در کدام موقعیت مسائل استثناپذیر می شوند: مالکیت مادی دربرابر مالکیت معنوی در تحقیقات زیست پزشکی دانشگاهی

On the basis of survey responses from 507 academic biomedical researchers, we examine the impact of patents on access to the knowledge and material inputs that are used in subsequent research. We observe that access to knowledge inputs is largely unaffected by patents. Accessing other researchers’ materials and/or data, such as cell lines, reagents, or unpublished information is, however, more problematic. The main factors associated with restricted access to materials and/or data include scientific competition, the cost of providing materials, a history of commercial activity on the part of the prospective supplier, and whether the material in question is itself a drug.

The patenting activity of American universities has grown almost an order of magnitude in 20 years, from 434 patents issued to universities in 1983 to 3259 in 2003. Nelson, 2006 and Nelson, 2004 and Dasgupta and David (1994), among others, argue that this growing “privatization of the scientific commons” may jeopardize scientific and technological progress, particularly by restricting access to upstream discoveries and understandings that are essential inputs to subsequent advance. Such restrictions come in the form of licensing fees, terms of exclusivity and other conditions of use, infringement liability, and transactions costs that potentially impose a significant burden on researchers.1 In addition to permitting the imposition of such restrictions, patents may also confer the incentive to do so by enabling academics to seek financial gain at the expense of the sharing of knowledge, data and materials (Blumenthal et al., 1997, Campbell et al., 2002 and Walsh and Hong, 2003).2 This concern over the impact of patenting on the free flow of knowledge in academic science remains of paramount concern even while numerous scholars acknowledge that academic patenting may strengthen firms’ incentives to invest in the downstream activities and resources necessary to commercialize discoveries of academic origin. This paper examines the impact of patent rights on academic researchers’ access to the knowledge and material inputs upon which their research depends—what are broadly termed, “research tools.” On the basis of a survey of 507 academic researchers in genomics and proteomics, we probe the determinants of project choice, and examine the question of access to research knowledge and material inputs, which is the main focus of our study. Our analysis relies on two samples of academic respondents. The first is a random sample of 414 academic researchers (including those in universities, non-profits or government labs). We also collected data from a second sample of 93 academic scientists who are conducting research on one of three important signaling proteins (CTLA-4, EGF and NF-kB), fields that were chosen because they all are the subject of extensive patenting activity by numerous actors and offer the promise of significant commercial gain; that is, they are characterized by conditions that are likely to spawn problems of research input access. The rationale for this more focused sample is that even if one finds little problem of access in a random sample, social welfare impacts could still be great if access is impeded in just one or two particularly important areas of research. This paper builds upon the authors’ prior work. Based on interviews with a limited number of biomedical researchers,3Walsh et al. (2003) found that, despite numerous patents on upstream discoveries, researchers have been readily able to access knowledge inputs. In addition to the typical solutions of contracting and licensing, biomedical researchers have implemented a variety of “working solutions” that commonly included the disregard – often unknowing – of patents on research tools. When questioned about possible infringement of research tool patents, academic researchers commonly suggested that they were protected by a “research exemption” from infringement liability. The Madey v. Duke decision of 2002 raised anew, however, the question of the impact of research tool patents on academic biomedical research by clarifying what many had argued had long been the case—that there was no general research exemption shielding academic researchers in biomedicine or any other field from infringement liability ( Eisenberg, 2003). This very visible decision, sample limitations on our prior work, and continuing concerns that the ever-growing number of patents may be impeding academic science prompted the current effort. While Walsh et al., 2005a and Walsh et al., 2005b presents a brief summary of our findings, the current paper examines more thoroughly the impact on academic biomedical research of patents and limits on access to tangible research inputs. For example, we consider whether the Madey v. Duke decision has affected access to patented discoveries, and also whether such restricted access causes delays, increased costs, or the redirection of research. We also examine: restrictions on access to material inputs broken down by type of input requested; the terms and impacts of material transfer agreements; and the extent to which patenting affects the ability to create the material input oneself. To the extent that we observe restricted access to either intellectual property or materials, we probe not only the role played by IP, but also the roles played by commercial incentives, burden of compliance, and scientific competition ( Hagstrom, 1974 and Walsh and Hong, 2003). Indeed, the policy implications attendant upon any social costs associated with restricted access will depend importantly on its source. To prefigure our main findings, we observe that access to knowledge inputs is largely unaffected by patents, even in our more focused sample. More problematic is access to materials and/or data possessed by other researchers, such as cell lines, reagents, genetically modified animals, unpublished information, etc. Restrictions on access, however, do not appear to turn on whether the material is itself patented. Rather, such restrictions are more closely associated with scientific competition, the cost of providing materials, a history of commercial activity on the part of the prospective supplier, and whether the material in question is itself a drug.

Our results suggest that academic biomedical researchers are engaged in a significant amount of commercial activity, including patenting and licensing. The results also suggest that patents in this field, while common, do not regularly prevent academic scientists from gaining access to the knowledge inputs that biomedical scientists require. None of our random sample of academics reported stopping a research project due to another's patent on a research input, and relatively few (1% of sample) reported delays or the redirection of their research, although some (3–10% depending on the question) did report that patents had a significant influence on their project choices. For researchers working on signaling proteins associated with important metabolic molecular pathways in areas that the literature suggests should be particularly susceptible to IP-induced frictions, we observe a slightly higher incidence of adverse effects (3% abandoning a project and 15% having some adverse effect). One important reason that the rates of adverse outcomes associated with intellectual property are not higher (given the large number of patents in this area) is that, notwithstanding the 2002 Madey v. Duke decision, academic researchers remain largely unaware of patents relevant to their research and typically proceed without considering them; only 8% of our random sample respondents reported awareness of using information or knowledge covered by a third party patent sometime in the prior 2 years. We do find, however, that those who are more engaged in commercial activity are more aware of third party patents, although, even for this group, only about 20% report knowing of relevant third party patents. We have no way of knowing what the true base rate is for the percentage of respondents who use others’ intellectual property. Given, however, the large number of biotech patents issued since 1990, we suspect that the number of academic researchers who are using others’ patented technology exceeds 8% of the total. Although such apparent disregard for IP may for the moment minimize the social costs that might otherwise emerge due to restricted access (Walsh et al., 2003), it remains an open question whether such disregard is sustainable. Indeed, an important question is why academic researchers seemingly disregard the possibility that the knowledge inputs they use may be patent protected. Is it just a matter of habit born of a time, not long ago, when upstream biomedical discoveries were not patented? Or, is it a matter of community norms and organizational and career incentives that place the highest value on getting the work of science done, without paying much attention to anything that might slow the work down? Or is it that, given the low likelihood thus far of academics’ being sued for patent infringement, the researchers have little incentive to change their behavior. There is the additional consideration that academic biomedical researchers are also not generally trained in how to conduct effective patent searches, so that the time spent searching the patent databases would unlikely allow the comprehensive identification of relevant patents, suggesting not searching may be the more rational strategy. No matter the explanation, however, our finding underscores Ellickson (1991) observation that the “law on the books” need not be the same as “law in action,” particularly if the law on the books contravenes a community's norms and interests. In contrast to the case of intellectual property, requests for tangible research inputs from other scientists are not fulfilled in a significant minority of cases. Almost 20% of our respondents report that their last request for a material or data was not fulfilled. Moreover, the incidence of non-compliance appears to be increasing. We also find that such non-compliance affects the research programs of individual researchers. For example, one in nine researchers report abandoning a promising line of research in a given year because he did not receive requested materials or data. This noncompliance with others’ requests for research inputs does not appear to be associated with a patent on the material, but is rather associated with a history of business activity by academics, scientific competition, the time and effort required to satisfy requests, and whether the material in question is a drug. But even if patents on a material do not affect compliance with a request, perhaps the fact that a material is patented affects whether a request is made to begin with. When asked why researchers do not make the research input themselves, we find, however, that patents are much less important than the cost/time involved or the lack of necessary capabilities in one's lab, suggesting that the likelihood of a request being made is not affected importantly by associated patents. Notwithstanding the reasons why a material is not shared, without more research, we cannot conclude that less sharing actually imposes a social welfare cost. Denied requests surely impose costs for individual researchers. And, social welfare is diminished to the extent that redirection of a scientist's research effort or reallocation across investigators impedes scientific progress. On the other hand, if such redirection reduces duplicative research, the social welfare loss may be minimal (Cole and Cole, 1972). There may even be a net welfare gain if redirection increases the variety of projects pursued (Dasgupta and Maskin, 1987). Aside from the welfare consequences of stopped or modified projects, it does appear that there are considerable frictions and costs associated with material transfers. Although MTAs are not universally required, about 40% of such requests require an MTA. Negotiating these MTAs can be time consuming, although only about 10% of all requests for research inputs led to a negotiation lasting more than 1 month, and in almost all cases there is no fee for the material. However, in a minority of cases (8% of requests), delays in accessing research inputs can stop the research for more than 1 month, which can represent a substantial delay in a fast moving research field. We find that MTAs (especially from industry suppliers) frequently include demands for reach-through rights of some form. They also often include terms that put restrictions on publication of research results. It is hard to know, however, what the social welfare implications of these terms are without a closer look at their specific content and the motivations for their inclusion. For example, one common reason for demanding restrictions on publication, such as the right to review papers before publication, or simply the right of advance notification of a pending publication, is to protect the supplier's ability to file patent claims on his own technology without fear that the consumer's publication will place the technology in the public domain. A modest delay in publication in exchange for access to the technology may be seen as a reasonable payment by the consuming scientist, even under NIH guidelines (Department of Health and Human Services, 1999). On the other hand, social welfare losses may be realized if such publication restrictions include the right to withhold publication of results entirely in order to achieve a competitive advantage through secrecy, or to ensure that unfavorable research results (such as adverse effects in clinical trials) are never disclosed.39 Given the modest response rate and the limitations of self-report data, we should be cautious in interpreting our findings. However, based on the data at hand, our results suggest that there is reason for concern about access to tangible research inputs. There is, however, little evidence that patent policy is the direct cause of restricted access to tangible research inputs (as opposed, for example, to scientific competition or prior business activity). Furthermore, the impact on scientific progress of this restricted access to research inputs is also not straightforward. In conclusion, debates that focus on the effects on academic research of the patenting of upstream biomedical discoveries may not be addressing the most pressing policy question. Although the patenting of knowledge inputs into academic biomedical research may impose significant social welfare costs in the future, academic biomedical research may for now be more effectively supported by addressing the transaction costs, competitive pressures and commercial interests that are impeding the sharing of data and material research inputs.