Appendix FCOMMISSIONED PAPER Sharing Knowledge for Global Health

Anthony D. So, MD, MPA, and Evan Stewart, BA*

The U.S. Commitment to Global Health identifies technological innovation and diffusion as the main drivers for improving health for all people by reducing avoidable disease, disabilities, and deaths. The sharing of knowledge is central to that vision, but involves far more than making a journal article open access, posting a database publicly to the web, or licensing a technology. These are all important building blocks to transferring technology effectively.

Sharing, as opposed to transferring, implies a two-way street. This is not to say that such exchanges are not asymmetric. Such exchanges slope along the steep gradients of disparities that separate industrialized and developing countries. Though only one-fifth of the world’s population, the developed world is home to over two-thirds of the world’s researchers, commands three-quarters of the gross expenditure on R&D, and originates over ninety percent of the patents granted by the patent offices in Europe, the United States, and Japan. The United States alone generates nearly twice the number of scientific publications (32.7% of the world total) than the whole of the developing world (17.3%).1

These asymmetries, of course, run in the other direction when examining where the global burden of disease falls. Increasingly, biomedicine is turning to the growing pools of talent in the developing world. The conduct of clinical trials is burgeoning in the developing world—no doubt lured, in part, by reports that a top-notch academic center in India charges a tenth per case report of what a second-tier medical center in the United States would in mounting a clinical trial.2 Pharmaceutical firms in the developing world may face different opportunity costs than large multinational corporations, and this may lead to gap-filling R&D investments, such as in more cost-effective processes for producing drugs. Shin Poong, a Korean firm that significantly lowered the costs of producing praziquantel, a drug to treat schistosomiasis, is a case in point.3 Some of these asymmetries are eroding away. From 2000 to 2006, the average annual growth rate in the number of patent filings originating from countries such as China and India outstripped that of all reported countries in Europe and North America.4

But the vibrancy of the scientific enterprise is only captured, in part, by traditional measures of innovation. Scientists trained, publications, patent filings, and revenues from health technologies highlight the disparities, but not the potential of research collaboration. Combination drugs effective for treating malaria may be produced by Northern pharmaceutical companies, but a core component is artemisinin, a Chinese traditional medicine. Without the collaboration of research centers in countries where SARS was endemic, the race to contain the threatening pandemic would have been crippled. Without the wild virus samples of avian flu from developing countries, steps to preparing a vaccine stockpile would slow. The interdependency of global health is clear from such examples. But these examples also underscore the importance of sharing knowledge for global health and of shaping effectively the enabling environment for doing so.

What should be the focus of the U.S. commitment to global health—leveling the slope or ensuring the flow of knowledge on that two-way street? Perhaps solutions need to address both. In a globalizing world, both knowledge and the human resources capable of applying that knowledge flow readily across borders. If the process of sharing knowledge only lures away the most talented to U.S. research laboratories, would it only exacerbate the brain drain from developing countries? If those from developing countries train here in the United States, will they return to settings where they can apply those skills? If the governance of product development partnerships represents the voices of donors but not those they purport to serve in developing countries, will the fruits of their work be effectively disseminated? As a recent study observed, U.S.-based companies increasingly have sponsored clinical trials in developing countries, but of the current Phase III clinical trials in these settings, not one in their sample focused on a disease endemic largely in developing nations.5 Is it really ethical or sustainable to mount clinical trials in developing countries that yield new treatments, but to fail to make these therapies affordable or more targeted to the public health needs of the populations in which they were tested?

Sharing knowledge in the context of the U.S. commitment to global health often emphasizes the North-South axis of collaboration. From the vantage point of the United States, the potential for collaboration and capacity building to improve global health is greatest along this axis. Of course, there are lessons from years of North-North collaborations across countries that might cross-apply. Not to overlook other axes though, South-South collaborations deserve particular note. The INDEPTH Network consists of 34 demographic surveillance sites in 18 countries, all in the developing world.6 Facilitating cross-site studies of longitudinal health and social studies in resource-limited settings, the network draws support from a range of Northern donors, including private foundations such as Gates, Rockefeller, the Wellcome Trust, and Hewlett; bilateral aid agencies such as CIDA, DFID, and Sida; and government research agencies such as IDRC and the U.S. National Institutes of Health (NIH). With its Secretariat based in Accra, Ghana, the network’s governance remains largely in the hands of researchers from developing countries. Partnerships though are welcomed with Northern institutions, from product development partnerships such as the Malaria Vaccine Initiative to universities such as the London School of Hygiene and Tropical Medicine and the Swiss Tropical Institute.

The rise of modern medicine has introduced another important dimension of knowledge sharing—the bidirectional exchange between industry and publicly funded institutions such as universities and government laboratories. Disproportionate to the level of corporate funding, the norms governing this exchange have reshaped the way universities share their inventions. Some have suggested that the commercialization of university research has corrupted the mission of higher education.7,8 Many of these concerns trace to the nature of the agreements struck between universities and corporations. These contracts affect the publishing of research, the sharing of data and research tools, and the licensing of patented inventions. Corporations bring complementary expertise, an ability to scale up products for delivery, and additional research resources. While such collaborations may bring value to university research efforts, the conditions under which they operate deserve greater scrutiny and transparency. Society has relied on the academy to contribute to knowledge in the public domain, to maintain the independence of inquiry with safeguards against conflict of interest, and to engage in “blue-sky” research and high-risk experimentation.

The market has failed to deliver diagnostics, drugs, and vaccines that meet the disproportionate burden of disease afflicting those in developing countries. Public-private partnerships have emerged over the past decade to fill this gap. Using public sector monies, product development partnerships have embarked on drug discovery programs for neglected diseases. Half of these partnerships involved multinational corporations that conducted these projects on a “no profit-no loss” basis. Of note though, the other half of these projects were conducted by small firms doing so on a commercial basis.9 The opportunity costs for these smaller firms may be different. There can be an important North-South dimension to these collaborations as well. A recent survey found that over half of private sector firms in health biotechnology in developing countries had ongoing collaborations with partners in developed countries.10

STEPS TO SHARING KNOWLEDGE FOR GLOBAL HEALTH

Sharing knowledge for global health involves generating knowledge relevant to the context of low- and middle-income countries, effectively transferring such knowledge and technologies to these settings, and ensuring that its intended beneficiaries can apply it on a sustained basis. Each of these steps presents its own set of challenges, but also affords new opportunities.

Knowledge must either be relevant or adapted to the context of low- and middle-income countries. Some of this knowledge will be relevant because the health problems are shared ones between North and South. Often thought to be diseases of affluence, noncommunicable diseases, in fact, comprise a growing share and already account for nearly half of the burden of disease in developing countries.11 Put in perspective, cardiovascular diseases, cancers, and diabetes comprise 16% of the burden of disease in low- and middle-income countries. By comparison, malaria is responsible for 4% of the disability-adjusted life years lost in these countries.12

As the 10/90 gap suggests though, the investment in global health R&D does not prioritize efforts focused on the burden of disease that disproportionately afflicts those in low- and middle-income countries. With the paying market being relatively small, some treatments rely on the spillovers from dual markets. The availability of eflornithine, the “resurrection drug” for treating sleeping sickness, has at times depended on its dual use as a treatment for the removal of facial hair in women. For those engaged in biodefense research, substantive review usually focuses on the dual use of such technologies for biodefense against emerging infections as well as for potentially nefarious purposes.13 Some of this research, including the platform technologies applied, might be evaluated for a third use—humanitarian applications to neglected diseases in developing countries. Not relying on the serendipity of finding incidental applications for neglected diseases, government and philanthropic funders have also invested in product development partnerships.

Sharing knowledge requires an enabling environment. The investment required to transfer information is a measure of the “stickiness” of that information.14 Stickiness is a function of the attributes of the information itself as well as that of the information seekers and providers. Intellectual property rights might make such knowledge costly to acquire while information technology has changed the speed and marginal cost of disseminating knowledge. Sometimes the skills are local to where that knowledge is being used. For example, laboratory apprenticeships may afford the firsthand experience necessary for performing certain procedures.

Several factors affect the sharing of knowledge: (1) the nature of the knowledge to be shared; (2) the norms for scientific exchange; and (3) its role in the innovation process. Today’s science has many ways of codifying knowledge, from study methods described in journal articles to patents disclosed. Tacit knowledge, on the other hand, is not well codified. A technology new to developing country firms—such as conjugation technology for vaccine production—may not easily transfer without technical assistance. Norms over the ownership of knowledge also influence the sharing of knowledge. These norms are rooted in statutes and regulations such as the Bayh-Dole Act, prevailing practices among research institutions, and guidance provided by funding agencies as well as competition among scientists.

The sharing of knowledge matters most if innovation and scientific progress are cumulative. By cumulative innovation, one might envision several types of arrangements of research inputs and outputs (see Figure F-1).15 A single innovation might spawn multiple, second-generation innovations. For example, a receptor target might lead to several promising new drugs. Alternatively a second-generation output might require the input of multiple first-generation inputs. Some of these inputs may eventually be incorporated into the second-generation product, but other needed inputs—research tools—will not be. Finally, the process of innovation may be a quality ladder, where successively better products build on the model of the previous one. Process innovations of drugs can lower the marginal cost of production, extend its shelf life outside the cold chain, or improve its bioavailability. Each pattern of cumulative innovation responds differently to the ways in which knowledge is shared or inventions are licensed. For example, a product patent on a drug effectively may block others who might otherwise pursue process innovations in the manufacture of that drug.

Those who benefit from this sharing of knowledge must have the absorptive capacity to apply and sustain its use. The transfer of technology depends on the absorptive capacity of the setting where it would be used. Technology has both hardware and software aspects. Hardware is the tool as embodied as a physical object while software is the information base for the tool.16 The capital costs for purchasing hardware may be out of reach, but so might be the maintenance costs. Variable costs such as reagents for diagnostic tests can be prohibitively expensive. The software side consists not just of the knowledge to use the tool, but also may require the human resource expertise to apply it.

ACCESS TO THE BUILDING BLOCKS FOR RESEARCH

From bench to bedside, the value chain of R&D consists of inputs and outputs at every stage, each dependent on the sharing of knowledge. Three stages in this value chain warrant closer scrutiny because decisions at these points significantly shape what knowledge is shared within the scientific community. These building blocks for research include access to scientific publications, the norms for data and material sharing, and patenting and licensing practices. Characterizing the obstacles and opportunities at each stage can help point the way to solution paths that lower the barriers to sharing knowledge and improve the scientific community’s ability to respond to the challenges of global health.

FIGURE F-1. A) Innovation may occur in several ways.

FIGURE F-1

A) Innovation may occur in several ways. One input may lead to a higher quality output, with each generation of innovation bringing a successively better product. Alternatively, a single input may spawn several outputs, as one target receptor may lead to several new drugs. Finally, several inputs may be required to produce one output; these inputs may be innovations themselves or simply research tools (adapted from Scotchmer, 2004).15 B) Tiering may segment the marketplace between a paying market and a resource-limited one that may receive a discounted price or other preferential access. C) Inputs may also be pooled, thereby reducing transaction costs to innovation and more readily enabling socially useful bundles. Such pooling—particularly when strategically done by the public and/or philanthropic sectors—may be structured to influence positively the norms and the licensing by which other inputs are also made available for innovation. Such an arrangement characterizes a technology trust.

Access to Scientific Publications

The challenges to sharing knowledge through scientific publication come both from the supply and the demand side. On the supply side, studies suggest that industry funding may not only occasionally introduce potential bias into the conduct of research, but also possible delays in its publication. Of those responding to a survey of life science faculties at universities receiving the most NIH funding, nearly a third of the investigators that benefited from corporate research-related gifts indicated that their industry sponsor wanted pre-publication review of journal articles resulting from the gift.17 A majority of the contracts struck between these scientists and life science companies also mandated a six-month period of confidentiality to give time for patenting of resulting inventions.18 By contrast, the NIH has provided guidance that such delays should not exceed a 30- to 60-day window.

On the demand side, subscription prices to journals may place access to some research out of reach. This problem not only faces some institutions in the developing world, but also among patients in the developed world. For many patients, especially those with rare diseases, the high cost of accessing individual journal articles can pose an obstacle to learning about one’s condition or treatment options. As a result, patient advocacy groups have recently joined the call on the U.S. government to embrace open access policies.19,20

To ensure greater access to scientific publications, several strategies have been deployed. One has involved tiered pricing, and the other, the pooling of published research in open-access journals or repositories. Particularly in developing countries, mailing hard copies of journals would be prohibitively costly. With the advent of the Internet, however, much of this access can now be provided electronically.

Launched in January 2002, the WHO-led Health InterNetwork Access to Research Initiative (HINARI) seeks to provide tiered access to more than 6,200 major journals in biomedicine and related social sciences. In collaboration with participating publishers, HINARI divides low- and middle-income countries into two groups: countries with a GNI per capita from US$1250-3500/year whose institutions can receive access for $1000/year and those below that cutpoint whose institutions receive free access via an online research portal.21 The publishing company Elsevier, whose journals are made available through HINARI, claimed in 2006 that the initiative contributed to raising the rates of publication by researchers in the 105 HINARI-eligible countries. In their analysis, researchers in HINARI countries increased their rates of publication by 63% while those in non-HINARI nations saw only a 38% increase.22 However, some problems have surfaced in gaining online access to these journals. In order to be eligible for HINARI access, researchers in developing nations must have an institutional affiliation, prohibiting nonaffiliated scientists, doctors, and government officials from accessing HINARI articles.23 Even for those with the correct institutional affiliation, investigators from a Peruvian university noted in 2007 that many of the highest impact journals were not available there.24 Those journals that were accessible via HINARI were often either open-access journals or those which already provided free access to low-income countries.

Across disciplines ranging from electrical engineering to mathematics, the free, online access of journal articles corresponded to higher mean citation rates.25 Several studies suggest that open access articles have a higher citation rate than closed-access articles.26,27This held true even when comparing open-access articles compared to non-open-access articles in the same journal.28 Importantly, the impact of open-access publication on citations in journal publications was twice as strong in the developing world.29

Open access can take several forms. By retaining copyright or nonexclusive license, authors can self-archive their work, oftentimes on their own websites or in a university repository. This is also known as the “green” road. In early 2008, Harvard University adopted its own open-access mandate through which members of the Faculty of Arts and Sciences will submit electronic copies of all completed articles to an institutional repository that will eventually be accessible worldwide via the Internet.30 Faculty members may opt out of the system if they choose, but it is expected that most will grant a nonexclusive license to the university to make use of their work. The approach of an institutional open-access repository has also spread: Harvard Law School and Harvard’s Kennedy School of Government recently adopted their own open-access initiatives as have the Stanford University School of Education, Boston University, and the Massachusetts Institute of Technology.31,32,33,34

Breaking with the approach to supporting journal access through subscriptions, open-access journals have offered an alternative model to scientific publishing, also known as the “gold” road. Open-access journals raise revenues from a variety of sources—endowments, institutional subsidies, membership dues, fundraising, advertising, or upfront submission or publication fees—or just depend on voluntarism. Of note, most open-access journals do not charge any publication fees.35 Open-access journals make published articles more broadly available online without subscriber fees. In so doing, open-access journals enable wider distribution of the research published in these outlets, and at the same time, the copyright licensing of these works allow greater potential of “remix.” For example, if a developing country research institution sought to pull together a compendium of key articles on schistosomiasis and to share such a resource with sister institutions, the transaction costs of assembling an open-access collection of journal articles are far lower than doing so with non-open-access articles, where reprint rights would have to be negotiated with each journal holding the copyright.

Open-access publishing has benefited from Creative Commons licensing. Such licensing enables artists, writers, and researchers to lift voluntarily some or all of the copyright restrictions upon their work. The family of Creative Commons licenses allows for different permutations of the conditions under which the work might be distributed, displayed, performed, or become the basis of a derivative work. These conditions may require attribution, limit subsequent use to noncommercial purposes, not allow derivative works, or allow sharing under condition that derivative works carry the same licensing.

In the biomedical sciences, much research is funded by governments, and given this support, the public understandably expects access to the findings from such research. The NIH estimates that 80,000 publications grew out of NIH-supported research in 2003.36 Initially making a nonbinding request of its researchers, NIH asked that all publications resulting, in whole or in part, from its funding to be deposited in PubMed Central, a publicly accessible archive of scientific publications, within 12 months after the study’s publication.37 However, the yield from voluntary compliance with this policy was very low: fewer than 5% of NIH-funded researchers submitted their articles.38 The failure of this policy prompted U.S. congressional action that mandated it as a requirement of NIH funding beginning in April 2008.39 The NIH Public Access Policy requires investigators to submit final, peer-reviewed journal manuscripts arising from NIH funding to PubMed Central upon acceptance for publication. Such papers must be available to the public through PubMed Central no later than 12 months after publication. Taking a green path, this approach mandates deposit of government funded research in an online archive broadly available to the public. By allowing grantees to use NIH funding for publication fees though, the NIH also supports, in part, the gold road.

Several prominent medical research funders have made open access a condition of grant support. The European Research Council (ERC), a funding body set up by the European Union (EU) to promote research in the region, has also put forward an open-access policy requiring its grantees to post all publications to a research repository within six months of publication.40 This marked the first EU-wide open-access policy and ERC has stated that it has interest in shortening the six-month window period in the future.41,42 The Wellcome Trust requires submission of scientific publications resulting from its grants into U.K. PubMed Central within six months of the publication date and even provides funding for the upfront fees associated with publishing in such outlets.43,44 Grantees of the Howard Hughes Medical Institute also face a similar requirement to deposit publications in PubMed within six months of the publication date.45 By contrast, NIH’s Public Access Policy remains at 12 months, twice the embargo period accepted by other leading funding agencies.

Access to Research Data and Materials

The sharing of research data and materials enables the scientific community to confirm study findings and also to build upon the work of others. Access to these building blocks of research, however, may also be encumbered for reasons similar to those encountered over scientific publications. The difference is that access to data and materials enriches immensely the pursuit of new hypotheses that derive or go substantially beyond its original research use.

Competing public policy concerns set some limits on the sharing of research data and materials. For example, some data may risk the personal privacy of human subjects, and the disclosure of other data may compromise the confidentiality of privileged proprietary information. Unlike the electronic distribution of journal articles or data, the marginal cost of disseminating research materials may not be negligible, and these transaction costs also may pose barriers to sharing. Dual use of technologies have the potential both to advance scientific knowledge and to pose threats to public health or the environment, and such research activities as well as resulting data and materials require governmental oversight.46 However, denying data access not only imposes additional costs and barriers to research along these lines, but also can place patients at risk of redundant or unnecessary clinical trials.

Slow responses to material transfer requests resulted in project delays of greater than a month among one out of six biomedical researchers surveyed in universities, government or nonprofit institutions.47 Noncompliance with these material transfer requests resulted in 1 out of 14 scientists giving up a line of research on at least one of their projects each year. While noncompliance with these requests were not reported to relate to the patent status of the requested material, key reasons given for noncompliance included the costs and effort involved in providing the sample and protecting the ability to publish. Negotiating MTAs with industry often came with conditions, such as reach-through claims, royalties, and publication restrictions. This was particularly common for requests for drugs.

The role of government in facilitating access to data and research materials is bounded, in part, by statute and regulations. For example, the U.S. Copyright Act of 1976 prevents the federal government from claiming copyright protection of its publications, and OMB Circular A-130 mandates that government-produced data should be made available at the marginal cost of disseminating it. OMB Circular A-76 prohibits the government from entering into direct competition with the private sector in providing information products and services. Tensions exist between treating scientific data as a public good as opposed to a private one, and there are important implications for the research commons.48

As with publications, open access may also multiply the impact of research data. For example, in a 2007 study of 85 cancer microarray clinical trial publications, the public sharing of available data contributed to a 69% increase in citations.49 While half the trials in the study made their data publicly available, they comprised 85% of the total citations.

As suggested by findings in the genetics research community, there are the familiar reasons for denying access to data and research materials. When making requests for information, data, or materials related to published research, nearly half of geneticists reported that at least one of their requests had been declined over the previous three years.50 Consequently, investigators said they could not confirm research that had been published. Among the reasons most frequently given for denying such requests, geneticists cited the high costs of producing materials or information, the need to protect their own or their colleagues’ ability to publish, and the commercial value of the data or material.

In the setting of emerging infectious diseases, the need for rapid and freer exchange of information and materials has become most clearly evident. The WHO’s Global Influenza Surveillance Network played a key role in linking the world’s leading laboratories and experts with real-time information during the SARS outbreak in 2003.51 In the race to identify the coronavirus as the cause of SARS, 11 laboratories recruited by the WHO regularly and voluntarily shared samples of the unknown virus and held conference calls to discuss their results.52 Without this level of collaboration and sharing, the transmission of SARS might not have been halted within four months. For other diseases that might not unfold as infectious disease outbreaks, would not freer exchange norms also help speed the race to a cure?

Funding agencies again have played an important role in setting norms for sharing data and materials. Providing guidance to its grantees in 2003, the NIH requires applicants for grants greater than $500,000 to provide a plan for “timely release and sharing of final research data from NIH-supported studies for use by other researchers.”53 The ERC requires that “primary data” such as nucleotide or protein sequences or epidemiological data must be submitted to a database within six months.54

Led by the Wellcome Trust and the NIH, leading sequence centers involved in the Human Genome Project pledged to deposit completed gene sequences of every 1,000 base pairs within 24 hours of completion into a publicly available database, GenBank. Called the “Bermuda Rules,” these rules were created to prevent the patenting of DNA sequences through defensive publishing.55 Providing further incentive to follow the Bermuda Rules, the NIH subsequently suggested that the patenting of work emerging from the publicly funded Human Genome Project would negatively impact the likelihood of receiving future grants.56 Data sharing has also been supported by other initiatives since the adoption of the Bermuda Rules—by the Merck Gene Index,57 the International Nucleotide Sequence Database Collaboration,58 and the Worldwide Protein Data Bank among others.59

Traditionally, the sharing of data and materials involves both informal and formal norms. Informally researchers sometimes bypass negotiation over material transfer agreements (MTAs), but such practices may place the institution at some risks that would otherwise be lessened by use of MTAs. Informal transfers of materials among investigators circumvent institutional management of the intellectual property and give advantage to some researchers better connected than others.60 Increasingly though, informal sharing has given way to formal agreements on data or material sharing that cover concerns such as attribution, protection of patient confidentiality, the right to publish resulting research findings, and intellectual property rights (IPRs).

Various groups have sought to lower the costs of such transactions. The first strategy involves harmonizing the formal agreement form used among institutions. The Uniform Biological Material Transfer Agreement (UBMTA) offers a standard approach for transferring materials for noncommercial, research pur poses, and the simple letter agreement (SLA), for transferring nonproprietary biological materials among public and nonprofit research institutions.61 However, challenges remain, particularly in striking such MTA agreements between academia and commercial entities. Science Commons has more recently elaborated an MTA with modular contract options for transfers between academia and industry.62

A second strategy is to lower the transaction costs at the level of the organization or even a research consortium involving multiple institutions. Research consortia may also build in preferential arrangements for sharing research materials among participating institutions. Of note, the original NIH guidance suggests adoption of the UBMTA at the organizational level, and nearly 350 institutions have signed the Master Agreement pledging to accept this standard form without modification when their scientists send materials to other nonprofit or public institutions.63 In guidance to its grantees, the NIH suggested using the SLA as a means to transfer unpatented materials arising out of its funded research. It also asked that funded investigators use terms no more restrictive than those of the SLA when transferring materials to other NIH grant recipients.64 This approach carries the promise of creating a limited public domain among these institutions. However, many of these signatories have, in practice, substituted their own agreement forms in place of the UBMTA.65 In so doing, their practices create a collective action problem, where an individual university would have little motivation to forego what it might gain from a more restrictive and perhaps more remunerative MTA approach.

A third strategy is to create institutions specifically dedicated to the sharing of these data or materials. This is not a new strategy. The American Tissue Culture Collection, a nonprofit bioresource center, has provided a depository for biological materials since 1949 and now contains over 20,000 specimens. Lowering the transaction costs of securing reagents for research on HIV and other retroviruses, the NIH’s AIDS Research & Reference Reagent Program has grown to over 8,560 reagents since 1988, distributed over 11,000 reagents last year, and has participating scientists in 65 countries.66

As seen through the efforts of the Broad Institute, the sharing of data can come under the aegis of various sponsors. Among them, the Broad Institute and the pharmaceutical company Novartis have collaborated to share freely genetic data about diabetes online as part of the Diabetes Genetic Initiative.67 In addition to this public-private partnership, Broad has partnered with a disease-based foundation in order to create the Multiple Myeloma Genomics Portal, which publishes the sequence of the myeloma genome,68 and with several other research teams to create the Tuberculosis Database Project.69 The Multiple Myeloma Genomics Portal prohibits patenting of any DNA sequences discovered, and all data must be posted to a public site upon completion of the analysis. The TB Database Project allows for both the options of posting data for public access and as private data pending publication or the resolution of intellectual property claims.

While much of the R&D capability to bring a vaccine to market exists in the developed world, avian flu cases have occurred primarily in the developing world. Thus, in order to research the virus, researchers in the industrialized world are dependent upon developing nations to supply them with wild virus samples. The need for reciprocal benefits for these developing countries to share has recently become very evident. Patenting of avian flu wild virus samples sent to developed world laboratories and the potentially high costs of any resulting vaccines created from those samples have created friction in the Global Influenza Surveillance Network. The refusal of Indonesia to share virus samples to WHO Collaborating Centers without assurances of benefit sharing demonstrates the importance of a bidirectional flow of benefits in the sharing of data and materials.70

Obtaining data on flu virus sequences from its network of laboratories, WHO can only release the data with permission from the country of origin. WHO had provided much of the data to 17 labs in a password-protected database out of Los Alamos National Laboratory. Responding to complaints from some scientists about these barriers to broader data access, the Global Initiative on Sharing Avian Influenza Data (GISAID) was launched in 2006.71 The Initiative “is open to all scientists, provided they agree to share their own data, credit the use of others’ data, analyze findings jointly, and publish results collaboratively.”72 Curating the data, GISAID pledges to deposit these sequences, following analysis and validation, to one or more publicly available databases with a delay no longer than six months.73 The sharing of information poses potential complications that need to be worked out, from intellectual property rights to verification by specialized reference labs.74 Some voiced concerns that sharing data immediately jeopardized their ability to publish first on these findings after considerable investment of time and resources. Others question whether publishing the paper should receive priority over the benefits of earlier data release for public health.

As seen with MTAs, there is value in aggregating efforts and thereby lowering transaction costs to sharing the building blocks of research. Pooling research data and materials has other benefits. A registry of clinical trials allows patients and providers to find treatments undergoing testing or uncover negative findings that might otherwise remain unpublished or hidden. A pool of compound libraries might diversify the spectrum of available druggable compounds, allow researchers to pursue novel compounds from parts of the genome considered “undruggable,” and bring useful data and annotation information to a larger group of researchers, some pursuing neglected diseases otherwise without the benefit of such resources.

While there are significant challenges both to creating such repositories and to sharing the knowledge from them, there are promising developments. Different but complementary approaches to broadening compound library access have emerged, both in the rare disease and neglected disease spaces.

The European Rare Diseases Therapeutic Initiative (ERDITI) facilitates access to compounds, developed by pharmaceutical firms, for academic teams. Enlisting the participation of 4 major companies in their efforts (Aventis, GlaxoSmithKline, Roche, and Servier) and 10 European research institutions, ERDITI screens requests from academic researchers interested in evaluating the therapeutic potential of compounds in preclinical studies for treating rare diseases.75 Of note, academic teams broker agreements with companies on a case-by-case basis and in confidentiality. The ERDITI arrangement though precludes high-throughput screening or assembling a common pool of “non-used compounds” drawn from the various companies contributing to these efforts.

Tackling a range of neglected diseases, the Special Programme for Research and Training in Tropical Diseases (TDR) has launched a web portal, TDR Targets, to bring together data and annotation in an open-access database on tropical disease pathogens. Users can undertake searches ranging from genomic or protein structural data to information on target druggability on neglected diseases from leprosy and filariasis to Chagas disease and leishmaniasis. In the first 16 months since the database’s launch, the site has logged more than 10,000 visits, with more than 30% coming from developing countries or regions where these neglected diseases are endemic.76 This web-based initiative complements efforts to bring together the partnerships and multidisciplinary networks needed for drug discovery for neglected diseases.77 Now the vision for TDR is considerably more ambitious: a virtual drug discovery network with negotiated access and screening of proprietary compound libraries on a contractual and confidential basis, sponsored scientists to work in pharmaceutical companies on these neglected disease projects, and a clearinghouse to help coordinate these efforts.78

While pharmaceutical and biotechnology firms focus on the “druggable genome,” the work of the NIH Molecular Libraries Initiative will illuminate the majority of the human genome thought to be “undruggable” in hopes of coming up with biologically useful substances and novel drug leads.79 To support this line of work, the NIH has established the Molecular Libraries Screening Centers Network (MLSCN) comprised of 10 centers, each with particular expertise and technology. The Network as a whole conducts 20 assays on more than 100,000 compounds each year. Each Center must deposit its screening results in PubChem, an NIH-supported, publicly accessible database with more than 8 million compounds. Sharing such knowledge is not without its challenges. Patenting of probes developed under this initiative would be discouraged as they may be the source of multiple chemical analogs that offer improved properties. As one of the lead investigators in MLSCN acknowledged, publication and attribution, the tension between the data release policy and timely submissions of assay results to MLSCN, and the critical path to optimizing and synthesizing biologically useful products remain challenges.80 As models like this emerge, resolving these issues will also require investment.

Findings of publication bias and the nondisclosure of unfavorable clinical trial results have stoked efforts to ensure the sharing of this information. Looking at efficacy trials for approved New Drug Applications for new molecular entities, a recent study concluded that many of these trials remain unpublished even five years after FDA approval. More disconcerting, there were differences between the trial information reviewed by FDA and that found in publications of these trial results. Nearly half of the unfavorable findings found in trials submitted for FDA approval did not make it into the published papers of these clinical trials.81 When post-marketing studies found increased suicide among children using Paxil82 and of increased incidence of heart failure from the use of Vioxx,83,84 companies delayed the release of clinical trial data that reflected the risks associated with their products.

Responding to this need, the NIH has developed a clinical trial registry and results data bank for both federal and privately supported clinical trials conducted around the world. In the wake of the public outcry over the non-disclosure of clinical trial results in the Vioxx case, the International Committee of Medical Journal Editors (ICMJE) announced that their journals collectively would “require, as a condition for consideration for publication, registration in a public trials registry.”85 Demonstrating the power of such norm-setting changes, ClinicalTrials.gov registered a 73% increase in the number of trials and a 195% increase in the number of data providers that registered trials over the 4.5 months before the ICMJE’s September 2005 deadline, with an enormous spike of new registrations in the two weeks leading up to and around the final deadline.86 The FDA Amendments of 2007 strengthened these reporting requirements by requiring clinical trial results completed before product approval to be submitted to ClinicalTrials.gov not later than 30 days after the drug or device has received FDA approval.87 Building upon the momentum of these efforts, WHO has sought to provide a forum for developing best practices for clinical trial registration, and a number of countries now maintain prospective trial registries.88

Access to Patented Inventions

The patenting of inventions and their licensing influence significantly the sharing of knowledge. Patenting this knowledge enhances the potential commercial value to this work, and this can help mobilize needed private sector resources for further research and development. In a survey of those involved in biomedical R&D, licensing was routine, worthwhile projects were almost never stopped because of patents on research tools, but infringement of research tool patents was frequent.89 In fact, the study found that one out of three respondents from industry and all nine of the government lab or university respondents admitted to using patented research tools occasionally without a license. In a larger survey of biomedical researchers in universities, nonprofit institutions and government, only 8% of respondents believed that they had conducted research involving patented inventions over the past two years, but even fewer regularly checked for patents on inputs to their research.90 None of the respondents reported abandoning research as a consequence of third-party patents, and though delays or workarounds were reported, they were infrequent.

In several research areas, however, problems over patenting have surfaced. Particularly in the field of genomics, several studies suggest that concerns over patenting may hinder research. One survey found that the license granted on the patent needed for clinical testing of hemochromatosis prompted 30% of laboratories to discontinue or not develop genetic tests for this disease.91 A broader survey of genetic laboratory directors found that over half had decided not to develop one or more genetic tests as a consequence of the underlying patent or license held on it.92 Controversy over the patenting of genetic tests continues to brew. On behalf of key professional research societies, genetics researchers and patents, the American Civil Liberties Union and the Public Patent Foundation recently filed suit against Myriad Genetics over their patents on break cancer genes. The legal complaint argues that Myriad has used the patents in a way that restricts access to the diagnostic tests for breast cancer.

The research enterprise yields both tools and products. Both may serve as inputs to follow-on research, and both may receive protection under the intellectual property rights system as inventions. Patents reward the inventor with time-limited market exclusivity, but licensing might be handled exclusively or nonexclusively. The approach to licensing shapes the conditions of access and the sharing of knowledge. The passage of the U.S. Bayh-Dole Act in 1980 encouraged nonprofit institutions to patent inventions emerging from government-funded research. Companies also may expect ownership of patentable inventions arising from sponsored research at the university.93

The U.S. Copyright Act of 1976 and the NIH Public Access policy lower barriers to the sharing of knowledge through scientific publications, and OMB Circular A-130, similarly kept the price of government databases to the marginal cost of dissemination. By contrast, the Bayh-Dole Act accelerated patenting, licensing and associated revenues at universities. These practices have stirred concerns over patenting and exclusive licensing over upstream research tools. At times unnecessary for adoption by industry, these practices contribute to patent thickets that complicate bargaining and broader use. Nearly 30 years after the Act’s passage, U.S. universities, hospitals and research institutions only derive 5% of total academic research dollars from licensing revenues.94 The reality seems to overstate the benefits of Bayh-Dole on commercialization of federally funded inventions and have prompted questions over the emulation of this statute in developing countries.95

Funders have sought to mitigate the concerns over unnecessary patenting and exclusive licensing of inventions. In 1999, NIH released “Principles and Guidelines for Sharing of Biomedical Research Resources,” in which grantees were advised not to license exclusively “a broad, enabling invention that will be useful to many scientists (or multiple companies in developing multiple products), rather than a project or product-specific resource.”96 Various foundations have also issued guidance that encourages greater sharing of inventions resulting from their research.

Some have built such conditions into their grant agreements. In funding point-of-care diagnostics for monitoring AIDS, the Doris Duke Charitable Foundation assessed how existing intellectual property affected the ability of their grantees to make good on the charitable objective of ensuring the technology’s availability at an affordable cost in developing countries. Their grant agreements went further and retained a nonexclusive, royalty-free license to any patents filed in developing countries. This would allow the Foundation to sublicense rights to make and distribute the product if the grantee failed to deliver on the charitable objective.97 The NIH itself has used “White Knight Clauses” (named after the company with which these clauses were first used) in its licensing to ensure the provision of products at cost in the developing world.98 To make possible the low-cost production of conjugate meningitis vaccine A targeted to strains in developing countries, the U.S. Food and Drug Administration transferred conjugation technology to the Serum Institute in India and SynCo Bio Partners.99

As with scientific publications, data and material transfers, tiering and pooling also can apply to patents and their licensing. The simplest approach to tiering is the use of two tiers. By setting limits of geography or use, licenses may offer lower royalty-free rates or reduced pricing for the invention’s application in the developing world. For example, the Institute for OneWorld Health has secured exclusive license from the University of Washington and Yale University to develop azole compounds that might help treat Chagas disease in the developing world.100 Similarly, the University of British Columbia licensed an oral formulation of Amphotericin B to iCo Therapeutics for treating blood-borne fungal infections in the developed world on condition that the company provides subsidized pricing of the drug to treat leishmaniasis in the developing world.101 Such licenses often promise little revenue return from the developing world, but by reserving rights for application in the industrialized world, revenues from paying markets remain possible.

Pooling patents can help lower the transaction costs associated with assembling the tools needed to conduct research on a health technology. The Wellcome Trust recognized the need for ready access to single-nucleotide polymorphisms (SNPs) as tools to map the human genome. With its support, a consortium of corporate, academic and funding partners came together to ensure the overall intellectual property (IP) objective “to maximize the number of SNPs that (1) enter the public domain at the earliest possible date, and, (2) to be free of third-party encumbrances such that the map can be used by all without financial or other IP obligations.”102 Begun in 1999, the Consortium significantly exceeded its initial goals. Instead of releasing 300,000 SNPs by 2001, the SNPs Consortium successfully placed 1.4 million SNPs into the public domain.

REENGINEERING THE VALUE CHAIN

The value chain of R&D represents the inputs and outputs at each stage from discovery to delivery of a health technology. The sharing of knowledge constitutes a key input all throughout this value chain. The process of investigation and invention is often a cumulative one. Scientific exchange speeds its evolutionary progress, opens new directions for research, and enables interdisciplinary and cross-institutional collaborations.

Across the value chain, however, the ownership of knowledge adds friction to the process of sharing. With industry collaborations and sponsorship, proprietary control does sometime limit the dissemination of the knowledge produced. In a survey of biomedical researchers in nonprofit institutions and government laboratories, over a quarter of MTAs carried reach-through claims, and a quarter also placed restrictions on publication. For MTAs involving drug requests, 70% carried such a restriction.103 Scientific competition also contributes to this situation as has the increasingly common practice of patenting and exclusively licensing inventions from universities. While commercial entities have an IP strategy to harness proprietary technologies to bring inventions to market, the public sector has not given as much thought as to how it might apply IPRs to protect the public good of scientific R&D.

At each link in the chain, various approaches to sharing knowledge have been discussed. Whether scientific publications, data or material transfer, or the licensing of patented inventions are the critical input to innovation, the scientific community has gravitated to solution paths that share some common elements. To be sure, there are contextual differences at each link in the chain. The dissemination of scientific publications can approach zero marginal cost, but data and material transfer involve costs to prepare or transfer, and patented inventions have at least opportunity costs. Understanding how these different solution paths contribute to scientific innovation might inform how to leverage best the U.S. commitment to global health.

At least three solution paths emerge from examining the interventions in the value chain—tiering, pooling, and open-source collaboration. The prime means by which each of these approaches ensures the sharing of knowledge is in its collective management of the ownership of knowledge. Certifying that ownership, patents give incentive to the R&D process by providing time-limited market exclusivity and enabling a tradable commodity in the market. Of course, the disadvantage of intellectual property ownership in a market comes as deadweight loss from monopoly pricing. Deadweight loss results when people are excluded from use of a good even when their willingness to pay exceeds the marginal cost of providing it. Through price discrimination, one can mitigate some of the inefficiency that comes from monopoly pricing, and tiering takes steps in this direction by making resources more available to lower-income groups. Pooling assembles research inputs in ways that lower the transaction costs of conducting studies or reveal socially useful bundles of research tools or technologies. Pooling may help build the research commons. These solution paths are not mutually exclusive, and in fact, hybrid approaches might have significant promise. For example, pooling for neglected diseases combines aspects of both approaches. Tiering and pooling address how we organize the inputs of research. Open source focuses more on the means of knowledge production. In open-source collaborations, the locus of control shifts from the owner of knowledge to its users.

By tiering, the market is segmented between those receiving preferential treatment and those not receiving such treatment. Such preferential treatment typically means lower access costs. Tiering can occur at different points in the value chain—when pricing the final product, licensing the underlying intellectual property, or making other research inputs available. Tiering sometimes distinguishes the private market from the public sector in a developing country. If the difference in tiered prices is steep, preventing arbitrage between the public sector and the private market may be more difficult or costly to implement. Sometimes those implementation costs fall on those in developing countries where the resources and infrastructure are already stretched to their limits.

Some practices of tiering remain challenging to resolve. Selecting what countries belong to which tier is a key consideration. With the support of Gates Foundation funding, the University of California, Berkeley, provided royalty-free licenses for the microbial synthesis of artemisinin to Amyris Biotechnologies and the Institute for One World Health.104 The University limited the field of use to the nonprofit production of artemisinin for treating malaria in the developing world. The University’s royalty-free license covers the developing world and does so in return for a commitment from its partners to produce the drug at no profit for the developing world. By contrast, under HINARI, middle-income countries such as China, India, Indonesia, or Thailand do not even qualify either for Band 1 (Free Access) or Band 2 (Low-Cost Access) despite having a GNI per capita that falls within HINARI’s bands.105 Still tiering may provide more equitable pricing scaled to the resources available in developing countries.

One of the more inspired business models also takes advantage of dual markets. A nonprofit firm, Global Vaccines, Inc., proposes to undertake vaccine development with public financing for developing country markets first. The University of North Carolina has provided a royalty-free license to technology for this purpose. When the technology reaches the proof-of-concept stage, it would hopefully have promise for commercial sublicense in industrialized countries. At that point, the commercial sublicense would return revenues for both the company and for the university. This potential model places priority on diseases in developing countries with the support of government or philanthropic funding, transfers the technology from South to North, and seeks to generate revenues from the commercialization of such technology in industrialized countries.106

Tiering can, however, also be divisive, particularly in regions like Latin America where the countries have sought price concessions for antiretrovirals. In the region, countries range from large middle-income country markets like Brazil to smaller, least developed country markets like Haiti. Tiered pricing available to Haiti may not be so for Brazil. Under the Accelerating Access Initiative, five pharmaceutical manufacturers offered lower prices for HIV medicines by brokering agreements on a drug-by-drug, country-by-country basis. With only five countries in Latin America and the Caribbean initially participating in this Initiative, the countries of the Caribbean started a subregional negotiation with Accelerating Access Initiative partners. Central American countries soon followed suit en bloc, and then ten other Latin American countries started collective negotiations.107 Each subregional negotiation improved upon the country-by-country negotiations with the Accelerating Access Initiative. The important lesson from this experience is the monopsony power of collective negotiation—or pooling—for tiered pricing.

Apart from organizing demand, pooling can facilitate access to the supply side by constructing a research commons. Such a step can lower the transaction costs of assembling these research inputs. Pools can come together by various means. Upstream in the R&D pipeline, pooling can build upon a more robust public domain of research tools and other inputs. The entanglements of IPRs might be fewer over the building blocks of knowledge. Downstream in the R&D pipeline, commercializable inventions will play a more important role in the pool, so the mix of incentives to contribute and disincentives to leave the pool may be more complicated to structure than in upstream pools.

By applying the Creative Commons Attribution License, open access publications create pools of journal articles that permit “unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.” Open-access repositories for data and journal articles posted online at universities similarly pool research resources for broad, public availability. Normsetting approaches like the adoption of the UBMTA have the same potential, but when universities substitute their own more restrictive MTAs in lieu of following the UBMTA, the fragility of such pooling arrangements becomes clear.

While private sector pools provide access to patents comprising MPEG-2, DVD and other standards in the electronics industry, the creation of pools in biomedicine has been slower in coming. There certainly have been fledgling efforts to create a SARS patent pool,108 to develop UNITAID’s proposed patent pool for HIV/AIDS drug products,109 and to seed a technology trust for neglected diseases.110 Recently though, GlaxoSmithKline stirred renewed interest in this approach with its announced commitment to donate more than 800 patents to a pool open to researchers working on developing treatments for neglected diseases.111 Going beyond patent pools, the “technology trust” model explores the potential for pooling across the value chain, from open-access databases to pooling of patented inventions. Using various arrangements for collectively managing intellectual property, it emphasizes the normative role that public sector pooling and its strategic use of IPRs can play in encouraging greater scientific exchange and innovation.112

Significant start-up costs exist for creating a pool. Organizers have to consider how to define what patents are essential to a pool or not; set valuation and remuneration, if any, for patented inventions or copyrighted materials in the pool; establish incentives for joining the pool and disincentives for leaving it; and seek antitrust guidance to ensure the pool is pro-competitive. In agricultural biotechnology, the Rockefeller and McKnight Foundations along with 10 universities created Public Sector Intellectual Property Resources in Agriculture (PIPRA) in 2003. Its stated goals were to “overcome the fragmentation of public-sector IP rights and re-establish the necessary FTO [freedom to operate] in agricultural biotechnology for the public good, while at the same time improving private-sector interactions by more efficiently identifying collective commercial licensing opportunities.”113 Acting more as a clearinghouse than a pool, its public database comprised of patented inventions from member institutions makes it easier to identify socially useful bundles of intellectual property for commercialization.

Public sector and philanthropic funders seldom foot these transaction costs for pooling in biomedical R&D. NIH grantees cannot charge legal fees for patenting or licensing as a direct cost to their project. However, it can be built into the indirect facilities and administrative costs of a grant. This certainly limits the means by which nonprofit research institutions might be willing to use IP strategically to protect the public domain. After all, patenting involves both legal, filing and maintenance fees, and protecting IP for the public domain does not promise a financial return on this investment. Nonetheless, the willingness of some funders and even some universities to support upfront fees for publication in open-access journals is a promising step in this direction, perhaps one that might be emulated when patenting to protect public access is at stake.114,115

Not paying these transaction costs for pooling, however, can be problematic, particularly for emerging infectious diseases. In these cases, the spread of the epidemic may outpace the prosecution of patents at the Patent Office. As a result, developers of diagnostics and treatments for the disease receive little certainty from the patent system as the epidemic unfolds. The U.S. Centers for Disease Control and Prevention and the British Columbia Cancer Agency argued that the rationale for rushing to the Patent Office during the SARS outbreak was to maintain the freedom to operate for potential innovators in this space.116 This example reveals the perceived need by public agencies to patent in order to protect researcher access and the public’s interest in areas of critical public health concern. With patents still issuing years after the initial SARS epidemic has been contained, pooling may help resolve the uncertainty faced by pharmaceutical firms working on emerging infectious diseases during the outbreak.

Effectively used, tiering and pooling efforts can contribute to greater openness in the sharing of knowledge. Open-source science focuses more on the way in which the resulting collaboration is organized. Taking a page from the free software movement, the philosophy is embodied in the General Public License that allows a copyright owner to license a user to use his or her work, examine the underlying source code, modify it, and redistribute modified or unmodified versions of the work. The license provides this right without paying a fee in return to the owner, but stipulates that the same conditions must be passed along to any subsequent user of that work.

This open-source approach turns the traditional model of innovation on its head. Open-source production empowers end users in the innovation of a technology, and in so doing, emphasizes transparency as well as peer review and feedback.117 Attribution of contributions in such communities is more difficult to trace than the authorship of scientific publications. With the successful experiences of open source in software, would such an approach apply in biomedicine? Perhaps bioinformatics might be, by analogy, a good starting point. The advent of the Internet has certainly changed the costs of open-source production. Distributed computing projects such as Folding@Home involve nonscientists and scientists alike in contributing desktop computing power to solving computationally intensive problems like protein folding. Moving from distributing computing projects to peer-based production among scientists may be more challenging.

Still some have applied similar open-source principles to biomedical science. Initially the Haplotype Map (HapMap) Project required users of its database to agree to a license, whereby investigators committed “not to use the data in any way that will restrict the access of others, and will only share the data obtained with others who have accepted the same license.”118 While such a license reaches virally through to subsequent users of the data, it may pose problems for those seeking to commercialize inventions in a marketplace where secure IP holdings can spell the difference between access to venture capital or not. Some have proposed the possible application of open source to finding cures for tropical diseases, where there is not a large paying market.119 The adoption of such an approach among wet lab scientists has been slower in coming.

However, the Open Source Drug Discovery (OSDD) project, launched by India’s Council of Scientific and Industrial Research in 2008, is a promising model to watch. The online platform allows a community of scientists to share and collaborate on projects, from gene sequencing to new drug development, on Mycobacterium tuberculosis. Backed by US$38 million in commitments from the Indian government, this open-source website has already engaged 700 participants from 130 cities across 56 active projects.120 OSDD differs from previous open-source drug discovery projects in that it has the support of a leading research institution in a major developing country, promises to adopt 30 colleges throughout India where students will have the opportunity to contribute research to this initiative, and importantly, has substantial financial resources to leverage research collaborations. Public financing may be the key to applying open-source production in biomedicine, both paying for what cannot be volunteered and supporting the open exchange important for collaboration.

By reengineering the value chain of R&D, alternative models for innovation may emerge and potentially better meet the needs of global public health. The approaches of tiering, pooling, and open source point to potential ways in which the sharing of knowledge might be improved. While some of these efforts will emerge spontaneously from the scientific community, others will require targeted and strategic public and philanthropic investment. Unlike the private sector, the public and philanthropic sector does relatively little to manage collectively or strategically its IPRs to seek fair returns from its investment.

Yet arguably if publicly funded research were not freely available, the taxpayers would have paid for the results several times over—grants for the academic research, salaries for those academics giving their time for peer review, and subscriptions for such journals.121 For drugs, diagnostics, and vaccines, taxpayers pay for much of the basic science and some of the clinical research, the academic training of research scientists, and of course, for the final product. Some have argued for the federal government to pay for clinical trials, so that the results would be treated as a public good.122

This calculus of “pay now or pay more later” might guide where the public ought to direct its investments to maximize the returns to the health care system. For example, in the value chain of scientific journal publication, paying the publication fees for open-access journals is one way of supporting a business model that encourages the sharing of knowledge. Going further, the U.S. government could develop a system of supporting open-access journals that publish peer-reviewed, publicly funded research. For those open-access journals that charge publication fees, it could build support into the direct or indirect cost structure of grants. For those open-access journals that do not charge fees, it could provide direct or indirect subsidies. Either way, it could support journals that provide open access rather than impose subscription fees on patients, providers, and universities. This support could factor in transition costs, the citation impact factor of the journal in that field, the rejection rate, and the number of publicly funded research articles published by the journal.

For clinical trials, greater public funding could also reap significant benefits. If structured appropriately, such support might result in improved data transparency and access, the sharing of clinical trial information on shelved products, the removal of financial conflict of interest in the conduct of clinical trials, priority placed on trials addressing major public health concerns, and transparency of R&D costs that might allow policy makers to assess reasonable pricing of the resulting products. The recently approved NIH funding for comparative effectiveness trials is a useful first step in this direction.123

Reengineering the value chain might also involve investing in alternative business models, one that might lower the cost of R&D for neglected diseases. The Gates Foundation grant to the Institute for One World Health, the University of California, Berkeley, and Amyris Biotechnologies to produce artemisinin at no profit for the developing world is one such example. Another example comes from the work of Global Vaccines, Inc., a nonprofit firm that seeks to develop affordable products for developing country markets with the support of public funding and then to disseminate this technology through commercial sublicenses for markets in industrialized countries. With Wellcome Trust and UK government funding, investigators from Imperial College and the London School of Pharmacy reengineered not only the existing version of hepatitis C treatment, pegylated interferon, but also the approach to help ensure its scale-up as a product affordable to the many afflicted with this disease in the developing world.124,125 Through a university spin-off, they licensed the drug to Shantha Biotechnics, bypassing the more customary route of licensing it to a multinational pharmaceutical firm. Facing different clinical trial costs, Shantha Biotechnics will try to produce a more affordable treatment than the one currently available.

Sharing knowledge from bench to bedside is critical to bringing about innovation the world—and particularly its poor—need from the biomedical sector. Overcoming the disparities between industrialized and developing countries sometimes seems like a Sisyphean challenge, but strategic steps taken by the public and philanthropic sector can help create an environment that enables both North and South to work together towards improved innovation and greater access to health technologies.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the support of the Institute of Medicine and a grant from the National Human Genome Research Institute (Grant 5R01HG003763, Building a Technology Trust in Genomics) as well as insightful feedback from Sarah Scheening on the IOM Program staff, the IOM Committee, and several experts, including Professor Peter Suber, Professor Arti Rai, Shaoyu Chang, MD, MPH, and Joseph S. Ross, MD, MHS.

Footnotes

*

Anthony D. So, MD, MPA, is with the Program on Global Health and Technology Access and the Center for Strategic Philanthropy and Civil Society, Terry Sanford Institute of Public Policy, Duke University and the Duke Global Health Institute, Durham, North Carolina. Evan Stewart, BA, is with the Program on Global Health and Technology Access, Terry Sanford Institute of Public Policy, Duke University, Durham, North Carolina.

This paper is made available as an open-access document distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are credited.

1

UNESCO (United Nations Educational, Scientific and Cultural Organization). 2005. UNESCO Science Report 2005. Paris, France: UNESCO. Available at: http://www​.unesco.org​/science/psd/publications/sc_rp_05​.shtml.

2

Glickman, S.W., J.G. McHutchison, E.D. Peterson, C.B. Cairns, R.A. Harrington, R.M. Califf, and K.A. Schulman. 2009. Ethical and Scientific Implications of the Globalization of Clinical Research. N Engl J Med 360(8): 816-823.

3

Michael R. Reich, editor. 1998. International strategies for tropical disease treatments: Experiences with praziquantel. Geneva, Switzerland: World Health Organization, Action Programme on Essential Drugs, Division of Control of Tropical Diseases. Available at: http://apps​.who.int/medicinedocs​/collect​/medicinedocs/pdf/whozip48e/whozip48e​.pdf.

4

WIPO (World Intellectual Property Organization). 2008. World Patent Report: A Statistical Review. Geneva, Switzerland: WIPO. Available at: http://www​.wipo.int/export​/sites/www/ipstats​/en/statistics/patents​/pdf/wipo_pub_931.pdf.

5

Glickman, S.W., J.G. McHutchison, E.D. Peterson, C.B. Cairns, R.A. Harrington, R.M. Califf, and K.A. Schulman. 2009. Ethical and Scientific Implications of the Globalization of Clinical Research. N Engl J Med 360(8): 816-823.

6

INDEPTH Sites. 2009. Available at: http://www​.indepth-network​.org/index.php?option​=com_content&task​=view&id​=50&Itemid=136.

7

Washburn, J. 2005. University Inc.: The Corporate Corruption of Higher Education. New York: Basic Books.

8

Krimsky, S. 2003. Science in the Private Interest: Has the Lure of Profits Corrupted Biomedical Research. New York: Rowman & Littlefield Publishers, Inc.

9

Moran, M. 2005. The New Landscape of Neglected Disease Drug Development: Pharmaceutical R&D Policy Project. London, United Kingdom: LSE Health and Social Care. Available at: http://www​.thegeorgeinstitute​.org/shadomx​/apps/fms/fmsdownload​.cfm?file_uuid=F2B06396-EEA0-851E-3049-C9A030AEDE0F&siteName=iih.

10

Melon, C.C., M. Ray, S. Chakkalackal, M. Li, J.E. Cooper, J. Chadder, W. Ke, L. Li, M.A. Madkour, S. Aly, N. Adly, S. Chaturvedi, V. Konde, A.S. Daar, P.A. Singer, and H. Thorsteinsdóttir. 2009. A survey of South-North health biotech collaboration. Nature Biotechnology 27(3): 229-232.

11

Global Burden of Disease and Risk Factors. 2006. Edited by A.D. Lopez, C.D. Mathers, M. Ezzati, D.T. Jamison, and C.J.L. Murray. New York: Oxford University Press and the World Bank. Available at: http://www​.dcp2.org/pubs/GBD/3/Table/3​.C1.

12

Lanjouw, J.O. 2001. A Patent Policy Proposal for Global Diseases. Brookings Policy Brief. Washington.

13

Davidson, E.M., R. Frothingham, and R. Cook-Deegan. 2007. Practical Experiences in Dual-Use Review. Science 316: 1432-1433.

14

Von Hippel, E. 1994. “Sticky Information” and the Locus of Problem Solving: Implications for Innovation. Management Science 40 (4): 429-439. Available at: http://web​.mit.edu/evhippel​/www/papers/stickyinfo.pdf.

15

Scotchmer, S. 2004. Standing on the Shoulders of Giants: Protecting Cumulative Innovators. Innovation and Incentives. Cambridge, Massachusetts: MIT Press, pp. 132-133.

16

Rogers, E.M. 2003. Diffusion of Innovations. New York: Free Press, p. 13.

17

Campbell, E.G., K.S. Louis, and D. Blumenthal. 1998. Looking a gift horse in the mouth: Corporate gifts supporting life sciences research. JAMA 279: 995-999.

18

Blumenthal, D., N. Causino, E. Campbell, and K.S. Louis. 1996. Relationships between academic institutions and industry in the life sciences: An industry survey. New England Journal of Medicine 334(6): 368-374.

19

J.B. Weitzman, ed. 2004. Interview: Sharon Terry. Patient advocate calls for Open Access. Available at: http://www​.biomedcentral​.com/openaccess/archive​/?page=features&issue=21.

20

Alliance for Taxpayer Access. Enhanced Access to the Published Results of NIH Research Will Benefit Science, the Economy and Human Health. Available at: http://www​.taxpayeraccess​.org/resources/Benefits_of_Access​.pdf.

21

About HINARI. 2009. World Health Organization Website. Available at: http://www​.who.int/hinari/about/en/.

22

Nightingale, K. 2008. Subsidised access “helps boost scientific output.” SciDev.Net. Available at: http://www​.scidev.net​/en/news/subsidised-access-helps-boost-scientific-output-​.html.

23

Cockerill, M.J., and B. G. J. Knols. 2008. Open Access to Research for the Developing World. Issues in Science and Technology. Available at: http://www​.issues.org/24​.2/cockerill.html.

24

Villafuerte-Gálvez, J., W.H. Curioso, and O. Gayoso. 2007. Biomedical Journals and Global Poverty: Is HINARI a step backwards? PLoS Med 4(6): e220.

25

Antelman, K. 2004. Do Open Access Articles Have a Greater Research Impact? College & Research Libraries News 65(5): 372-382.

26

Lawrence, S. 2001. Free online availability substantially increases a paper’s impact. Nature. Available at: http://www​.nature.com​/nature/debates/e-access​/Articles/lawrence.html.

27

Hajjem, C., S. Harnad, and Y. Gingras. 2005. Ten-Year Cross-Disciplinary Comparison of the Growth of Open Access and How It Increases Research Citation Impact. IEEE Data Engineering Bulletin 28 (4).

28

Eysenbach, G. 2006. Citation Advantage of Open Access Articles. PLoS Biol 4(5): e157.

29

Evans, J.A., and J. Reimer. 2009. Open Access and Global Participation in Science. Science 323 (5917): 1025.

30

Harvard Faculty Adopts Open-Access Requirement. 2008. Chronicle of Higher Education News Blog. Available at: http://chronicle​.com​/news/article/3943/harvard-faculty-adopts-open-access-requirement.

31

Harvard Kennedy School Faculty Votes for Open Access for Scholarly Articles. 2009. Harvard Kennedy School of Government. Available at: http://www​.hks.harvard​.edu/news-events/news​/press-releases/open-access-vote.

32

Suber, P. 2009. OA Mandate at the Stanford School of Ed. Open Access News. Available at: http://www​.earlham.edu​/~peters/fos/2008/06​/oa-mandate-at-stanford-school-of-ed​.html.

33

Jahnke, A., and J. Ullian. 2009. University Council Approves Open Access Plan. BU Today. Available at: http://www​.bu.edu/today/node/8320.

34

Taylor, M. 2009. MIT Moves Toward Open Access. The Wall Street Journal. Available at: http://blogs​.wsj.com​/digits/2009/03/25/mit-moves-toward-open-access/.

35

Suber, P. 2006. No-fee open access journals. SPARC Open Access Newsletter. Issue #103. Available at: http://www​.earlham.edu​/~peters/fos/newsletter/11-02-06​.htm#nofee.

36

FAQ. 2008. NIH website. Available at: http://publicaccess​.nih.gov/FAQ.htm#f4.

37

U.S. National Institutes of Health. 2005. Policy on Enhancing Public Access to Archived Publications Resulting from NIH-Funded Research. NIH Guide: Notices (NOT-OD-05-022). Available at: http://grants​.nih.gov​/grants/guide/notice-files​/NOT-OD-05-022.html.

38

PubMed Central National Advisory Committee. 2006. Summary Minutes of Meeting. Available at: http://www​.pubmedcentral​.nih.gov/pmcdoc/mins-2006oct.pdf.

39

NIH Public Access. 2008. Policy on Enhancing Public Access to Archived Publications Resulting from NIH-Funded Research. Available at: http://publicaccess​.nih.gov/.

40

European Research Council. 2007. ERC Scientific Guidelines for Open Access. Available at: http://erc​.europa.eu​/pdf/ScC_Guidelines_Open​_Access_revised_Dec07_FINAL.pdf.

41

Suber, P. 2008. OA mandate from the European Research Council. Open Access News. Available at: http://www​.earlham.edu​/~peters/fos/2008/01​/oa-mandate-from-european-research​.html.

42

European Research Council. 2007. ERC Scientific Guidelines for Open Access. Available at: http://erc​.europa.eu​/pdf/ScC_Guidelines_Open​_Access_revised_Dec07_FINAL.pdf.

43

Wellcome Trust. 2007. Conditions under which a Grant is Awarded. Available at: fdp.866620xtw/tnemucod_bew/stnemucod/nimda_stnarg_lartnec_fs@/etisetaroproc/spuorg/tnellets/ku.ca.emocllew.www//:ptth.

44

Wellcome Trust. 2007. Wellcome Trust Position Statement in Support of Open and Unrestricted Access to Published Research. Available at: http://www​.wellcome.ac​.uk/doc_WTD002766.html.

45

Howard Hughes Medical Foundation. 2007. Public Access to Publications. Available at: http://www​.hhmi.org/about​/research/sc320.pdf.

46

Davidson, E.M., R. Frothingham, and R. Cook-Deegan. 2007. Practical Experiences in Dual-Use Review. Science 316: 1432-1433.

47

Walsh, J.P., C. Cho, and W.M. Cohen. 2005. View from the Bench: Patents and Material Transfers. Science 309: 2002-2003.

48

Reichman, J.H., and P.F. Uhlir. 2003. A Contractually Reconstructed Research Commons for Scientific Data in a Highly Protectionist Intellectual Property Environment. Law and Contemporary Problems 66: 315-462.

49

Piwowar, H.A., R.S. Day, and D.B. Fridsma. 2007. Sharing Detailed Research Data Is Associated with Increased Citation Rate. PLoS One 2(3): 1-5.

50

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. JAMA 287(4): 473-480.

51

Heymann, D.L., and G. Rodier. 2003. Global surveillance, national surveillance, and SARS. Emerg Infect Dis [serial online].

52

Surowiecki, J. 2004. The Wisdom of Crowds: Why the Many are Smarter than the Few and How Collective Wisdom Shapes Business, Economies, Societies, and Nations. New York: Doubleday.

53

U.S. National Institutes of Health. 2003. Final NIH Statement on Sharing Research Data. Notice NOT-OD-03-032. Available at: http://grants​.nih.gov​/grants/guide/notice-files​/NOT-OD-03-032.html.

54

European Research Council. 2007. ERC Scientific Guidelines for Open Access. Available at: http://erc​.europa.eu​/pdf/ScC_Guidelines_Open​_Access_revised_Dec07_FINAL.pdf.

55

Marshall, E. 2001. Bermuda Rules: community spirit, with teeth. Science 291:1192.

56

National Human Genome Research Institute. 1996. NHGRI Policy Regarding Intellectual Property of Human Genomic Sequence. Available at: http://www​.genome.gov/10000926.

57

Williamson, A.R. 1999. The Merck Gene Index project. Drug Discov Today. 4(3):115-22.

58

International Nucleotide Sequence Database Collaboration. 2009. About INSDC. Available at: http://www​.insdc.org/page.php?page=home.

59

Worldwide Protein Data Bank. 2009. Welcome to the Worldwide Protein Data Bank. Available at: http://www​.wwpdb.org/.

60

Nguyen, T. 2007. Science Common: Material Transfer Agreement Project. Innovations Summer: 137-143.

61

U.S. National Institutes of Health. 1995. Uniform Biological Material Transfer Agreement: Discussion of Public Comments Received; Publication of the Final Format of the Agreement. Federal Register. Available at: http://www​.autm.net/AM/Template​.cfm?Section​=Technology_Transfer​_Resources&Template=​/CM/ContentDisplay​.cfm&ContentID=1406.

62

Nguyen, T. 2007. Science Commons: Material Transfer Agreement Project. Innovations. Summer: 137-143.

63

AUTM (Association of University Technology Managers). Signatories to the March 8, 1995, Master UBMTA Agreement. Available at: http://www​.autm.net/AM/Template​.cfm?Section​=Technology_Transfer​_Resources&Template=​/CM/ContentDisplay​.cfm&ContentID=2645.

64

U.S. National Institutes of Health. 1999. Principles and Guidelines for Recipients of NIH Research Grants and Contracts on Obtaining and Disseminating Biomedical Research Resources: A Final Notice. Federal Register 64(246): 72090. Available at: http://grants​.nih.gov​/grants/intell-property_64FR72090.pdf.

65

Rai, A.K., and R.S. Eisenberg. 2003. Bayh-Dole Reform and the Progress of Biomedicine. Law and Contemporary Problems 66: 289-314.

66

NIH AIDS Research & Reference Reagent Program. 2009. About the Program. Available at: https://www​.aidsreagent​.org/about_program.cfm.

67

Broad, Novartis announce diabetes initiative. 2004. MIT News. Available at: http://web​.mit.edu/newsoffice​/2004/diabetes2.html.

68

Multiple Myeloma Research Foundation. 2007. MMRF, MMRC Launch Multiple Myeloma Genomics Portal. Available at: http://www​.multiplemyeloma​.org/in_the_news/6.03.034.php.

69

Gates grant funds global tuberculosis database development at Stanford. 2007. Stanford School of Medicine News Release. Availability at: http://med​.stanford.edu​/news_releases/2007​/february/tuberculosis.html.

70

Khor, M., and S. Shashikant. 2008. Developing countries look to WHA for solution to flu virus issue. Third World Network. Available at: http://www​.twnside.org​.sg/title2/avian.flu/news​.stories/afns.004.htm.

71

Bogner, P., I. Capua, N.J. Cox, D.J. Lipman, and others. A global initiative on sharing avian flu data. Nature. 442(31).

72

Roos, R. 2006. Scientists launch effort to share avian flu data. CIDRAP News. Available at: http://www​.cidrap.umn​.edu/cidrap/content​/influenza/avianflu/news/aug2506data​.html.

73

Enserink, M. 2006. Pushed by an Outsider, Scientists Call for Global Plan to Share Flu Data. Science 313: 1026.

74

Enserink, M. 2006. As H5N1 Keeps Spreading, A Call to Release More Data. Science 311: 1224.

75

Fischer, A., P. Borensztei, and C. Rousse. 2005. The European Rare Diseases Therapeutic Initiative. PLoS Med 2(9): e243.

76

Agüero, F., B. Al-Lazikani, M. Aslett, M. Berriman, F.S. Buckner, R.K. Campbell, S. Carmona, I.M. Carruthers, A.W.E. Chan, F. Chen, G.J. Crowther, M.A. Doyle, C. Hertz-Fowler, A.L. Hopkins, G. McAllister, S. Nwaka, J.P. Overington, A. Pain, G.V. Paolini, U. Pieper, S.A. Ralph, A. Riechers, D.S. Roos, A. Sali, D. Shanmugam, T. Suzuki, W.C. Van Voorhis, and C.L.M.J. Verlinde. 2008. Genomic-scale prioritization of drug targets: the TDR Targets database. Nature Reviews: Drug Discovery 7: 900-907.

77

Senior K. “Web Initiative for Neglected Diseases.” The Lancet Infectious Diseases. June 2007; 7: 377.

78

Hopkins, A.L, M.J. Witty and S. Nwaka. 2007. Mission possible. Nature 449(13): 166-169.

79

Austin, C.P., L.S. Brady, T.R. Insel, and F.S. Collins. 2004. NIH Molecular Libraries Initiative. Science 306: 1138-1139.

80

Lazo, J.S. 2006. Roadmap or Roadkill: A Pharmacologist’s Analysis of the NIH Molecular Libraries Initiative. Molecular Interventions 6(5): 240-243.

81

Rising, K., P. Bacchetti and L. Bero. 2008. Reporting Bias in Drug Trials Submitted to the Food and Drug Administration: Review of Publication and Presentation. PLoS Medicine 5(11): 1561-1570.

82

US Food and Drug Administration. 2003. FDA Statement Regarding the Anti-Depressant Paxil for Pediatric Population. FDA Talk Paper. Available at: http://www​.fda.gov/bbs​/topics/ANSWERS/2003/ANS01230.html.

83

US Food and Drug Administration. 2008. Vioxx (rofecoxib) Questions and Answers. Available at http://www​.fda.gov/cder​/drug/infopage/vioxx/vioxxQA.htm.

84

Topol, E.J. 2004. Failing the Public Health—Rofecoxib, Merck, and the FDA. New England Journal of Medicine 351(17): 1707-1709.

85

World Association of Medical Editors. 2004. Clinical Trial Registration: A Statement from the International Committee of Medical Journal Editors. Available at: http://www​.wame.org/wame-listserve-discussions​/clinical-trials-registry.

86

Zarin, D.A., T. Tse, N.C. Ide. 2005. Trial registration at ClinicalTrials​.gov between May and October 2005. New England Journal of Medicine. 353:2779-2787.

87

United States Code. 2007. Title VIII—Clinical Trial Databases. Sec. 801. Expanded Clinical Trial Registry Data Bank. Public Law 110-85.

88

World Health Organization. 2009. The Register Network. Available at: http://www​.who.int/ictrp​/network/en/index.html.

89

Walsh, J.P., A. Arora and W.M. Cohen. 2003. Working Through the Patent Problem. Science 299: 1021.

90

Walsh, J.P., C. Cho, and W.M. Cohen. 2005. View from the Bench: Patents and Material Transfers. Science 309: 2002-2003.

91

Merz, J.F., A.G. Kriss, D.G.B. Leonard, and M.K. Cho. 2002. Diagnostic testing fails the test. Nature 415:577-579.

92

Cho, M.K., S. llangasekare, M.A. Weaver, D.G.B. Leonard, and J.F. Merz. 2003. Effects of patents and licenses on the provision of clinical genetic testing services. J Mol Diagn 5(1):3-8.

93

Campbell, E.G., K.S. Louis, and D. Blumenthal. 1998. Looking a gift horse in the mouth: Corporate gifts supporting life sciences research. JAMA 279: 995-999.

94

AUTM (Association of University Technology Managers). 2007. AUTM U.S. licensing activity survey: FY2006 survey summary, Data Appendix. Available at: http://www​.autm.net/events​/file/AUTM_06)US%20LSS_FNL.pdf.

95

So, A.D., B. Sampat, A.K. Rai, R. Cook-Deegan, J.H. Reichman, R. Weissman, A. Kapzynski. 2008. Is Bayh-Dole good for developing countries? Lessons from the US experience. PLoS Biol 6(10): e262.

96

U.S. National Institutes of Health. 1999. Principles and Guidelines for Recipients of NIH Research Grants and Contracts on Obtaining and Disseminating Biomedical Research Resources: A Final Notice. Federal Register 64(246): 72090. Available at: http://grants​.nih.gov​/grants/intell-property_64FR72090.pdf.

97

Doris Duke Charitable Foundation. 2004. Medical Research Program Bulletin. Available at: http://www​.ddcf.org/doris_duke_files​/download_files​/MRPBulletinOct04.pdf.

98

Ferguson, S. 2005. Presentation at Duke Conference on “Collective Action and Proprietary Rights: Promoting Innovation and Access in Health Care.”

99

Sanjit, B. 2004. New meningitis vaccine in the pipeline. Science in Africa. Available at http://www​.scienceinafrica​.co.za/2004/july/meningitis.htm.

100

Institute for OneWorld Health Licenses Potent Therapy from Yale and University of Washington to Treat Chagas, One of the Largest Parasitic Diseases in the World. 2003. Institute for OneWorld Health Press Release. Available at: http://www​.oneworldhealth​.org/media/details.php?prID=9.

101

UBC Global Access Agreement with iCo Therapeutics Offers Oral Drug to Help Millions in Developing World. 2008. University-Industry Liaison Office Press Release. Available at: http://www​.publicaffairs​.ubc.ca/media/releases​/2008/mr-08-054.html.

102

SNP (Single-Nucleotide Polymorphism) Consortium. Full Genome Representative SNP Map Program Summary. Available at: http://snp​.cshl.org/about/program.shtml.

103

Walsh, J.P., C. Cho and W.M. Cohen. 2005. View from the Bench: Patents and Material Transfers. Science 309: 2002-2003.

104

$42.6 Million, Five-Year Grant from Gates Foundation for Antimalarial Drugs Brings Together Unique Collaboration of Biotech, Academia, and Nonprofit Pharma. 2004. Institute for One-World Health Press Release. Available at: http://www​.oneworldhealth​.org/documents/DP_​%20Malaria%20Release%20121304.pdf.

105

Eligibility. HINARI Web Site. Available at: http://www​.who.int/hinari​/eligibility/en/.

106

Johnston, R.E. 2005. Academic science and the business of vaccines. Arch Virol Suppl. (19): 203-206.

107

Fitzgerald, J., and B. Gomez. 2003. An Open Competition Model for Regional Price Negotiations Yields Lowest ARV Prices in the Americas. Presentation to the 8th World STI/AIDS Congress, Punta del Este, Uruguay.

108

Simon, J.H.M., E. Claassen, C.E. Correa, and A.D.M.E. Osterhaus. 2005. Managing severe acute respiratory syndrome (SARS) intellectual property rights: the possible role of patent pooling. Bulletin of the World Health Organization 83: 707-710.

109

UNITAID moves towards a patent pool for medicines. 2008. UNITAID Press Release. Available at: http://www​.unitaid.eu/index​.php/en/NEWS/UNITAID-moves-towards-a-patent-pool-for-medicines.html.

110

So, A. 2008. Creating an Enabling IP Environment for Neglected and Rare Diseases. IOM Workshop presentation. Available at: http://www​.iom.edu/?id=54295 as cited in Wizeman, T., S. Robinson, and R. Giffin. 2008. Breakthrough Business Models: Drug Development for Rare and Neglected Diseases and Individualized Therapies: Workshop Summary. Washington, DC: The National Academies Press, p. 63. Available at: http://books​.nap.edu/openbook​.php?record_id​=12219&page=63.

111

Whalen, J. 2009. Glaxo Offers Patents to Aid Research. Wall Street Journal.

112

So, A.D. 2004. Enabling Conditions for the Scientific Commons. Presentation at Innovation in the Life Sciences: Intellectual Property and Public Investment for Pharmaceuticals and Agriculture at the Earth Institute, Columbia University. Available at: http://www​.earth.columbia​.edu/cgsd/events​/life_sciences_agenda.html.

113

Atkinson, R.C., R.N. Beachy, G. Conway, F.A. Cordova, M.A. Fox, K.A. Holbrook, D.F. Klessig, R.L. McCormick, P.M. McPherson, H.R. Rawlings, R.Rapson, L.N. Vanderhoef, J.D. Wiley, and C.E. Young. 2003. Public Sector Collaboration for Agricultural IP Management. Science 301(5630):174-175.

114

Wellcome Trust. 2008. Position statement in support of open and unrestricted access to published research. Available at: http://www​.wellcome.ac​.uk/About-us/Policy​/Spotlight-issues/Open-access​/Policy/index.htm.

115

Berkeley steps forward with bold initiative to pay authors’ open-access charges. 2008. SPARC enews. Available at: http://www​.arl.org/sparc​/publications/articles​/memberprofile-berkeley.shtml.

116

Brickley, P. 2003. Preemptive SARS patents. New Scientist 4(1): 20030509-02.

117

Hope, J. 2008. Biobazaar: The Open Source Revolution and Biotechnology. Cambridge, Massachusetts: Harvard University Press.

118

Cukier, K.N. 2003. Open source biotech: can a non-proprietary approach to intellectual property work in the life sciences? Acumen J of Life Sciences 1(1). Available in abridged form at: http://www​.cukier.com​/writings/opensourcebiotech.html.

119

Maurer, S.M., A.K. Rai and A. Sali. 2004. Finding Cures for Tropical Diseases: Is Open Source an Answer? PLoS Medicine 1(3): 183-186.

120

Menon, S., 2009. Sreelatha Menon: Researchers sans borders. Business Standard. New Delhi. Available at: http://www​.business-standard​.com/india/news​/sreelatha-menon-researchers-sans-borders/04/50/350429/.

121

Alliance for Taxpayer Access. Enhanced Access to the Published Results of NIH Research Will Benefit Science, the Economy and Human Health. Available at http://www​.taxpayeraccess​.org/resources/Benefits_of_Access​.pdf.

122

Lewis, T.R., J.H. Reichman, and A.D. So. The Case for Public Funding and Public Oversight of Clinical Trials. Economists’ Voice January 2007.

123

Loftus, P. 2009. Coming Soon: Comparative Effectiveness Research for Biotech, Wall Street Journal Health Blog. Available at: http://blogs​.wsj.com​/health/2009/03/27/coming-soon-comparative-effectiveness-research-for-biotech/.

124

Careers and Recruitment. 2007. Entrepreneurial experience. Nature Reviews Drug Discovery 6, 499.

125

Heilemann, J. 2007. A biotech pioneer takes on Big Pharma. CNNMoney.com.