American scientist online - bayh-dole reform and the progress of biomedicine

American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine Volume: 91 Number: 1 Page: 52
DOI: 10.1511/2003.1.52
Other Formats: PDF
Bayh-Dole Reform and the Progress of Biomedicine
Allowing universities to patent the results of government-sponsored research sometimes
works against the public interest
Arti K. Rai, Rebecca S. Eisenberg
Although the development of pharmaceutical compounds has long been a commercial enterprise, thebroader field of biomedical research has enjoyed a very different tradition, one that allows the freesharing of scientific knowledge. But the culture of open science has eroded considerably over the pastquarter-century. Proprietary claims have increasingly moved upstream, from the end productsthemselves to the ground-breaking discoveries that made them possible in the first place. Oneimportant reason for this change has been a narrowing of the gap between fundamental research andcommercial applications. Once largely a matter of serendipity or trial and error, drug discovery nowdepends critically on basic knowledge of genes, proteins and associated biochemical pathways. Inaddition, the practical payoffs of basic research have become easy to anticipate in many cases,making it straightforward to obtain patents for discoveries that in an earlier era would have seemedtoo far removed from useful application to warrant the effort.
This shift in patenting activity has met little resistance. For example, in 1980 the U.S. Supreme Court held that genetically engineered microorganisms wereeligible for patent protection. Shortly thereafter, Congress created a specializedcourt to hear appeals in patent matters, the Court of Appeals for the FederalCircuit, which has further extended the Supreme Court's expansive approach topatent eligibility. The Federal Circuit has also relaxed the standards normallyrequired for patent protection, such as proof of the practical utility of aninvention and of its lack of obviousness—standards that might otherwise haveprevented the patenting of incremental advances in biomedical research.
These changes in the economics of research and in the interpretation of thepatent laws have been important factors in the proliferation of intellectualproperty claims for discoveries of a fundamental nature. But perhaps even more significant has beenthe explicit U.S. policy of allowing grantees to seek patent rights for the results of government-sponsored research. This policy, which began in 1980 with passage of the Bayh-Dole Act, has turneduniversities into major players in the biotech business.
The Bayh-Dole Act was intended to promote the widespread use of federally funded inventions. Thesponsors of the legislation believed that permitting grantees to obtain patent rights and to conveyexclusive licenses for their inventions to private corporations would motivate investors to pick upwhere the government left off. This process, it was hoped, would produce commercial products from http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine discoveries that might otherwise languish in the halls of academe.
This goal is, of course, quite noble. But the law draws no distinction between inventions that leaddirectly to commercial products and fundamental advances that enable further scientific studies.
Universities have taken the opportunity to file patent applications on discoveries like new DNAsequences, protein structures and disease pathways—results that are primarily valuable because theyenable more investigation. Columbia University, for example, now holds a portfolio in which 50 percentof its licensed patents represent such research tools. And even when they do not seek patents,universities often try to preserve their expectations for profitable payoffs by imposing restrictions onthe dissemination of materials and reagents that might generate commercial value somewhere downthe line.
This frenzy of proprietary claiming has coincided with unprecedented levels of both public and privateinvestment in biopharmaceutical research and development—and magnificent progress in health care.
So for many people, it may be difficult to see that there is any problem. But in the long run thecurrent system may, paradoxically, hinder rather than accelerate biomedical research. Here we explorehow the current system emerged and what could be done to fix some of its problems.
Gold in Them There Halls
In 1979, U.S. universities were granted only 264 patents. But the statistics changed quickly after thepassage of the Bayh-Dole Act the following year. In 1997, for instance, U.S. universities received2,436 patents, a nearly 10-fold increase in 17 years. This rise was significantly greater than thetwofold increase in the overall rate of patenting during the same time period and also exceededgrowth in university research spending. Biomedical discoveries account for a large share of thesepatents, particularly in terms of licensing revenues.
The majority of this patented research was publicly funded. (Despite the increasingly intimate involvement of industry with universities, private companiesactually fund only a small percentage of university-based research in the lifesciences.) A prominent recent example involves embryonic stem cells. In the1990s, the National Institutes of Health (NIH) sponsored research at theUniversity of Wisconsin that succeeded in deriving such cells from rhesus monkeys and macaques. The NIH-sponsored research on primates yielded a broad patent for theWisconsin Alumni Research Foundation, the technology-transfer arm of the University of Wisconsin,which issued an exclusive license to Geron Corporation. This patent covers all lines of embryonic stemcells for primates, including humans (although for complicated reasons, Geron now holds rights to justthree types of differentiated human embryonic stem cells).
Clearly, NIH has a strong interest in ensuring the widespread dissemination of such broadly enabling research tools. But the Bayh-Dole Act significantly restrictswhat NIH can do. As long as the contractor is based in the United States,funding agencies may restrict patenting only in "exceptional circumstances,"when they determine that withholding title to the invention will better promotethe goals of the Act. The Bayh-Dole legislation also provides administrativeprocedures under which a grantee can challenge the determination of exceptional circumstances, witha right of appeal to the U.S. Claims Court. In addition, the agency must notify the CommerceDepartment, which has primary responsibility for administering this law, each time it claims exceptionalcircumstances, and it must provide an analysis justifying the action. If the Secretary of Commercedecides that "any individual determination or pattern of determinations is contrary to the policies andobjectives of [the Bayh-Dole Act]," he or she must advise the head of the agency and theAdministrator of the Office of Federal Procurement Policy and recommend corrective actions. Giventhese cumbersome procedures, it is perhaps not surprising that NIH declarations of exceptionalcircumstances have been extremely rare. Indeed, we are aware of only a single case.
http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine The Bayh-Dole Act also permits an agency to compel licensing of the patents that result from researchit had previously funded. But an agency can do so only if it determines that the university (or itsexclusive licensee) is not taking steps to achieve "practical application of the subject invention" or ifsuch licensing is necessary "to alleviate public health or safety needs or requirements for public usespecified by Federal regulations." Exercise of such rights is not subject to an overarching directive thatthe circumstances be "exceptional." Nonetheless, the Bayh-Dole Act seriously restricts the value of thisprovision by deferring such actions pending elaborate administrative proceedings and exhaustion ofcourt appeals. The administrative obstacles have proved sufficiently high that NIH has never onceexercised this option.
Out of Reach
Although the idea of private universities earning large sums of money from publicly sponsoredresearch may be troubling enough for many, the real problem with the Bayh-Dole Act is that it oftenputs such academic research advances out of the reach. Although one might imagine that patentholders don't enforce their patents for noncommercial uses, some have in fact been quite aggressive inthis regard, insisting that university investigators sign license agreements, especially when they seekto transfer materials covered by a patent rather than simply practicing a patented techniqueinconspicuously in their own labs. Given that patent law offers no significant exemption from liabilityfor experimental use and that the division between noncommercial and commercial research can beblurry, it is indeed foolhardy for academic scientists to rely on the forbearance of patent holders.
Thus some patents can stall scientific progress. This concern is particularly acute for claims to early-stage discoveries that open up entirely new fields. Suchpatents may be quite broad, permitting their owners to control a wide range ofsubsequent research. One reply to this argument is that profit-seeking owners ofpioneering patents will find it in their own best interest to disseminate theirdiscoveries to as many follow-on improvers as possible. History shows otherwise.
The Wright brothers, for example, refused to offer reasonable licensing terms forsome of their aeronautical innovations until compelled to do so by the government. One notable recentexample in the pharmaceutical industry is the controversy generated when DuPont imposed restrictionson academic investigators wishing to use its "oncomouse" technology, which DuPont controls under anexclusive license from Harvard University, the patentee.
Why would a company not want to license its technology as widely as possible? Isn't that how itmakes money? One reason is that issuing such licenses requires considerable time and effort. Giventhe imperfect information available to the parties involved, the disparate assessments of value to thetechnology and the danger that one side might misappropriate the research plans of the other oncethey are disclosed in the course of negotiations, the transaction costs associated with such bargainingare likely to be quite high. And these costs mount quickly when the basic research discoveriesnecessary for subsequent work are owned not by just one company, but by a number of differententities.
Concern about an "anticommons" or "property rights thicket" is quite pressing in contemporarybiomedical research, which often draws from many prior discoveries made by different scientists inuniversities and private companies. Exchanges of DNA sequences, laboratory animals, reagents anddata that were once shared freely are today subject to licenses, material-transfer and database-accessagreements. These arrangements have to be reviewed and negotiated before research may proceed.
A standard response to these fears is that market forces will motivate the emergence of patent poolsand other institutions for bundling intellectual property rights. But this prediction has not yet beenborne out. Indeed, when representatives of biopharmaceutical companies have seen the potential foran anticommons, they have reacted not by forming patent pools, but rather by strengthening the http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine The case of single nucleotide polymorphisms, or SNPs, provides an interesting example of this phenomenon. Collections of SNPs are found throughout thegenome and are a useful resource for scientists searching for genes involved inspecific diseases. These SNPs also promise to be useful in developing diagnosticand therapeutic products. In recent years, various biotechnology companies haveidentified and sought patents on large numbers of SNPs, provoking concern onthe part of both NIH and the pharmaceutical industry about the potential forbalkanization of intellectual property rights to this important resource.
Paradoxically, the pharmaceutical industry has enjoyed more latitude than NIH torespond to this threat by placing SNPs in the public domain. Pharmaceuticalcompanies have joined together with the nonprofit Wellcome Trust (a U.K.-basednongovernmental partner in the Human Genome Project, which is not bound bythe Bayh-Dole Act) in a consortium to sponsor an SNP-identification effort with explicit instructions to put the information in the public domain. The SNP Consortium has candidlyembraced a goal of defeating patent claims to SNPs. The willingness of private companies in a patent-sensitive industry to spend money to enhance the public domain is indeed curious. We think it ispowerful evidence of a perception in industry that claims to intellectual property rights for fundamentaldiscoveries can create significant barriers to subsequent research and product development.
Possible Fixes
One solution might involve changing the patent laws to restrict patents on fundamental research.
Congress or the courts might, for example, reinvigorate the "products of nature" limitation on patenteligibility so as to exclude discoveries of DNA sequences, proteins and biochemical mechanisms frompatent protection. Lawmakers and judges could also fortify the utility standard to limit the patenting ofresearch tools. Another much-discussed idea is to provide an exemption from infringement liability forresearch, particularly noncommercial research. Although such legal adjustments are worth considering,it is difficult to calibrate these changes accurately, and the consequences of overdoing it could begrave.
Patents clearly matter to the biopharmaceutical industry, and undue restrictions on them may indeeddeter private investment. Although it is possible that these companies—particularly those that makeend products—would benefit in the long term from limits on certain patents, many of these businessescontinue to insist that they need patents on their research to raise capital. Given that privateinvestment in biomedical research and development today exceeds public funding, the strong belief ofinvestors that patents are essential urges caution in changing the underlying legal rules.
When research is publicly sponsored, however, the argument for strong patent rights loses much of itsforce. The Bayh-Dole Act does not presume that patents are necessary to motivate grantees toperform research but rather that patents will promote subsequent utilization and development ofinventions. The reasoning that lurks behind this presumption is that patents and exclusive licenses areessential to attract the necessary private investment. Whatever the merits of this presumption forpatents on final products such as new drugs, it makes little sense for patents on broadly enablinginformation and techniques that are ready for dissemination to scientists in both public and privateinstitutions, advances that can be put to use in the laboratory right away, without any furtherinvestment.
A classic example is the Cohen-Boyer method for combining DNA from different organisms. Manyobservers attribute the rapid progress of the biotechnology industry to the fact that this technologywas made widely available rather than licensed exclusively to a single company. Although this pre–Bayh-Dole technology was, in fact, patented, it was offered nonexclusively and cheaply to encouragecompanies to purchase licenses rather than to challenge the patents. These nonexclusive licenses http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine generated some $300 million for the universities that owned the patents, but it is difficult to see howthey did anything to enhance product profitability or otherwise motivate subsequent research anddevelopment. If anything, the patent royalties imposed a modest tax on product development.
A greater concern is that the Bayh-Dole Act does little to ensure that a university will license suchpatents nonexclusively. To the contrary, Congress was careful in the terms of the Bayh-Dole Act andsubsequent legislation to give universities discretion to grant exclusive licenses, which may be morefinancially attractive than nonexclusive licensing. Exclusive licenses typically command higher royalties,and companies holding exclusive licenses are more willing to reimburse for patent costs and to provideadditional grant funding to the inventor. Indeed, the information available suggests that the majorityof university licenses to startups and small businesses are exclusive.
But it is not clear that such exclusive licenses are necessary to achieve the aims of the Bayh-Dole Act.
A recent case in which patenting and subsequent exclusive licensing were not necessary for productdevelopment involves federally funded research that identified the cell-signalling pathway for NF-kB(nuclear factor kappa B), which regulates genes that function during inflammation, cell proliferationand programmed cell death. This research (which scientists at Harvard, the Massachusetts Institute ofTechnology and the Whitehead Institute for Biomedical Research carried out in the 1980s) led to abroad patent claiming all drugs that work by inhibiting NF-kb cell signaling. Because the NF-kBpathway has been implicated in diseases ranging from cancer and osteoporosis to atherosclerosis andrheumatoid arthritis, the patent—which was issued just last year—may cover drug treatments for all ofthese diseases. Indeed, these academic institutions, together with their exclusive licensee, AriadPharmaceuticals, are now suing Eli Lilly & Co., claiming that Lilly's osteoporosis drug Evista and itssepsis drug Xigris work in a manner that infringes the NF-kB patent. Ariad has also sent letters tosome 50 other companies with products that work via the NF-kB pathway, demanding royalties onpresent or future product sales. Obviously, the companies that are now being asked to pay royaltiesdid not need an exclusive license from Harvard, MIT and Whitehead to motivate them to pursueproduct development; the prospect of obtaining patents on their own end products was sufficient. Inthis case, as in many others, pioneering patents issued to academic institutions only thwart innovation.
For many discoveries emerging from government-sponsored research, the benefits of patenting are lowrelative to its costs. But some discoveries, including some important research tools and enablingtechnologies generated in the course of publicly sponsored research, undoubtedly require substantialcommercial investment to become reliably mass-produced for widespread distribution. For example,technologies and machines for DNA sequencing and analysis, initially developed in academiclaboratories, required substantial follow-up investment by private companies to turn them into reliableand commercially available equipment. Patents and exclusive licenses may be crucial to motivate thissort of investment.
The policy challenge, then, is to devise a system that distinguishes cases in which proprietary claimsmake sense from cases in which they do not. The complexity of biomedical research makes this aformidable task, and the public interest in getting these determinations right demands assigning thisresponsibility to the most qualified body. Ideally, decisions about the dividing line between the publicdomain and private property should be made by institutions that are in a position to appreciate thetensions between widespread access and preservation of commercial incentives without being undulyswayed by motivations that diverge from the overall public interest.
Preserving the Commons
So where should these decisions be made? On first examination, one might think that universities,which reap the rewards of the proprietary restrictions they impose on others but also pay the costs ofrestrictions that others impose on them, might be interested in maintaining at least some research inthe public domain. The problem is that the costs to a university are largely borne by its scientists whocannot get prompt access to the proprietary technologies they seek, whereas the gains from licensing http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine revenues are much more salient to its technology-transfer officers, who are charged with generatingrevenue. So coming to a consensus might be difficult.
Even when universities recognize that the larger academic community might be better off if theyshared their research tools more freely, they face a serious problem: ensuring collective action. Solong as other institutions are staking out claims, no university is likely to abstain from asserting itsown rights. Appeal to the traditions of open science may not be sufficient, especially given that thescientists who hold those values don't usually make decisions regarding assertions of proprietary rights.
Left to their own devices, universities probably cannot mount the sustained community effort neededto preserve the research commons. But, interestingly, on a number of occasions NIH has been able touse sternly worded appeals to the norms of open science to convince academic institutions to keepbasic research in the public domain. For example, in 1996 leaders of the National Human GenomeResearch Institute (NHGRI), together with the Wellcome Trust and academic researchers at the majorhuman genome mapping centers, resolved that "all human genomic DNA sequence information,generated by centers funded for large-scale human sequencing, should be freely available and in thepublic domain in order to encourage research and development and to maximize its benefit tosociety." The NHGRI followed up with a policy statement making "rapid release of data into publicdatabases" a condition for grants for large-scale human genome sequencing. The NIH could not,however, go so far as to forbid its grantees from filing patent applications without relying on thecumbersome "exceptional circumstances" clause of the Bayh-Dole Act. Rather than take this step, NIHdeclared that, as a matter of doctrine and policy, raw human genomic DNA sequence informationshould not be considered patentable. The statement also warned that NHGRI would monitor whethergrantees were patenting "large blocks of primary human genomic DNA sequence" and threatened toinvoke the "exceptional circumstances" limitation in future grants. In the specific context of large-scalegenome mapping, universities were willing to embrace this policy.
Administrators at NIH undertook a similar strategy for SNPs. Before the SNP Consortium steppedforward to place this information in the public domain, NIH had decided to allocate public funds forSNP identification. Once again, NIH refrained from invoking the "exceptional circumstances" provisionof the Bayh-Dole Act. Instead, in its request for applications for SNP-related grants, the agencystressed the importance of making information about SNPs readily available to the researchcommunity and asked grant applicants to specify their plans for sharing data, materials and software.
The NIH also warned that it reserved the right to monitor their patenting activity.
The efforts of NIH to constrain its grantees in pursuing intellectual property rights have not beenlimited to genome projects. A more general statement of "Principles and Guidelines for Sharing ofBiomedical Research Resources," adopted by NIH in December 1999, also attempts to guide NIHgrantees regarding proprietary rights. These principles state that "the use of patents and exclusivelicenses is not the only, nor in some cases the most appropriate, means of implementing the [Bayh-Dole] Act. Where the subject invention is useful primarily as a research tool, inappropriate licensingpractices are likely to thwart rather than promote utilization, commercialization, and publicavailability." What NIH has sought to achieve through these various statements is broadly consistent with the intentof the Bayh-Dole legislation "to promote the utilization of inventions arising from federally supportedresearch or development." Arguably, however, at least with respect to patentable inventions, NIH hasacted outside the scope of its authority, leaving itself vulnerable to legal challenge.
Sound Footing
The time is ripe to fine-tune the Bayh-Dole Act to give funding agencies more latitude in guiding thepatenting and licensing activities of their grantees. We propose two modest reforms that would givethese agencies, which have the proper combination of knowledge and incentives, somewhat greater http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine discretion to determine when publicly funded discoveries should be put in the public domain.
First, the circumstances in which an agency may prevent its contractor from retaining title to aninvention should be liberalized. The current language of the law creates a clear presumption that anagency should exercise this power very infrequently. That should be changed. Once the "exceptionalcircumstances" language is deleted, the law could be more freely applied to achieve the goal ofpromoting widespread dissemination and use of research results. The process for review of"exceptional circumstance" determinations should be streamlined as well, with provisions for researchto proceed while examination of the decision runs its course.
Second, Congress should modify the requirement that a funding agency's authority to compel licensingof university patents be held in abeyance until all court appeals are exhausted. By the same token,however, an agency should not be given authority to act without some provision for judicial review.
Unlike a determination to restrict patenting, a subsequent exercise of the right to compel licensingdisturbs settled expectations. If business planning is too readily upset, industry could become wary ofinvesting in university-based technology.
It might be argued that restoring greater authority to agencies would return us to the unhappyposition that motivated Congress to pass the Bayh-Dole Act in the first place. This danger appearsquite small. In the intervening 23 years, NIH has embraced patenting and technology transfer infurtherance of its mission of improving public health. Moreover, our proposal to give agencies greaterauthority would not overturn the general presumption in favor of allowing government contractors topatent inventions. It would simply permit agencies to decide that patenting is not warranted inparticular cases, while streamlining procedures for making and reviewing these decisions. Givinggreater discretion to agencies would also correct a dangerous oversimplification of how best to achievethe important policies underlying the Bayh-Dole Act, by recognizing that patenting and exclusivelicensing are not always the best way to go.
Bibliography
Rai, A. K. 1999. Regulating scientific research: intellectual property rights and the norms ofscience. Northwestern University Law Review 94:77-152.
Campbell, E. G., B. R. Clarridge, M. Gokhale, L. Birenbaum, S. Hilgartner, N. A. Holtzman and D.
Blumenthal. 2002. Data withholding in academic genetics: evidence from a national survey.
Journal of the American Medical Association 287:473-480.
Eisenberg, R. S. 1996. Public research and private development: patents and technology transferin government-sponsored research. Virginia Law Review 82:1663-1683.
Gelijns, A. C., and S. O. Thier. 2002. Medical innovation and institutional independence:rethinking university-industry connections. Journal of the American Medical Association 287:72-77.
Heller, M. A., and R. S. Eisenberg. 1998. Can patents deter innovation? The anticommons inbiomedical research. Science 280: 698-701.
Mowery, D. C., R. R. Nelson, B. N. Sampat and A. A. Ziedonis. 2001. The growth of patentingand licensing by U.S. universities: an assessment of the effects of the Bayh-Dole Act of 1980.
Research Policy 30:99-119.
Mowery, D. C., and A. A. Ziedonis. 2002. Academic patent quality and quantity before and afterthe Bayh-Dole Act in the United States. Research Policy 31:399-418.
Article Tools
http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes American Scientist Online - Bayh-Dole Reform and the Progress of Biomedicine Of Possible Interest
FEATURE ARTICLE: Is Combustion of Plastics Desirable?
INTERVIEW: Lester Brown
BOOK REVIEW: A Plan for Preserving Civilization
BOOK REVIEW: Fundamentally Flawed
COLUMN: The Biotech Future
Related Internet Resources
NIH Report on Research Tools
NHGRI Policy on Intellectual Property of Human Genomic Sequence
1999 OncoMouse Agreement between the NIH and DuPont 2001 Agreement between the NIH and WiCell Research Institute (PDF)2001 NIH Plan to Ensure Taxpayers' Interests are Protected Related Sigma Xi Links
Forum: International Cooperation in Science and Technology
Pan-Organizational Summit on the U.S. Science Engineering Workforce
Global Change PublicationsScience Advocacy Programs Sigma Xi, The Scientific Research Society http://www.americanscientist.org/template/AssetDetail/assetid/16219?fulltext=true&print=yes

Source: ftp://ftp.cs.cmu.edu/usr/rwh/www/home/LeftOut/jan-30-07/bayh-dole.pdf