INTRODUCTION I. BIOMEDICAL DATA: CHALLENGES OF CONCENTRATION AND FRAGMENTATION A. Overlapping, Adjacent, and Fragmented Intellectual Property Rights B. Diagnostic Data Silos: Beyond Intellectual Property Rights II. AGGREGATING BIOPHARMACEUTICAL TRIAL DATA A. Biopharmaceutical Trial Data in the United States: Basic History, Law, and Policy B. Recent Developments at the FDA and EMA III. DIAGNOSTIC TESTING DATA A. The Basic Regulatory Scheme B. Looking to the Future CONCLUSION INTRODUCTION
Overlapping legal claims pervade data-based innovation. Because of its heavily regulated status, and its potential to improve human welfare, innovation based on biomedical data is a particularly notable example.
As in most other areas of innovation, (1) intellectual property rights can overlap. Patents and trade secrecy can cover the same inventive territory simply because the U.S. Patent and Trademark Office (USPTO), litigants, and courts find it excessively expensive to police the patent statute's complex disclosure requirements.
Moreover, precisely how much disclosure should be required is a policy question. Patent cases have sometimes allowed patent applicants to claim an invention that arises after patent filing without requiring disclosure of this after-arising invention. Meanwhile, the subsequent technology or information could itself be the subject of separate intellectual property, whether trade secrecy or an improvement patent.
While the caselaw trend has moved away from very broad rights, this movement has not precluded "adjacent" intellectual property rights. In these cases, the narrower patent claims do not formally cover the follow-on innovation. However, to the extent that even patent claims of moderate scope on a key research tool cannot be "invented around," and the claims can be enforced, the practical effect may be similar.
Depending on the precise nature of the follow-on innovation generated, adjacent rights may include subsequent patents, trade secrecy, or regulatory exclusivities administered by the Food and Drug Administration (FDA). In the case of biomedical data--the focus of this Article--the adjacent right typically involves either trade secrecy or regulatory exclusivity. Both overlapping and adjacent rights have generated controversy regarding the possibility of excessive control by the initial rights holder.
In 2012 and 2013, when cases involving diagnostic testing patents ultimately brought these controversies to the U.S. Supreme Court, the Court not only struck down the individual patents involved in those cases but also suggested more generally that patents on diagnostic tools might confer undue control over future research. (2) The Court used as its cudgel the argument that the territory covered by the patents failed to constitute patent-eligible subject matter under the so-called "law of nature" or "product of nature" doctrines. (3)
While these Supreme Court cases address concern about excessive control by a single rights holder, there is an obvious flip side. Concentrating patent ownership in a single firm allows that firm to aggregate all of the "adjacent" data generated through use of the patent. With patents no longer serving as "data aggregators" (4) in the area of diagnostic testing, the problem of data fragmentation in this area may become even more acute. (5) In diagnostic testing, as in other areas of biomedicine, large data sets promote cumulative innovation. (6)
Notably, once biomedical data becomes fragmented, overcoming fragmentation creates its own legal overlap challenge. Even with the recent Supreme Court cases, biomedical data pooling--like pooling of scientific data in other areas (7)--will face multiple intellectual property obstacles. These include both formal intellectual property--trade secrets and narrow patents--as well as reluctance on the part of academic researchers to share what they view as their competitive advantage. (8)
Biomedical data pooling also has to overcome relatively unique challenges posed by health privacy law. Privacy law--and related regimes of informed consent--produce an additional set of claimants, leading some to worry that these legal regimes will create a "tragedy of the anticommons" as formidable as that postulated by Michael Heller and Rebecca Eisenberg in their discussion of the challenge of assembling fragmented patent rights. (9)
Scholars have suggested both private- and public-sector responses to the pooling challenge. On the private-sector side, some have argued in favor of stronger individual rights. On this view, individual consumers accorded control or even property rights over "their" health data might be incentivized to create pools. (10) While proposals for some level of consumer control are interesting, according millions of individuals property rights in data is likely to exacerbate transaction cost problems, not resolve them.
Other scholars have highlighted the potential role of institutional private-sector players like insurance companies. (11) The increasing interest shown by insurance firms in using claims data to determine the therapeutic value of medical interventions they cover is a positive development. However, it is not clear that insurance firms would necessarily be motivated to create pools across firm boundaries.
As for the public sector, many scholars (myself included) have discussed mechanisms by which public funders of data resource creation can exercise informal or formal regulatory power to promote data pooling. (12) But public-sector funding is always a scarce resource.
One set of regulatory actors that have received less attention than they merit is risk regulators. (13) In the data pooling context, risk regulators are part of the problem. But they are also part of the solution.
On the one hand, aggressive moves by risk regulators to control perceived risks to privacy may impair possibilities for pooling. More positively, for a significant set of biomedical data--all pre-approval biopharmaceutical data and a significant subset of diagnostic data--risk regulators like the FDA are the default data aggregators. They can use this status to promote innovation based on pooled data.
The recent history of the FDA's role in addressing rights-encumbered biopharmaceutical clinical trial data silos highlights both the agency's promise and its limitations as a promoter of data-based innovation infrastructure. This history also highlights the status of data as a global public good, with FDA action being supplemented, and arguably superseded, by the actions of its European counterpart, the European Medicines Agency (EMA). (14)
The data exclusivity administered by risk regulators in the United States and other advanced economies emerges as an attractive mechanism for balancing the interests of initial and subsequent data generators. Indeed, for purposes of promoting data pooling, the temporal overlap between patents and data exclusivity is a feature, not a bug. A similar regime of exclusivity for diagnostic data would be worth considering.
Data exclusivity does not address legal claims that emerge from the law of health privacy and informed consent. These regimes raise major transaction cost challenges, particularly to the extent data holders cannot guarantee that de-identified or anonymized data is impervious to re-identification. Unfortunately, U.S. risk regulators' recent tool of choice for addressing potentially identifiable data--broad consent to any future research use--provides a suboptimal response. Broad consent is superior to any attempt to confer property rights on raw data inputs that will be provided by millions of different individuals. But it is inferior to a pragmatic approach that emphasizes sanctions for illegitimate re-identification and protecting those whose data is used from harm.
On balance, the contribution of risk regulators has been mixed. To some extent, risk regulators, particularly in Europe, have capitalized on their pivotal aggregation role to address the challenges posed by rights-encumbered data silos. However, even in the political environment of the Obama Administration--where the White House was a prominent advocate of open government data, (15) and generally viewed regulatory activity as compatible with, or even a promoter of, innovation (16)--U.S. risk regulators were cautious.
In elucidating the role of risk regulators, this Article contributes another chapter to the rich legal and economic literature discussing whether, and how, risk and social regulation can promote socially valuable innovation. (17) As contrasted with much of this literature, however, the focus here is not on innovation induced by efforts to comply with regulation but instead on regulation as a force for creating a data infrastructure for future innovation.
Part I of this Article briefly summarizes the history of overlapping and adjacent intellectual property rights in biomedical innovation. It then discusses the manner in which the Supreme Court's reaction to such rights concentration may exacerbate legally-encumbered diagnostic data silos. Part II outlines the basic history of biopharmaceutical trial data silos as well as the core legal and policy arguments in favor of increasing access to the aggregated data held by risk regulators. It then discusses recent developments, including the stance taken by European risk regulators. Part III discusses lessons from the biopharmaceutical trial data experience and how some of these lessons may play out in current debates over diagnostic testing silos and overlapping rights.
BIOMEDICAL DATA: CHALLENGES OF CONCENTRATION AND FRAGMENTATION
Overlapping, Adjacent, and Fragmented Intellectual Property Rights
Both inadvertently and by design, overlapping and adjacent intellectual property rights pervade biomedical innovation. As an initial matter, patents and trade secrecy can overlap. To be sure, the text of the patent statute--which requires that a patent...