the ability to keep trade secrets, and often rely on the ability to sell the rights to their invention in order to commercialize it. The likelihood of securing a patent motivates not only inventors, but lawyers. Small-time inventors already have limited access to legal council, and attracting patent attorneys to the products of large-firm risky research by increasing the likelihood of successful prosecution further burdens small firms.

Ideologically, patent law should treat all industries, all inventive entities, and all research investments equally. To legally require certain kinds of inventions to clear a higher inventive hurdle than others sends the message that certain industries do not deserve a level playing field. The greatest good for society is achieved when each inventor is incentivized to make the best use of his education and talent, rather than to seek out higher-paying risks (even if this leads inventors to arts of vegetable processors, hair creams, or exercise equipment, instead of cancer research). Although drug design is less predictable than vacuum cleaner innovation, it does not follow that a particular drug is less obvious than a particular cleaning system. The quid pro quo of patent protection is based on the quality of the inventor’s contribution to the progress of his art, not the cost, risk, or uncertainty of the path he took.

A Doctrine of Equivalents that Performs Substantially the Same Function and Achieves a Substantially Different Result

The decisions of Graver Tank and Warner-Jenkinson introduced the modern understanding of the doctrine of equivalents in the context of chemistry. Graver Tank identified the test for equivalence between two inventions as performing “substantially the same function in substantially the same way to obtain the same result” even if the two inventions “differ in name, form or shape.” The court went on to explain that this test is not employed in a vacuum, but is determined against the context of the circumstances of the case. While the chemical processes at issue in Graver Tank and Warner-Jenkinson fit within the limitations of this test, it is not clear that biological processes and sub-cellular “devices” are appropriately suited for the modern equivalence framework.

One of the best illustrations for challenges to this framework may be its potential application to restriction enzymes. Restriction enzymes are endonucleases that recognize and cut DNA at a particular sequence of nucleotides. The nucleotide sequence corresponding to a particular restriction enzyme may naturally occur any number of times in a given source or construct of DNA, or may be intentionally included in the sequence during the creation of the construct. By cutting DNA at specific locations and mixing two or more samples of these DNA fragments together with the enzyme DNA ligase, scientists can engineer recombinant DNA constructs to meet their specifications. The use of recombinant DNA has been essential to almost all genetic research and medical therapies using human proteins. Some examples of the most popular restriction enzymes are AluI and HaeIII, which make even cuts through DNA leaving blunt ends, and BamHI, HindIII, and EcoRI, which make offset cuts through DNA leaving exposed base-pairs at the ends of the fragments anxious to bind again.

[diagram omitted]

Each restriction enzyme optimally functions within a specific temperature and pH range, and is listed in large directories from which scientists choose the DNA cutting tool best suited for their purposes. If a certain cutting sequence is too numerous, or too close to the functional domain of a desired coding region, scientists can simply substitute an equivalent enzyme. This need for options is reflected in the commercial viability of more than 400 restriction enzymes. Under Graver Tank and Warner-Jenkinson, alternative biochemical processes using restriction enzymes would be considered equivalent in almost every instance. They all serve the same function (to cut DNA at chosen locations) in the same way (using an endonuclease that recognizes short, usually palindromic sequences, and operates within certain environmental conditions) to produce the same result (DNA fragments with blunt or sticky ends capable of being ligated with other DNA fragments to produce new constructs). Any person having ordinary skill in the pertinent art would describe restriction enzymes as being equivalent, indeed, many restriction enzymes derive their entire commercial value from being equivalent to popular enzymes. This result seems consistent with the facts and analysis in Graver Tank.

In Graver Tank, welding compositions were found to be equivalent when one formula used magnesium silicates and the other used manganese silicates. It was found that in the context of welding, magnesium and manganese compositions were equivalent because chemists understood the two metals to be interchangeable in the patented composition. Likewise, restriction digests are equivalent when molecular geneticists understand two restriction enzymes to be interchangeable in a given reaction. The problem with this framework is that the process of securing and experimenting with chemical ingredients like magnesium and manganese is vastly different than the process of securing and experimenting with restriction enzymes, and implicates a different system of incentives. Although restriction enzymes are among the most stable and predictable proteins, identifying them from their bacterial or other host organism, isolating, purifying, and preparing them for use in a laboratory setting involves an enormous experimental effort. This is yet another example of how the high level of uncertainty and cost-intensive experimentation in biotechnology research and development challenges uniformly applied patent doctrines.

In addition to the difference in effort and optimal incentive level for creating processes using restriction enzymes, restriction enzymes themselves constitute patentable inventions. Trying to distinguish restriction enzymes by result via comparing blunt vs. sticky ends is unavailing, because this only separates the enzymes into a few large categories of hundreds of otherwise equivalent enzymes. The only avenue for distinguishing restriction enzymes under Graver Tank and Warner-Jenkinson is by arguing that restriction enzymes do not work in substantially the same way because they have a unique cutting sequence, and require different environmental conditions. This suggestion only goes so far, because many enzymes share cutting sequences, and it is not entirely convincing that changing a few base pairs is enough to substantially alter the way in which an endonuclease is considered to achieve its function. For example, magnesium is in Group IIA, period 3 of the periodic table, meaning it has two electrons in the outer binding orbital, and manganese is in Group VIIB, period 4 of the periodic table, meaning it has seven binding electrons and is a much bigger and heavier atom than magnesium. The two elements bind atomically to form silicates in different ways, but this is not an important difference in the context of welding, wherein manganese (a transition metal) is interchangeable for practical purposes with alkaline earth metals. Likewise, enzymes will often function outside of their optimal temperature and pH ranges (sometimes unexpectedly), optimal conditions are common to many enzymes, and often overlap. In the context of DNA restriction digests, the differences in how the enzymes cut is similar to differences in how metals bind when forming silicates.

Basic patent policy does not support treating restriction enzymes as equivalent. Despite restriction enzymes being equivalent under standard function-way-result analysis, they represent valuable inventions and should be incentivized under the patent system (even though restriction enzymes are “naturally occurring” they are patentable in their useful form under Parke-Davis) to promote future invention and development in recombinant DNA. Situations where one restriction enzyme may have an advantage over another are not foreseeable at the time of patenting or invention, and may never be encountered, but have they the potential to be immensely valuable if ever encountered by anyone. On the other hand, the policy underlying the doctrine of equivalents as discussed in Graver Tank and Warner-Jenkinson cuts somewhat in the other direction. Allowing new restriction enzymes that perform the same function and achieve the same result as patented enzymes to be found non-infringing devalues existing restriction enzyme patents.

This restriction enzyme problem and all of the preceding argumentation is also applicable to other sub-cellular tools for making recombinant DNA, as well as affinity beads (used in protein purification and purification of many nuclear components in tandem), antibodies, radiolabels, and other common reaction components of molecular biology, biochemistry, and genetics. There are two possible responses to this problem. The first is to construe the function, way, and result of these inventions very narrowly in the context of biotechnology. For example, the function of EcoRI is to cut at the sequence 5’ …GAATTC… 3’, the way it does this is confined to its optimal temperature, pH range, and the particular phosphodiester bonds broken, and the result is a DNA fragment with four overhanging single-stranded base-pairs reading 5’ …AATT… 3’. This narrow interpretation produces a better outcome (a finding against equivalence) in this situation, but cannot be applied to other industries effectively or desirably.

The second possible response to the problem would be to develop a separate equivalence test for biotechnological inventions. While such a bold move would unlikely be well received by courts, it avoids the slippery slope and industry characterization problems of interpreting function, way, and result with vastly different degrees of broadness across different contexts. Biotechnology as already been singled out in obviousness analysis (§ 103(b)) without extending this exception to any other industry, making the case for extending customized equivalence tests to industries beyond biotechnology weak, and the case for singling out biotechnology less problematic. How this new and improved test would be worded is subject to numerous considerations and possibilities.

I think such a test would also need to further distinguish between process claims and device claims within biotechnology. One possible suggestion for such a test would be that biotechnological processes are equivalent if they perform substantially the same function in a substantially compatible way, to achieve effectively the same result and a researcher having ordinary skill in the art would reasonably expect this equivalence. Key word changes in this test are compatible and effective. Whether a function is achieved in a compatible way is determined in light of how extensively reactions must be redesigned to accommodate the different processes. This includes, but is not limited to, redesign of buffers, temperatures, concentration of active ingredients, timescale of reaction, and calibration of instruments. Results are effectively the same if they are interchangeable without undue biochemical adaptation in subsequent experiments. Sub-cellular devices are equivalent if they are interchangeable without undue biochemical adaptation. Undue biochemical adaptation is similar to undue experimentation, but the threshold for what kind of work counts as adaptation is slightly lower and in a narrower scope.

As applied to restriction enzymes, affinity beads, antibodies, radiolabels, etc. they would still be considered to perform substantially the same function. Some inventions, such as restriction enzymes and affinity beads would be found not equivalent for functioning in incompatible ways, although not all enzymes would definitely have this protection. Other inventions, such as labels and processes for translating proteins would be found not equivalent for producing results that require undue biochemical adaptation before becoming suitable for use in the same subsequent application.

As a whole, this test raises the standard of equivalence for biotechnology, reflecting both the unique uncertainty of the field and the unique advantage of having many seemingly equivalent options from which to design experiments (thus justifying the incentives for developing these options). It creates a wider gap between legal equivalence for the purposes of patenting, and practical equivalence for the purpose of research. This will lower the value of patents in biotechnology, but create more incentives in other ways and preserve more flexibility in research. Although this test is only slightly different than a well-informed application of the current test, I think it would have substantially different outcomes in certain highly valuable aspects of biotechnology.

Tick... tick... tick... Time's Up: On Your Patent Rights

In /Symbol Technologies, Inc. v. Lemelson Medical, Ed. & Res. Fdn., L.P./, the Federal Circuit held that the equitable doctrine of laches could be used to bar enforcement of patent claims which have been delayed in prosecution for an unreasonable or unexplained period of time. This bar applies even if the applicant complied with pertinent statutes and rules. The policy behind this doctrine is that unreasonable delays can grant certain advantages to the patentee and disadvantages for the public, i.e. less time for public to have access to the invention and greater monopoly period for the inventor. The equity in this doctrine is easily seen by looking at the circumstances of /Symbol Technologies, Inc./ Lemelson had filed patents in the 1950s, which remained pending and unpublished for eighteen to thirty nine years after they were filed. In 1998, Symbol Technologies’s customers began receiving letters from Lemelson stating that the use of Symbol’s products infringed several of Lemelson’s patents. This was a time period in which bar code technology had already progressed and been in practice for years, so it was a surprise for an industry already established. Allowing Lemelson to proceed in this infringment suit would have been giving him protection for a patent filed nearly fifty years beforehand.

A series of pushes and pulls determine the boundaries of patent laws; on the one hand we want to provide ample protection and monopoly for an inventor as an incentive to create, but on the other hand we want the public to benefit from these inventions for the greater good. Currently, the time period of protection from the date of issuance is twenty years, but prosecution laches is a defense which serves to bar enforcement in order to uphold the policies behind patent laws. Other time limitations within these twenty years can also serve to enforce the goals of patent law in other circumstances. I will discuss another laches doctrine in United States patent law which is similar to a statute of limitations and also the concept of “local workings” which is not currently in practice in the United States.

Laches is an equitable defense available in patent infringement litigation under 35 U.S.C. Section 282. (Szendro, Peter. /Doctrine of Laches and Patent Infringement Litigation/, http://www.converium.com/2103.asp). If the patent holder has delayed too long before commencing litigation, laches will typically limit recovery but usually not serve as a complete defense to infringement. Courts have generally considered six years to be the period of time in which a presumption of laches arises. However, there is no set time period, rather the inquiry is many times related to the facts of the case, i.e. whether the infringement was open and notorious, whether the patentee policed its rights by surveying the marketplace, etc. The laches defense is very important in ensuring that a patentee will not just sit and wait for an infringer to accumulate large success before filing a suit, hoping to get more in damages or more leverage because the product is already widely distributed. A recently settled suit with NTP and RIM, although not directly related to laches, illustrates this concept of delayed suit filing very well and the problems it can cause for the alleged infringer and the public. NTP was the owner of wireless email patents, which were issued in the early 1990s. The first Blackberry email device, produced by RIM, came out in 1998, and NTP subsequently filed suit in 2001. (Hickey, Andrew R., /Blackberry Dodges Blackout Bullet for Now/, http://searchcio.techtarget.com /originalContent/ 0,289142,sid19_gci1169025,00.html ). However, due to a series of delays resulting from litigation, a final settlement wasn’t reached until 2006. The following short timeline illustrates the possible rise in stakes which can occur over the course of five years.

* 2002, Nov. 21 - A federal jury awards NTP $23.1 million in damages. * 2005, Mar. 16 - RIM agrees to pay NTP a settlement of $450 million. * 2006, Mar. 3- In final settlement, RIM pays NTP one time sum of $612.5 million

On top of the exponential rise in damages, the threat of a possible injunction on Blackberry devices caused hysteria within the public, who were frightened (or relieved) at the prospect of not being able to check email on their “Crackberries.” Again, this case serves not as an example of laches being applied as a defense, but rather as an illustration of the difference in damages that a few years can make in addition to the reliance of an industry or the public. The intention of patent law is not to protect inventor’s rights to the point where the overall good of society can be harmed.

Another example of a time limitation on the rights of patent holders is the “local workings” requirement, which exists in many countries outside of the United States. (Amir Attaran and Paul Champ, /Patent Rights and Local Working Under the TRIPs Agreement: An Analysis of the US-Brazil Patent Dispute/, 27 Yale J. Int’l. L. 365 (2002)). Local workings authorize compulsory licensing on the grounds of the patentee’s failure to work locally, which typically means a failure to manufacture the product in the territory of the country issuing the patent. The usual time period is three years after a patent’s issuance before someone else can apply for compulsory licensing from the original patentee. Once this right is granted by the governing body, the applicant may practice the invention without the consent of the patentee. The benefits of local workings to the country as a whole are obvious; no one can create an innovation and then not allow the public to at least get the benefit of the product. Once again, the goals of patent law are to balance the rights of the inventor with the benefit to the public, so in the United States we do not use local workings in order to respect the wishes of the inventor. However, given the harm to the public, it is my opinion that in most cases it will outweigh any harm which may come to the inventor from not practicing his invention. It is also important to note that local workings involves an application for compulsory licensing, so it is not a certainty that it will be granted. There may be instances where a patentee does not want to practice the invention, and by excluding other people from practicing it is indirectly helping the patentee’s business. In these cases we can reject the applications for compulsory licensing since there is a material purpose behind the patentee’s actions, or lack thereof.

All of these time limitations on a patentee’s rights contribute to the overall policy objectives of patent law. Granting a strict twenty year period of exclusion rights from the date of issuance is not sufficient in itself to uphold these objectives. On top of prosecution laches and the statute of limitations, I believe we should also incorporate local workings into US patent law. Not only would this serve the public good of receiving the benefits of innovation, it would partly deter the behavior of patent trolls. A stricter definition of patent trolls is a person or business who is in the possession of a large number of patents and makes a living from filing infringement suits on other companies. This behavior is a recognized problem in patent law, constantly threatening to put enough weight on the negative aspects of patent law to make it outweigh the positive. Often times a patent troll will not have any research or manufacturing of its own, thus not leaving itself open to counter-suits and more importantly not contributing any innovation of its own to the scientific community. As its name suggests, the patent troll will figuratively hide in waiting and leap out at unsuspecting infringers to bring them to court, often times leveraging them for a settlement. If there were a local workings requirement, the value of unworked patents would drop, since the time limitation of a patent would practically be three years (assuming vigilant searching by would-be infringers). Patent trolls would have a harder time coming up with a valuable portfolio of patents, and given a less lucrative business there wouldn’t be as much incentive to troll.

Should we implement a local workings requirement in the US? How does the balance of the goals of patent law come out with regards to this?