Order without Intellectual Property Law : Open Science in Influenza

Cornell Law Review

,)1*"

//1" September 2017

.0& )"

Order without Intellectual Property Law : Open

Science in Inuenza

Amy Kapczynski

Yale Law School

,)),30%&/+!!!&0&,+)3,.(/0 %7-// %,)./%&-)3 ,.+"))"!1 ).

.0,#0%" +0"))" 01).,-".043,**,+/

6&/.0& )"&/.,1$%00,4,1#,.#.""+!,-"+ "//40%",1.+)/0 %,)./%&-,.+"))3&$&0)"-,/&0,.40%/""+ "-0"!#,.

&+ )1/&,+&+,.+"))3"2&"34+10%,.&5"!"!&0,.,# %,)./%&-,.+"))3&$&0)"-,/&0,.4,.*,."&+#,.*0&,+-)"/" ,+0 0

'*- ,.+"))"!1

" ,**"+!"!&00&,+

*4- 54+/(& Order without Intellectual Property Law : Open Science in Inuenza ,.+"))  "2

2&))"0%7-// %,)./%&-)3 ,.+"))"!1 ).2,)&//

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 1 29-SEP-17 12:59

ORDER WITHOUT INTELLECTUAL PROPERTY

LAW: OPEN SCIENCE IN INFLUENZA

Amy Kapczynski†

Today, intellectual property (IP) scholars accept that IP as

an approach to information production has serious limits. But

what lies beyond IP? A new literature on “intellectual produc-

tion without IP” (or “IP without IP”) has emerged to explore this

question, but its examples and explanations have yet to con-

vince skeptics. This Article reorients this new literature via a

study of a hard case: a global influenza virus-sharing network

that has for decades produced critically important information

goods, at significant expense, and in a loose-knit group—all

without recourse to IP. I analyze the Network as an example

of “open science,” a mode of information production that dif-

fers strikingly from conventional IP, and yet that successfully

produces important scientific goods in response to social need.

The theory and example developed here refute the most pow-

erful criticisms of the emerging “IP without IP” literature, and

provide a stronger foundation for this important new field.

Even where capital costs are high, creation without IP can be

reasonably effective in social terms, if it can link sources of

funding to reputational and evaluative feedback loops like

those that characterize open science. It can also be sustained

over time, even by loose-knit groups and where the stakes are

high, because organizations and other forms of law can help

to stabilize cooperation. I also show that contract law is well

suited to modes of information production that rely upon a

“supply side” rather than “demand side” model. In its most

important instances, “order without IP” is not order without

governance, nor order without law. Recognizing this can help

† Professor of Law, Yale Law School. For their helpful input, I thank Bruce

Ackerman, Ian Ayres, Jack Balkin, Nancy Cox, Robert Ellickson, Heather Gerken,

David Grewal, Henry Hansmann, Oona Hathaway, Dan Kahan, Issa Kohler-Haus-

mann, Daniel Markovits, Tracey Meares, Lisa Larrimore Ouellette, Robert Post,

Carol Rose, Scott Shapiro, Reva Siegel, Talha Syed, John Witt, and commentators

at the Dalhousie Health Law Workshop, the Columbia Law School Legal Theory

Workshop, the Buffalo Law School faculty workshop, the University of Toronto

Innovation Law and Policy Workshop, and the Tri-State IP Scholars Workshop.

All errors, of course, are my own. I also thank Jimmy Zhuang and Patrick Lauppe

for their invaluable research assistance. Finally, my deep appreciation goes to the

many individuals associated with the Flu Network who generously shared their

time and insights with me, and whose work contributes so importantly to our

understanding and well-being.

1539

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 2 29-SEP-17 12:59

1540 CORNELL LAW REVIEW [Vol. 102:1539

us better ground this new field, and better study and support

forms of knowledge production that deserve our attention, and

that sometimes sustain our very lives.

NTRODUCTION

.......................................... 1540

I. I

NFLUENZA

, IP,

AND

PEN

CIENCE

................. 1551

A. Influenza: A Primer .......................... 1551

B. The Economics of Information Production in

Influenza .................................... 1556

II. F

RACKERS

................................... 1562

A. History and Structure ....................... 1562

B. Motivation in the Network ................... 1564

1. Government Funding and Motivation ...... 1566

2. Scientific Motivation in the Network ....... 1571

C. Crisis ....................................... 1578

D. Reconstruction and Rules ................... 1584

III. T

ODEL OF

PEN

CIENCE

..................... 1590

A. The Basic Model of Open Science ............ 1591

B. Open Science in Practice: Allocation ......... 1595

C. Open Science in Practice: Collaboration ...... 1599

IV. R

EORIENTING THE

IP W

ITHOUT

IP L

ITERATURE

........ 1607

ONCLUSION

........................................... 1612

NTRODUCTION

In August 1997, a three-year-old boy was admitted to a

hospital in Hong Kong, critically ill. Tests showed that he had

influenza, a virus that has circulated in humans for thousands

of years.

Far more striking was the type of flu that he had

contracted—an avian type, H5N1, that had never before been

seen in humans.

While Ebola and Zika have more recently captured the

headlines, there is no existing virus more dangerous than in-

fluenza.

In 1918-1920, a flu pandemic killed an estimated 50

to 100 million people around the world, most of them young

adults.

A similar pandemic today could take the lives of hun-

Christopher W. Potter, Chronicle of Influenza Pandemics, in T

EXTBOOK OF

NFLUENZA

3, 3 (Karl G. Nicholson et al. eds., 1998).

J.C. de Jong et al., A Pandemic Warning?, 389 N

ATURE

554, 554 (1997).

Michael T. Osterholm, Preparing for the Next Pandemic, 84 F

OREIGN

. 24,

26 (2005) (“[O]f the more than 1,500 microbes known to cause disease in humans,

influenza continues to be the king in terms of overall mortality.”). See also Sonia

Shah, Is Bird Flu Back?, N.Y. T

IMES

, Feb. 7, 2016, at 6 (noting that the H5N1

pandemic of 2009 killed an estimated 200,000 people).

See, e.g., K. David Patterson & Gerald F. Pyle, The Geography and Mortality

of the 1918 Influenza Pandemic, 65 B

ULL

. H

IST

. M

. 4, 19 (1991).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 3 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1541

dreds of thousands and perhaps millions of people in the

United States alone.

Influenza pandemics occur when radically new strains that

evade our immunities emerge and become transmissible be-

tween humans.

The new avian strain, therefore, was of major

concern to influenza scientists.

The one institution most criti-

cal to our ability to respond to it is a little-known network

called the World Health Organization’s “Global Influenza Sur-

veillance and Response Network,” or “GISRS”

(here, simply

“the Flu Network” or the “Network”).

The flurry of activity that followed the emergence of H5N1

shows the work of the Network at its most dramatic. Immedi-

ately after the boy was hospitalized, scientists from two Net-

work Collaborating Centers were sent as a rapid response team

to Hong Kong.

Working with local authorities and a WHO

Task Force, the team traced the virus to poultry in northern

Hong Kong, and traced the boy’s contacts to determine whether

he had recently had contact with sick chickens.

More cases

soon emerged, and new evidence suggested that the virus pro-

voked a catastrophic immune response in humans, frequently

causing organ failure and death.

Investigators worked ex-

haustively, putting in eighteen-hour days, knocking on doors,

tracing contacts, and gathering blood, sputum, and chicken

U.S. D

’

TOF

EALTH

& H

UMAN

ERVS

, HHS P

ANDEMIC

NFLUENZA

LAN

(Nov. 2005).

See infra Part I.

Michael Rosenwald, The Flu Hunter, S

MITHSONIAN

, Jan. 2006, at 183,

184–85.

Typically pronounced “gis–ris.”

See Gretchen Reynolds, The Flu Hunters, N.Y. T

IMES

, Nov. 7, 2004, at

41–42. See also Ren´e Snacken et al., The Next Influenza Pandemic: Lessons from

Hong Kong, 1997, 5 E

MERGING

NFECTIOUS

ISEASES

195, 197 (1999) (“Investigation

of the circumstances surrounding each case was undertaken by the local authori-

ties with assistance from the World Health Organization Collaborating Centers in

the United States and Japan.”); id. (noting that a WHO Task Force initiated a

“technical investigation and evaluation of the Hong Kong situation,” and that a

small staff from Japan and the US “join[ed] local authorities in collecting informa-

tion needed for risk assessment”); see also infra p. 126 (describing the role of

Collaborating Centers in the Network).

See Reynolds, supra note 9, at 42 (describing the investigation); Snacken et

al., supra note 9, at 197 (describing the Task Force). Contact tracing helps deter-

mine whether the virus has become transmissible between people, which can

inaugurate a new pandemic. See Reynolds, supra note 9, at 39.

K. Y. Yuen et al., Clinical Features and Rapid Viral Diagnosis of Human

Disease Associated with Avian Influenza A H5N1 Virus, 351 L

ANCET

467, 469–70

(1998).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 4 29-SEP-17 12:59

1542 CORNELL LAW REVIEW [Vol. 102:1539

droppings.

They shipped virus samples on dry ice to the

Network’s labs in Atlanta, Melbourne, London, and Tokyo for

analysis. The scientists at those labs too worked late into the

night, analyzing the thousands of samples shipped to them and

beginning the work needed to develop a vaccine.

The virus was not yet transmissible between humans, but

given how deadly it was, it was critical to slow the process by

which it could reproduce and mutate. The scientific team rec-

ommended that every chicken in Hong Kong be slaughtered,

and the government agreed.

In four days, 1.5 million chick-

ens were killed, effectively ending the epidemic in poultry, and

H5N1 in humans, in the city.

The moment was a dramatic

success for the Flu Network.

Though it is almost unknown, the WHO’s Flu Network is

critically important to global health. It provides our global ar-

chitecture for predicting and responding to pandemic flu, and

it also plays a central role in our response to seasonal flu,

which is a major public health concern in its own right.

The

Flu Network collects and examines millions of virus samples a

year, chooses the virus strains that go into flu vaccines, and

helps to refine seed viruses that are used to manufacture the

seasonal flu vaccine every year. If you have been vaccinated

against the flu, the Network, in a sense, is a part of you. Your

immune system carries its traces around.

A study of the Flu Network is valuable not only because of

its critical importance to global health. The Network also pro-

vides a pivotal example for current intellectual property (IP)

scholarship. IP scholarship has for decades been centered on a

simple account: IP is necessary to achieve the information pro-

Patricia Guthrie, CDC Virus Fighters Go Where the Action Is, A

TLANTA

ONST

., http://www.cdc.gov/excite/careers/fighters.htm [https://perma.cc/

UX7J-R9DD] (last visited Aug. 2, 2014); Reynolds, supra note 9, at 42.

Guthrie, supra note 12.

Reynolds, supra note 9, at 38; Rosenwald, supra note 7, at 184; Miriam

Shuchman, Improving Global Health—Margaret Chan at the WHO, 356 N

J. M

. 653, 655 (2007).

Kelvin K.W. To et al., The Emergence of Influenza A H7N9 in Human Beings

16 Years After Influenza A H5N1: A Tale of Two Cities, 13 L

ANCET

NFECTIOUS

ISEASES

809, 818 (2013).

Aleksandr S. Lipatov et al., Influenza: Emergence and Control, 78 J. V

IROL-

OGY

8951, 8955 (2004); Shuchman, supra note 14, at 654–55.

In the United States, for example, seasonal influenza causes, on average,

tens of thousands of deaths each year. See William W. Thompson et al., Mortality

Associated with Influenza and Respiratory Syncytial Virus in the United States,

289 JAMA 179, 185 (2003); William W. Thompson et al., Influenza-Associated

Hospitalizations in the United States, 292 JAMA 1333, 1333–34 (2004).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 5 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1543

duction that we as a society desire.

But over the last few

years, the field has come to recognize that IP as an approach

has both significant costs and substantial limits.

In re-

sponse, an important new scholarly literature on “intellectual

production without intellectual property,” or “IP without IP”

has emerged.

This new literature has two primary strands. The “norms-

based” strand argues that norms sometimes provide a viable

alternative to formal intellectual property law.

Case studies

See, e.g., R

OBERT

P. M

ERGES ET AL

., I

NTELLECTUAL

ROPERTY IN THE

ECH-

NOLOGICAL

11–17 (6th ed. 2012) (summarizing the standard efficiency-based

justification for IP); see also id. at 11 (describing this as “the dominant paradigm

for analyzing and justifying the various forms of intellectual property protection);

infra Part I. I use the term “information” here in the sense in which it is used in

the IP and economics literature, to refer to a broad range of immaterial resources,

from scientific formulas and data, to cultural products in their intangible form (for

example, musical compositions or the text of a novel).

For example, IP can be costly because exclusive rights in information can

be extraordinarily hard to define, trace, and value, making transaction problems a

still more serious issue here than they are in the realm of tangible property. It is

limited because it covers only some intellectual goods and indeed cannot be

extended symmetrically across different kinds of information goods. See infra

Part I.

Mario Biagioli coined the term “IP Without IP,” but it was made salient by

Rochelle Dreyfuss. See Rochelle Cooper Dreyfuss, Does IP Need IP? Accommodat-

ing Intellectual Production Outside the Intellectual Property Paradigm, 31 C

ARDOZO

L. R

. 1437, 1437 n.*, 1439 (2010).

See, e.g., Emmanuelle Fauchart & Eric von Hippel, Norms-Based Intellec-

tual Property Systems: The Case of French Chefs, 19 O

. S

. 187, 187 (2008)

(arguing that norms-based IP is used by French chefs to protect new recipes);

Dotan Oliar & Christopher Sprigman, There’s No Free Laugh (Anymore): The Emer-

gence of Intellectual Property Norms and the Transformation of Stand-Up Comedy,

94 V

. L. R

. 1787, 1809–12 (2008) (arguing that comedians’ social norms offer

significant protection for creators’ incentives to produce new content); see also

David Fagundes, Talk Derby to Me: Intellectual Property Norms Governing Roller

Derby Pseudonyms, 90 T

. L. R

. 1094, 1146 (2012) (discussing extralegal

regulation in roller derbies); Jacob Loshin, Secrets Revealed: Protecting Magicians’

Intellectual Property Without Law, in L

AW AND

AGIC

: A C

OLLECTION OF

SSAYS

123

passim (Christine A. Corcos ed., 2010) (describing the role of norms in sustaining

creation among magicians); Aaron Perzanowski, Tattoos & IP Norms, 98 M

INN

. L.

. 511, 577–84 (2013) (explaining why tattooers “have opted consistently for

informal social norms rather than the formal property-like rules of copyright

law”). This strand has been strongly influenced by the work of law and norms

scholars. See, e.g., Lisa Bernstein, Opting Out of the Legal System: Extralegal

Contractual Relations in the Diamond Industry, 21 J. L

EGAL

TUD

. 115, 124–30

(1992) (discussing the diamond industry’s reliance on internal norms and proce-

dures to handle disputes); R

OBERT

C. E

LLICKSON

, O

RDER

ITHOUT

: H

EIGH-

BORS

ETTLE

ISPUTES

176–78 (1991) (discussing how neighbors often settle

disputes outside of courts). The norms-based IP without IP literature should be

distinguished from a contiguous literature that questions the need for IP in vari-

ous industries, such as fashion, but that does not invoke norms as an alternative

means to support creative work. See Christopher J. Buccafusco, On the Legal

Consequences of Sauces: Should Thomas Keller’s Recipes Be Per Se Copyright-

able?, 24 C

ARDOZO

RTS

& E

. L. J. 1121, 1155 (2007); Kal Raustiala & Christo-

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 6 29-SEP-17 12:59

1544 CORNELL LAW REVIEW [Vol. 102:1539

in this vein have shown that groups such as magicians, stand-

up comedians, and French chefs successfully create immate-

rial goods, relying on informal norms that typically resemble IP

rights and that are backed up by reputational sanctions.

second “commons-based” strand of this literature has emerged

from Yochai Benkler’s work.

Drawing on examples such as

open source software and Wikipedia, Benkler argues that the

advent of networked digital computers has unleashed the po-

tential for more open and collaborative forms of information

production.

He describes commons-based production as

driven primarily by social motivation and/or strategies of indi-

rect appropriation.

It operates, he argues, via emergent coop-

eration and involves “rare use of formal processes; never of law

or managerial fiat.”

pher Sprigman, The Piracy Paradox: Innovation and Intellectual Property in

Fashion Design, 92 V

. L. R

. 1687, 1718–32 (2006). These studies focus not on

norms, but on special qualities of particular creative goods or markets that make

low-IP environments a stable equilibrium.

For example, one such account focuses on magicians, who invent new

magic tricks with regularity, although IP law provides little effective protection for

such tricks. The secret, the author argues, is in the informal norms that stand in

for—and seem to closely track—formal IP law. These norms, for example, forbid

the copying of tricks created by others, unless the tricks have been widely shared,

published, or sold. There is also a strong norm against revealing the secrets of

magicians to non-magicians. Violations of these norms can result in serious

reputational harm, and forms of retaliation, such as “not be[ing] invited to give

lectures, . . . perform in magic competitions, or [feature] in trade publications.”

Loshin, supra note 21, at 138.

See Y

OCHAI

ENKLER

, W

EALTH OF

ETWORKS

(2006); Yochai Benkler, Coase’s

Penguin, or, Linux and The Nature of the Firm, 112 Y

ALE

L.J. 369 (2002). The

banner of the commons has been taken up by many, and most prominently by

Madison, Frischmann, Strandburg. See Michael Madison, Brett M. Frischmann

& Katherine J. Strandburg, Constructing Commons in the Cultural Environment,

95 C

ORNELL

L. R

. 657 (2010).

ENKLER

, supra note 23, at 99–102. Benkler also refers to this mode as

“peer production,” which is characterized by “peers who interact and collaborate

without being organized on either a market-based or a managerial/hierarchical

model.” Benkler, Coase’s Penguin, supra note 23, at 381.

See id. at 424–25; B

ENKLER

, supra note 23, at 99. Indirect appropriation

facilitates the more open forms of production that Benkler describes in market

settings, such as when lawyers or open-source software firms sell services without

asserting exclusive rights over the information produced in the process. Benkler,

Coase’s Penguin, supra note 23, at 424–25.

Yochai Benkler, Commons and Growth: The Essential Role of Open Com-

mons in Market Economies, 80 U. C

. L. R

. 1499, 1554 (2013). Elsewhere he

recognizes that open source software licenses may play an important role in

stabilizing cooperation. See Benkler, Coase’s Penguin, supra note 23, at 445

(“This commitment would require specific licenses that secure access to work over

time to everyone, including contributors.”). Implicitly, he suggests that this is not

reliance on law in some sense, because it seeks only to secure the commons from

incursions from IP, so that in a world without IP, open source software might not

require even this form of recourse to law. See id.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 7 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1545

Skeptics, however, have remained unpersuaded that this

emerging literature reveals a robust or important alternative to

IP. They argue that the norms-based strand focuses on “niche”

settings, where the value of information is low and where

groups are close-knit.

If groups are large or loose-knit, and/

or where asset values or capital costs are high, in turn, they

argue that “IP without IP” will “tend to be replaced by property-

based arrangements.”

They contend, drawing on the existing

law and norms literature, that a system that relies on norms

backed up by reputational sanctions will work well only if peo-

ple are closely enough connected to one another to observe

rule-breaking and to care about their reputation in the group.

If the information goods being produced have more value, then

participants may also have stronger incentives to defect. If

capital costs are high, self-funding will be difficult. But without

property, groups may have difficulty accessing external fund-

ing. If, however, “IP without IP” can only operate in these set-

tings, then it truly may be a “niche” phenomenon that does not

deserve substantial attention.

The “commons-based” strand identifies more valuable

forms of creation that often emerge from large and dispersed

groups. But here too, capital costs tend to be low. With its

focus on intrinsic motivation and technology, the commons-

based literature also arguably offers a too-rosy conception of

unmediated social cooperation.

Critics point out that such

See Robert P. Merges, Economics of Intellectual Property Law, in O

XFORD

ANDBOOK OF

AW AND

CONOMICS

, at *7 (Francesco Parisi ed., forthcoming), http:/

/papers.ssrn.com/sol3/papers.cfm?abstract_id=2412251 [https://perma.cc/

EAE4-UFWW]; see also Jonathan M. Barnett, The Illusion of the Commons, 25

ERKELEY

ECH

. L.J. 1751, 1755 (2010) (arguing that “the sharing model works in

settings that are small in scale, size, value, and diversity”). The term “close-knit”

is important to the law and norms literature, because it describes the conditions

under which individuals, acting informally, are thought to be able to enforce

norms successfully. While it sometimes is taken to describe the small size of a

group, more accurately understood, a group is close-knit “when informal power is

broadly distributed among group members and the information pertinent to infor-

mal control circulates easily among them.” E

LLICKSON

, supra note 21, at 177–78;

see also Lior Jacob Strahilevitz, Social Norms from Close-Knit Groups to Loose-Knit

Groups, 70 U. C

. L. R

. 359, 359–360 (2003) (citing Ellickson’s definition and

contrasting close-knit groups with loose-knit groups).

Barnett, supra note 27, at 1755; see also Merges, supra note 27, at *7

(“[O]nce groups grow beyond a certain size, and the economic value of their activi-

ties passes some threshold, informal norms cease to work.”).

See, e.g., David A. Hoffman & Salil K. Mehra, Wikitruth through Wikiorder,

59 E

MORY

L.J. 151, 155 (2009) (arguing that work on commons-based peer pro-

duction “generally ha[s] not articulated a mechanism that would coordinate such

altruistic production”); Lior Jacob Strahilevitz, Review: Wealth Without Markets?,

116 Y

ALE

L.J. 1472, 1493–95 (2007) (arguing that Benkler does not “adequately

confront” the problem that commons-based communities are vulnerable to mali-

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 8 29-SEP-17 12:59

1546 CORNELL LAW REVIEW [Vol. 102:1539

systems do not address the power of bad actors—or “trolls”—to

sabotage cooperation.

They argue that market-proprietary

production is better able to discipline such actors and so is

more durable.

Contemporary empirical accounts of several of

the leading examples from the commons literature also show

them as more complex and fraught than first supposed.

Wikipedia today, for example, is governed by an extraordinary

array of rules and sanctions, and has an active and formal

internal dispute settlement system.

Finally, both strands of the “IP without IP” literature have

not clearly confronted the most analytically powerful aspect of

the market-based account: the claim that markets not only

produce information but also direct investments appropriately

toward collective social aims. The canonical defense of IP is

that it directs investment “efficiently,” because it permits dis-

persed market signals of preference to drive investments in new

information.

The strong version of this argument is clearly

problematic, given both the long-standing critiques of efficiency

as a value

and the pervasive problem of market failures in

information production.

But the canonical defense of IP does

at least respond to the normative sensibility that underpins the

field: innovative activities should both be reasonably effective

and serve socially defined priorities.

Magicians and comedians produce creative works, to be

sure, but do they produce enough of them, or the right kind,

from a social perspective? The norms literature has largely

cious users). This is perhaps particularly so given what we know about the

preconditions of cooperation in the tangible commons. See the Ostrom discus-

sion infra Part IV.

Strahilevitz, supra note 29, at 1495; Hoffman & Mehra, supra note 29, at

160.

Strahilevitz, supra note 29, at 1495.

See D

ARIUSZ

EMIELNIAK

, C

OMMON

NOWLEDGE

? A

THNOGRAPHY OF

IKIPEDIA

17–22, 29–84 (2014).

See infra notes 74–75; 81–82.

See, e.g., Ronald M. Dworkin, Is Wealth a Value?, 9 J. L

EGAL

TUD

. 191,

200–01 (1980) (providing a normative critique of wealth maximization).

See Amy Kapczynski, The Cost of Price: Why and How to Get Beyond Intel-

lectual Property Internalism, 59 UCLA L. R

. 970, 981–93 (2012).

These values are typically understood as welfarist in nature. See William

Fisher, Theories of Intellectual Property, in N

SSAYS IN THE

EGAL AND

OLITICAL

HEORY OF

ROPERTY

168, 177 (Stephen R. Munzer ed., 2001). In the United States

in particular, the field is structured by collective priorities rather than, say, con-

ceptions of authors’ rights. See, e.g.,M

ERGES ET AL

., supra note 18, at 11–12

(making this point and suggesting that this value inheres in the U.S.

Constitution).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 9 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1547

ignored the question.

If commons-based peer production is

driven in significant part by individual motives, may it some-

times produce the wrong kinds of goods from a social perspec-

tive? The commons literature tends to assume this question

away by describing commons-based creation as “intrinsically”

or “socially” motivated.

To counter critics, the field needs examples of “IP without

IP” that are capital intensive and produce valuable social

goods. The most informative examples will involve loose-knit

groups that are able to sustain cooperation over time, despite

threats from within and without. The most normatively attrac-

tive instances of “IP without IP” will also have mechanisms to

link investment to social aims.

This Article describes just such a case.

The Flu Network

is capital intensive, produces goods of enormous social value,

and has operated successfully for decades without any signifi-

cant recourse to intellectual property rights. It has succeeded

despite its global reach and loose-knit nature,

and despite

significant internal conflict, including a recent crisis so signifi-

cant that the Network nearly collapsed. That crisis triggered

Many of the central contributions in the literature do not address alloca-

tion at all. See Fagundes, supra note 21; Fauchart & von Hippel, supra note 21;

Loshin, supra note 21. But see Oliar & Sprigman, supra note 21, at 1856–57

(noting that the informal IP regime encourages jokes that are “point-of-view”

driven and so harder to steal). The leading examples in the literature focus on

information goods produced in market settings, such as comedy clubs and res-

taurants, and presumably would draw upon the same normative defense that the

market-exclusionary account draws on, that links market signals to social value.

But at times they also invoke intrinsic motivations, and could be read to suggest

that there are other values at stake in the system (the transcendence of art, for

example) that also shape its productive output. They also tend to say little about

the potential problem that without property, the goods in question may be pro-

duced but in inadequate supply. But see, e.g., Oliar & Sprigman, supra note 21,

at 1858–59 (implying that a greater role for formal IP in comedy might have little

net effect on the amount of comedy produced, because it would increase returns

but also transaction costs).

This tends to obscure the reality of social conflict and hierarchy, as well as

a tendency for individuals and groups to have particular—and sometimes exclu-

sive—interests. Many more of Wikipedia’s editors are men than women, and from

the global North rather than the South, likely contributing to significant gendered

and geographic disparities in coverage. See, e.g., Tom Simonite, The Decline of

Wikipedia, 116 MIT Tech. Rev. 51, 52 (2013) (“Among the significant problems

that aren’t getting resolved is the site’s skewed coverage: its entries on Pok[´e]mon

and female porn stars are comprehensive, but its pages on female novelists or

places in sub-Saharan Africa are sketchy.”); Wikipedia, Gender Bias, https://

en.wikipedia.org/wiki/Gender_bias_on_Wikipedia [https://perma.cc/CYH7-

GHLS] (last visited Sept. 8, 2016).

For more on the value of the case study approach and the methods that I

use here see Appendix A.

See infra subpart III.B.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 10 29-SEP-17 12:59

1548 CORNELL LAW REVIEW [Vol. 102:1539

years of high-level diplomatic negotiations and led to legal

agreements that provide an exceptionally good window into the

norms and rules that are central to the Network’s success.

The Flu Network, as I will describe, works via a distinct

model of information production that we can call “open sci-

ence.” Although a schematic description of open science has

existed for several decades, the model has received almost no

attention in the legal literature.

In open science, those mak-

ing new discoveries freely share them rather than excluding

others. In exchange, they earn reputational credit, which in

turn can lead to increased funding and salary. Information

sharing allows scientists to judge the quality of one another’s

This conflict also provides methodological advantages, because it facili-

tates “process tracing.” See Appendix A.

The foundations of the open science system were first described by the

sociologist Robert Merton, who famously emphasized the importance of priority

and sharing to scientific work. See generally Robert K. Merton, Priorities in Scien-

tific Discovery: A Chapter in the Sociology of Science, 22 A

. S

. R

. 635, 645–46

(1957) (making these points). Drawing directly on Merton’s work, Partha Das-

gupta and Paul David proposed a schematic model of open science two decades

ago. See Partha Dasgupta & Paul A. David, Toward a New Economics of Science,

23 R

. P

’

487, 499 (1994). Their model has been invoked sporadically in the

legal literature, but has never received sustained attention. See, e.g., Bernardita

Escobar Andrae, Scientific Productivity and Gender Performance Under Open and

Proprietary Science Systems: The Case of Chile in Recent Years, 19 A

. U. J.

ENDER

. P

’

& L. 799, 799–800 (2011) (briefly describing Dasgupta and

David’s account of open science); Jorge L. Contreras, Data Sharing, Latency Vari-

ables, and Science Commons, 25 B

ERKELEY

ECH

. L.J. 1601, 1622 (2010) (citing

David’s concept of open science in passing); Dan M. Kahan, The Logic of Reciproc-

ity: Trust, Collective Action, and Law, 102 M

ICH

. L. R

. 71, 91 (2003) (invoking

Dasgupta and David in a broader discussion of theories of collective action). The

importance of public funding to science has been recognized by some in the IP

literature, but as will become clear, understanding open science—what makes it

work well, or fail—requires far more than simple recognition of the role of public

funding in science. While the existing literature has taken note of Merton’s ac-

count of scientific norms, it also has focused dominantly on the obligation to

share, and not engaged with the more elaborate open science model—and with

potential failures within open science—the way I do here. See, e.g., Rebecca S.

Eisenberg, Proprietary Rights and the Norms of Science in Biotechnology Research,

97 Y

ALE

. L.J. 177, 180–81 (1987) (relying on Merton to argue that there is a

tension between norms in biotechnology and the recent expansion of patents in

that domain); Arti Kaur Rai, Regulating Scientific Research: Intellectual Property

Rights and the Norms of Science, 94 N

. U. L. R

. 77, 90 (1999) (invoking

Mertonian “communism” in science, and arguing that scientific sharing is plausi-

bly efficient and should be reinforced by law). Robert Merges and Kathy

Strandburg have sought to develop an understanding of science as distinct from

the market-exclusionary system, in ways compatible with the account I develop

here. See Katherine Strandburg, Norms and the Sharing of Research Materials

and Tacit Knowledge, in W

ORKING

ITHIN THE

OUNDARIES OF

NTELLECTUAL

ROPERTY

85, 92 (Rochelle C. Dreyfuss et al. eds., Oxford Univ. Press. 2010); Robert P.

Merges, Property Rights Theory and the Commons: The Case of Scientific Research,

13 S

. P

HIL

. & P

’

145, 146 (2009). Neither, however, engages with the

processes, cycles, and tensions of open science as I do here.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 11 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1549

work. Financial support typically comes from governments,

which rely significantly upon scientific reputation and priori-

ties to inform their decisions. Open science works very differ-

ently than the market-exclusionary model of information

production but, like that same system, has plausibly effective

mechanisms to mobilize decentralized information and to dy-

namically link investment to social aims.

The case study developed here helps validate the open sci-

ence model, but it also offers some important correctives to it.

For example, rules in open science—such as those mandating

credit or forbidding exclusive appropriation through patents—

are often more difficult to enforce than the basic model sug-

gests. There is also more potential for misalignment between

states’ interests and scientists’ interests than the basic model

predicts. The basic model also can give us no reason to think

open science will respond adequately to global aims or needs,

given the potential for conflict between funders and the simple

fact that many states lack the resources and/or the structures

of political accountability that are presupposed in the basic

model. Open science, as I describe, will in fact have difficulty

generating investment in accordance with global need, and will

tend to respond more to polities with more resources and with

more accountable governments.

A closer look at the example offered by the Network, how-

ever, also allows us to better understand how conflicts such as

these are managed in open science. Addressing these tensions

in the Network has required not merely norms, intrinsic moti-

vations, and technology—the resources stressed in the existing

“IP without IP” literature. It has also required recourse to orga-

nizations and to law. By linking an empirical analysis of the

Network to the most sophisticated body of literature that ana-

lyzes the informal cooperation of groups—the literature

spawned by Elinor Ostrom’s work on the commons—I show

that the Network has used organizations and law to serve many

of the same functions that are important to cooperation in

groups managing common-pool resources. Organizations and

law help clearly define group and resource boundaries, facili-

tate the monitoring and enforcement of rules, manage the in-

terpretation and revision of norms, and mediate the group’s

relationship to the “outside.” They can also be critical to the

cultivation of a shared ethos and of trust, which are also cru-

cial for cooperation, especially in loose-knit groups where en-

forcement is inevitably imperfect. Finally, organizations and

law in the Network have also helped those within it cultivate a

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 12 29-SEP-17 12:59

1550 CORNELL LAW REVIEW [Vol. 102:1539

sense of purpose. A close study of the Network, in fact, requires

us to move beyond a simplified rational actor model of open

science and “IP without IP,” as we begin to recognize the role

that law and organizations play in sustaining both.

The account developed here decisively refutes the most im-

portant criticisms of the “IP without IP” literature. First, it

shows that capital-intensive production without IP can be sus-

tained over time to produce information goods of extraordina-

rily high value. Generalizing from the Network’s experience, I

suggest that even where capital costs are high, creation with-

out IP can be effective and responsive to social aims, if reputa-

tional and evaluative circuits can be linked to a source of

capital, as in open science. Second, the account here shows

that loose-knit groups can cooperate, even under extraordinary

strain. But criticisms of the “IP without IP” literature have

some merit: that cooperation is not likely to be sustained under

strain by norms alone. Rather, where stakes are high and

groups are loose-knit, cooperation without IP likely requires

recourse to organizations and law. In its most importance in-

stances, order without IP will not be order without law, nor will

it be order without constraint or coordination.

This is emphatically not to say, as some critics have, that

“IP without IP” inevitably relies upon forms of exclusion that

“operate with an approximately equivalent effect” to IP itself.

In the Network, organizations and law are being used instead to

produce the kinds of governance and interpersonal goods de-

scribed by scholars like Ostrom and Elizabeth Anderson, in the

support of a production system that is more oriented toward

sharing rather than exclusion. The law deployed here is, in

addition, not conventional IP law but rather contract law. Al-

though contract law lacks some of the power of IP, it may be

perfectly adequate to sustain “IP without IP,” especially when it

supports not a demand-side model but a supply-side model

that is contingent upon cooperation among an identifiable

(even if not close-knit) group.

In the pages that follow, Part I provides background on

influenza, and describes the key information production tasks

required to respond to the threat of seasonal and pandemic flu.

Part II develops a detailed empirical account of the Flu Net-

work, based upon three-dozen interviews, as well as patent

analysis, citation analysis, analysis of sample and sequence

sharing, and archival research. Part III describes the system of

Barnett, supra note 27, at 1754.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 13 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1551

production that sustains the Network, drawing on the existing

model of open science. It then returns to the example of the

Network to clarify some of the allocative mechanisms (and po-

tential problems) of the basic model, and to describe how the

Network has succeeded, particularly in periods of enormous

strain. Part IV derives lessons from the case study, arguing

that the next wave of this literature should focus on a new set

of examples and also return to old examples, to see them anew

in the wake of the lessons that the Flu Network has to teach us.

NFLUENZA

, IP,

AND

PEN

CIENCE

A. Influenza: A Primer

Colloquially, we often call any sudden and severe respira-

tory illness “the flu.” True influenza, however, is caused by a

distinct family of RNA viruses.

It usually has mild effects in

its seasonal form—fever, aches, and respiratory symptoms that

last up to a week—but it can be extraordinarily lethal when a

new pandemic strain emerges.

A pandemic like the one in

1918-1920 could take the lives of nearly two million people in

the United States alone.

Influenza pandemics have emerged with some regularity

over the last century,

and new influenza pandemics today are

considered inevitable, given the nature of the virus and condi-

tions of contemporary agriculture and travel.

The core prob-

lem is the volatile nature of the influenza virus itself. Influenza

mutates constantly. Small mutations in the virus cause pre-

EORGE

EHNER

, I

NFLUENZA

: A C

ENTURY OF

CIENCE AND

UBLIC

EALTH

E-

SPONSE

23 (2012).

See supra notes 4–5.

See U.S. D

’

TOF

EALTH

& H

UMAN

ERVS

, supra note 5.

de Jong et al., supra note 2, at 554.

See W. Bruine De Bruin et al, Expert Judgments of Pandemic Influenza

Risks, 1 G

LOBAL

UBLIC

EALTH

178, 189–90 (2006) (discussing the likelihood of an

influenza outbreak); Richard J. Webby & Robert G. Webster, Are We Ready for

Pandemic Influenza?, 203 S

CIENCE

1519, 1520–21 (2003) (“Influenza experts agree

that another influenza pandemic is inevitable and may be imminent.”); see also

RESIDENT

’

OUNCIL OF

DVISORS ON

. & T

ECH

., E

XEC

. O

FFICE OF THE

RESIDENT

EPORT TO THE

RESIDENT ON

EENGINEERING THE

NFLUENZA

ACCINE

RODUCTION

N-

TERPRISE TO

EET THE

HALLENGES OF

ANDEMIC

NFLUENZA

, at v–vi (2010) [hereinafter

PCAST R

EPORT

], http://www.whitehouse.gov/sites/default/files/microsites/

ostp/PCAST-Influenza-Vaccinology-Report.pdf [https://perma.cc/7SWF-VACE]

(framing the importance of effective vaccine supply as a matter of when, not if, the

next influenza pandemic occurs); Patrick Adams, The Influenza Enigma, 90 B

ULL

ORLD

EALTH

. 250, 251 (2012) (describing risk factors thought to increase

the risk of pandemic outbreaks, such as intensification of agriculture).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 14 29-SEP-17 12:59

1552 CORNELL LAW REVIEW [Vol. 102:1539

dictable waves of sickness almost every year.

Seasonal flu is

less serious than pandemic flu, but still can be deadly to the

vulnerable, and particularly the elderly.

The virus can also

change in more radical ways, reshuffling its genome, typically

in an animal host, resulting in strains that are radically differ-

ent from any that have circulated before.

Pandemics occur

when radically new strains that evade immunities emerge, and

become transmissible between humans.

Two kinds of measures are critical to the public health

response to both pandemic and seasonal flu: “social counter-

measures” that aim to contain its spread, and “medical coun-

termeasures” such as vaccines and treatments that can

prevent infection and improve survival rates.

Social countermeasures range in intensity from, for exam-

ple, hand-washing to school closings and travel restrictions.

Their efficacy depends on characteristics of the pandemic

strain, as well as on how early the measures can be enacted.

The Network’s rapid reaction to H5N1 via the poultry cull in

Hong Kong was critical to stopping human cases there, for

example, and significantly slowed the advance of new human

cases to other regions.

Once a new strain emerges that is

transmissible between humans, models suggest it might still

be contained to a particular region, if it is only moderately

infectious.

A highly infectious strain, however, will inevitably

spread quickly around the world, even in the face of highly

intrusive restrictions such as bans on air travel.

Medicines that directly combat the influenza virus exist,

but recent studies suggest that they may alleviate some symp-

Alan J. Hay, The Virus Genome and Its Replication, in T

EXTBOOK OF

NFLU-

ENZA

43, 43 (Karl G. Nicholson et al. eds., 1998).

See Thompson et al., Mortality Associated with Influenza, supra note 17, at

185; Thompson et al., Influenza-Associated Hospitalizations, supra note 17, at

1333.

ILSCHUT

& J

ANET

E. M

LHANEY

, I

NFLUENZA

45–46 (2005). Pandemic

strains like the one that struck in 1919 have been traced to flu viruses that

circulated in birds, so scientists are particularly concerned when they see a new

avian flu that is capable of infecting a human. de Jong et al., supra note 2, at 554.

They are still more concerned when that flu can also infect other animal hosts,

and shows signs of being very lethal. See Rosenwald, supra note 7, at 183.

Arin Dutta, The Effectiveness of Policies to Control a Human Pandemic: A

Literature Review 40–44 (World Bank Dev. Research Grp., Working Paper No.

4524, 2008).

See infra text accompanying notes 192–93.

Dutta, supra note 52, at 40. Social control measures such as school clos-

ings may also slow the rate of spread of influenza. See, e.g., Gerargo Chowell et

al., Characterizing the Epidemiology of the 2009 Influenza A/H1N1 Pandemic in

Mexico, P

S M

., May 24, 2011, at 6–7.

Dutta, supra note 52, at 42.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 15 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1553

toms yet have no significant effect on flu-related complica-

tions.

Flu vaccines, therefore, are the most important

medical countermeasure against influenza. Indeed, they are

often described as the single most important tool in this con-

text,

because they are the only one that can provide pro-

longed individual protection against infection.

Influenza vaccines have been around since the 1940s, and

are still largely made in the same way that they were then—by

growing live influenza viruses in eggs, purifying and killing

them, and administering them by injection.

Such vaccines

appear to be reasonably or even highly effective, but only if they

are well matched to circulating strains.

The traditional mode of vaccine production, however, may

be inadequate in a pandemic. Egg-based production of vac-

cines is difficult to scale up quickly.

It is also slow, taking at

least five months, and at times up to a year.

With seasonal

flu, early strain selection leaves eight to nine months to pro-

Tom Jefferson et al, Neuraminidase inhibitors for preventing and treating

influenza in adults and children, Cochrane Database of Systematic Reviews 2014,

Issue 4. Art. No.: CD008965. DOI: 10.1002/14651858.CD008965.pub4.

See, e.g., Key Facts About Seasonal Flu Vaccine, C

TRS

FOR

ISEASE

ONTROL

& P

REVENTION

, https://www.cdc.gov/flu/protect/keyfacts.htm [https://perma

.cc/KL3Z-TXPD] (“An annual seasonal flu vaccine is the best way to reduce your

risk of getting sick with seasonal flu and spreading it to others.”).

Questions and Answers on Pandemic Influenza Vaccine, W

ORLD

EALTH

. (May 9, 2007), https://www.cdc.gov/flu/protect/keyfacts.htm [https://per

ma.cc/ VP57-S2ZA].

PCAST R

EPORT

, supra note 48 at 11; Interview with John McCauley, Direc-

tor, WHO Collaborating Ctr. for Reference & Research on Influenza, U.K. (Nov. 18,

2011).

See, e.g., Kristin L. Nichol, Efficacy/Clinical Effectiveness of Inactivated

Influenza Virus Vaccines in Adults, in T

EXTBOOK OF

NFLUENZA

358, 361 (Karl G.

Nicholson et al. eds., 1998) (collecting studies that report between 70 and 90

percent efficacy for years when the vaccine is well matched); Tom Jefferson et al.,

Vaccines for Preventing Influenza in Healthy Adults, 2010 Cochrane Database of

Sys. Rev. 1, 7 (2010) (estimating 73% efficacy in healthy adults); Michael T.

Osterholm et al., Efficacy and Effectiveness of Influenza Vaccines: A Systematic

Review and Meta-Analysis, 12 L

ANCET

NFECTIOUS

ISEASES

36, 39 (2012) (estimat-

ing 59% efficacy in healthy adults); Vaccine Effectiveness - How Well Does the Flu

Vaccine Work?, C

TRS

FOR

ISEASE

ONTROL

& P

REVENTION

, https://www.cdc.gov/

flu/about/qa/vaccineeffect.htm [https://perma.cc/CZ4R-XHW8] (“[R]ecent stud-

ies show vaccine reduces the risk of flu illness by about 50% to 60% among the

overall population during seasons when most circulating flu viruses are like the

vaccine viruses.”).

Egg-based production, for example, relies upon the availability of millions

of fertilized chicken eggs of a precise age, raised in special, sterile environments.

The number of eggs required is also enormous. The third-largest consumer of

eggs in the UK is a flu vaccine manufacturing plant. See PCAST R

EPORT

, supra

note 48, at 34.

A Description of the Process of Seasonal and H5N1 Influenza Vaccine Virus

Selection and Development, W

ORLD

EALTH

. 10 (Nov. 19, 2007), http://www

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 16 29-SEP-17 12:59

1554 CORNELL LAW REVIEW [Vol. 102:1539

duce the vaccine.

In a serious pandemic, however, this delay

significantly undermines the ameliorative potential of vac-

cines.

Egg-based production may also not work for a pan-

demic avian strain, because avian strains can kill the eggs in

which they are grown.

In the wake of H5N1, governments have therefore poured

enormous sums into improving vaccine manufacture, much of

this going to the private sector.

The new technologies under

experiment are numerous and complex,

but nearly all of

them would still depend upon the use of circulating virus

strains as the basis for the vaccine, and have received substan-

tial public sector support.

.who.int/influenza/resources/documents/influenza_vaccine-virus_selection/en

[http://perma.cc/8R27-G8UD] [hereinafter WHO Selection Description].

PCAST R

EPORT

, supra note 48, at 2.

For example, it was nearly six months after the identification of the H1N1

outbreak that the first doses of the pandemic vaccine became available, which

was too late to provide protection for most of the population in the United States.

Id. at 2. It would have taken nearly a year to produce enough vaccine to protect

the whole U.S. population against H1N1. Id. at vi.

See Anatole Krattiger et al., Intellectual Property Management Strategies to

Accelerate the Development and Access of Vaccines and Diagnostics: Case Studies

on Pandemic Influenza, Malaria and SARS, 2 I

NNOVATION

TRATEGY

ODAY

67, 76–77

(2006).

The U.S. Congress, for example, appropriated nearly $6 billion to improv-

ing pandemic preparedness in 2007. C

HARLES

E. J

OHNSON

, A

SSISTANT

’

Y FOR

. & T

ECH

., D

’

TOF

EALTH

& H

UMAN

ERVS

., R

EPORT TO

ONGRESS ON

ANDEMIC

NFLUENZA

REPAREDNESS

UNDING

1–2 (2007), https://www.medi-

calcountermeasures.gov/barda/documents/hhspanfluspending-0706.pdf

[https://perma.cc/8ZQP-JE7X]. Some of this money went to improve global sys-

tems such as the Flu Network. Id. at 2. About $1.5 billion went directly to private

sector companies, to help these companies improve manufacturing facilities and

techniques, and to meet regulatory demands, particularly in order to speed along

non-egg based technologies for pandemic vaccines. Id.; see also Kevin Freking, 5

Drug Companies Get Flu Vaccine Funding, W

ASH

. P

OST

(May 4, 2006, 8:40 PM),

http://www.washingtonpost.com/wp-dyn/content/article/2006/05/04/

AR2006050400893.html [https://perma.cc/C7G4-4QM8] (reporting that several

drug manufacturers received more than $1 billion in Federal funding in 2006 to

fund development for speedier mass production of virus vaccines).

For an accessible overview of recent influenza vaccine technology develop-

ments, see generally PCAST R

EPORT

, supra note 48. The most important of these

include cell-based manufacture (which is increasingly viable and could eliminate

the need for eggs), live attenuated vaccines (which are old but only recently used

more widely), and dose-sparing adjuvants. See id., at 34–35, 46.

There is some hope that a universal vaccine may be possible, which would

be based upon parts of the virus that do not change, and so provide broad and

perhaps long-term protection, and not rely on access to this year’s strains for the

vaccine. PCAST R

EPORT

, supra note 48, at 51–52. Recent research suggests that

this pathway may have promise, but it is still highly uncertain whether it will

succeed. Id. at 53–55; Antionetta Impagliazzo et al., A Stable Trimeric Influenza

Hemagglutinin Stem as a Broadly Protective Immunogen, 349 S

CIENCE

1301, 1301

(2015). Notably, public sector funding has also been key here: researchers at the

U.S. National Institute of Allergy and Infectious Disease are leading one of the

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 17 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1555

Synthesizing this, we can identify two key types of informa-

tional needs where influenza is concerned. The first is surveil-

lance. New strains of influenza emerge constantly, and

circulate around the world quickly.

To have effective vac-

cines, as well as to prepare hospitals and other health agencies

for particularly severe seasons or new pandemics, scientists

must constantly monitor the virus as it evolves around the

world. New viruses must be typed, so scientists can isolate

which strains are circulating and quickly identify any radical

new strains that might result in a pandemic outbreak. Scien-

tists also must track the health effects of different circulating

strains (identifying how many cases are showing up in hospi-

tals, how severe they are, and the like), to develop an accurate

picture of the dangers of different strains. As I will describe,

these are among the key tasks of the Flu Network.

A second set of informational needs relate to medical coun-

termeasures. Vaccines, in particular, are considered the most

urgent and promising domain of development. To generate an

effective vaccine for seasonal and pandemic flu, scientists must

select the appropriate strains to be included in vaccines, isolate

corresponding viral strains, and modify them to grow well in

industrial settings.

Scientists also need reagents that allow

evaluation of the potency of new vaccines—a critical regulatory

step, particularly because vaccines are produced in slightly

different form each year.

Last, there is a need for forward-

looking research that can help speed vaccine production, move

us away from egg-based production, and—though there is a

major new initiatives, and another involves key collaborators from the public

sector. See Hadi M. Yassine et al., Hemagglutinin-Stem Nanoparticles Generate

Heterosubtypic Influenza Protection, 21 N

ATURE

. 1065, 1065 n.1 (2015) (NIAID

team); Impagliazzo et al., supra note 68, at 1301 (involving researchers from

public sector).

Colin A. Russell et al., The Global Circulation of Seasonal Influenza A

(H3N2) Viruses, 320 S

. 340, 341 (2008). New strains have long been thought to

typically emerge in East and South-East Asia (where influenza circulates continu-

ously), and spread westward across the world, making surveillance in this region

a high priority. Id; but see Patrick Adams, The Influenza Enigma, 90 B

ULL

. W

ORLD

EALTH

. 250, 251 (2012) (providing evidence that viral patterns of circulation

may not be this predictable).

WHO Selection Description, supra note 62, at 7. To make seed strains for

avian strains (which kill eggs), scientists must use synthetic methods, via a pro-

cess known as “reverse genetics.” WHO Selection Description, supra note 62, at 9;

PCAST R

EPORT

, supra note 48 at 24–25.

See Conversations with the Director: Michael Shaw, C

TRS

FOR

ISEASE

ON-

TROL

& P

REVENTION

(Apr. 30, 2012), http://www.cdc.gov/about/cdcdirector/con-

versations/shaw.html [https://perma.cc/T5MT-YCY2] (last visited Aug. 2, 2014).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 18 29-SEP-17 12:59

1556 CORNELL LAW REVIEW [Vol. 102:1539

great deal of uncertainty about the viability of this approach—

ideally develop vaccines with more universal potential.

B. The Economics of Information Production in Influenza

The economics of information goods begin with the insight

that information has the qualities of a public good: it is non-

rivalous and also difficult to exclude in the absence of strong

legal entitlements (or, IP rights).

As a result, if we want “effi-

cient” levels of production of information goods in markets, we

may need exclusive rights regimes.

This forms the basis of the canonical justification for IP.

If information is expensive to produce but cheap to reproduce,

actors in competitive markets will not produce as much of it as

we wish from a social perspective. IP rights help create dy-

namic incentives for the production of information, because

they allow investors to exclude competitors for a period of time

and so recoup their investment.

It is also widely recognized, however, that IP generates inef-

ficiencies from both static and dynamic perspectives.

The

non-rivalrous nature of information means that its marginal

cost—the cost needed to produce an additional unit for con-

sumption—is zero.

In static perspective, only at a price of

zero, then, will information be consumed at socially optimal

amounts. IP also creates dynamic inefficiencies because infor-

mation is an input and output of its own production.

Supra-

marginal cost pricing suppresses uptake by producers.

This

is why leading information economists, harkening back to the

foundational work of Kenneth Arrow, actively debate the com-

See supra note 68.

See Kenneth J. Arrow, Economic Welfare and the Allocation of Resources for

Invention, in T

ATE AND

IRECTION OF

NVENTIVE

CTIVITY

: E

CONOMIC AND

OCIAL

ACTORS

609, 623–24 (Nat’l Bureau Comm. for Econ. Res., Comm. on Econ.

Growth of the Soc. Sci. Res. Council ed., 1962).

Efficiency in this context is usually defined via either the Kaldor-Hicks or

wealth-maximizing criterion. Fisher, supra note 36, at 177.

See M

ERGES

, M

ENELL

& L

EMLEY

, supra note 18, at 12–13; Mark A. Lemley,

Property, Intellectual Property, and Free Riding, 83 T

. L. R

. 1031, 1033,

1037–40 (2005).

See, e.g., B

ENKLER

, supra note 23, at 36–40 (outlining such inefficiencies);

UZANNE

COTCHMER

, I

NNOVATIONS AND

NCENTIVES

140 (2006) (same).

See Arrow, supra note 73, at 614–15.

For more on this problem, and additional drawbacks such as the potential

for racing, see Kapczynski, supra note 35, at 982.

See Suzanne Scotchmer, Standing on the Shoulders of Giants: Cumulative

Research and the Patent Law, 5 J. E

CON

. P

ERSPS

. 29, 30–31 (1991).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 19 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1557

parative efficiency of IP and various alternatives, such as gov-

ernment grants and prizes.

The most powerful defense of the superiority of IP over

these other approaches is based on a simple challenge: How

will governments know which information resources we want,

and how much we should spend?

The main virtue of IP over

other alternatives is said to be its allocative advantage: because

it relies on prices and markets, it is thought to be a good tool for

mobilizing decentralized information about demand.

Alloca-

tion, it turns out, is the key to economic arguments about

information production. Because information is non-rivalrous,

concerns about conservation or depletion that are important in

the tangible property context are here irrelevant. You cannot

deplete information, but you can waste the resources you in-

vest in developing information, or make poor decisions about

which information goods to invest in.

The allocative argument for the market-exclusionary ap-

proach is powerful. But recent scholarship on IP has sharp-

ened our understanding of its limits. For example, property

regimes require frictionless transactions if efficient allocation is

to occur. But transaction costs are particularly acute where

information is concerned, because of intense uncertainties re-

lated to the bounds of IP rights, as well as their value.

There

See Arrow, supra note 73, at 623; Brian D. Wright, The Economics of

Invention Incentives: Patents, Prizes, and Research Contracts, 73 A

. E

CON

. R

691, 691, 697–98, 701 (1983).

See Harold Demsetz, Information and Efficiency: Another Viewpoint, 12 J.L.

& E

CON

. 1, 19–20 (1969).

See id. at 12–13; W

ILLIAM

M. L

ANDES

& R

ICHARD

A. P

OSNER

, T

CONOMIC

TRUCTURE

NTELLECTUAL

ROPERTY

24 (2003).

For example, entitlements in information are often unclear, because the

metes and bounds of information are far more difficult to map than the metes and

bounds of tangible resources. See Brett M. Frischmann & Mark A. Lemley, Spil-

lovers, 107 C

OLUM

. L. R

. 257, 274–75 (2007). The bounds of associated legal

entitlements are also often unclear. See, e.g., id. at 275 (noting that “[i]t is diffi-

cult—and in many cases impossible—to know whether one is ‘trespassing’ upon

another’s IP right,” and describing the uncertainties generated by the idea/ex-

pression distinction and fair use in copyright, as well as the lack of an indepen-

dent invention defense in patent law). To this can be added major problems in

identification of owners, especially in copyright. See, e.g., David R. Hansen et al.,

Solving the Orphan Works Problem for the United States, 37 C

OLUM

. J.L. & A

RTS

4–14 (2013) (discussing the pervasive problem of being unable to locate, through

a reasonably diligent search, the owners of old copyrights). Information produc-

tion is also characterized by high levels of uncertainty, making informational

imperfections and asymmetries in bargaining particularly likely. See, e.g., Robert

Merges, Intellectual Property Rights and Bargaining Breakdown: The Case of

Blocking Patents, 62 T

ENN

. L. R

. 75, 83–84 (1994) (noting that strategic bargain-

ing problems may be especially common in the context of patent law because

advances in technology are particularly hard to value).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 20 29-SEP-17 12:59

1558 CORNELL LAW REVIEW [Vol. 102:1539

are many settings, additionally, where even if market actors

have strong IP rights, they will underinvest in information

goods. Externalities are one issue here: IP rights will under-

value solutions to pollution problems, for example, because of

pervasive externalities in the “market” for pollution.

Dis-

counting and the unpredictable nature of innovation are an-

other: IP works poorly to incentivize basic R&D, for example,

because of the high degree of uncertainty and long periods of

time before it yields commercial application.

Already implicit in this account are several important rea-

sons that market exclusion will not be adequate to produce

most of the information goods we need to address influenza.

The facts, data, and virus samples that are essential to high

quality flu surveillance are classic “basic” information goods.

They must be able to be accessed by many parties and recom-

bined readily to be useful, and it is hard to assign a value in

advance to any particular fact, datum, or sample. IP law also

provides very little protection for these kinds of information

goods.

This may not be a coincidence because an IP law that

did extend to these goods would function poorly to produce the

goods that we need. This is not only because of discounting

See Ian Ayres & Amy Kapczynski, Innovation Sticks: The Limited Case for

Penalizing Failures to Innovate, 82 C

. L. R

. 1781, 1812–22 (2015). On the

externality problem in IP more generally, see Frischmann & Lemley, supra note

83.

Economists have long concluded that markets are likely to underproduce

“basic” science. See Richard R. Nelson, The Simple Economics of Basic Scientific

Research, 67 J. P

. E

CON

. 297, 304 (1959); see also Dasgupta & David, supra

note 42, at 490 (discussing the economics of basic research and the divergence

between private and social returns to basic research outlays). The more upstream

a set of research inputs are, for example, the more significant transaction costs

problems are likely to be. Early stage research is likely to be unusually uncertain

in value, making licensing negotiations particularly difficult. See L

ANDES

& P

OS-

NER

, supra note 82, at 307; see also S

COTCHMER

, supra note 76, at 141 (emphasiz-

ing the problems of information asymmetry in licensing).

Patent law only covers new inventions and processes, and excludes from its

scope products of nature, such as wild-type viruses. See 35 U.S.C. § 101 (2016)

(making eligible for patents “any new and useful process, machine, manufacture,

or composition of matter”). Also unpatentable are abstract ideas, natural correla-

tions, and discovered facts. See Mayo Collaborative Servs. v. Prometheus Labs.,

Inc., 566 U.S. 66, 70–71 (2012). Finally, genetic sequences identical to those

found in nature are increasingly considered unpatentable. See Ass’n for Molecu-

lar Pathology v. Myriad Genetics, Inc., 133 S. Ct. 2107, 2116–19 (2013); D’Arcy v

Myriad Genetics [2015] HCA 35 (Austl.). Copyright law does not cover facts, and

provides no protection for databases absent some level of creativity in their crea-

tion. Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 348 (1999). Euro-

pean law protects databases under a sui generis regime in certain circumstances,

but this is unusual and not a very robust form of protection. See Directive 96/9/

EC of the European Parliament and of the Council of 11 March 1996 on the Legal

Protection of Databases, 1996 O.J. (L 77) 20, 25–26.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 21 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1559

and transaction cost dynamics, but also because much of this

information is highly non-excludable even in the presence of

property rights.

The second set of information goods, related to medical

countermeasures and vaccines in particular, are also difficult

to produce in sufficient amount in markets, but for different

reasons. Here, there is much more scope within existing IP law

for exclusive rights. For example, patent claims may cover

modified genetic material,

diagnostic reagents,

processes of

making vaccines or therapeutics,

and therapeutic agents

such as small molecule drugs or the viral particles that could

be important to universal vaccines.

For seasonal flu, patents

are a poor incentive, nonetheless, because the existing technol-

ogies work fairly well (given access to the surveillance goods

described above).

Pandemic flu is a far graver concern, but its unpredictabil-

ity creates significant barriers to market-led investment. A se-

rious new flu pandemic is considered inevitable, but rather like

a major earthquake, its timing is highly uncertain.

One

study surveying experts revealed that they predicted a 100%

chance that a new flu pandemic will occur in the next thirty

years.

But patents last only twenty years, and the dis-

See, e.g., Amy Kapczynski & Talha Syed, The Continuum of Excludability

and the Limits of Patents, 122 Y

ALE

L.J. 1900, 1902–05 (2013) (providing a theory

and examples of non-excludability and analyzing its implications for patent schol-

arship). One of the Network’s most important roles is to announce a new pan-

demic. Such information is extraordinarily valuable, but is not in any practical

way excludable, even if law permitted it to be treated as property.

See, e.g., Ass’n for Molecular Pathology, 133 S. Ct. at 2115 (concluding that

modified DNA is patentable subject matter); see also Directive 98/44, of the

European Parliament and of the Council of 6 July 1998 on the Legal Protection of

Biotechnological Inventions, 1998 O.J. (L 213) 13, 13 (stating in Article 3(2) that

“biological material which is isolated from its natural environment or produced by

means of a technical process may be the subject of an invention even if it previ-

ously occurred in nature”).

See U.S. Patent No. 5,156,949 (filed Dec. 24, 1987) (claiming reagent for

detecting HIV); U.S. Patent No. 6,074,816 (filed Sept. 16, 1994) (claiming reagent

for detecting hepatitis C virus).

35 U.S.C. § 101 (2012) (making eligible for patents “any new and useful

process, machine, manufacture, or composition of matter”).

Two recent major papers on advances toward universal vaccines, for exam-

ple, disclose patent applications related to the research in question. See Impag-

liazzo et al., supra note 68; Yassineet al., supra note 68.

Externalities play a role in undermining incentives for improvements as

well: vaccination confers benefits on others, and individuals may not take these

benefits fully into account when deciding whether to be vaccinated.

See supra note 48.

Bruine De Bruin et al., supra note 48, at 183–84.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 22 29-SEP-17 12:59

1560 CORNELL LAW REVIEW [Vol. 102:1539

counted present value for firms of profits many years away

quickly diminishes.

Moreover, in a true pandemic setting, with millions criti-

cally ill, patent holders would be under unthinkable political

pressure to voluntarily permit use of the patent. And in the

alternative, consistent with international law, countries could

simply override patents.

For example, at the highpoint of the

concern about a new avian flu pandemic, it became clear that

one suite of patents could be an impediment to public scien-

tists working to create reference strains for H5N1 vaccines.

The company holding the patents immediately publicly an-

nounced that it would license them to government organiza-

tions and developing countries at no cost, to develop vaccines

for public health purposes.

In fact, IP will plausibly be least efficacious for some of the

most important kinds of information goods. We might call this

an example of the “bell curve” of property law. Where property

is of low value, or very high value, private property systems

may be difficult to establish or administer. At low value, the

costs of the formal system will often be higher than the re-

wards.

But where the value of property is exceptionally

high—as would be a patent on a key vaccine component in a

global pandemic—states are unlikely to tolerate property-as-

sovereignty and will act on their own sovereign authority to

override it.

Therefore, it is of little surprise that, as implied above and

described in more detail below, the vast majority of the scien-

tific goods we need for influenza have long been produced not

See Agreement on Trade-Related Aspects of Intellectual Property Rights

art. 31, Apr. 15, 1994, Marrakesh Agreement Establishing the World Trade Or-

ganization, Annex 1C, 1869 U.N.T.S. 299 (providing processes for compulsory

licensing of patents).

NITIATIVE FOR

ACCINE

ESEARCH

, W

ORLD

EALTH

., M

APPING OF

NTELLEC-

TUAL

ROPERTY

ELATED TO THE

RODUCTION OF

ANDEMIC

NFLUENZA

ACCINES

(2007) (hereinafter IVR R

EPORT

2007).

Id.

See Harold Demsetz, Toward a Theory of Property Rights, 57 A

. E

CON

. 347, 356 (1967).

See, e.g., Dealing with Anthrax: Patent Problems Pending, T

CONOMIST

Oct. 27, 2001, at 14 (describing how, after the anthrax attack in the United States

in 2001, the Canadian government sought to buy a stockpile of generic anthrax

treatments, only relenting when the patent-holding company dramatically re-

duced the price and increased supply); Jill Carroll & Ron Winslow, Bayer Agrees

to Slash Price for Cipro Drug, W

ALL

. J., Oct. 25, 2001, at A3 (describing Bayer’s

dramatic reduction of the price for its anthrax drug to the U.S. government after

the HHS secretary threatened to procure generics despite Bayer’s patent after the

2001 anthrax attack).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 23 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1561

via markets but via government funding.

100

That funding

might be considered the answer to how information is pro-

duced without IP, but funding alone does not answer any of the

most acute questions posed by the IP literature: How can such

funding work well? What is its logic, when and how might it

succeed or fail, and how, in particular, can it address the criti-

cal allocative question at the center of contemporary informa-

tion economics, and the field of IP law?

The Flu Network provides an excellent case for the explora-

tion of these questions. It is a critical element of our global

scientific and public health infrastructure. It has proven re-

markably stable and successful over time.

101

It has operated

for decades without any significant recourse to intellectual

property. How then, has the Network motivated and organized

its work? It is to our case study of the Network’s long history,

crisis, and recent reconstruction that we now turn.

100

See supra note 66 (describing government funding of new vaccine technol-

ogies); see also infra subpart II.B (describing the activities and government fund-

ing of the Flu Network). There is some private financing, and a great deal of

private sector patenting (see Appendix B) in influenza research, focused almost

uniformly on medical counter-measures. But much of this receives public fund-

ing, for the reasons described above.

101

ORLD

EALTH

., S

TRENGTHENING

ESPONSE TO

ANDEMICS AND

THER

UB-

LIC

-H

EALTH

MERGENCIES

: R

EPORT OF THE

EVIEW

OMMITTEE ON THE

UNCTIONING OF

THE

NTERNATIONAL

EALTH

EGULATIONS

(2005)

AND ON

ANDEMIC

NFLUENZA

(H1N1)

2009 at 75–76 (2011) (in a report by an independent review committee, describing

the Network’s rapid response to the H1N1 swine flu outbreak, concluding that the

Network “had worked well and facilitated the timely detection, identification, ini-

tial characterization and monitoring of the pandemic (H1N1) 2009 virus,” and

noting that this was “the first time that a worldwide laboratory initiative was well-

coordinated for an extended period of time”). See Nancy J. Cox et al., Influenza:

Global Surveillance for Epidemic and Pandemic Variants, 10 E

. J. E

PIDEMIOLOGY

467, 469 (1994) (describing the Network as “quite successful”); Interview with

Alan Hay, Former Director, WHO Collaborating Ctr. for Reference & Research on

Influenza, U.K. (Nov. 18, 2011) (commenting that the Network “is still held up

as . . . a prime example[ ] of an international network that works”); see also J.M.

Wood, Selection of Influenza Vaccine Strains and Developing Pandemic Vaccines,

20 V

ACCINE

B40, B40 (2002) (noting the Flu Network as “one of the most success-

ful of the WHO programmes”). For evidence that its strain selection has worked

relatively well, see L. Steinbr¨uck, T.R. Klingen & A.C. McHardy, Computational

Prediction of Vaccine Strains for Human Influenza A (H3N2) Viruses, 88 J. V

IROLOGY

12123, 12128 (2014) (66% accuracy); see also Colin A. Russell et al, Influenza

Vaccine Strain Selection and Recent Studies on the Global Migration of Seasonal

Influenza Viruses, 26 V

ACCINE

D31, D32 (2008) (noting that “substantial mis-

match” between WHO strain selections and circulating strains is “infrequent” but

that strain selection could be further improved through better understanding of

global flu migration patterns).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 24 29-SEP-17 12:59

1562 CORNELL LAW REVIEW [Vol. 102:1539

RACKERS

A. History and Structure

When the World Health Organization was launched in

1947, one of its first initiatives was the creation of the Flu

Network.

102

With 1919 still very much in living memory, the

impetus was clear—by tracing and collecting new flu strains, if

a dangerous new strain emerged, “it might perhaps be possible

to prevent world-wide spread of the disease by means of pro-

phylactic immunization.”

103

The Network began, as it would operate for decades, in

largely informal fashion.

104

Several early participants were lu-

minaries in the scientific world, such as Jonas Salk, who in-

vented the polio vaccine.

105

Through personal contacts, a

small network of elite scientists was developed to share viruses,

techniques, and reagents across borders.

106

Since then, the Network has grown substantially and be-

come much more formal in nature. But its basic structure has

remained the same for decades. “National Influenza Centers”

(which I will call “national labs”) are responsible for the ground-

level work needed to create the “surveillance-related” informa-

tion goods described above. Today, there are more than 140 of

these labs, in more than 110 countries

107

—a number that has

been significantly and very deliberately increased over time.

108

These national labs work with local doctors and hospitals,

or with sentinel surveillance sites, to collect nose and throat

swabs of sick patients, and to test them for the influenza virus.

102

M.M. Kaplan, The Role of the World Health Organization in the Study of

Influenza, 288 P

HIL

. T

RANSACTIONS

OYAL

’

ONDON

ERIES

B B

IOLOGICAL

417, 417 (1980).

103

World Health Org., Priority Comm., Proposal for the Setting Up of a Com-

mittee on Influenza Made by the Representative of the Netherlands, U.N. Doc. No.

WHO.IC/P/1 (Apr. 3, 1947) (on file with author).

104

The Network’s institutional home was just “a couple of laboratory rooms

and some animal quarters” in London. Kaplan, supra note 102, at 418.

105

In the 1940s, Salk helped create the first successful influenza vaccine, and

was an early collaborator with the Network. C

HARLES

ERBERT

TUART

-H

ARRIS

EOFFREY

C. S

CHILD

, I

NFLUENZA

: T

IRUSES AND THE

ISEASE

165 (1976). Other

famous scientists involved with the Network included Sir Christopher Andrewes,

whose team had first isolated the influenza virus, and who led the London lab.

See Kaplan, supra note 102, at 129; MRC N

’

NST

FOR

. R

ESEARCH

, A C

EN-

TURY OF

CIENCE FOR

EALTH

233, 238–39 (2015).

106

Kaplan, supra note 102, at 417–19.

107

See National Influenza Centres, W

ORLD

EALTH

. (Jan. 12, 2017), http:/

/www.who.int/influenza/gisrs_laboratory/national_influenza_centres/list/en

[https://perma.cc/T753-P7BR] [hereinafter Network NICs].

108

Kaplan, supra note 102, at 419.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 25 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1563

Where the virus is present, they perform preliminary genetic or

antigenic analysis to identify the particular sub-strains circu-

lating.

109

They also collect relevant epidemiological data (for

example, about flu hospitalization rates) and provide weekly

updates on the influenza situation in their country during their

flu season.

110

After their preliminary analysis, national labs forward both

representative strains and novel strains on to a WHO “Collabo-

rating Center”—labs that also play a critical role in surveil-

lance, but that in addition generate information goods critical

to medical countermeasures to the flu. For decades, there were

only two CCs—one in London and the other in the US (now, at

the CDC)—which constituted the main scientific leadership of

the Network.

111

Today, the group is more diversified, with CCs

in Melbourne, Tokyo, Memphis, and most recently Beijing. The

CCs can analyze influenza viruses in much more depth, and

each is a very substantial research institution in its own

right.

112

Their terms of reference are also much more demand-

ing than for the national centers. They are required to do more

in-depth analysis of the viruses and data sent on by national

labs, and to share this information with the WHO and other

Network labs.

113

When outbreaks such as H5N1 occur, CCs

are required to help WHO member states by investigating and

responding.

114

They also create standard reagents and an-

tisera for labs around the world, which helps standardize diag-

nostic and laboratory practices and facilitates the regulation of

vaccines.

115

109

See WHO Global Influenza Program, Terms of Reference for National Influ-

enza Centers, W

ORLD

EALTH

., http://www.who.int/influenza/gisrs_labora

tory/national_influenza_centres/terms_of_reference_for_national_influenza_cen

tres.pdf [https://perma.cc/M8WD-N5CK].

110

WHO Selection Description, supra note 62, at 4.

111

Interview with Alan Hay, supra note 101; Kaplan, supra note 102, at 420.

112

For example, all CCs must have Biosafety Level 3 laboratories. See Core

Terms of Reference for WHO Collaborating Centres for Reference and Research on

Influenza, W

ORLD

EALTH

. (Oct. 12, 2006), http://www.who.int/entity/influ-

enza/gisrs_laboratory/collaborating_centres/whocccoretor2006.pdf?ua=1

[https://perma.cc/9RQS-46PF] [hereinafter WHO CC Terms].

113

See WHO CC Terms, supra note 112.

114

These duties are formalized in terms of reference, and labs can be de-

designated if they fail in their assigned tasks. For more on enforcement of these

rules, see infra pp. 168–70.

115

See WHO Selection Description, supra note 62, at 8; WHO CC Terms, supra

note 112. The CCs provide antisera and reagents for free to WHO labs, but may

ask for cost-reimbursement if for-profit clinics want to use them. See Conversa-

tions with the Director: Michael Shaw, supra note 71. Standardization of diagnos-

tic kits and other reagents can dramatically increase efficiency, much as

interchangeable parts on an assembly line do.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 26 29-SEP-17 12:59

1564 CORNELL LAW REVIEW [Vol. 102:1539

Perhaps most remarkable is the extensive involvement that

the CCs have in the creation of flu vaccines. Twice a year, the

CCs meet with other Network experts, with no manufacturers

in the room, and decide which strains should be included in

seasonal flu vaccines.

116

Their recommendations are made

public,

117

and while countries are not required to follow them,

in practice they almost always do.

118

When potential pandemic

strains emerge, a similar but much accelerated process is fol-

lowed, in an attempt to protect people and regions that are later

affected with a vaccine matched to the pandemic strain.

CCs and the Network’s “essential regulatory labs” (ERLs)

also play a critical role in the creation of information needed to

facilitate influenza vaccines. They help to make the modified,

“high-growth” virus strains that grow well in industrial set-

tings.

119

These seed strains are then transferred, free of

charge, to vaccine manufacturers (typically in the private

sector).

120

B. Motivation in the Network

For the Network to function well, governments must sup-

port it financially. And, its scientists must collect high-quality

samples, rapidly and accurately analyze them, and quickly for-

ward the relevant strains to the WHO Network. Also critical,

116

In a normal year, they meet in February to discuss the Northern hemi-

sphere, and in September to discuss the Southern hemisphere. See WHO Selec-

tion Description, supra note 62, at 6; Reynolds, supra note 9, at 39.

Manufacturers are excluded because there are possible conflicts of interest, for

example because companies may favor strains that are easier to produce or that

they have existing experience with. See Interview with Masato Tashiro, Director,

WHO Collaborating Ctr. for Reference & Research on Influenza, Japan, in Geneva,

Switz. (Nov. 16, 2011).

117

WHO Selection Description, supra note 62, at 6.

118

Interview with Terry Besselaar, WHO Global Influenza Surveillance and

Response Sys. (Nov. 8, 2011); Interview with Alan Hay, supra note 101; Interview

with John McCauley, supra note 59; see also WHO Selection Description, supra

note 62, at 9 (noting that sometimes a national regulatory agency will instead

require an antigenically similar strain).

119

WHO Selection Description, supra note 62, at 7 (describing three labs, one

of which is a WHO lab, and another that is closely linked to the Australian CC,

that undertake this work). When pandemic strains are involved, the process may

require synthetic production of seed strains, as happened with H7N9. Summary

of Status of Development and Availability of Avian Influenza A(H7N9) Candidate

Vaccine Viruses, W

ORLD

EALTH

. (May 25, 2013), http://www.who.int/entity/

influenza/vaccines/virus/candidates_reagents/summary_a_h7n9_cvv_201305

25.pdf [https://perma.cc/J8UM-AERP].

120

Interview with Terry Besselaar, supra note 118; Interview with Masato

Tashiro, supra note 116. Today, vaccine manufacturers benefitting from such

materials are expected to make a financial contribution to facilitate pandemic

preparedness. See infra note 153.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 27 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1565

and demanding, is the work done in the CCs to select the best

strains for vaccines, optimize seed strains, design reagents for

diagnosis, and evaluate vaccine quality. When a new pandemic

strain emerges, work in the Network requires enormous ef-

fort.

121

Why do countries support the Network? And why do

scientists in the Network, who are not compensated directly for

participating in the Network, nonetheless participate?

122

How

do they coordinate their work and mobilize effort with very little

centralized hierarchy? What generates the intensity of effort in

pandemic periods, what facilitates the broad collaboration that

the Network requires for its success, and what stimulates the

path-breaking scientific work that occurs in the most sophisti-

cated labs in the Network?

Here, and when discussing scientific motivation in the Net-

work, I rely significantly on interviews with Network partici-

pants.

123

Humans, of course, are complicated, and do not

always understand or accurately report their own motives.

124

Where possible, I therefore supplement interviews with data

from other sources, to determine if other evidence is consistent

with what is described by Network scientists.

125

Notably,

where other-regarding motives and norms are consistently re-

ported, we can also discern a social logic at work that demands

concern for others. Values in the Network, as I will later de-

scribe, are important, even though they do not invariably coin-

cide with behavior.

121

See, e.g., Interview with Michael Shaw, Senior Advisor for Lab. Sci., WHO

Collaborating Ctr. for Reference & Research on Influenza (Feb. 5, 2014) (describ-

ing long hours for CDC scientists when new pandemic strains emerge).

122

Network scientists typically are salaried employees whose compensation is

not keyed in a granular way toward their performance, or even to the hours that

they work. See Interview with Michael Shaw, supra note 121 (indicating that

scientists are not paid overtime at the CDC, though certain technicians are eligi-

ble for overtime). Scientists working for firms that operate in the market-exclu-

sionary mode may be compensated in similar ways, raising interesting questions

about whether R&D firms, on the inside, follow something more like a market

exclusionary model or something more like open science.

123

See Appendix A for details of the interview methodology.

124

People may, for example, have poor access to their true motivations, or if

they are aware of their motives, may be inclined to misreport them. They might

emphasize other-regarding concerns, for example, and minimize the degree to

which they act out of self-interest. For more on this, see Appendix A.

125

For example, if facts show (as they do) that scientists in the Network are

not directly compensated from the publication of a paper or sharing of a sample,

this fact is important to show that another form of motivation likely supports

these activities. Similarly, if scientists report open sharing of their data, it matters

that such reports are validated in the norms recently codified by the Network, and

in the databases that host such data.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 28 29-SEP-17 12:59

1566 CORNELL LAW REVIEW [Vol. 102:1539

1. Government Funding and Motivation

The Network’s activity does not correlate to a market-ex-

clusionary mode of information production. Formal intellec-

tual property rights have played no role in supporting the

Network’s activities, from its inception through to the present

day. Network labs have only rarely sought patents related to

influenza, and this only recently.

126

There is no evidence that

any of these patents have been licensed, and some Network

members have perceived the few Network-relevant patents that

they hold as “defensive” in nature.

127

As will soon become

clear, these patents did nonetheless have implications for the

Network. In the crucible of the 2005 return of H5N1, they

became a focal point for accusations that some labs were be-

traying the ethos and rules of the Network.

128

Instead of relying on IP, funding for Flu Network labs has

historically come almost exclusively from public sources—over-

whelmingly government, with modest additional support from

other non-profit sources such as foundations.

129

Indeed, gov-

ernment support is a precondition for designation as a Network

lab.

130

Because the vast majority of funding flows from na-

tional governments directly to their laboratories, there is no

centralized accounting of the annual running costs of the Net-

work. It has been estimated, however, at about $56 million.

131

The WHO’s budget for Network activities is much smaller,

around $2 million in 2011.

132

As this reflects, the WHO has

126

See Appendix B at 5–6.

127

Id. at 7.

128

See infra subpart II.C.

129

ORLD

EALTH

., S

TRENGTHENING THE

WHO G

LOBAL

NFLUENZA

URVEIL-

LANCE

ETWORK

(GISN), R

EPORT OF THE

EETING WITH

ATIONAL

NFLUENZA

EN-

TRES

(NIC

) H

ELD IN

AMMAMET

, T

UNISIA

13 (2011) [hereinafter S

TRENGTHENING

GISN

EPORT

], http://www.who.int/influenza/gisrs_laboratory/GISN_Meeting_Report_

apr2011.pdf [https://perma.cc/Y2CA-DBBP]; see also Interview with Alan Hay,

supra note 101 (“[M]ost of the national labs are being funded nationally.”); Inter-

view with Wenqing Zhang, Team Lead, Virus Monitoring, Assessment & Vaccine

Support (VMV), WHO (Oct. 1, 2013) (explaining that those that are not supported

directly by governments get funding through their status as academic institu-

tions, or through foundations).

130

Interview with Wenqing Zhang, Team Lead, Virus Monitoring, Assessment

and Vaccine Support (VMV), WHO (Nov. 15, 2011).

131

Id.; Pandemic Influenza Preparedness Framework for the Sharing of Influ-

enza Viruses and Access to Vaccines and Other Benefits, World Health Assembly

Res. WHA 64.5, at 21 n.1 (May 24, 2011) [hereinafter PIP Framework], http://

apps.who.int/gb/pip [https://perma.cc/SV9V-YCMK] (last visited Aug. 2, 2014).

132

Interview with Wenqing Zhang, supra note 130. Before H5N1 the WHO

provided no direct funding to national labs, but after H5N1 more resources were

provided, allowing the WHO limited funds for capacity building and targeted

training courses in certain developing countries. Id. Recent contributions from

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 29 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1567

historically been responsible for coordinating the Network,

while national governments directly support the labs.

133

Why do governments pay these costs, and why do they

participate in the Network? Critically, many governments, par-

ticularly in developed and wealthier developing countries, are

concerned about flu surveillance nationally and want high

quality vaccines.

134

They want national capacity, and de-

ploying this capacity for the purposes of the Network is a low

cost way to improve vaccines.

135

Being part of a global scien-

tific network is understood to increase the skills and informa-

tion-base of local scientists, and to give states a voice in the

transnational activities of the Network.

136

Developing countries that do not use the seasonal flu vac-

cine, but where pandemic strains have been active recently,

may see local influenza surveillance as an important invest-

ment as well.

137

This is less true, however, in countries with

industry have given the WHO additional resources that are now primarily directed

to capacity building in local labs. See infra note 153.

133

See W. Chas. Cockburn, The Programme of the World Health Organization

in Medical Virology, in 6 P

ROGRESS IN

. V

IROLOGY

175, 177–78 (J.L. Melnick ed.,

1964) (describing the Network in the 1960s as comprised of “national centres

[that] freely give their voluntary collaboration” and identifying WHO’s role as coor-

dination, standardization, and collection of information); Interview with Alan Hay,

supra note 101 (“WHO may have had some little startup money or things, but in

general, they don’t [fund the labs;] all the time that I’ve known, WHO didn’t give

anything really.”).

134

See, e.g., Interview with Anne Kelso, Director, WHO Collaborating Ctr. for

Reference & Research on Influenza, Austl. (Oct. 29, 2013) (describing Australia’s

motivation as including interest in the southern strain selection process and

pandemic preparedness).

135

See, e.g., Interview with Ian Gust, Former Member, WHO Expert Comm. on

Virus Diseases, Former Head, Austl. CC (Oct. 16, 2013) (“[E]ach of the laborato-

ries which is involved in the WHO Network is already involved in diagnostic or

public health work in their own countries for which they are being funded by their

national or state government. And it’s simply a matter of provision of information

that they’re already collecting or strains of viruses that they’re already collecting

and passing those onto a third party . . . [Participating in the GISRS is] not

without cost. But it’s usually built into the cost of running those centers.”).

136

See, e.g., Interview with Anne Kelso, supra note 134 (“[I]t’s to the country’s

advantage to be part of this Network, to get the inside information, to get the extra

assistance that might flow”); id. (citing the benefits of “having a seat at the table”);

Interview with John McCauley, supra note 59 (“I would suspect the benefits of

joining in [for countries] is actually just being part of the community and the

network and to actually know what’s going on and therefore be able to advise

when something happens.”). This is consistent with what others have observed:

governments tend to be willing to support transnational scientific collaborations

where they see the local benefits as sufficiently large to justify the local expendi-

ture. See C

AROLINE

S. W

AGNER

, T

NVISIBLE

OLLEGE

: S

CIENCE FOR

EVELOP-

MENT

107 (2008).

137

Interview with Keiji Fukuda, Assistant Director-General for Health, Sec.&

Environ., World Health Org., in Geneva, Switz. (Nov. 7, 2011) (“[T]his system has

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 30 29-SEP-17 12:59

1568 CORNELL LAW REVIEW [Vol. 102:1539

very limited human and financial resources, and where there is

no emergent pandemic strain or clear impact of seasonal flu.

138

For these reasons, national labs in some developing countries,

and particularly in sub-Saharan Africa, tend to be less well-

funded for influenza-related activities and concomitantly less

able to participate in some of the Network’s activities.

139

His-

torically, the WHO office provided little direct financial support

to national labs and instead provided support via the provision

of some reagents as well as training.

140

This has frustrated

some in the national labs, especially where they feel they have

insufficient capacity, for example in terms of the necessary

laboratory equipment, to provide what the WHO requests.

141

In the aftermath of the H5N1 crisis, the WHO began to

receive funds from donor governments for capacity building,

and so has begun to provide modest support to national labs,

primarily through a virus shipment fund where labs can send

influenza viruses to the WHO CCs for free as well as training

courses.

142

Donor governments also have at times funded na-

tional labs directly, or provided the WHO with earmarked funds

to meet particular needs of national labs.

143

The US CDC, for

example, has funded training and shipping fees for national

labs in the Network for a number of years.

144

Countries thus

are differently situated with respect to their own perceived na-

some role to play in detecting and preparing for a pandemic, so that affects

everybody, and that brings in developing countries as well as developed coun-

tries”); see also Interview with Yuelong Shu, Director, WHO Collaborating Ctr. for

Reference & Research on Influenza, China (Nov. 26, 2013) (explaining that China

is more concerned with pandemics than the seasonal flu).

138

For example, in tropical countries there is no seasonality to influenza,

making its burdens less obvious. See Interview with Masato Tashiro, supra note

116; C´ecile Viboud et al., Influenza in Tropical Regions, 3 PL

S M

. 468, 468

(2006).

139

See Interview with Terry Besselaar, supra note 118 (describing the capacity

of the Network in Africa as a significant issue, ameliorated to some degree with

new funds made available after H5N1); see also Interview with Ian Gust, supra

note 116 (describing some national labs historically as having provided very few

samples); Interview with JM Heraud, Head, Virology Unit & Nat’l Influenza Ctr.,

Madag. (Feb. 6, 2014) (describing the funding for his lab in Madagascar, which

comes almost entirely from donors, with very little from the government itself).

140

See Interview with Terry Besselaar, supra note 118.

141

Id.

142

Id. Post H5N1, resources from donor governments have increased, and

have been used by the WHO to address such gaps, for example by funding modest

capacity-building in labs particularly in Africa, and covering the shipping costs for

national labs that are sending viruses to CCs (paid for by a fund created by the

CDC). See Interview with Terry Besselaar, supra note 118; Interview with Wenq-

ing Zhang, supra note 130.

143

Interview with Terry Besselaar, supra note 118.

144

Id.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 31 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1569

tional interest in influenza itself, and that has historically gen-

erated a certain divide within the Network, with countries that

perceive little for themselves at stake providing less support for

their own scientists to participate in Network activities.

145

Network scientists also report that funding for Network

activities has been inconsistent. In the early years, WHO oper-

ated on “a shoestring,” with insufficient funds to keep staff on

long-term contracts and to consistently fund Network meet-

ings.

146

Funds for Network activities sharply increased after

the H5N1 outbreak, but decreased thereafter, in the shadow of

the global financial crisis.

147

The Network has also been af-

fected by the general funding situation of the WHO, which has

been extremely precarious in recent years.

148

Since the 1980s,

member states have withdrawn much of the WHO’s institu-

tional funding, leaving the agency—the Network’s WHO office

included—reliant on so-called “extra-budgetary” contribu-

tions.

149

These voluntary contributions by states and donors,

earmarked for particular projects, now make up nearly 80% of

the agency’s budget.

150

To address needs for funding for activities in periods where

state support is insufficient, Network scientists have some-

times proposed contributions from private companies.

151

For

most of the history of the Network, these were rejected, because

of concerns about conflicts of interest.

152

The Network’s new

legal Framework—the agreement brokered in 2011 to resolve

145

Interview with Keiji Fukuda, supra note 137 (describing not all countries as

sharing the same “level of interest” in influenza); Interview with Ian Gust, supra

note 135 (describing some national labs as historically having provided few sam-

ples to the Network).

146

Interview with Alan Hay, supra note 101.

147

Interview with Terry Besselaar, supra note 118.

148

See Kate Kelland, The World Health Organization’s Critical Challenge: Heal-

ing Itself, R

EUTERS

(Feb. 8, 2016, 11:55 AM), http://www.reuters.com/investi-

gates/special-report/health-who-future/ [https://perma.cc/P5CU-9P4S]

(describing the “budget pressures” facing WHO since the global financial crisis).

149

World Health Org., Sixty-Fourth World Health Assembly, Proposed Pro-

gramme Budget 2014–2015, U.N. Doc. No. A66/7, at 12 (April 19, 2013).

150

Id. The WHO also now relies heavily on non-state actors for support. The

Gates Foundation, for example, provides 10% of the WHO’s voluntary contribu-

tions, making it the third largest contributor after the US (16%) and the UK (11%).

World Health Org., Voluntary Contributions by Fund and by Contributor, 2015,

U.N. Doc. No. A69/INF./3, at 2, 8–9 (May 13, 2016).

151

For example, in the early 2000s, a proposal to obtain Network funding from

industry was rejected. Interview with Ian Gust, supra note 135.

152

Id.; see also Interview with Masato Tashiro, supra note 116 (describing

complex negotiations to ensure that the Australian CC was fully publicly funded,

after a vaccine manufacturing company with which it had ties was privatized).

There were some very minor exceptions. For example, the CC in London in recent

years had received a small amount of funding from industry to help support work

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 32 29-SEP-17 12:59

1570 CORNELL LAW REVIEW [Vol. 102:1539

the conflict—broke new ground by for the first time demanding

a significant financial contribution from industry. But it also

deliberately fenced this funding off to tasks involving pandemic

preparedness, with the intent that industry funds not displace

government funding of the Network.

153

Scientists in the Network have, over the years, found ways

to deliberately cultivate support from states for their work, for

example, by “showing [countries] that the lab capacity devel-

oped for flu is applicable for other diseases as well,” and “get-

ting across the message that developing capacity for influenza

helps you in other ways.”

154

The WHO has also supported and

facilitated burden-of-disease studies, so that scientists could

then “go into a country and show that this many people are

dying [and] you’re also having this dollar amount of impact in

your economy due to people missing work, having to stay home

and take care of sick children—you know, things that they

probably hadn’t even thought about.”

155

At a macro level, then, the priorities of the Network are

established by governments, when they make decisions about

how and whether to fund the labs. These decisions are deeply

informed by scientists, however, and scientists also make the

more granular decisions about Network priorities and prac-

tices. Particularly in its early years, the day-to-day governance

of the Network was delegated to scientists themselves. Still

today, scientists, who run the national labs and who occupy

positions in the WHO office, make key decisions.

156

The Net-

to make viruses adapted to their vaccine-making process. McCauley interview,

supra note 59.

153

See PIP Framework, supra note 131, art. 6.14.3, at 21; see also Interview

with Wenqing Zhang, supra note 130 (noting that the reasoning was that Network

labs “should be supported by the government”). Contributions began in 2011; the

total contributed by the end of 2015 was nearly $31 million. WHO, P

ARTNERSHIP

ONTRIBUTION

NNUAL

EPORT

2015 63 WHO/OHE/PED/2016.01. The funds are

administered by the WHO, and have been spent primarily to enhance laboratory

and surveillance capacity, with the remainder for activities such as risk-commu-

nication. Id. at 65. The partnership contribution has the potential, over time, to

both substantially increase the resources available to the Network, and to invite

more industry influence over the Network. The structure of disbursement, how-

ever, appears to be designed to minimize industry influence, with WHO deciding

on allocations and priorities, guided by an advisory board that includes no repre-

sentation from industry. Id. at 51.

154

Interview with Michael Shaw, supra note 121.

155

Id.

156

For example, significant decisions, such as the one to undertake a separate

annual recommendation for vaccines for the Southern Hemisphere, have histori-

cally always been made “by the WHO or through consultation with representatives

from the network and from the CCs, ERLs and National Influenza Centers.” Inter-

view with Wenqing Zhang, supra note 132. The Network also has an “expert

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 33 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1571

work itself is also uniformly described as a “voluntary” one.

157

As Hay described it, until the emergence of H5N1 and the re-

cent watershed negotiations, there were also very few formal

rules inside of the Network.

158

Decisions were not made hier-

archically, from the governmental level down, but organically

and horizontally, from scientist to scientist.

159

For example, in

the 1990s, the Network decided to begin making formal recom-

mendations for the composition of the flu vaccine twice a year

rather than once a year, to better serve the needs of Southern

countries (whose seasonal flu occurs during their winter,

which is our summer).

160

The decision was made by Network

scientists, led by emerging scientific practice and consensus

rather than a top-down directive from states.

2. Scientific Motivation in the Network

In interviews, Network scientists cited a variety of factors

when asked to describe why scientists contribute to the Net-

work, and why they exert such effort to advance influenza sci-

group” from the CCs that facilitates technical decision-making. Id. The most

significant outputs of the Network, such as the annual vaccine strain recommen-

dations, are made entirely by scientists, who meet annually at meetings convened

by the WHO. See WHO Selection Description, supra note 65, at 6.

157

See, e.g., Interview with Wenqing Zhang, supra note 132; see also Inter-

view with Alan Hay, supra note 101 (describing the Network as an informal ar-

rangement involving “anyone who wished to participate”). As experts in influenza,

scientists in the GISRS labs may also have influence on a state’s decision to

participate in the GISRS Network. See, e.g., Interview with Marilda Siqueira,

Head, WHO Nat’l Influenza Ctr., Braz. (Oct. 8, 2013). Historically, the Network’s

reach has extended across geopolitical lines. Both China and Iran, for example,

are longtime members, and even North Korea today has a national lab. See

Network NICs, supra note 107.

158

See Interview with Alan Hay, supra note 101; see also Interview with

Masato Tashiro, supra note 116 (describing the informal process of norm-making

or rule-making in the Network, as “traditional—not [a] rule but an agreement”).

159

For example, as Alan Hay, one of the longest-serving and most widely

respected influenza scientists in the Network, put it: “there [were] one or two cases

where their Ministry [of Health] might say [to a National lab], ‘Well, why do we

continue to support you to do this?’ And I have written one or two, three, I can’t

remember, letters of support to those labs as to why what they were doing was

important for the Network . . . . [B]ut . . . in many cases what these labs did were

determined by the people running the labs.” As Hay described it, “you take

responsibility for doing something, so you have to decide what is necessary for

you to do it . . . within the constraints of the budget. . . . And that’s just the nature

of the science that you’re trying to do on a day-to-day basis.” Interview with Alan

Hay, supra note 101. Hay has worked on influenza for decades, and ran the WHO

CC in London from 1993 to 2009. See Alan Hay, Director of the WHO Collaborat-

ing Centre for Reference and Research on Influenza at the Francis Crick Institute,

UK, I

’

Y FOR

NFLUENZA

& O

THER

ESPIRATORY

IRUS

ISEASES

, https://isirv

.org/site/index.php/board-members/10-board/55-alan-hay [https://perma.cc/

E3S4-7HJY].

160

Interview with Alan Hay, supra note 101.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 34 29-SEP-17 12:59

1572 CORNELL LAW REVIEW [Vol. 102:1539

ence, if not in the pursuit of financial gain. Three forms of

motivation or allegiance in particular were cited regularly. A

commitment to public health and an affinity for scientific in-

quiry were often described first, and a desire for credit and

scientific reputation added later. Scientists often described

several of these motivations together, one merging into the

next. They also often characterized these values as a historical

achievement, and as shared by some scientists more than

others.

Consider, for example, these statements from scientists

that are or were important in the CCs and in the Network’s

Geneva office:

It’s worked because you have a sufficient number of individu-

als that are keen to make it work. It doesn’t work because

WHO is telling people what to do, okay? It works because

people understand this is important and they want to be part

of it. . . . I think it partly is linked to the mentality of that time

[when the Network was created], which is different from the

mentality of today. . . . [W]hen we graduated, we didn’t think

about money. . . . [I]t was more[,] what were you going to

do.

161

[T]here was quite clear understanding in the past that what

we were doing was a public health role, a scientific role. You

know, part of it was fed by the scientific interest. If it didn’t

have interest, I wouldn’t have done it. You know, you weren’t

getting anything [i.e., compensation] out of it. And monitor-

ing what’s going on, there’s always some twists of scientific

interest that come out of it. And very often just at the techni-

cal level, in the assays you’re using that stop functioning and

things like that. So your continuing it feeds into your re-

search interests.

162

If you’re involved in public health your ultimate ambition is

to preserve or improve the health of the public. And this is

just one of those ways in which you can do it.

163

161

Interview with Alan Hay, supra note 101. For similar statements, see Inter-

view with Julian Druce, Head, Nat’l Influence Ctr., Austl. (Oct. 16, 2013) (“[I]t is all

for the greater good at the end of the day. And I think that everyone that partici-

pates in it does sense that and feel that.”); Interview with John McCauley, supra

note 59 (invoking the “global good” when describing why scientists share and

publicize their findings even before publications are accepted); Interview with

Masato Tashiro, supra note 116 (“Our motivation was only the contribution inter-

nationally for . . . vaccine development and selection. . . .”); id. (describing the key

motivation to contribute to the GISRS being “global public health safety”).

162

Interview with Alan Hay, supra note 101.

163

Interview with Ian Gust, supra note 135. Gust is also a long-time flu

scientist, who ran the CC in Australia. See Professor Ian Gust, L

ONDON

YGIENE

& T

ROPICAL

., http://www.lshtm.ac.uk/alumni/survey/profes-

sor_ian_gust.html [https://perma.cc/47HW-YM85] (last visited Feb. 11, 2017).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 35 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1573

[The Network started from the spirit of] public health needs,

because their memory of the Spanish Flu in 1918 still was

vivid at that time in 1950, or 1940. . . . [P]ublic health good,

collaboration, this type of spirit start[ed] this Network. . . .

[P]eople really were working together out of good willingness

and collaboration. So a lot of our work actually was con-

ducted automatically. . . . There was no legal document,

binding document . . . or people talking about “I’ll give you

this, what I could get in return?” . . . . So we were lucky it

started that way.

164

I think in a lot of cases you see it[—]especially in the develop-

ing countries[—]you see very dedicated people are actually

trying to do something good[.] [T]hey’re trying to make a con-

tribution to society. They’re trying to improve the health of

their fellow citizens. . . . [I]t’s an internal drive that’s hard to

define. It’s the excitement of facing a challenge and

succeeding.

165

[T]he genesis of it was pure public health. There’s no rea-

son . . . the UN would have gotten together to single out

influenza if it was simply a scientific or technical issue. It

was really the public health concerns about influenza, which

started the discussions. Nonetheless, the maintenance of the

system, particularly through much of the early decades, I

think, relied a lot upon scientific interest, interpersonal rela-

tions, institution to institution relations, and so on.

166

Scientists in the national centers likewise described motiva-

tions that merged the advancement of public health and the

language of scientific interest. As one Australian scientist de-

scribed it, sharing viruses in the Network provides a means

both to protect “the greater good of the population,” as well as

to allow scientists to “analyz[e] what’s around” and what might

go into the next vaccine.

167

A scientist in Brazil described simi-

larly both a sense of public importance of their virus sharing

work, but also a pleasure in the science, noting:

[I]n terms of benefits, . . . it’s a very interesting Network

because we can have in more or less real time what happens

in different parts of the world. . . . [T]hat’s a very dynamic way

to work. [You asked me] what are your benefits as a scientist

[to participating in the Network]. I think that sharing infor-

mation is, for me, is absolutely an amazing way to work.

168

164

Interview with Wenqing Zhang, supra note 130.

165

Interview with Michael Shaw, supra note 121.

166

Interview with Keiji Fukuda, supra note 137. For a brief bio, see Reynolds,

supra note 9, at 38.

167

Interview with Julian Druce, supra note 161.

168

Interview with Marilda Siqueira, supra note 157.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 36 29-SEP-17 12:59

1574 CORNELL LAW REVIEW [Vol. 102:1539

Notably, very similar accounts of the importance of both a

sense of public mission and of scientific inquiry to the Network

were offered by Network members decades ago.

169

When asked about why individuals participated in the Net-

work, scientists would refer to credit and self-interest, but also

to a sense of community, as well as the importance of values of

respect and fairness, particularly for those in the national labs.

As one long-time CC head put it,

[I]f you want people to cooperate in doing something which is

purely altruistic and not necessarily in their own day to day

interest and [that] causes them extra work, you need to make

sure that they realize that there’s value in what they’re doing,

that the information that they’re creating is an effort globally

and that you think they are wonderful and you hope that

they’ll continue doing it. . . . [I]t’s like any endeavor in life. If

you want people to do things for you, you have to be nice to

them, and make them feel wanted and trusted and valued.

170

Many others also noted that a sense of community among the

various labs was critical to the Network’s success.

171

The CCs

also described personal connections with national labs espe-

cially in their region, as well as trainings and technical sup-

port, as an important part of their role.

172

Through this

process, as one scientist described, CCs and national labs be-

come “intimately connected.”

173

It was clear from interviews

that many of the scientists in the Network, and particularly in

the CCs, knew each other well. One described the Network as

feeling “a bit like a family.”

174

Scientists also connected participation in the WHO system

with their ability to do their own scientific work, because WHO

accreditation, or credit from the scientific system of publication

169

Cockburn, supra note 133, at 184. The then-head of the Network at WHO

said this when describing its operation: “Virologists will always want as precise an

identification as possible of the viruses they isolate, not only because of the

epidemiological importance of such identification, but also because of the fact

that this is essential for the satisfaction of the natural curiosity of the scientist.”

Id.

170

Interview with Ian Gust, supra note 135.

171

See Interview with Terry Besselaar, supra note 118 (“[W]e try and keep in

touch with [Network scientists and] thank[ ] them for their contributions, it

doesn’t matter how small they are, . . . to make them feel part of the Network[.]

[A]nd we do try and invite them to training workshops etc., so at least they’re

getting something out of the system.”); Interview with Alan Hay, supra note 101

(describing the importance of “good personal relationship[s]” and “personal inter-

action,” as well as the mutual sharing of information).

172

See Interview with Masato Tashiro, supra note 116.

173

Id.

174

Interview with Terry Besselaar, supra note 118.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 37 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1575

more generally, helped support their efforts to fund their labs.

The Australian head of a national lab, for example, said: “You

do get kudos, you are able to say you’re a WHO National Influ-

enza Center. . . . [That has value for the individual, and] when

we apply to the government for this or that or the other, it

always helps . . . .”

175

Scientists in labs in Madagascar and

Brazil both noted that there were important connections be-

tween scientific credit, for example as marked by authorship or

acknowledgement in scientific papers, and government sup-

port for their work.

176

Both also cited examples of cases where data from their

labs had been used in publications without acknowledgement

of their efforts to gather the data. When asked why this was

problematic, one replied:

I think that sometimes they don’t realize the work that has

been [done] by the people in the country that collect data. . . .

I was really involved in setting up of the Network in Madagas-

car. Now I know how [much] work [and] negotiation with

people [it is]. . . . [S]ince it is a lot of work . . . all the people

who are working in the Network [should be able to say],

“Look, it’s my name, I am acknowledged.” . . . [P]eople are

proud of that. . . . And also we have to justify to the Ministry

of Health what we are doing. So imagine [that] all the papers

regarding Madagascar were published by other people with-

out acknowledgment. Look, when I’m going to the Minister of

Health and telling him we are doing this, we are doing that—

these [acknowledgements] are showing the value.

177

Quickly in conversations about motivation, then, issues of

credit emerge, in a language that reflects its importance both

as a reward in itself, and as a signal to funders, who use such

acknowledgement as evidence of the importance or success of

the lab they are funding.

178

A lead CDC scientist, when asked

why credit was important for scientists in the Network, said:

“Well, there are two drivers; one is just the personal satisfac-

tion of being recognized for what you’ve done. The other is of

course professional advancement; being able to put something

on your CV when you’re asking for promotion or applying for a

different position.”

179

While labs in the Network are supposed

175

Interview with Julian Druce, supra note 161.

176

Interview with JM Heraud, supra note 139; Interview with Marilda Si-

queira, supra note 157.

177

Interview with JM Heraud, supra note 139.

178

See Interview with Yuelong Shu, supra note 137 (describing credit as a

“fundamental or a basic principle” of the Network, because credit provides a form

of compensation to national labs for their hard work).

179

Interview with Michael Shaw, supra note 121.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 38 29-SEP-17 12:59

1576 CORNELL LAW REVIEW [Vol. 102:1539

to have committed financial support from their governments,

and thus are often less dependent upon grants than other

scientists, publication records and peer recognition can impact

promotions or the level of support a lab receives.

180

Beyond

that, as one scientist put it, “I think everybody likes to have a

nice fleshy CV. I mean, you being at Yale, I’m sure you under-

stand that.”

181

In the archive, one can find traces of scientists engaged in

close reading of the reports of the Network, and issuing correc-

tions when they believe that their contributions are insuffi-

ciently recognized.

182

Other documents also reflect the deep

importance of reputation, for example, with references to scien-

tists getting “very hot under the collar” if possible mistakes in

their work are aired publicly.

183

Finally, as discussed below,

one key complaint that brought the Indonesian crisis to a head

in 2007 was that the work of scientists in the national labs was

not being adequately recognized.

184

Some of the strongest evi-

dence of the importance of credit to Network scientists comes

from complaints that the basic scientific obligation to give

credit and to include those who provide materials in your col-

laborations and papers was not always scrupulously

respected.

185

As these sources reflect, there are a range of ways that

credit and recognition are granted in the Network. Some are

informal, as in the archival example noted above, where un-

180

Id.

181

Id.

182

See, e.g., World Health Org., Second World Health Assembly, World Influ-

enza Centre, Provisional Agenda Item 8.16.4.2, at 5, U.N. Doc. No. A2/62 (June

15, 1949), http://apps.who.int/iris/bitstream/10665/98934/1/WHA2_62_eng

.pdf [http://perma.cc/WR77-T9XY] (issuing a correction that acknowledges that

certain virus samples had been “isolated at the Institut Pasteur, Paris”),

183

Letter from C.H. Andrewes to Dr. C. Klimt (Nov. 6, 1950) (on file with

author) (describing the sensitivities of publicly suggesting that a scientists’ results

were the result of contamination).

184

See infra note 204 and accompanying text.

185

See Interview with Ian Gust, supra note 135 (describing appropriate credit

as something that sometimes people in the Network “overlooked or forgot” and

noting that “occasionally that caused resentment”); Interview with Anne Kelso,

supra note 134 (describing “cases where laboratories and the individual scientists

felt that they’d been, if you like, done over by other center people, Collaborating

Centers who might have then published data based on their viruses without

acknowledgement or without including them in the research. And this is particu-

larly awkward at the interface between surveillance and research. If a National

Influenza Center has sent viruses to a Collaborating Center for surveillance pur-

poses, and then the Collaborating Center does work that the National Influenza

Center sees as research and publishes a paper without them on it, then that’s

understandably disturbing.”); see also infra notes 203–12 and accompanying

text.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 39 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1577

published reports of the Network were in dispute.

186

Speakers

at Network events are also chosen in part to acknowledge and

reward excellence in the national labs.

187

Interviewees also

described a formal system of credit and accreditation that was

of great importance. Several noted that the WHO system of

designating official national labs and CCs was valuable as a

signal to other scientists, as well as governments, of the scien-

tific capacity of the lab.

188

Most frequently, however, refer-

ences to credit and reputation were linked to the system of

scientific publication. The current head of the London CC,

John McCauley, for example, described scientists in general as

wanting “recognition for observations and discoveries,” which

he described as “done through publication.”

189

The most in-

tense disputes about credit described above were about credit

given in published work, allocated either through co-author-

ship or through references in the acknowledgements section of

a paper. A “fleshy CV,” of course, also measures value largely

in terms of publications.

Scientists in the Network indeed publish regularly. A

search of the leading scientific citation index shows that over

80 percent of the scientists that led a Network lab in 2016 have

authored scientific articles, with over six in ten having pub-

lished articles on influenza.

190

This is particularly impressive

given the vast geographic diversity present in the Network, and

the fact that many of the national labs are funded more for

surveillance than research, and include flu only as one disease

among many that they study.

191

Publications records also ap-

pear to be in fact linked to funding in the way Network partici-

pants suggest: scientists in more well-resourced labs, and

186

See supra note 185 and accompanying text.

187

See Interview with Terry Besselaar, supra note 118. Often, in interviews,

scientists would also praise a particular scientist as especially skilled, or having

contributed in particularly important ways to the Network, offering an example of

the importance of informal reputational circuits to scientific reputation. For ex-

ample, several scientists went out of their way to praise the efforts of the Chinese

CC, in particular for quickly releasing all of its data on H7N9, on the same day

that the outbreak was publicly announced. See Interview with Anne Kelso, supra

note 134; Interview with Michael Shaw, supra note 121. As this suggests, I was

myself becoming part of the reputational circuit.

188

A key WHO official described the main motivation for scientists in the

Network as “recognition of involvement in public health,” which she noted comes

in significant part through both official recognition of Network labs. Interview

with Terry Besselaar, supra note 118.

189

Interview with John McCauley, supra note 59.

190

Appendix C, Table 3.

191

See id. For reasons described in the Appendix, this is also likely an under-

estimate of their contributions.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 40 29-SEP-17 12:59

1578 CORNELL LAW REVIEW [Vol. 102:1539

especially in the research-intensive CCs, tend to publish

more.

192

Both descriptions by Network participants and pub-

lishing patterns, then, confirm the importance of scientific

publishing to the modality of work undertaken in the Network.

C. Crisis

When avian influenza emerged, however, the normal prac-

tices of the Network were put under intense new pressure. The

rapid response to H5N1 in Hong Kong significantly slowed its

spread to other regions. But as expected,

193

H5N1 evolved and

re-emerged. It underwent reassortments and mutations that

made it tougher, and more deadly to a range of animals, includ-

ing humans.

194

By early 2004, it had killed millions of birds in

Thailand and Vietnam, and dozens of people.

195

WHO warn-

ings about the virus became increasingly ominous, particularly

as human cases turned up that could not immediately be

traced to poultry.

196

The symptoms in humans were also wor-

ryingly similar to those of the 1919 flu—not just fever and

coughing, but also encephalitis, acute respiratory distress, and

internal bleeding.

197

Indonesia reported its first case in 2005, and soon became

the epicenter of the disease in humans, with an average of five

new cases reported each month.

198

The fatality rate was

thought to be above 80%.

199

The median age of victims was

just twenty years.

200

The result was acute fear of a very deadly

pandemic that quickly began to radiate into political circles.

201

The Indonesian government scrambled to respond, and

soon found—to its great frustration—that it was unable to buy

192

The mean number of influenza publications for heads of Network labs

generally is 17.5, but for CC heads it is 176. Appendix C, Table 2; see also id. at

Figure 1. The number of influenza publications by scientists based in North

America and Europe is also much higher than, for example, the number by

scientists based in Latin America and Africa. See id. at Table 4.

193

Rosenwald, supra note 7, at 183 (citing Robert Webster, then head of the

CC at St. Jude).

194

See Laurie Garrett, The Next Pandemic?, 84 F

OREIGN

. 3, 11–12 (2005);

Osterholm, supra note 3, at 25.

195

Reynolds, supra note 9, at 38; see also Garrett, supra note 194, at 12.

196

Garrett, supra note 194, at 13–14.

197

Id. at 14.

198

Endang R. Sedyaningsih et al., Towards Mutual Trust, Transparency and

Equity in Virus Sharing Mechanism: The Avian Influenza Case of Indonesia, 37

NNALS

CAD

. M

. S

ING

. 482, 483 (2008).

199

Id. at 484. The fatality rate for H5N1 is now thought to be closer to 50%.

See Garrett, supra note 194, at 3.

200

Sedyaningsih et al., supra note 198, at 484.

201

See, e.g., Garrett, supra note 194, at 3–4.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 41 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1579

antiviral medicines, or H5N1 vaccines (to ship when they were

completed), because wealthy countries had pre-purchased all

the available supply of both.

202

In addition, the Indonesian

Health Minister discovered that one H5N1 vaccine being pre-

pared by an Australian company was based upon an Indone-

sian strain that had been contributed to the Network—and that

Indonesia was considered to have no rights either to the sam-

ples or to the resulting vaccines.

203

Finally, the Minister also

learned that research on Indonesian H5N1 strains had been

presented at international scientific meetings, with neither the

permission nor participation of Indonesians.

204

Though Indo-

nesia had been contributing virus samples to the Network

since the 1960s, in January 2007, the country announced that

it would cease its contributions to the Network until these con-

cerns were addressed.

205

As the Indonesian crisis quickly revealed, there were no

mechanisms in place to ensure adequate global supply of vac-

cines in the case of a pandemic, and nearly all manufacturing

capacity was concentrated in the global North.

206

This virtually

ensured that developing countries would have little if any ac-

cess to vaccines in a pandemic. In the firestorm of intense

advocacy that emerged thereafter, concern about patents, in-

cluding in the Network, also became a key focus. Though some

of those at the center of the Flu Network seem to have known

that sporadic patenting was happening before, that fact was

not widely known until now. Particularly influential was a re-

port commissioned by an NGO that was advising developing

countries, that identified many patent applications related to

H5N1, including some coming from WHO labs, that made use

202

See S

ITI

ADILAH

UPARI

, I

’

IME FOR THE

ORLD TO

HANGE

, I

N THE

PIRIT OF

IGNITY

, E

QUITY

AND

RANSPARENCY

: D

IVINE

AND

EHIND

VIAN

NFLUENZA

40–42

(2008); Sedyaningsih et al., supra note 198, at 486.

203

UPARI

, supra note 202, at 35–37.

204

Sedyaningsih et al., supra note 198, at 485. A WikiLeaks cable from John

Heffern, then Deputy Chief of Mission in Jakarta, confirms the importance of the

disrespect perceived by Indonesian scientists. See Indonesia - Avian Influenza

Sample Sharing Update, W

IKI

EAKS

(Apr. 13, 2007, 08:51AM), http://www.wiki

leaks.org/plusd/cables/07JAKARTA1053_a.html [https://perma.cc/N785-

ZYGX] (“Endang explained that NIHRD researchers want peer respect in the inter-

national research community.”)

205

UPARI

, supra note 202, at 25, 34–35; Sedyaningsih et al., supra note 198,

at 486.

206

World Health Org., Mapping the Global Vaccine Manufacturing Workforce:

Preliminary Results of a Survey Among Vaccine Manufacturers 5 (2011) (unpub-

lished draft report), http://www.who.int/phi/news/Draft_Survey_Report_Phases

1-2.pdf [https://perma.cc/K376-6RJD] (last visited Aug. 4, 2014).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 42 29-SEP-17 12:59

1580 CORNELL LAW REVIEW [Vol. 102:1539

of strains contributed by developing countries.

207

The report

warned that such patents “are resulting in a much more com-

plex and limiting field of intellectual property claims than has

ever before existed for influenza vaccine,”

208

and that flu-re-

lated patents had already shown the potential to hinder gov-

ernment efforts to prepare for a pandemic.

209

Existing patents

could in fact restrict or slow access to vaccines and therapeu-

tics in a pandemic,

210

though only one identified patent suite

held by a Network lab could have this power.

211

In the wake of

the recent campaigns to overcome patent barriers to HIV/AIDS

medicines, and patent mappings such as this one, developing

countries nonetheless found much reason for suspicion. As

the Indonesian Health Minister saw it, if the Network’s benefits

were distributed according to market logic, Indonesia would be

foolish to not to take its samples directly to companies in ex-

change for vaccines. She in fact explored such a deal, to give

viruses to a private company, and obtain vaccines for the state

in exchange.

212

There was an immediate outcry when Indonesia suspended

its cooperation with the Network, because of the dangerous

consequences for both flu surveillance and the collection of

vaccine viruses.

213

What followed was a five year long, very

high-profile global negotiation at the WHO to create a new

framework for the sharing of possible pandemic viruses,

214

in-

207

DWARD

AMMOND

, T

HIRD

ORLD

ETWORK

, S

OME

NTELLECTUAL

ROPERTY

S-

SUES

ELATED TO

H5N1 I

NFLUENZA

IRUSES

, R

ESEARCH AND

ACCINES

9 (2009). The

report was commissioned by Third World Network, which was a key NGO provid-

ing support to developing country governments during the negotiation. See Inter-

view with Sangeeta Shashikant, Legal Advisor, Third World Network, in Geneva,

Switz. (Nov. 4, 2011). For more on these patents see Appendix B.2.

208

AMMOND

, supra note 207, at 2.

209

For example, the report recounts an incident in which the U.S. government

was forced to threaten to use its march-in rights to get a license from MedImmune

for Sanofi to produce an H5N1 influenza vaccine, a privilege that other govern-

ments would not have had. H

AMMOND

, supra note 207, at 28.

210

See Appendix B.2.

211

See Appendix B.1 at 6–7 (describing St. Jude’s reverse genetics patents,

and the fact that the public sector was permitted to freely use this suite of pat-

ents—licensed to a company called Medimmune—during the avian flu scare).

212

See, e.g., S

UPARI

, supra note 202, at 27–28.

213

See, e.g., Richard C. Holbrooke & Laurie Garrett, Op-Ed., ‘Sovereignty’

That Risks Global Health, W

ASH

. P

OST

, Aug. 10, 2008, at B7.

214

Discussions were early on limited to viruses of pandemic potential, exclud-

ing those related to seasonal flu. See, e.g., World Health Org., Director-General,

Sharing of Influenza Viruses and Access to Vaccines and Other Benefits: Interdis-

ciplinary Working Group on Pandemic Influenza Preparedness, 3, U.N. Doc. No.

A/PIP/IGM/4 (Oct. 9, 2007) [hereinafter IGM-4], http://apps.who.int/gb/pip/

pdf_files/PIP_IGM_4-en.pdf [http://perma.cc/33N7-VPFL] (last visited Aug. 4,

2014). This was in order to focus attention on the issue of greatest public health

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 43 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1581

volving hundreds of health officials and high-level government

officials, including Ambassadors to the U.N.

215

Country positions of course differed from one another and

evolved over time, but broadly speaking, developing and devel-

oped countries were divided by two key issues. The first was

the nature of any benefit sharing. Developing countries in-

sisted that companies who benefitted from the Network should

be legally bound to contribute a certain amount of vaccines and

treatments to countries in need, and (for some) also to facilitate

the production of vaccines in developing countries through

transfer of technology and licensing of any relevant IP.

216

De-

veloped countries supported the creation of a fund and stock-

pile through WHO to improve access to vaccines, but insisted

that any benefit sharing through the Network should be

voluntary.

217

The second issue of major contention regarded intellectual

property. Developing countries took the position that Network

labs (and initially also any third-party recipients) should be

broadly forbidden to seek IP on Network materials, and also on

concern, and to limit the complexity of the negotiations. Interview with Keiji

Fukuda, supra note 137.

215

See, e.g., World Health Org., List of Participants of the Intergovernmental

Meeting on Pandemic Influenza Preparedness: Sharing of Influenza Viruses, U.N.

Doc. No. A/PIP/IGM/DIV/2 Rev. 1 (Nov. 22, 2007), http://apps.who.int/gb/pip/

pdf_files/PIP_%20IGM_DIV2Rev1.pdf [https://perma.cc/Z8WM-AE6S] (listing

representatives from 109 countries, the United Nations, and other agencies and

organizations); World Health Org., List of Participants in Open-Ended Working

Group of Member States on Pandemic Influenza Preparedness: Sharing of Influ-

enza and Access to Vaccines and Other Benefits, U.N. Doc. No. A/PIP/OEWG/

DIV/1 Rev. 1 (May 11, 2010), http://apps.who.int/gb/pip/pdf_files/OEWG1/

PIP_OEWG%20_DIV1Rev1.pdf [https://perma.cc/4UBT-A9Y8] (last visited Aug.

4, 2014) (listing representatives from 79 countries).

216

See, for example, Indonesia’s Proposal requiring benefit sharing, and tech-

nology and know-how transfer. World Health Org., Sharing of influenza viruses

and access to vaccines and other benefits: Interdisciplinary Working Group on

Pandemic Influenza Preparedness, 3–4, U.N. Doc. No. A/PIP/IGM/5 (Nov. 19,

2007) [hereinafter Indonesia Proposal], http://apps.who.int/gb/pip/pdf_files/

PIP_IGM_5-en.pdf [https://perma.cc/8HY2-CMW6] (last visited Aug. 4, 2014).

See also, for example, Thailand’s Proposal requiring legally binding obligations on

manufacturers to contribute vaccines, as well as suggesting technology transfer

and know-how requirements. World Health Org., Sharing of influenza viruses

and access to vaccines and other benefits: Interdisciplinary Working Group on

Pandemic Influenza Preparedness, 6–8, U.N. Doc. No. A/PIP/IGM/6 (Nov. 19,

2007) [hereinafter Thailand Proposal], http://apps.who.int/gb/pip/pdf_files/

PIP_IGM_6-en.pdf [https://perma.cc/7YQE-3VG2] (last visited Aug. 4, 2014).

217

See, e.g., World Health Org., Secretariat, Pandemic Influenza Prepared-

ness: Sharing of Influenza Viruses and Access to Vaccines and Other Benefits, 3,

U.N. Doc. No. EB 126/4 (Dec. 10, 2009) [hereinafter Secretariat Report], http://

apps.who.int/gb/ebwha/pdf_files/EB126/B126_4-en.pdf [https://perma.cc/

5NUV-B8P8] (last visited Aug. 5, 2014).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 44 29-SEP-17 12:59

1582 CORNELL LAW REVIEW [Vol. 102:1539

broadly defined derivatives thereof.

218

Developed countries, in

turn, resisted any restrictions on IP either within or outside the

network.

219

A third set of tensions that emerged during the negotia-

tions related to the practices of scientists within the Network.

Developing countries wanted clear commitments that Network

labs and third-party recipients would acknowledge the contri-

butions of the scientists in national labs, and seek to involve

these scientists in subsequent work.

220

Scientists also became

more aware of the existence of patents inside of the Network

during the negotiations, and the matter raised concerns. As

Alan Hay described, suspicions emerged that “some people

were making a lot of money out of” patents related to their

Network work.

221

Others reported that there were suspicions

from some in developing countries that the CCs were selling

viruses to industry.

222

Although there is no evidence that ei-

ther was correct,

223

it spoke to an emerging mistrust in the

Network—particularly coming from developing countries and

their scientists—that began to threaten its practices of free

information sharing.

With so many contentious issues on the table, the negotia-

tions were protracted and difficult. Commentators expressed

deep concern that the process would fail, leaving the Network

218

See, e.g., Thailand Proposal, supra note 216, at § 12(a)(viii) (prohibiting

recipients of GISRS materials, inside or outside the Network, to assert IP rights on

any products derived from, or that incorporate, GISRS materials); Indonesia Pro-

posal, supra note 216, at 4 (same).

219

Developed countries, for example, insisted on a narrow definition of GISRS

materials, and rejected the notion that anything except wild-type viruses—which

would not generally be subject to patents as such—could be excluded from IP

protection. World Health Org., Director-General, Pandemic influenza prepared-

ness: sharing of influenza viruses and access to vaccines and other benefits,

Outcome of the resumed Intergovernmental Meeting, 12, U.N. Doc. No. A62/5

Add.1 (May 18, 2009), http://apps.who.int/gb/ebwha/pdf_files/A62/A62_5A

dd1-en.pdf [https://perma.cc/24LL-YTNL] (last visited Aug. 5, 2014); IGM-4,

supra note 214, at app. 3 paras. 4, 7, 30.

220

See, e.g., Indonesia Proposal, supra note 216, at 4 (discussing credit and

involvement); World Health Org., Sharing of Influenza Viruses and Access to

Vaccines and Other Benefits: Interdisciplinary Working Group on Pandemic Influ-

enza Preparedness, 3–23, U.N. Doc. No. A/PIP/IGM/7 (Jan. 4, 2008) [hereinafter

Africa proposal], http://apps.who.int/gb/pip/pdf_files/PIP_IGM_7-en.pdf

[https://perma.cc/QA89-N5E9] (same).

221

Interview with Alan Hay, supra note 101.

222

Interview with Masato Tashiro, supra note 116.

223

As noted in Appendix B, the only evidence I have found of any Network lab

earning income from patents involves the reverse genetics patents at St. Jude,

which are not specific to influenza and arguably stem from basic research that

was much broader than the CC’s Network activities. See Appendix B at 7.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 45 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1583

fundamentally undermined.

224

The 2009 “swine flu” pan-

demic, caused by an H1N1 virus, however, played a critical role

in reigniting the negotiations. It increased the salience of the

risk of a new flu pandemic, and generated renewed political will

to find a resolution.

225

Two new co-chairs were appointed at

this point, who brought the parties back together and engaged

in unprecedented outreach to industry, who were brought in to

speak directly to negotiators.

226

A key breakthrough in the negotiation was described by

one co-chair as “almost philosophical.”

227

As he recounted, it

came at a moment when all participants came to agree “that

the principle of both virus sharing and benefits sharing should

go together,” and that the “idea was . . . to make the system a

fair system,” and to ensure that the benefits—mainly vac-

cines—to be shared would go not to individual countries, but to

the WHO.

228

At a key moment, the logic of reciprocity that

characterized the Network, then, was reestablished.

The final breakthrough came in April 2011, when with the

final hours of the negotiation running out, the new co-chairs

gathered together key countries—the United States, United

Kingdom, France, Germany, Finland, Australia, Canada,

China, Brazil, Indonesia, India, Egypt, and Turkey—to propose

a “take-it-or-leave it” package with three key elements.

229

Two

were “music to the ears of the developing countries”: Industry

would contribute half of the running costs of the Network to the

224

See, e.g., David P. Fidler, Negotiating Equitable Access to Influenza Vac-

cines: Global Health Diplomacy and the Controversies Surrounding Avian Influenza

H5N1 and Pandemic Influenza H1N1, PL

S M

., May 2010, at 1, 3 (“The negotiat-

ing path that could lead to a new global access framework for influenza vaccines is

not apparent . . . .”); Rachel Irwin, Indonesia, H5N1, and Global Health Diplomacy,

LOB

. H

EALTH

OVERNANCE

, 2010, http://eprints.lse.ac.uk/28272/1/Irwin_Indo

nesia_and_Global_Health_Diplomacy.pdf [https://perma.cc/L6 TQ-PPD2] (noting

that “[a]s the virus-sharing negotiations have continued for three years without

resolution, some of the urgency and political will has been lost”).

225

Interview with Juan Jos´e G ´omez Camacho, Permanent Representative of

Mex. to the United Nations, in Geneva, Switz. (Nov. 18, 2011) (remarking that

“after H1N1 . . . the awareness and the pressure for these processes to be finished

obviously increased, because it was clear that we were not playing with hypotheti-

cal games”); Interview with Gaudenz Silberschmidt, Head of Int’l Aff., Swiss Fed.

Office of Pub. Health, in Geneva, Switz. (Nov. 17, 2011) (remarking that H1N1

“made it easier to move to real commitments in the . . . negotiations”).

Silberschmidt has worked for the WHO since 2012, and gave the interview in his

previous role working for the Swiss government.

226

Interview with Juan Jos´e G ´omez Camacho, supra note 225; Interview with

Gaudenz Silberschmidt, supra note 225.

227

Interview with Juan Jos´e G ´omez Camacho, supra note 225.

228

Id.

229

See id.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 46 29-SEP-17 12:59

1584 CORNELL LAW REVIEW [Vol. 102:1539

WHO, and would donate a certain amount of real-time produc-

tion of pandemic vaccines to the WHO.

230

The third was criti-

cal for developed countries: the Framework would include no

compulsory transfers of patents or know-how from companies,

which developing countries had demanded.

231

While IP restric-

tions are imposed inside the Network, there are no mandatory

IP-related conditions imposed by the Network on outsiders.

232

One by one, each country announced that they were willing to

take the deal.

233

The lone hold-out was the United States. The

next morning, the United States announced that it too would

agree. According to sources intimately involved in the negotia-

tions, but who did not wish to be named, the about-face was

the result of late-night phone calls between negotiators and the

CEOs of various multinational companies, who in turn reached

out to U.S. officials in Washington who had not been aware

that the industry supported the deal.

The result was the path-breaking Pandemic Influenza

Preparedness Framework, or “PIP Framework,” officially

adopted by the WHO General Assembly a few weeks later.

234

The Framework included important new rules, but also codified

existing rules inside of the Network. As such, it is an excellent

place to look to understand what was critical to the Network’s

reconstruction, as well as what aspects of the Network’s rules

are most important to its operation.

D. Reconstruction and Rules

The crisis in 2007 was triggered primarily by tensions be-

tween developing and developed countries, rather than among

Network scientists, and its resolution required significant inno-

vations in the Network’s legal framework. The informality that

had once characterized virus exchanges in the Network was

replaced by a highly formal system requiring licenses to accom-

pany the transfer of all potential pandemic strains. The Frame-

work developed two “standard material transfer agreements,”

which operate as running contracts that mandate certain re-

sponsibilities for those who receive viruses from the Network,

one for insiders, and another for Network outsiders.

230

See id.

231

See id.

232

As I will note, firms have an option, in fulfilling their benefit sharing obliga-

tions, to contribute benefits via voluntary licensing instead of vaccine or drug

donations, but this is not mandatory.

233

See Interview with Juan Jos´e G ´omez Camacho, supra note 225.

234

PIP Framework, supra note 131.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 47 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1585

All transfers of virus material to non-Network entities are

governed by the outsider license. Over substantial opposition

from developed countries, these contracts were rendered le-

gally enforceable: disputes are to be referred to binding inter-

national arbitration.

235

The most important obligation of the

license applies to manufacturers of vaccines that benefit from

the Flu Network, who are obliged to commit to benefit sharing

in return. While firms have options, it is widely believed that

they will chose to meet their obligations by donating approxi-

mately 10% of their real time production in a pandemic to the

WHO, and reserving another 10% of such production “at af-

fordable prices” to the WHO.

236

These goods are to be distrib-

uted by the WHO according to public health need.

237

Industry

that benefits from the Network also must contribute financially

to the Network, to “improv[e] global pandemic influenza

preparedness and response,” to the tune of approximately $23

million a year.

238

For insiders, a separate contract was drafted, supple-

mented by standard terms of reference for Network labs.

239

These agreements largely track practices in the Network that

had evolved informally. In their emphasis, however—and in

particular in a much-debated provision over IP—they are re-

vealing. Three salient obligations for labs emerge from the

Framework.

The first highlights the importance to Network scientists of

recognition and collaboration. According to their model terms

of reference, labs in the Network must “actively seek the partic-

ipation of scientists . . . from originating laboratories and other

235

PIP Framework, supra note 131, annex 2, art. 5, at 35; see also Interview

with Sangeeta Shashikant, supra note 207 (discussing the United States’ resis-

tance to binding arbitration).

236

PIP Framework, supra note 131, annex 2, art. 4.1(A1), at 34. Other options

include licensing and technology transfer to developing country manufacturers.

Manufacturers of anti-retroviral medicines that rely on the GISRS will have a

similar legal obligation, either for licensing or the donation of “at least X treatment

courses of needed anti-retroviral medicine.” Id. annex 2, art. 4.1(A3), at 34.

237

Id. art. 1.8, at 3 (“[T]he benefits arising from the sharing of H5N1 and other

influenza viruses with human pandemic potential should be shared with all Mem-

ber States based on public health risk and need.”).

238

Id. art. 6.14.3.1, at 22. The amount is articulated at 50% of the running

cost of the Network, so should increase as those running costs increase. Id. art.

6.14.3, at 21–22. Counter-intuitively, given the metric chosen, the “contribution”

was intended not to help defray the existing running costs of the GISRS, but

rather to “improv[e] global pandemic influenza preparedness and response.” Id.

art. 6.14.3.1, at 22. For more on its implementation, see supra note 153.

239

Id. annex 1, art. 1.1, at 29 (designating as subject to the license “influenza

laboratories that have been designated or recognized by WHO and have accepted

to work under agreed WHO terms of reference”).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 48 29-SEP-17 12:59

1586 CORNELL LAW REVIEW [Vol. 102:1539

authorized laboratories, especially those from developing coun-

tries, in scientific projects [and publications] associated

with . . . clinical specimens . . . from their countries . . . .”

240

The second highlights the tensions around IP that emerged

in the Network. This rule was new, though it may not require

significant departure from previous practice. It provides that

Network labs “should” not seek to obtain IP “on” a carefully

defined set of Network “biological materials.”

241

The definition

of biological materials includes not only wild viruses, but also

virus isolates, modified candidate vaccine viruses, and certain

cDNA.

242

The restriction thus clearly reaches certain patent-

eligible subject matter.

243

Though couched in the discretion-

ary language of “should,” there is some evidence that the WHO

is committed to treating the rule as mandatory.

244

Third, the Framework imposes obligations on labs in the

Flu Network to share materials and information, both with one

another and with the general public.

245

National labs are obli-

gated specifically to “maintain active communication and col-

laboration with other members of the [Network],” as well as to

inform WHO, national authorities, and the public quickly when

potential pandemic strains emerge.

246

CCs are given special

obligations to share information back to national labs, and “to

ensure that up-to-date information and findings of public

health significance are rapidly exchanged . . . .”

247

240

Id. annex 1, art. 5.2, at 30–31; see also id. annex 4, at 40–41 (same); id.

annex 5, at 45, 48, 52 (same). Outsiders are also required to “appropriately

acknowledge,” for example in publications and presentations, the contributions of

WHO laboratories that provide biological materials to the Network. Id. annex 2,

art. 4.3, at 35; see also id. annex 1, art. 5.3, at 31 (requiring insider recipients to

“acknowledge in presentations and publications” contributions by other Network

scientists and laboratories).

241

Id. annex 1, art 6.1, at 31; see also id. art. 4.1, at 8 (defining “biological

materials”).

242

See id. art. 4.1, at 8.

243

Less clear is whether it bars patents of the sort that were previously ob-

tained or sought by the Network. This depends on interpretation of both the

Framework and the complicated claims of the relevant patents or applications.

For an overview of existing evidence on Network patents and patent applications,

see Appendix B.

244

Compare PIP Framework, supra note 131, annex 1, art. 6.1, at 31 (“Neither

the Provider nor the Recipient should seek to obtain any intellectual property

rights (IPRs) on the Materials”) (emphasis added), with Interview with Sangeeta

Shashikant, supra note 206 (noting that WHO Director General Chan reportedly

gave countries personal assurances that GISRS labs would not seek such IPRs).

245

See PIP Framework, supra note 131, annex 4, at 41; id. annex 5, at 45.

246

Id. annex 5, at 51–52.

247

Id. annex 5, at 45; see also id. annex 5, at 46 (obliging CCs to share

information about H and N subtyping, as well as gene sequences, not only with

other CCs but also with originating labs).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 49 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1587

All Network labs are also obliged to submit genetic se-

quence data to publicly available databases in a timely man-

ner.

248

This codified a norm that emerged only in 2006, when a

prominent influenza scientist who was not a member of a Net-

work lab objected to the earlier practice of depositing Network

sequence information in a password protected database at Los

Alamos.

249

Only a limited number of labs, most of them WHO

labs, had access to this site, perhaps reflecting the concerns of

contributing countries, and perhaps arising out of the desires

of some scientists to keep their information close to allow them

to benefit through exclusive publications. Just a few months

later, a powerful coalition of scientists, including the heads of

the CCs, announced a new initiative to ensure that all H5N1

sequences would be quickly entered into public-access

databases, housed in an organization created for the purpose,

called the GISAID.

250

The new database, EpiFlu, binds all

users of the data with a click-wrap contract to two key condi-

tions, designed to address the fairness concerns raised by de-

veloping countries and their scientists in the negotiations.

Recipients are first obliged to give credit to originating labs and

make “best efforts” to collaborate with them, and second are

required not to seek patents on any data or fraction of data

obtained from EpiFlu.

251

EpiFlu has become the repository of

choice for many Network labs, particularly in developing coun-

tries.

252

Some Network labs, including the CDC, however, pre-

fer to use a public database that imposes no restrictions on

users, when compatible with their obligations to those who

share samples with the Network.

253

248

See id. annex 4, at 41; see also id. annex 5, at 46 (obliging CCs to share H

and N and other gene sequences of potential pandemic viruses “to a publicly

accessible database in a timely manner but no later than three months after

sequencing is completed, unless otherwise instructed by the laboratory or country

providing the clinical specimens and/or viruses”).

249

Martin Enserink, As H5N1 Keeps Spreading, A Call to Release More Data,

311 S

. 1224, 1224 (2006).

250

See Martin Enserink, Pushed by an Outsider, Scientists Call for Global Plan

to Share Flu Data, 313 S

CIENCE

1026, 1026 (2006). The Indonesians were also

early adherents to the GISAID model. See S

UPARI

, supra note 202, at 20–21.

251

See Registration Form for Individual Users, GISAID, http://platform.gisaid

.org/epi3/frontend#4b51bf [https://perma.cc/6WGL-SLZR]. This too is struc-

tured as a running contract, so that data may not be transferred to anyone not

bound by the same agreement.

252

See Lisa Schnirring, Pandemic Reveals Strengths of New Flu Database,

FOR

NFECTIOUS

ISEASE

ESEARCH

& P

OLICY

(June 25, 2009), http://www

.cidrap.umn.edu/news-perspective/2009/06/pandemic-reveals-strengths-new-

flu-database [https://perma.cc/7N45-X7XC].

253

See Interview with Michael Shaw, supra note 121.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 50 29-SEP-17 12:59

1588 CORNELL LAW REVIEW [Vol. 102:1539

Distilling all of this and combining it with other informa-

tion we have about the Network, we can discern a key set of

rules that govern information exchange and use within and

outside the Network, that, as far as evidence permits us to see,

are often followed. Inside of the Network, labs are reciprocally

obliged to share data and information with one another. These

obligations reflect what leading scientists in the Network de-

scribe as longstanding practices of free exchange within the

Network.

254

As Hay put it: “if somebody calls me up and asks

me a question, I give them an answer. . . . Now who else gives

information away freely? . . . [I]t’s just a different way of think-

ing.”

255

This norm was articulated also by others, who de-

scribed rapid sharing of samples and information as “one of the

most important rules” of the Network.

256

The archives also

reveal that the expectation that viruses and data will be shared

freely has a long history in the Network. Accusations that a

scientist has failed adequately to share virus samples with

others are taken very seriously, and with some umbrage.

257

Information exchange is also seen as a primary benefit by

scientists themselves. John McCauley, head of the London CC,

254

See Interview with Alan Hay, supra note 101 (noting that giving “informa-

tion freely” was the Network’s responsibility); Interview with John McCauley,

supra note 59 (remarking that “there has always been free exchange of human

influenza viruses and exchange of other influenza viruses within the Network”).

255

Interview with Alan Hay, supra note 101.

256

Interview with Marilda Siqueira, supra note 157.

257

Consider, for example, this exchange between C.H. Andrewes, a leading flu

scientist at the World Influenza Centre at Mill Hill in London, and the coordinator

of the Network in Geneva at the time, Dr. A. Payne. Andrewes writes, of a third

influenza scientist, Dr. Magill: “I take great exception to Magill’s statement that he

has not had access to our strains or data. I have sent information on representa-

tive strains across the Atlantic and have offered others . . . . Over several years we

got no tittle of information about his own serological studies—however, I won’t

rub that in.” Letter from C.H. Andrewes to Dr. A.M.M. Payne, M.D. (June 8, 1953)

(on file with author). At times, Andrewes took such complaints to the broader

community of scientists. For example, in the summary influenza report produced

by the Network in 1952, which would have been sent to all of the labs in the

Network, Andrews included this statement: “In no instance, other than those

mentioned above, has material been sent to the World Influenza Centre [from the

United States], nor has any information come in as to the type of virus.” C.H.

Andrewes, Summary Report on Influenza 1951–52 to Dr. A.A.M. Payne (Mar. 12,

1952). Andrewes had underlined it for emphasis. Alarmed, Payne had written

back noting that this sentence had to be omitted in the final report, since “the

cooperating laboratories are entirely independent and are cooperating on a volun-

tary basis, and that therefore we have to be rather careful to retain their good

will.” Letter from Dr. A.M.M. Payne, M.D., to Dr. Andrewes (Mar. 25, 1952). When

he learned that Andrewes had already sent these sentences as part of the draft

report directly to the offending scientist, Payne followed up with a letter to that

scientist seeking to smooth things out and to suggest better ways to report their

information. See Letter from A.M.M. Payne to Dorland Davis (Apr. 7, 1952).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 51 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1589

when asked about the benefits of being part of the Network

particularly for those in countries that benefitted little or not at

all from the seasonal flu vaccine, said: “I would suspect the

benefit of joining in is actually just being part of the community

and the Network and to actually know what’s going on and

therefore be able to advise when something happens.”

258

Net-

work scientists are also obliged to provide one another with

credit and to affirmatively seek collaboration. Over and over

again in interviews, Network scientists affirmed the importance

of these rules of information sharing and acknowledgement,

because they help ensure that all in the Network feel that their

work is respected,

259

and avoid a “two-speed” system for infor-

mation production or exchange that would leave some labs

behind.

260

Self-reported norms requiring sharing of information or

credit might of course not be followed. Indeed, in interviews

scientists noted that these norms are not always followed,

though they tended to suggest that they were followed more

often inside of the Network than outside of it.

261

There is no

easy way to track whether credit is allocated appropriately in

publications, because a baseline would be difficult to establish.

We can, however, use empirical sources to trace the routine

sharing of data and samples by Network labs and by influenza

scientists more generally. Network labs routinely report their

influenza virus sample collection in a WHO database, showing

that they have collected many millions of virus samples over

258

Interview with John McCauley, supra note 59.

259

Interview with JM Heraud, supra note 139; Interview with John McCauley,

supra note 59; Interview with Marilda Siqueira, supra note 157; Interview with

Masato Tashiro, supra note 116.

260

Interview with Catherine Thompson, Healthcare Scientist, Respiratory Vi-

rus Unit, Virus Reference Department, Public Health England, United Kingdom,

in London, U.K. (Nov. 18, 2011).

261

See Interview with John McCauley, supra note 59 (describing examples of

rapid sharing of important data and viruses on the Network before publication,

and expressing the view that inside the Network, although not outside of it, it will

“always be the case” that the “global good”—namely, sharing—will take prece-

dence over more selfish motivations); see also Interview with Masato Tashiro,

supra note 116 (in the Network our “first priority is to share, to give services [to

promote] international health issues, rather than our private publication”).

Others were somewhat less sanguine. For example, Alan Hay remarked dryly that

“most people want to get something out of all the work they’ve done rather than

pouring their information for nothing into the lap of someone else who can then

write a very nice paper on it.” Interview with Alan Hay, supra note 101. “Scien-

tists,” he concluded, “are just as bad as anyone else.” Id. See also Nicholas

Zamiska, How Academic Flap Hurt World Effort on Chinese Bird Flu, W

ALL

. J.,

Feb. 24, 2006, at A6 (“All the scientists should collaborate, but there’s still a lot of

competition. . . . Scientists are human.”).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 52 29-SEP-17 12:59

1590 CORNELL LAW REVIEW [Vol. 102:1539

the last two decades, and made basic information about these

samples available to the public broadly.

262

Influenza scien-

tists, including many in Network labs, also regularly upload

influenza virus sequences on GISAID, with the number of

shared samples from each WHO region quite high.

263

Thus,

empirical sources can verify that there is a great deal of infor-

mation sharing by the Network, and among influenza scientists

generally.

III

ODEL OF

PEN

CIENCE

The existing IP literature offers us no model that can ex-

plain how and why the Flu Network operated well for decades.

Here, I describe a model that can begin to, that of “open sci-

ence.”

264

Two decades ago, Partha Dasgupta and Paul David

modeled open science as a coherent system of information pro-

duction, one with currencies and mechanisms that allow it to

produce high quality information in a manner that serves not

only scientists’ aims but also social aims. The Network’s expe-

rience offers a validation of this basic thesis, along with some

important correctives in our understanding of its allocative

processes, particularly as open science moves beyond the sin-

gle-state setting.

The basic model of open science, like the existing “IP with-

out IP” literature, describes cooperation as a relatively simple

affair. In the Network, it has been anything but. Cooperation

here has not been a mere byproduct of scientific self-interest

(as the open science model at times suggests). Nor can it be

explained as the result of the norms, intrinsic interest, and

technology that are critical to the IP without IP literature.

Rather, sustaining it has required recourse to organizations

and to law. In fact, the Network has used organizations and

law to serve many of the functions that have been demon-

strated to be important to cooperation among groups managing

common-pool resources. Its experience lays the foundation for

a more plausible account of how IP without IP can be sus-

262

See Appendix D.

263

See id. Regional distinctions in the level of samples and sequences shared

exist, as we would expect, but even in Africa, Latin America, and especially Asia,

there is very extensive collection of influenza viruses and information. For details,

see id.

264

See Dasgupta & David, supra note 42, at 499. It might also be called

“public science,” a term which would reflect the prominent role of the state in

funding open science. “Reputational science” is another possible designation, to

signal the importance of reputation in this mode of scientific production.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 53 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1591

tained, even under strain and where undertaken by loose-knit

groups.

A. The Basic Model of Open Science

A basic account of open science as a system of production

begins with the foundational work of Robert Merton.

265

Merton, an influential sociologist, described science as a sys-

tem of institutional control governed largely by norms.

266

One

of the most important norms, he argued, was the rule that

scientific results are the common property of all scientists.

267

Merton also proposed that science is characterized by a partic-

ular reward structure, which places a premium on the priority

of discovery, and that awards recognition accordingly, so that

“[r]ecognition and fame then become symbol and reward for

doing one’s job well.”

268

Partha Dasgupta and Paul David connected Merton’s work

to information economics, proposing that what Merton called

“science” was in fact a particular model of public or open sci-

ence. In this system, priority of discovery is the basis for repu-

tation, and reputation in turn “is the fundamental ‘currency’ in

the reward structure that governs the community of academic

scientists.”

269

Funding comes from patrons, typically govern-

ments.

270

When deciding which projects to fund, patrons rely

heavily on scientific reputation, amassed through the decen-

tralized processes of peer review and publication.

271

Compen-

265

Id. at 487, 510. Michael Polanyi’s account of the “Republic of Science” is

an important precursor to their model. See Michael Polanyi, The Republic of

Science: Its Political and Economic Theory, 1 M

INERVA

54, 54 (1962).

266

OBERT

K. M

ERTON

, The Normative Structure of Science, in T

OCIOLOGY OF

CIENCE

: T

HEORETICAL AND

MPIRICAL

NVESTIGATIONS

267, 269 (Norman W. Storer

ed., 1973).

267

Id. at 274–75. The other three norms central to the Mertonian account are

universalism (which requires that scientific claims are evaluated objectively); dis-

tinterestedness (which involves commitment to following the rules of science,

rather than acting in narrow self-interest), and organized skepticism (which re-

quires detached scrutiny of results and theories). Id. at 270, 275, 277.

268

ERTON

, Priorities in Scientific Discovery, in T

OCIOLOGY OF

CIENCE

: T

HE-

ORETICAL AND

MPIRICAL

NVESTIGATIONS

, supra note 266, at 286, 294.

269

Dasgupta & David, supra note 42, at 498; see also Strandburg, supra note

42, at 92–95 (showing reputational reward to be an important factor in a scien-

tist’s decision to share a discovery).

270

But compare William J. Broad, Billionaires With Big Ideas Are Privatizing

American Science, N.Y. T

IMES

(Mar. 15, 2014), http://www.nytimes.com/2014/

03/16/science/billionaires-with-big-ideas-are-privatizing-american-science.html

[https://perma.cc/8ZME-AR4X], for a discussion of the significant, possibly

growing role for foundations and philanthropists.

271

Dasgupta & David, supra note 42, at 491–92. These may be embodied, for

example, in scientific peer review panels that score grant proposals, as well as

influenced by the decentralized processes that generate publication records and

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 54 29-SEP-17 12:59

1592 CORNELL LAW REVIEW [Vol. 102:1539

sation “consists of something like a flat salary for entering

science, supplemented by rewards to winners of scientific com-

petitions, with the proviso that the better is the performance,

the higher will be the reward.”

272

The system is “open” because

scientists share their knowledge with one another freely and

without charge.

273

In open science, then, “the use of others’

output is encouraged and relatively cheap, with the cost being

appropriate citation and possibly some reciprocity in sharing

knowledge.”

274

Open science also has its own “cycle of investment and

conversion,”

275

that can be visualized as a cycle in which capi-

tal is converted into reputation, and then back again, via in-

termediaries including publications and peer review. To show

the distinctions, we can sketch market-exclusionary science as

a parallel cycle that converts capital into patents, and back to

capital, via intermediaries that include markets and patent

offices.

276

IGURE

YCLES OF

PEN VS

. M

ARKET

-E

XCLUSIONARY

CIENCE

Open Science

scientific

priority

grant

(peer review)

reputation

publication

(peer review)

scientific views about the relative importance and interest of different scientific

questions. The allocation of NIH monies (about $30 billion per year), for example,

relies on a peer-review process like this. For a good description, see Bhaven N.

Sampat, Mission-Oriented Biomedical Research at the NIH, 41 R

. P

’

1729,

1732 & n.9 (2012).

272

Dasgupta & David, supra note 42, at 499.

273

Id. at 510.

274

Alfonso Gambardella & Bronwyn H. Hall, Proprietary Versus Public Domain

Licensing of Software and Research Products, 35 R

. P

’

875, 877 (2006).

275

See B

RUNO

ATOUR

& S

TEVE

OOLGAR

, L

ABORATORY

IFE

: T

ONSTRUCTION OF

CIENTIFIC

ACTS

200 (1979).

276

This is adapted from a more materialist depiction of the cycles of credibility

in science sketched by Latour and Woolgar. See id. at 201 fig.5.1.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 55 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1593

Market-Exclusionary Science

legal

priority

managerial

allocation

market

sales

patent

(agency

review)

Open science can this way be seen as a highly articulated

system of information production, with processes that help it

achieve the two primary values associated with markets: infor-

mation gathering and investment allocation.

Much like the price system, cycles of credibility in open

science (funding

→

publication

→

peer review

→

funding) help

assemble and actuate widely dispersed information at rela-

tively low cost. Scientists share their discoveries, and credit

rather than permission is the prerequisite for building on the

work of others. Because of this, the cost of identifying projects

and collaborators is low, and scientists are free to invest their

effort wherever they think that it will, via the cycles of open

science, lead to the greatest scientific and reputational re-

turns.

277

Desire for scientific priority encourages both diligent

effort and the disclosure of discoveries, for to earn reputational

points in open science it is important both to be first and to be

seen to be first.

278

Disclosure, in turn, increases the chances

that the scientists best suited to the task can evaluate and

build upon the advances made by others.

279

In this way, the

system addresses the central challenge to the efficiency of gov-

ernment funding: the quality of government information.

280

In-

deed, when compared to a system of market exclusion, which

277

See Benkler, Coase’s Penguin, supra note 23, at 405, 414–15. For exam-

ple, a scientist need not pay or negotiate permission to write an article on H5N1.

(She may need funding to pursue the research, and here peer review serves an

important validation function.) Open science thus takes advantage of the effi-

ciency advantages that open systems, such as free software, also have. See id. at

414 (describing how systems that permit self-nomination for creative work can

plausibly allocate human creativity more effectively than more centralized sys-

tems, such as firms).

278

Dasgupta & David, supra note 42, at 499.

279

Id. at 500.

280

See supra notes 81–82 and accompanying text.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 56 29-SEP-17 12:59

1594 CORNELL LAW REVIEW [Vol. 102:1539

generates more secrecy about projects and collaborators, and

which often requires licenses for particular lines of research,

the open science system has plausible advantages in its ability

to gather and make use of decentralized information.

Can open science appropriately direct investment toward

social aims? David and Dasgupta argue that it can, because

scientists operate “collectively as an ‘agent’ for the society at

large,” which is otherwise “incapable of screening scientists . . .

[and] equally incapable of evaluating the relative importance of

scientific discoveries.”

281

As long as scientists do not abuse

that trust by acting as a cartel, and as long as the political

process respects scientific autonomy, open science can func-

tion “rather well in satisfying the requirement of social effi-

ciency in the allocation of resources.”

282

Indeed, they argue

that it has an allocative advantage over the market-exclusion-

ary system if research is highly uncertain.

283

An examination of the Flu Network helps to validate the

basic model, because the Network operates much as the model

of open science suggests that it should. Scientists in the Net-

work report sharing to be a central norm, and in fact share

extensively. They describe functional systems of reputational

rewards and sanctions that are linked to both scientific credi-

bility and to the following of scientific norms. They also de-

scribe their own funding as relying upon reputation, as well as

publications whose value is established by their peers.

The Network’s experience also provides some support for

the claim that open science “functions rather well in satisfying

the requirement of social efficiency in increasing the stock of

reliable knowledge.”

284

Markets cannot reliably produce most

of the kinds of goods that the Network produces, for the rea-

sons described in Part I. The Network, though, has generated

these goods both reliably and well over many decades. At cer-

tain critical moments, states have also responded to scientists’

expert judgment that more resources are needed, and in-

281

Dasgupta & David, supra note 42, at 505.

282

Id.

283

As described above, the market-exclusionary system is known to work

poorly for basic R&D. See supra Section I. Open science, in contrast, may be

particularly suited to such research. Expert judgment about the value of a line of

work is likely a better signal than market returns where such returns are highly

uncertain, far in the future, or unavailable because of the institutional limits of IP

law. See Dasgupta & David, supra note 42, at 490.

284

Id. at 487.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 57 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1595

creased the resources available to the WHO and the national

labs.

285

The Network’s experience, however, also offers some im-

portant correctives to the basic model of open science. The

Network’s experience requires us to refine our understanding

of the allocative processes of open science, and gives us a better

understanding of its difficulties, particularly as we move be-

yond the single-state setting. It also demonstrates the critical

role that organizations and law have played in sustaining the

Network, particularly in times of strain.

B. Open Science in Practice: Allocation

The basic open science model offers a rudimentary account

of how open science responds dynamically over time to social

aims. Dasgupta and David assume that scientific interest co-

incides with social interest,

286

but do not explain why this

must be. We might instead suspect that the two can diverge,

leading scientists to prioritize projects that have high intrinsic

interest (or that may provide the easiest path to striking re-

sults), but that do little to benefit the public. In the basic

version of open science, states are also imagined as simply

deferring to scientific judgment. Dasgupta and David recognize

that states may not in fact defer, but have little to say about

why states would defer, nor do they theorize mechanisms by

which states could be held accountable for not so doing.

287

In the Network, however, we can observe two subtle dy-

namics that may facilitate allocative success in open science.

First, scientists describe motivation in the Network as not

solely responsive to scientific interest. When asked why scien-

tists contribute to the Network, many scientists described not

only its scientific interest but also its public health impor-

tance.

288

And they reported that scientists in the Network gen-

erally cared about and were motivated by its public health

consequences, perhaps particularly in the national labs that

had fewer resources for research.

289

If both public health and

scientific interest are values that dynamically constitute the

research agenda of its scientists, then open science has an

internal means to discipline scientists whose research verges

285

See supra notes 142–46 and accompanying text.

286

Dasgupta & David, supra note 42, at 487–90.

287

Id. at 514–15 (noting that funders may be myopic, may fail to appreciate

the importance of scientific autonomy to the system, and may fail to invest

enough to allow open science to thrive).

288

See Interview with Alan Hay, supra note 101.

289

See id.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 58 29-SEP-17 12:59

1596 CORNELL LAW REVIEW [Vol. 102:1539

too far from work that could benefit the public. Scientific repu-

tation, that is, could be expected not merely to reward scien-

tists merely for scientific breakthroughs, but to reward

scientists more if their breakthroughs were of more social

significance.

290

The status of public health as a value in the Network also

helps explain how scientists gain the support of states. Scien-

tists do not merely assert their scientific authority, but instead

argue that their work is of social value. Sometimes they take

their case directly to states. Scientists in the Network, for ex-

ample, described designing and publishing scientific studies to

help show states that influenza is a significant local health

concern.

291

They also described the WHO accreditation of na-

tional labs as a means of gaining state buy-in, and reported

directly communicating with states to stress the importance of

the national labs’ work when that was needed.

292

If states fail

to respond, scientists also have some ability to discipline

states, because their open publication practices can also help

influence public opinion. The Network was, as described

above, for many years very poorly funded.

293

Scientists only

secured adequate resources for their work when states became

more acutely concerned about the public health threat posed

by influenza. In the wake of H5N1, though, prominent journal-

istic accounts—informed by influenza experts—were critical to

gaining the attention of states.

294

As archival sources reveal,

Network scientists have sometimes provided and urged the

publication of information that their states wish to keep secret

to spur them to action.

295

Here too, the fact that scientific

290

The prominence of public health as a value in the Network may not be

reflective of open science more generally. A particle physicist would not explain

her work in similar terms. Would she nonetheless perceive and describe a form of

public interest that disciplines her work? Questions such as this can produc-

tively govern future inquiries into open science, and IP without IP generally, as

Part IV describes.

291

See supra notes 154–55 and accompanying text.

292

See supra notes 159, 175 and accompanying text.

293

See supra note 146 and accompanying text.

294

See Garrett, supra note 194, at 3–4; Osterholm, supra note 3, at 31.

295

See Telegram to Dr. Payne from Dr. Mulder (July 13, 1957) (on file with

author) (reporting influenza activity, and asking the WHO office to make such

information public to provide “authoritative assistance against officials denying

[the] epidemic spread of Asiatic influenza in Holland,” which Dr. Mulder attributes

to “economic reason[s]” and an “irresponsable [sic] attitude” toward the “out-

land”). Governments of course sometimes succeed at preventing reporting of

outbreaks. China’s reluctance to admit to the scale of its SARS epidemic was

widely reported and criticized, and was a key part—along with the outbreak of

H5N1—that led to the strengthening of a core WHO agreement that requires

national governments to share with the WHO information about “all events which

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 59 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1597

values are linked with public values is important to the success

of the Network.

The publicity of open science thus anchors a second mech-

anism that can facilitate allocative success in open science.

Scientists can pressure states that fail to respond by appealing

to public reason and public concern about the effectiveness of

state action. This of course will be most effective at times and

in states where public reason and political accountability are

closely linked. More generally, open science presumes that

states are themselves responsive to social aims via systems of

political accountability. Mechanisms such as voting are a

widely recognized means of providing such accountability and

of setting social priorities.

296

In systems that are democratic,

voting helps direct, and indeed define, social aims.

297

But what of political systems that are not democratic?

Here we come up against limits of the basic model. Open sci-

ence seems unlikely to command the same support where

states are not politically accountable, and/or where their legiti-

macy does not turn on their ability to provide research goods to

their citizens. Many states do not, as described, provide much

support for the Network’s national labs.

298

This can be—and

typically is—explained by Network participants as the result of

simple resource constraints. But at least some of the time it

may instead result from failures of political accountability at

the national level. This is hardly a condemnation of the open

science model. All systems of information production—includ-

ing market exclusionary production—rely on a responsive and

effective state.

299

But it suggests an important limitation on

the claim that open science can in any simple fashion be relied

upon to advance social aims in global perspective.

may constitute a public health emergency of international concern within its

territory.” W

ORLD

EALTH

., I

NTERNATIONAL

EALTH

EGULATIONS

, art. 6 (2d ed.

2005), http://apps.who.int/iris/bitstream/10665/43883/1/9789241580410_

eng.pdf [https://perma.cc/L4YT-UZ2S].

296

In democratic states, voting is a key mechanism, for example. See, e.g.,

ENNETH

J. A

RROW

, S

OCIAL

HOICE AND

NDIVIDUAL

ALUE

1 (1951) (contrasting “vot-

ing” and “markets” as different means for making social choices). The mecha-

nisms to hold states accountable in non-democratic states are more obscure—

forming part of the reason that the open science model faces challenges as it

becomes more global.

297

See E

LIZABETH

NDERSON

, V

ALUE IN

THICS AND

CONOMICS

158–59 (1990).

298

See supra notes 138–39 and accompanying text.

299

See, e.g., Amy Kapczynski, Intellectual Property’s Leviathan, 77 L. & C

ON-

TEMP

. P

ROBS

. 131, 140–44 (2014) (exploring the essential role of a “capable state”

in IP law); see also Douglass C. North, Institutions, Ideology, and Economic Per-

formance, 11 C

ATO

J. 477, 477–81 (1991–1992) (exploring the role of institutions

in facilitating efficient markets).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 60 29-SEP-17 12:59

1598 CORNELL LAW REVIEW [Vol. 102:1539

Open science also faces another difficulty when it moves

beyond the single state setting. If politically accountable states

invest in open science because they are responding to public

opinion and democratic priority-setting, how can open science

function to serve global social aims in the absence of something

like a global state? The question begs the answer, one that is

also etched into the contours of the Network. The crisis of the

Network reflected some of the coordination problems that can

occur in a multi-state setting: states want assurances that

others will contribute, and that they will have access to bene-

fits, if they are to contribute. Historically, the Network has also

been funded predominantly by wealthy countries, who were

also the recipients of the vaccines that are its most tangible

benefit. The Network has served aims that can plausibly be

called “global” in part because they were coterminous with the

aims of wealthy countries. Frequently, however, countries are

situated differently with respect to health problems.

300

In the

absence of more global governance, open science will have diffi-

culty generating investment in truly “global” priorities where

the gains would be in the South but not the North.

This is in an important sense confirmed by the Network’s

experience. First, only because of our rudimentary system of

global governance—in particular the organization of the WHO,

and the availability of the international legal order that made

the PIP Framework (and binding international arbitration) pos-

sible—could cooperation in the Network survive the H5N1 cri-

sis. Second, that process managed not only to stabilize

cooperation, but also to go some way toward ensuring that the

benefits of the Network were more available around the world.

The 10–20% share of pandemic vaccine allocated to developing

countries in the PIP Framework is far less than what global

justice would seem to require—but is also more than any ex-

isting transnational agreement has produced.

301

Open sci-

ence, in sum, relies critically on international organizations

and law to sustain cooperation between states when under

strain. And it cannot claim to meet the demands of “global

social aims” in a systematic and reliable way, without stronger

300

This leads to the problem of so-called “neglected diseases,” or those that

primarily affect countries in the global South that are chronically underfunded.

See Peter J. Hotez et al., Eliminating the Neglected Tropical Diseases: Translational

Science and New Technologies, PL

S N

EGLECTED

ROPICAL

ISEASES

, Mar. 2, 2016,

at 1–2.

301

See Meena Krishnamurthy & Matthew Herder, Justice in Global Pandemic

Influenza Preparedness: An Analysis Based on the Values of Contribution, Owner-

ship and Reciprocity, 6 P

. H

EALTH

THICS

272, 280–82 (2013).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 61 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1599

organizations and stronger legal frameworks at the interna-

tional level that could help both define and pursue such

aims.

302

C. Open Science in Practice: Collaboration

The basic open science model, like the early IP without IP

literature, describes cooperation in open science as fairly sim-

ple. It notes that open science is susceptible to tensions and

failures, but sees these as largely self-correcting. For example,

because researchers care intensely about priority, they may

engage in wasteful races,

303

or costly fights over who has prior-

ity.

304

Scientists also may hoard data to preserve future publi-

cation opportunities,

305

or defect to the private market system

if they are allowed to patent.

306

All but the last problem, Davis

and Dasgupta predict, can be addressed by scientists via a

kind of “outcasting,” if they are in small-enough groups.

307

They will do this because “cooperative behavior furthers their

self-interest in the race for priority, and denial of access to

pools of shared information would place them at a severe dis-

advantage vis-a-vis competitors.”

308

Defection to private sci-

ence is more difficult to control, they suggest, presumably

because exit dampens the power of reputational sanctions.

This risk, they argue, can be prevented in the short term only

302

The chronic underfunding of the WHO as an entity (as distinct from the

Network as a sub-unit) is here a significant problem. The WHO itself has few

resources that it can directly manage, and major donors can dictate priorities via

earmarked contributions. See supra notes 148–50 and accompanying text.

303

Dasgupta & David, supra note 42, at 506–07. Scientists may, for example,

see other scientists with a promising project and seek to duplicate it to win the

race, when from a social perspective more diversity in project selection would have

greater benefits. Id. at 507.

304

Id. at 501.

305

Id. at 500. Dasgupta and David also emphasize the risk of racing, wherein

competition for a prize (here, priority), causes those involved to expend more effort

than is justified by the value of the accelerated advance. See id. at 506–09. The

problem of secrecy in science is well-documented. See Wesley M. Cohen & John

P. Walsh, Real Impediments to Academic Biomedical Research, 8 I

NNOVATION

’

& E

CON

. 1, 6, 15–16 (2007); Eisenberg, supra note 42, at 216; Strandburg, supra

note 42, at 91. It also has a long pedigree. Galileo and his compatriots, for

example, sometimes reported their discoveries through the use of anagrams, so

that they could establish priority while still reserving time to work out the details

or to develop valuable extensions of their insights. See Merton, Priorities in Scien-

tific Discovery, supra note 42, at 654.

306

Dasgupta & David, supra note 42, at 513.

307

Id. at 504; see also Oona Hathaway & Scott J. Shapiro, Outcasting: En-

forcement in Domestic and International Law, 121 Y

ALE

L.J. 252, 308–10 (2011)

(theorizing “outcasting” as a means of enforcing international law).

308

Dasgupta & David, supra note 42, at 504.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 62 29-SEP-17 12:59

1600 CORNELL LAW REVIEW [Vol. 102:1539

via a “pre-commitment” to sharing,

309

and over the longer term

via cultural conditioning that would encourage young scien-

tists to forego more lucrative private sector work.

310

Science studies scholars have criticized Davis and Das-

gupta’s account of scientific behavior as “idealized”

311

and as

divorced from the actual everyday experience of scientists.

312

The Network’s experience, in particular, suggests that scientific

collaboration in open science can be far more fraught than the

basic account suggests.

This is in a sense not surprising, because open science in

the Network cannot be described as occurring in a “close-knit”

group. This is not primarily because scientists operate at a

distance, but because the information that scientists need to

accurately enforce open science norms is not always readily

available throughout the Network.

313

Reputation circuits are

imperfect, particularly for those at the periphery of the Net-

work. A scientist in a national lab in Madagascar, for example,

reported that his data had been used in published papers with-

out proper acknowledgement, and that he had no effective

means to retaliate.

314

Similar complaints were raised by Indo-

nesian scientists

315

and by Chinese scientists.

316

While scien-

309

Id. at 513.

310

Id. at 514–15 (describing, for example, conditioning to “value scientific

inquiry for its own sake”).

311

Philip Mirowski & Esther-Mirjam Sent, Introduction, in S

CIENCE

OUGHT AND

OLD

: E

SSAYS IN THE

CONOMICS OF

CIENCE

50 (Philip Mirowski & Esther-Mirjam

Sent eds., 2002). As they also recognize, David’s more recent work is more

nuanced. See id.; see also Paul A. David, The Republic of Open Science, SIEPR

Discussion Paper No. 13-037 (June 14, 2014) (a more recent elaboration that

treats open science in more historical and contingent fashion).

312

Mirowski & Sent, supra note 311, at 51; see also Stephen Turner, Scien-

tists as Agents, in S

CIENCE

OUGHT AND

OLD

: E

SSAYS IN THE

CONOMICS OF

CIENCE

supra note 311, at 380 (criticizing Merton and arguing that the empirical practices

of scientists are more diverse, and involve more deviation from Merton’s norms,

than his account allows); D

OMINIQUE

INCK

, T

OCIOLOGY OF

CIENTIFIC

ORK

: T

UNDAMENTAL

ELATIONSHIP BETWEEN

CIENCE AND

OCIETY

50–54, 111–19, 184–87

(2010) (similar); see also S

TEVEN

HAPIN

, T

CIENTIFIC

IFE

: A M

ORAL

ISTORY OF A

ATE

ODERN

OCATION

113–14 (2008) (similar).

313

Cf. Strahilevitz, supra note 27, at 365 n.31 (“[I]t is the community mem-

bers’ ability to monitor instances of noncooperation and communicate with fellow

members about each member’s reputation . . . rather than group size [that] frame

the likely mechanisms by which cooperation might arise. Thus, the eBay auction-

ing network exhibits extremely impressive levels of user cooperation, despite its

millions of members, thanks to an ingenious mechanism for tracking each user’s

reputation.”).

314

See Interview with JM Heraud, supra note 139.

315

See supra note 204 and accompanying text.

316

See Zamiska, supra note 261, at A1 (describing a scientist in the Chinese

Ministry of Agriculture who was angered by lack of credit for his work in a paper

by a scientist from the U.S. CC at St. Jude).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 63 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1601

tists in these examples could refuse to cooperate with those

responsible, this sanction is costly to open science, and was

clearly not always sufficient to change behavior.

317

Decades ago, Merton coined the term the “Matthew effect”

to designate the problem that eminent scientists receive dispro-

portionate credit as compared to unknown researchers.

318

argued that in science the reputationally rich tend to get richer,

for example because they have memorable names, so tend to

receive more credit than their co-authors. Plausibly, the ability

to effectively sanction via reputational harms is also not evenly

distributed. In the Network, the problem is amplified by its

vast resource divides and divisions of labor. While scientists in

national labs do engage in publishing, they are far less able to

accrue reputational capital by publishing in top journals than

are those in the CCs, so may be both less likely to get credit for

their important data gathering work, and also less likely to

exert the influence needed to successfully sanction others.

Scientists in the Network might be able to easily identify (if

not sanction) failures to give credit to their work, because they

routinely read papers in their field. But violations of other

norms, such as those against patenting, are exponentially

more difficult to detect. Before the H5N1 crisis, Network scien-

tists began to develop suspicions about who held patents, but

none understood precisely what labs in the Network held pat-

ents, much less exactly what they covered. Patent analysis was

important because norms in the Network are not simple and

binary. They require not that scientists associated with the

Network never seek patents, but that they not seek patents on

Network-related work.

319

But patents are hard to interpret, and Network scientists

have neither the time nor expertise to evaluate them. When the

avian flu crisis ignited, NGOs and the WHO initiated patent

mappings to help identify the bounds of the Network’s patent-

ing activities. These efforts took many months, and are still far

from comprehensive.

320

The protracted and bitter fight that

317

See id. at A6 (describing the difficulty of continuing cooperation); Interview

with JM Heraud, supra note 139 (saying, of those who do not properly credit work,

“my philosophy is that I know these people and I won’t work anymore with these

people. I am careful when I am sharing something.”).

318

INCK

, supra note 312, at 114.

319

Some Network scientists defended patents if they were on their own scien-

tific work, for example, as opposed to the work of the Network. See, e.g., Interview

with Alan Hay, supra note 101. Indeed, the rules of the new Framework only

exclude patents on Network materials, carefully defined.

320

See Appendix B.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 64 29-SEP-17 12:59

1602 CORNELL LAW REVIEW [Vol. 102:1539

began in 2007 can be attributed in part to the fact that scien-

tists have extreme difficulty knowing when others have de-

fected from open science via patents. Where scientists cannot

detect violations of open science norms, suspicion may take

over.

Despite these difficulties, collaboration in the Network has

persisted and expanded over time, and survived the challenges

posed by the H5N1 outbreak. It thus provides us with a valua-

ble opportunity to theorize the means by which IP without IP

can be sustained in larger, loose-knit groups.

The two problems described above—the difficulty tracing

and punishing violations of norms—have been managed in the

Network not merely through the use of norms, nor via intrinsic

motivations or technology. Critically, fights over both credit

and patenting were solved in the Network with recourse to

organizations and law.

The Network scientist who led the national lab in Madagas-

car, and who was not accurately credited, for example, turned

to the WHO.

321

The problem was “fixed,” he recounted, by the

new click-wrap contracts that now mandate credit-sharing in

the GISAID database.

322

The WHO, and codified agreements,

have played a similar role in the credit disputes described by

China.

323

The concerns raised over failure to credit were also a

major component of the PIP Framework negotiations, which

concluded with codified obligations to share built into standard

material transfer agreements.

324

321

See Interview with JM Heraud, supra note 139; see also supra notes

250–53 and accompanying text (providing details on GISAID and EpiFlu’s require-

ment for credit). For an example of scientists using the WHO to address disputes

over sharing, see supra note 257.

322

See Interview with JM Heraud, supra note 139.

323

See Zamiska, supra note 316.

324

Medical and scientific journals have played an important clearinghouse

role as well, by both codifying and enforcing norms about acknowledgement and

appropriate sharing. In the Chinese example, the scientific journal in which the

offending paper was also a key intermediary: the complaining party brought their

concern to the editors, and the journal published a formal erratum, acknowledg-

ing the contributions of the local lab. See Elena A. Govorkova et al., Author’s

Correction: Lethality to Ferrets of H5N1 Influenza Viruses Isolated from Humans

and Poultry in 2004, 80 J. V

IROLOGY

6195, 6195 (2006). As this reflects, journals

self-consciously play an intermediary role around issues of credit. For example,

the International Committee of Journal Medical Editors has since 1978 published

guidelines for scientific manuscript submission that includes an elaborate defini-

tion of who should count as an author. See Defining the Role of Authors and

Contributors, I

’

. C

OMM

. J

OURNAL

DITORS

, http://www.icmje.org/rec-

ommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-

and-contributors.html [https://perma.cc/KT2L-MRF4]. Recently, the ICMJE

also proposed that journals create obligations on authors to share their datasets.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 65 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1603

The WHO office also consciously facilitates reputational re-

wards for national labs that are more fine-grained than publi-

cation alone.

325

These help smooth over the resource

inequities in the Network, and keep reputational rewards flow-

ing to those activities that are essential to the Network’s work.

Similarly, the tensions caused by the incursion of patents

into the Network could not be resolved without the help of the

WHO. The WHO facilitated a patent mapping that helped to

identify how patents had been used, and then convened negoti-

ations that resulted in a legal commitment that precluded pat-

ents on Network materials.

326

Among insiders, law is used

with a light touch, to buttress what remains primarily informal

and reputationally driven cooperation. It is the WHO, for ex-

ample, that mediates any disputes about the terms of reference

and PIP Framework between Network labs.

327

The Network’s

formal organizational structure, and its use of legal tools, have

both intensified over time, as conflicts became more intense,

and as the Network has grown in size and complexity.

Organizations and law help create infrastructure for more

accurate monitoring and enforcement of norms in the Network.

But their role reaches still broader. They help not just to “en-

force” norms but also to adjust them, and to resolve disputes

about their application. They redress concerns about fair-

ness—such as those raised by the national labs—that are cor-

rosive to cooperation over time. Critical to the reconstruction

of the Network was also the benefit sharing agreement, which

controls not Network participants, but the third-party firms

that produce vaccines and medicines.

See Darren B. Taichman et al., Sharing Clinical Trial Data — A Proposal from the

International Committee of Medical Journal Editors, 374 N

NGL

. J. M

. 384,

384 (2016).

325

See, e.g., Besselaar interview, supra note 118 (noting that more active

national labs are rewarded with speaking slots at the biannual meetings, in part

to “encourage[ ] the others that are possibly not as well developed to make efforts

to reach that sort of stage,” and describing the importance of thanking national

labs and providing them with opportunities for training).

326

Like the relationship to patents in the Network, the relation to private

funding is not a simple one, but one that must constantly be managed, via a kind

of boundary-work that constructs a separation between open science and market-

exclusionary science. See supra note 151 and accompanying text. This is not

reflected in the basic model, which imagines that open science is exclusively

funded by governments.

327

PIP Framework, supra note 131, annex 1, art. 7.2. This makes sense, if we

consider the delicate balance of a system based upon reciprocity, and the possibil-

ity that too much formal legal obligation may disrupt forms of trust that are

important to the workings of the open science system.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 66 29-SEP-17 12:59

1604 CORNELL LAW REVIEW [Vol. 102:1539

All of this suggests that an emphasis on “close-knitted-

ness” leaves out functions that have been important to the

success of open science in the Network. We can understand

these functions better if we turn to the well-developed literature

on the preconditions of successful cooperation in the material

commons. This literature has been developed over decades by

a group that has empirically tested Elinor Ostrom’s theories of

the conditions of sustainable community management of re-

sources such as fisheries and grazing pastures.

328

The Ostrom

group has identified a series of factors that are important to

effective group management of a common pool resource: clearly

defined boundaries for both the resource and the group; rela-

tively low-cost monitoring; graduated sanctions (e.g., gossip for

mild infractions but more serious consequences for more seri-

ous infractions); fair distribution of benefits and costs; group

decision-making procedures; conflict resolution mechanisms;

internal authority to organize to influence the rules; and “ap-

propriate coordination” with outside groups where the group

interacts with a multi-level system.

329

In the Network, many of the factors that Ostrom and her

followers stress as critical to group management of resources—

which reach significantly beyond the simple monitoring and

enforcement described in the open science model—would be

absent without the support of the WHO, bolstered at key mo-

ments by legal agreements. It is the WHO, supported by con-

tractual terms of reference and material transfer agreements,

that creates the group boundaries of the Network. The WHO

office also creates subtle means of recognition and rebuke, to

supplement the sometimes crude and weak sanctions that

328

See D

IGITAL

IBRARY OF THE

OMMONS

, http://dlc.dlib.indiana.edu/dlc/

[https://perma.cc/WV98-YEGL].

329

David Sloan Wilson, Elinor Ostrom & Michael E. Cox, Generalizing the Core

Design Principles for the Efficacy of Groups, 90S J. E

CON

. B

EHAV

. & O

. S21,

S21–S22 (2013). Ostrom first theorized a set of factors similar to this in 1990.

LINOR

STROM

, G

OVERNING THE

OMMONS

: T

VOLUTION OF

NSTITUTIONS FOR

OL-

LECTIVE

CTION

90–102 (1990). These were refined into the factors above via doz-

ens of case studies. Michael Cox, Gwen Arnold & Sergio Villamayor-Tom´as, A

Review of Design Principles for Community-Based Natural Resource Management

15 E

COLOGY

& S

’

38, 38 (2010), http://www.ecologyandsociety.org/vol15/

iss4/art38/ [https://perma.cc/7BNU-3YNH]. This account has been very influ-

ential in the resource management field, but it can also profitably be linked to a

much broader literature on the evolution of governance, or the move from dyadic

to triadic relations. See Alec Stone Sweet, Judicialization and the Construction of

Governance, 32 C

OMPAR

. P

. S

TUD

. 147, 148–51 (1999). Ostrom and her follow-

ers, we might say, have been developing an account of when and why groups need

governance, focusing on one specific context: the management of tangible

resources.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 67 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1605

scientists alone can mobilize. Informal rule-making and dis-

pute settlement has long occurred in the Network, but as the

Network grew, and particularly during the recent tensions, the

elaboration of rules and resolution of disputes have been criti-

cally facilitated by the WHO, here too with important reliance

on law. The WHO also plays an important role in modulating

the terms on which outsiders can access the Network’s materi-

als and data. Formal law is especially important as regards

this last function. The “outsider licenses,” after all, are en-

forceable not via informal WHO dispute settlement, but via

binding international arbitration.

Even this broader account, however, fails to account fully

for what Network participants see as essential to their success-

ful collaboration. As described in Part II, participants also reg-

ularly described the important role that the Network played in

deliberately cultivating a sense of community, equality, and

trust.

330

The WHO office does this in part by coming in to help

resolve disputes and ensure norms are enforced. Also impor-

tant, though, may be the normative power of its negotiations

and legal agreements. The PIP Framework, for example, makes

numerous references to the value of a “trust-based system,”

articulates the primacy of values such as “the protection of all

people of the world from the international spread of disease,”

and describes the importance of a “fair, transparent, equitable

and efficient framework” for virus sharing.

331

The WHO also

facilitates trust through more mundane day-to-day activities

that help keep the more distant national labs in personal con-

tact with one another.

332

We can make sense of this via the sizeable literature

describing the difficulties of explaining collective action solely

through rational actor models and the logic of self-interest.

These models have long been known to have problems account-

330

See supra notes 170–74 and accompanying text. The Flu Network’s

archive also suggests that activities to better connect the various nodes of the

Network have long been a central component of its work, even at times when

funding was very meager. Much of the archive is comprised of letters and memos

about arrangements for visits that WHO officials or CC members are making to

various laboratories, as well as efforts to bring scientists to WHO for training.

331

PIP F

RAMEWORK

, supra note 131, arts. 1.4, 1.9, at 3. See also id. art. 7.2.1,

at 23–24 (references to a “trust-based system”); id. annex 3, arts. 1.1, 1.2, at 37

(same); id. art. 1.4, at 3 (universal protection from disease).

332

Interview with Wenqing Zhang, supra note 130 (describing how the WHO

office holds global meetings for the national labs every two years in part for this

purpose). It has also created a secure website dedicated to the GISRS where

discussions occur, and regularly facilitates communication between the CCs,

national labs, and ERLs. Id.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 68 29-SEP-17 12:59

1606 CORNELL LAW REVIEW [Vol. 102:1539

ing for stable collective action except under the most narrow of

circumstances.

333

Trust, for example, is often critical to suc-

cessful cooperation, because it permits individuals to contrib-

ute in the belief that others too will contribute, even where

rules are sometimes broken and violations sometimes go un-

punished.

334

And though the open science model might not

predict it, participants in the Network see enforcement of

norms in open science as inevitably imperfect. As one scientist

put it, “if you don’t trust [people], . . . and you are going to

check where they are at each moment, then they cannot work

anymore.”

335

The Network’s success, as this all shows, was not sus-

tained via norms and decentralized action alone. Rather, orga-

nizations and law stand in at crucial moments to help

scientists in this loose-knit group not only monitor and enforce

norms, but also to interpret and revise rules, resolve disputes,

and define the boundaries of the group and exert control over

the outside. Finally, organizations and law have also helped to

foster trust, as well as what Elizabeth Anderson calls the

“normativity of norms”—the “understanding of group members

that they all ought to obey the standard of conduct defined by

[the] norm . . . .”

336

The Network’s experience, in fact, demon-

strates the need to move beyond a simplified rational actor

model of open science and IP without IP, as we begin to recog-

nize the role that law and organizations play in sustaining

both.

333

See Elizabeth Anderson, Beyond Homo Economicus: New Developments in

Theories of Social Norms, 29 P

HIL

. & P

. A

. 170, 177–81 (2000); Ernst Fehr &

Simon G¨achter, How Effective Are Trust- and Reciprocity-Based Incentives?, in

CONOMICS

, V

ALUES

AND

RGANIZATION

337, 337–38 (Avner Ben-Ner & Louis Put-

terman eds. 1998).

334

See Anderson, supra note 333, at 175 (“Trust appears to be a key factor

behind the willingness to cooperate. The norm of trust tells people to act as if they

believe others will reciprocate their own cooperation. It is expressed in a persis-

tent willingness to put oneself at risk, even in the face of short-term losses due to

failures to reach cooperative equilibria with one’s group.”); see also Fehr &

G¨achter, supra note 333, at 338–40 (showing that trust in reciprocal behavior

promotes the enforcement of contracts); Kahan, supra note 42, at 74 (same).

335

Interview with Sylvie Briand, Project Leader, Disease Monitoring, Assess-

ment & Control, Global Influenza Programme (Nov. 11, 2011).

336

Anderson, supra note 333, at 171. See also Jane Mansbridge, Starting

With Nothing: On the Impossibility of Grounding Norms Solely In Self-Interest, in

Ben-Ner & Putterman eds., supra note 333, at 151–66 (concluding that rationality

and self-interest alone cannot explain behavior).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 69 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1607

EORIENTING THE

IP W

ITHOUT

IP L

ITERATURE

The Network’s example offers important lessons not only

for those interested in open science but also for those con-

cerned with information production more broadly. As Part I

describes, critics have argued that there is no persuasive evi-

dence that IP without IP can be sustained under pressure,

where information value and capital costs are high, and/or

where creators are loose-knit. To the contrary, as this study

decisively shows, IP without IP can succeed under all of these

circumstances.

The Network’s example reveals that capital-intensive infor-

mation production can occur without any recourse to IP, yet

plausibly be reasonably effective and responsive to social aims

if it combines a source of capital with processes like those

described in the open science model. The two key processes

here are 1) reputational circuits that help generate reliable sig-

nals of quality and value for funders, and 2) mechanisms that

render the system accountable to public priorities.

These criteria are general, but will be helpful in structuring

future inquiries into information production settings that are

funded by government or other patrons. The point is not that

such systems are inevitably effective and responsive, but

rather that we can begin, by leveraging the model of open sci-

ence, to inquire systematically into the conditions under which

capital-intensive IP without IP systems can claim to be effective

and normatively attractive.

The Network’s example also shows that IP without IP can

succeed in groups that are loose-knit, with the support of orga-

nizations and law. The commons-based strand of the literature

has downplayed the need for management or compulsion in IP

without IP, pointing out that because information is non-

rivalrous, it escapes the problems of congestion and scarcity

that afflict the material commons.

337

This however misses

something important. The Flu Network produces a myriad of

immaterial goods, but cannot be said to operate beyond scar-

city, or without need for management or structures of com-

mand. It has characteristics of both significant openness—its

information products are almost all freely shared with the pub-

lic—and significant governance. The governance exists not to

prevent the depletion of information, but to address tensions

337

Benkler, Commons and Growth, supra note 26, at 1554.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 70 29-SEP-17 12:59

1608 CORNELL LAW REVIEW [Vol. 102:1539

that can undermine the cooperation needed to produce high-

quality, and socially oriented open science.

This is not to say, as some have suggested, that IP without

IP inevitably relies upon forms of exclusion that “operate with

an approximately equivalent effect” to IP itself.

338

Organiza-

tions and law in the Network operate not to support IP-like

exclusions, but rather to support what we might call “Ostrom

goods”—clear group boundaries, low-cost monitoring and ef-

fective sanctions, norm interpretation and revision, dispute

settlement, and management of the boundary with the outside.

They also help facilitate the cultivation of trust and public val-

ues, and normativity itself, working to support sharing despite

a continued risk of defection, rather than to support exclusion.

Moreover, the Network did not reach for conventional IP

law to stabilize cooperation. Rather, it made use of contract

law, loosely with respect to insiders and more assertively with

respect to outsiders. Conventional wisdom in the field of IP

suggests that contract law is inadequate to support the pro-

duction of information because it does not bind third parties

generally.

339

This is a problem for the market-exclusionary

model, because it depends on the ability to capture value

broadly from third parties. But the Network exemplifies a dif-

ferent system of production—not a demand-side model that

requires that users be prevented from free riding, but a supply-

side model that requires the maintenance of incentives among

a community of producers. Contract may be especially useful

for such systems. Its flexibility means it can be used to codify a

vast range of different rules, helping to clarify expectations and

permitting legal recourse if cooperation ultimately breaks

down.

340

Relations among Network insiders were anchored,

338

Barnett, supra note 27, at 1754.

339

See, e.g., Jane C. Ginsburg, Creation and Commercial Value: Copyright

Protection of Works of Information, 90 C

OLUM

. L. R

. 1865, 1917–21 (1990) (posit-

ing that contract law provides the original creator of a work with inadequate

postdelivery control); Mark A. Lemley, Intellectual Property and Shrinkwrap Li-

censes, 68 S. C

. L. R

. 1239, 1286 (1995) (“Intellectual property is only margin-

ally susceptible to protection by contract alone, because it is very easy for third

parties to duplicate an idea once it has become public. . . . Patent and copyright

law both impose liability on third parties who could not have been expected to

contract with intellectual property owners ex ante.”).

340

As Karl Llewellyn described many years ago, “the major importance of legal

contract is to provide a frame-work for well-nigh every type of group organization

and for well-nigh every type of passing or permanent relation between individuals

and groups, up to and including states—a frame-work highly adjustable, a frame-

work which almost never accurately indicates real working relations, but which

affords a rough indication around which such relations vary, an occasional guide

in cases of doubt, and a norm of ultimate appeal when the relations cease in fact

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 71 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1609

particularly over time, by contracts such as these—contracts

where individuals and groups bilaterally promised one another

to adhere to certain rules.

Contracts can also be used to govern outsiders, if those

outsiders can be identified and attracted into contractual rela-

tionships ex ante—as exemplified by the Network’s “outsider

license.” These licenses require outsiders to the Network to

follow certain rules, most prominently regarding benefit shar-

ing. The result is a kind of control over outsiders that is less

stringent than would be available if the Network held and ex-

erted intellectual property rights (because those would bind

third parties without prior agreement). But contract has been

sufficient to sustain the Network, in part because the Network

reliably generates benefits over time, access to which outsiders

wish to maintain.

341

The “click-wrap” licenses that condition

access to data from databases like GISAID similarly use con-

tract to bind outsiders who access data from the Network.

342

Here too, contracts provide less comprehensive control than a

more robust property right in this data would do, but have

been enough to stabilize the Network.

The case study offered here confirms the value and viability

of IP without IP, and can also help direct the next wave of

developments in the field. This wave should focus its attention,

at least initially, on other examples of IP without IP that are

capital intensive and produce valuable social goods, especially

if they are also loose-knit and exhibit sustained cooperation

over time, despite threats from within and without. There are

many other possible examples in the vein of open science. The

most interesting will be those that in one way or another chal-

lenge the model described here.

343

Examples drawn from other

to work.” Karl N. Llewellyn, What Price Contract? An Essay in Perspective, 40 Y

ALE

L.J. 704, 736–37 (1931).

341

If patents were not so difficult to apply to the Network’s work, and not so

corrosive to its open science practices, then these might have been leveraged into

still more assertive power over outsiders. One of the more intriguing possibilities

that the Network’s experience raises regards the potential for what we might call

“common IP” to stabilize IP without IP practices.

342

See supra note 251.

343

For example, the open science model predicts that decentralized decision-

making and open publication of research are critical to success. How then can we

explain successful public scientific enterprises that require secrecy, or that em-

ploy far more centralized government control? Defense research is an example

where secrecy may be standard, and where governments may exert more central

control over allocation. National laboratories in the United States—which are

directly government controlled, but some of which have produced an extraordi-

nary number of Nobel prizes—would also make excellent subjects of study. A

good example here is the Lawrence Berkeley Lab. Supported by the Department

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 72 29-SEP-17 12:59

1610 CORNELL LAW REVIEW [Vol. 102:1539

domains—for example, involving state and philanthropic fund-

ing for the arts, journalism, and scholarship in the academy—

would also be extraordinarily valuable.

We should also return to familiar examples in the field to

consider them anew. For example, we should develop a better

understanding of how examples studied earlier do—or do not—

link creative effort to social aims. For example, does the reli-

ance on in-kind labor bias Wikipedia to production of informa-

tion that satisfies the relatively well-off, who have more leisure

time—and if so, does the organization cultivate public values

that could discipline this tendency, for example by insisting

that priority be given to comprehensiveness?

Revisiting these examples will also allow us to consider

whether mechanisms and tools used in open science also help

to explain some of the success of these other communities. For

example, creative groups that have been described as close-

knit may, upon closer examination, rely on organizations and

law in a manner not dissimilar from the Network.

344

Even

information producers that appear to be closely connected may

be “loose-knit” in the proper sense of the term, if it is difficult

of Energy and managed by the University of California, it has thirteen Nobel Prizes

associated with its work. See About the Lab, B

ERKELEY

, http://www.lbl.gov/

about/ [https://perma.cc/5WNJ-5SU4]. Because the theory of open science sug-

gests that it is most suited to basic research, we should also study examples

where open science successfully reaches into the domain of “technology.” The

Network has some of this aspect, but still more interesting would be cases where

the public sector takes on more conventionally technological tasks, such as late-

stage drug development. We should also consider examples where the private

sector undertakes basic R&D on a large scale. See S

HAPIN

, supra note 312, at

132–45 (describing basic R&D in Bell Labs, Eastman Kodak, and General Elec-

tric). Finally, the open science model, and the example of the Flu Network, sug-

gests that the most significant problems in open science may relate to the ability

of scientists to command adequate support from the state—particularly where

support from multiple states is demanded. Studies of other transnational scien-

tific networks that have been less successful than the Flu Network are important.

Contrasting the Network to WHO’s networks on foodborne infections, human

African trypanosomiasis, and the more multi-purpose Global Outbreak Alert and

Response Network (GOARN)—to name just a few possibilities—would be illumi-

nating. Finally, studies of capital-intensive open science projects supported by

multiple states would also be useful. Here, entities such as CERN, the large

particle collider that is supported by twenty-one countries, would make excellent

subjects. For a basic description of CERN, also known as the European Organiza-

tion for Nuclear Research, see About CERN, http://home.web.cern.ch/about

[https://perma.cc/EY7P-HD8Y].

344

See, e.g., Loshin, supra note 21, at 138 (describing examples in which

magicians who have revealed tricks to outsiders were ejected from illustrious

magician organizations); see also id. at 125 (noting that societies such as the

“International Brotherhood of Magicians” are important venues for sharing of

knowledge among insiders); Oliar & Sprigman, supra note 21, at 1815 (noting that

comedy clubs and booking agents refuse to book joke-stealers).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 73 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1611

for creators to cheaply monitor and enforce norms. It would in

fact not be surprising if this were often the case with respect to

information goods, for example because the boundaries of in-

formation goods are notoriously hard to define.

Returning to consider Wikipedia and open source and free

software anew, we can now see the significance of the organiza-

tional and legal structures that have emerged to sustain them.

Wikipedia has developed a welter of specialized norms and or-

ganizationally mediated rules, backed up by formal dispute

settlement systems and sanctions.

345

It is also supported by

the Wikimedia Foundation, a legally constituted non-profit that

employs close to 200 people,

346

with an annual budget of over

$50 million.

347

Most prominent open source software projects

are also supported by organizations.

348

These organizations

play a variety of roles that resonate with the account of the

Network offered here: they contract with participants and out-

siders over rights and responsibilities, they set standards and

facilitate collective decision-making, and they manage the as-

sets and licensing strategies of these enterprises.

349

Open source software and Wikipedia also can now be seen

as relying heavily upon both organizations and law to help

produce “Ostrom goods.” Licenses help reinforce group norms

and control outsiders by mandating sharing and sometimes

credit,

350

thus helping to prevent forms of defection that would

destabilize the community. They also help create group bound-

aries and prevent incursions from the outside. They are sup-

345

See, e.g., J

EMIELNIAK

, supra note 32, at 7–8 (2014) (describing Wikipedia as

relying upon “hundreds of rules, norms, policies, and guidelines”); id. at 17–22,

29–84 (describing ten different formal organizational roles, and active systems for

monitoring, internal adjudication, and different levels of sanctioning); id. at 8

(describing Wikipedia as characterized by a degree of “regulation [that] is much

higher than in many even explicitly bureaucratic organizations”).

346

Id. at 129.

347

See Wikimedia Foundation 2014-15 Annual Plan, W

IKIMEDIA

OUND

. 3,

http://upload.wikimedia.org/wikipedia/foundation/e/e0/2014-15_Wikimedia_

Foundation_Plan.pdf [https://perma.cc/8DUT-WFBP] (last visited Aug. 30,

2014).

348

Jyh-An Lee, Organizing the Unorganized: The Role of Nonprofit Organiza-

tions in the Commons Communities, 50 J

URIMETRICS

275, 288 (2009).

349

Id. at 290–96.

350

See, e.g., GNU General Public License, GNU O

PERATING

. (last updated

June 29, 2007), https://www.gnu.org/licenses/gpl-3.0.en.html [https://perma

.cc/GH22-8QNG] (the most prominent free software license, which requires con-

tributors to share their developments under the same terms); Wikipedia Copy-

rights, https://en.wikipedia.org/wiki/Wikipedia:Copyrights [https://perma.cc/

W9Q6-LV9K ] (describing the license used by Wikipedia, which permits free repro-

duction and revision of the covered works, but requires attribution and the licens-

ing of derivative works under the same licensing terms).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 74 29-SEP-17 12:59

1612 CORNELL LAW REVIEW [Vol. 102:1539

ported by organizations that help lower the cost of monitoring,

enforcing, and revising group norms, and also—very promi-

nently in the case of both Wikipedia and organizations like the

Free Software Foundation—cultivate the values of the

group.

351

ONCLUSION

The importance of the case study here is not that it shows

us another point on the spectrum of IP without IP. Rather, it

allows us to see the field—and some of its most canonical ex-

amples—differently. It also decisively both proves the impor-

tance of, and reorients, this new literature.

It can no longer be said that IP without IP is limited to

“niche” fields or products of low social value, nor that valuable

forms of IP without IP are eventually replaced with a property

model. The open science model described here identifies a sus-

tainable, well-configured system that works very differently

than does the market-exclusion model. The system has been

durable over time, and can claim to do as well—and indeed at

times far better—than IP at effectively producing information

goods in response to social priorities. For more than six de-

cades, the Network has produced data, analysis, and standard-

ized inputs to critically important vaccines and diagnostics—

when for reasons described in Part I, none of these could be

well produced in markets. IP without IP does not only work,

but it is essential to our collective well-being. Indeed, as this

single example amply shows, it is essential to our very lives.

We must do more to understand open science, and the

structures that permit it to work well, and that cause it to fail.

The Network is a remarkable example of success, for example,

but it has also struggled at times for state support, and has

been buffeted by tensions from within and without. Even to-

day, it cannot claim to be truly responsive to social aims at a

global level—as in truth no system of information production,

including the market exclusionary one, can. With other global

diseases on the horizon, and other emergent global scientific

priorities—climate change, for example—we urgently need a

new literature that better maps the potential and limits of open

science, particularly at the global level.

351

See, e.g., W

IKIMEDIA

OUND

. (last modified Nov. 26, 2016), https://wiki

mediafoundation.org/wiki/Home [https://perma.cc/5NWW-Q886] (describing

the function and goals of the Wikimedia Foundation); F

REE

OFTWARE

OUND

http://www.fsf.org/ [https://perma.cc/G248-M76V] (stating the mission of the

Free Software Foundation).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 75 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1613

The example and model described here can also help us

better understand production of non-scientific goods, whether

magic tricks, online encyclopedias, or legal scholarship itself.

It offers us a window into how reputational systems, especially

when linked to public-minded values, can help guide socially

responsive information production. We can also move beyond

a conception of IP without IP as requiring small communities,

or, at the other end the spectrum, emerging in large online

cohorts that operate with no governance at all. Many instances

of IP without IP—plausibly the most important ones (i.e., that

are under strain, and producing goods of high value)—will re-

quire concerted support from organizations and from law. We

will see the kinds of solutions that have helped sustain the Flu

Network much more broadly in high-stakes IP without IP, I

believe, if we look for them. And only by training our eyes

differently to look for these interventions, can we learn better

how to sustain them.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 76 29-SEP-17 12:59

1614 CORNELL LAW REVIEW [Vol. 102:1539

APPENDIX A

ESEARCH

ETHODOLOGY

The case study method, like other methods, has certain

advantages and disadvantages.

352

A case study, or “an inten-

sive study of a single unit for the purpose of understanding a

larger class of units,”

353

is particularly well-suited to generat-

ing theories and to “clarifying previously obscure theoretical

relationships.”

354

The main purpose of the case study here is an exploratory

one. I use the example of the Network to theorize the condi-

tions under which IP without IP may be viable, particularly in

high-stakes settings, where information production is expen-

sive and temptations to defect substantial. The primary

sources of data for understanding the operations, rules, and

motivations of actors in the Network, as well as the recent crisis

and reconstruction, are interviews, textual material on the Net-

work (from the WHO archives,

355

as well as current documents

available on the WHO website), and databases and secondary

materials that facilitated tracing of the activity of Network

members.

I conducted thirty-six semi-structured interviews, recruit-

ing participants through a supplemented snowball method.

356

Interviews were conducted in November 2011, and in October

2013 to June 2014. Each interview typically lasted between

352

For an account of the impossibility of achieving all analytic goals at the

same time, and a general framework for understanding the value and limits of

different quantitative and qualitative social science methods, see A

DAM

RZEWORSKI

& H

ENRY

EUNE

, T

OGIC OF

OMPARATIVE

OCIAL

NQUIRY

20–23 (1970) (describing

trade-offs among accuracy, generality, parsimony, and causality).

353

John Gerring, What Is a Case Study and What Is It Good For?, 98 A

. P

. R

. 341, 342 (2004).

354

Timothy J. McKeown, Case Studies and the Limits of the Quantitative

Worldview, in R

ETHINKING

OCIAL

NQUIRY

: D

IVERSE

OOLS

, S

HARED

TANDARDS

153

(Henry E. Brady & David Collier eds., 2004). Put another way, case studies have

an advantage in “exploratory” as opposed to “confirmatory” research. Gerring,

supra note 353, at 349–50.

355

The WHO archives are accessible through the WHO library in Geneva, and

include a selection of internal papers from WHO offices from the 1940s to the

1980s. Library and Information Networks for Knowledge, About Us, W

ORLD

EALTH

., http://www.who.int/library/en/ [https://perma.cc/4KGY-GJGE].

356

Snowball sampling is “a method for generating a field sample of individuals

possessing the characteristics of interest by asking initial contacts if they could

name a few individuals with similar characteristics who might agree to be inter-

viewed.” J

OHN

OFLAND ET AL

., A

NALYZING

OCIAL

ETTINGS

: A G

UIDE TO

UALITATIVE

BSERVATION AND

NALYSIS

43 (2006). I did not rely on referrals alone, but used the

WHO’s list of Network laboratories, and the assistance of an NGO participant

closely involved in the negotiations, to seek some representativeness among the

interviewees, for example on basis of geography and type of participant.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 77 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1615

forty-five minutes and an hour, and all but five (according to

the wishes of the subjects) were recorded and transcribed. In-

terviewees included seventeen scientists in Network labs (eight

who worked in CCs, eight who worked in national labs, and one

with experience in an ERL), six WHO officials with responsibili-

ties related to the Network, three private sector actors familiar

with the Network, eight ambassadors or members of missions

that participated in the PIP Framework negotiations, and two

NGO participants closely involved in the Framework negotia-

tions. Each interviewee was asked whether they were comfort-

able speaking on the record and being cited and quoted in

research. Nearly all agreed to speak on the record, with some

also asking to review quotations before publication (which I

agreed to do).

Once I developed an understanding of the basic dynamics

of the Network and negotiations, I sought representation from

important constituent interests, which for the Network meant

seeking scientists working in different regions of the world, and

in national labs as well as CCs and the WHO central office. In

seeking interviews with negotiators, I focused on those who

were acknowledged by others as important to, or particularly

involved in, the negotiations, and I also sought to speak to

representatives from both the global North and South. I was

relatively successful in both regards, speaking to scientists in

Network labs from Latin America, Africa, East and Southeast

Asia and Australia, and the US and Europe.

Important limitations on the sample should be noted. It is

neither a random nor strictly representative sample. Inter-

views were conducted in English, limiting interviewees to those

who are very comfortable in English (which many but not all

Network scientists are), and for all scientists except those in

Geneva and London, were conducted on the telephone. Finally,

I particularly sought out scientists who had been involved in

the Network for a long time, in positions that gave them an

overview of the Network’s activities and rules. It was relatively

more difficult to interview heads of national labs, reflecting

both language barriers and some of the structure of the Net-

work itself. The Network is most interconnected at the CC

level, and has some national labs that are only loosely inte-

grated into its activities. This—along with the limits on who

was willing and able to speak with me—meant that the sample

includes disproportionate representation from those involved

with the CCs and the WHO office, who may in turn be better

resourced and less likely to emphasize problems affecting na-

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 78 29-SEP-17 12:59

1616 CORNELL LAW REVIEW [Vol. 102:1539

tional researchers, perhaps particularly in resource-poor set-

tings. Nonetheless, participants in both national labs and CCs

did acknowledge similar problems in the Network, perhaps

aided by the recent negotiations which made many of those

problems a matter of public record.

One major limitation of interview-based research is that

people may inaccurately report facts, their own views, or their

motivations. Memories are inaccurate, people do not always

clearly perceive events or their own state of mind, and people

are more likely to report more socially acceptable behavior and

less likely to report less socially acceptable behavior.

357

Where

possible, I therefore sought to triangulate interview data with

other evidence, for example from the PIP Framework or WHO

archives, or from the Web of Science

358

and the databases that

track and make visible the sharing activities of the Network.

359

I also rely on the Network’s recent crisis and renegotiation to

help trace the processes and conditions that are most essential

to the Network.

360

Because the negotiations were intense and

contentious, the resulting rules can be thought of as a kind of

revealed preference, which can usefully be triangulated with

the historical and experiential accounts of how the Network

operates.

Finally, generalizing from case-based research is challeng-

ing.

361

Recognizing this, I primarily use the example here to

refute deterministic theories (such as those suggesting that IP

without IP cannot succeed where information is high value or

357

See H. R

USSELL

ERNARD

, R

ESEARCH

ETHODS IN

NTHROPOLOGY

: Q

UALITATIVE

AND

UANTITATIVE

PPROACHES

247 (4th ed., 2006). This may be particularly so

when interviewees are speaking “on the record.” Cf. id. at 245–50 (discussing

several of the difficulties in maintaining factual accuracy that may arise from

interview-based research). Notably, however, the accounts of those few interview-

ees who wished to remain anonymous did not differ in significant ways from those

who were willing to be publicly cited.

358

See infra Appendix C.

359

See infra Appendix D.

360

Studying examples that involve failure or crises facilitate what case study

methodologists call “process tracing,” which relies on within-case variance to help

to isolate the causes of an outcome in that case. See David Collier, James Maho-

ney & Jason Seawright, Claiming Too Much: Warnings about Selection Bias, in

ETHINKING

OCIAL

NQUIRY

: D

IVERSE

OOLS

, S

HARED

TANDARDS

92–93 (Henry E.

Brady & David Collier eds., 2004). The goal of process tracing is to connect the

various phases of a policy process or change, in a way that permits us to deter-

mine why a particular outcome occurred. See Alexander L. George & Timothy J.

McKeown, Case Studies and Theories of Organizational Decision Making, in 2

DVANCES IN

NFORMATION

ROCESSING IN

RGANIZATIONS

34–41 (1985).

361

See Collier, Mahoney & Seawright, supra note 360, at 96–98.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 79 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1617

cannot succeed over time and under pressure),

362

and to elabo-

rate new theories (for example of certain potential failures of

open science, and certain conditions of success for high-stakes

“IP without IP”).

362

Case studies can empirically refute deterministic theories, because a the-

ory that asserts that “X is not possible” can be disproven if a single instance of X

can be shown to exist. See Gerring, supra note 353, at 350.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 80 29-SEP-17 12:59

1618 CORNELL LAW REVIEW [Vol. 102:1539

APPENDIX B

ATENTS IN THE

ETWORK AND IN

NFLUENZA

ORE

ENERALLY

This Appendix gathers and analyzes evidence of patenting

practices related to influenza in the Network, and in the public

and private sectors more generally. Patent searches are re-

source- and time-intensive, but fortunately, several published

reports have mapped patents related to pandemic influenza. I

draw significantly upon them, because they provide good evi-

dence of the Network’s orientation toward patenting. Given the

importance of and interest in recent pandemic strains, if the

Network is patenting extensively, it would almost certainly be

visible in this domain. A broader secondary literature provides

some additional evidence of patenting, particularly about pat-

ents that may hinder responses to a pandemic.

1. Patenting in the Network.

There is very little evidence of patenting activity among the

labs that make up the Network. Before getting into the evi-

dence, it is worth noting—both on methodological grounds,

and for reasons that are important to the problems that pat-

ents created for the Network—that it is in fact extraordinarily

difficult to determine with precision the extent of patenting in

the Network, or in influenza generally. Methods for searching

patents, particularly globally, are highly imperfect, and there

are particular difficulties with searching for patents in the ge-

nomics area.

363

When looking contemporaneously, one often

only has patent applications rather than examined patents to

review, and applications themselves are not published immedi-

ately.

364

But applications are commonly rejected or substan-

tially narrowed upon examination.

365

Patent claims are

363

For example, as one of the attempts to map genomic pandemic influenza

patents described, digital patent searches are only possible in a small number of

jurisdictions. See I

NFLUENZA

ENOME

ECHNOLOGY

ANDSCAPE

EPORT

, P

ATENT

ENS

http://www.bios.net/daisy/influenza/ext/navaggregator/navaggregator [https:/

/perma.cc/QZ5Z-SJAN] [hereinafter P

ATENT

ENS

EPORT

]. And “[e]ven for those

jurisdictions with some searching facility, finding DNA or protein sequences dis-

closed in or claimed in patents is extraordinarily difficult if not impossible,” be-

cause of the variety of ways that such sequences can be claimed, and the lack of

search infrastructures adequate to track these variable techniques of claiming.

Id.

364

See, e.g., 35 U.S.C. § 122(b) (2012) (“[E]ach application for a patent shall be

published, in accordance with procedures determined by the Director, promptly

after the expiration of a period of 18 months from the earliest filing date for which

a benefit is sought under title.”).

365

See, e.g., P

ATENT

ENS

EPORT

, supra note 363 (documenting the narrowing

of influenza sequence patent claims in examination); W

ORLD

NTELL

. P

ROP

. O

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 81 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1619

difficult to interpret,

366

and in the genomics area often tend to

be especially so, not only because of the limits of language and

ambiguity of doctrine, but also because it may be very difficult

to predict the possible consequences of a patent until later

scientific developments emerge.

367

If it is hard for experts in patent law to determine the

import of patenting in the Network, it is still more difficult for

scientists in the Network. In interviews, some Network scien-

tists exhibited their understanding of patents by questioning

the obviousness of certain patents,

368

expressing concerns

about patent quality that mirror recent debates,

369

and making

distinctions between the legitimacy of different kinds of pat-

ents; for example, patents on organisms (disfavored) versus

patents on scientific methods or products such as monoclonal

antibodies (more acceptable or even desirable, though only for

non-Network work).

370

But it was also clear that even those

scientists most comfortable discussing patents were rarely if

ever intimately familiar with the claims of particular patents.

Interviews, while a good source of information about how scien-

tists perceive patents in the Network—an issue that turns out

to be critical—are a poor source of the fact of the matter.

Fortunately, the task of determining the role that patents

have played in the Network is facilitated by four publicly availa-

ble patent mappings that were conducted in the aftermath of

WIPO P

ATENT

EARCH

EPORT ON

ANDEMIC

NFLUENZA

REPAREDNESS

(PIP)-

ATENTS AND

ATENT

PPLICATIONS

8 (2011) (explaining the challenges presented for

mapping by the withdrawal of applications, the rejection of patent applications,

and the narrowing of the scope of claims in patent applications) [hereinafter WIPO

2011 R

EPORT

366

See Dan L. Burk & Mark A. Lemley, Fence Posts or Sign Posts? Rethinking

Patent Claim Construction, 157 U. P

. L. R

. 1743, 1748, 1791–92 (2009).

367

On some of these points, see Lori B. Andrews & Laura A. Shackelton,

Influenza Genetic Sequence Patents: Where Intellectual Property Clashes with Pub-

lic Health Needs, 3 F

UTURE

IROLOGY

235, 238–39 (2008) (describing unresolved

issues in the interpretation of genomics patents on highly changeable viruses).

368

Interview with John McCauley, Director of the WHO Collaborating Centre

for Reference and Research on Influenza, United Kingdom (Nov. 18, 2011).

369

Id. (criticizing a patent as obvious, and noting the fact that often “you just

let everything go till it’s challenged. Nobody challenges because it’s too

expensive.”).

370

Id.; Interview with Alan Hay, former Director of the WHO Collaborating

Centre for Reference and Research on Influenza, United Kingdom (Nov. 18, 2011);

Interview with Michael Shaw, Senior Advisor for Laboratory Science, WHO Collab-

orating Centre for Reference and Research on Influenza (Feb. 5, 2014) (expressing

a policy at CDC against patenting “a naturally occurring virus,” but distinguish-

ing these from “patents on our techniques”); see also Interview with JM Heraud,

Head, Virology Unit and National Influenza Center, Madagascar (Feb. 6, 2014)

(objecting to patents on virus sequences as “against science”); Interview with

Julian Druce, Head, National Influenza Centre, Australia (Oct. 16, 2013) (similar).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 82 29-SEP-17 12:59

1620 CORNELL LAW REVIEW [Vol. 102:1539

the avian flu outbreak.

371

Most of them limit their searches to

patents and patent applications relevant to recent pandemic

influenza strains, and all reflect the difficulty of finding patents

outside of those few jurisdictions that make patents available

online. They each also use slightly different techniques to de-

fine and find relevant patents. The result is that even after

these substantial reviews, no definitive accounting of Network-

held patents and their implications is possible.

However, by triangulating between these reports and inter-

views, one can develop a clear, if general, picture of the role of

patents in the Network over time. Scientists with a long history

in the Network report that patents were unheard of in most of

its history.

372

The patent mappings show that as early as the

mid-1990s, a few Network labs did begin to seek patents re-

lated to their influenza work.

373

Over the years, such patents

371

The World Intellectual Property Organization undertook an extensive pat-

ent mapping in 2011 to identify patents and patent applications “in connection

with the H5N1 and H1N1 pandemic virus.” WIPO 2011 R

EPORT

, supra note 365, at

3. This review searched U.S., European, and PCT patents and patent applications,

dividing potentially relevant patents into two categories. The first were patents

that clearly claimed either parts of the H5N1 and H1N1 viruses themselves or a

“derivative of the virus, for diagnostic, therapeutic or prophylactic purposes.” Id.

at 3; see also id. at 17 (describing “Group 1” patents in further detail). Note that

patent applications may not have been granted or may have been narrowed if

granted. Id. at 4–5, 8. A second report by CAMBIA focused more narrowly on U.S.

patents and patent applications claiming either isolated genetic sequences or

expressed proteins of H5N1 viruses. See P

ATENT

ENS

EPORT

, supra note 363.

This study used search techniques complementary to those used by WIPO, relying

on searches based upon flu genetic sequences themselves (instead of keyword

searches, which would miss patents that did not use key terms, such as “H5N1”).

A third report was prepared by Edward Hammond for the Third World Network,

and searched the WIPO international patent application database for claims that

mention H5N1. E

DWARD

AMMOND

, S

OME

NTELLECTUAL

ROPERTY

SSUES

ELATED TO

H5N1 I

NFLUENZA

IRUSES

, R

ESEARCH AND

ACCINES

3 (2009) [hereinafter H

AMMOND

TUDY

]. The patent applications were included in the study if the claims could be

classified as covering medicines, vaccines, microbes, peptides, nucleic acids, or

immunoassays. Id. at 4. A fourth report prepared by the WHO Initiative for

Vaccine Research cast the broadest net, looking for patents related to many

dimensions of vaccine technology, such as processes involved in influenza vaccine

production, and adjuvants used to make vaccines more potent. M

APPING OF

NTEL-

LECTUAL

ROPERTY

ELATED TO THE

RODUCTION OF

ANDEMIC

NFLUENZA

ACCINES

ORLD

EALTH

RGANIZATION

NITIATIVE FOR

ACCINE

ESEARCH

(Oct. 23, 2007) [here-

inafter IVR R

EPORT

2007].

372

See Interview with Ian Gust, former member of WHO expert committee on

virus diseases (Oct. 16, 2013); Interview with Anne Kelso, Director of the WHO

Collaborating Centre for Reference and Research on Influenza, Australia (Oct. 29,

2013); Interview with John McCauley, supra note 368.

373

Two of the earliest patents found in the mappings are: U.S. Patent No.

5,824,536 (filed June 17, 1996) (assigned to St. Jude) and U.S. Patent No.

5,976,551 (filed June 7, 1995) (assigned to Pasteur Institute in Paris). This does

not provide definitive evidence of the date when patenting by Network labs began.

Existing mappings focus largely on existing patents. But because patents expire

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 83 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1621

emerged only sporadically, sought by individual labs rather

than the Network as an entity, with just a few labs and very few

patents involved. While each mapping found at least one pat-

ent or patent application stemming from a Network lab in their

relevant pool of patents, collectively they show patenting activ-

ity in only five out of the more than 100 labs associated with

the Network, with fewer than 20 total patents and patent appli-

cations, and more than half of these stemming from one influ-

enza lab with close ties to the animal sector, at St. Jude.

Patents stemming from the Network are dwarfed, in every case,

by the much larger number of patents from other entities. Only

two patent applications by Network labs identified in these

mappings include in their claims natural nucleotide sequences

of Network viruses (an important issue when controversies

emerged over whether patents were being obtained on “Net-

work material” itself).

374

Interpreting and generalizing the claims of the 20 pan-

demic-strain related patents and patent applications identified

as associated with the Network is difficult, but they tend to fall

into two subsets. One subset claims selections or modifica-

tions of influenza virus components designed to elicit a better

immune response for vaccines or therapeutics,

375

and the

other claims ways to stabilize or maximize the production of

vaccine viruses.

376

The latter category includes one notable

suite of patents held by St. Jude that are important to a pro-

after a term of twenty years, 35 U.S.C. § 154(a)(2) (2012), these mappings are not

a good source of information about patenting practices in the 1980s or earlier.

374

See H

AMMOND

TUDY

, supra note 371, at 5, 15 (within discussion of patent

applications including H5N1 genetic material, identifying two patent applications

by Network labs that included natural virus sequences, one in the context of

claims involving DNA vaccines and the other in the context of claims involving

monoclonal antibodies and potent HA molecules); see also P

ATENT

ENS

EPORT

supra note 363 (within mappings of patents and patent applications claiming

influenza nucleotide sequences, listing no patents or patent applications by Net-

work labs that exclusively claim natural virus sequences).

375

For example, a patent application by St. Jude, WO 2007/019094 A3, uses

reverse genetics to design specific changes to HA molecules that induce greater

vaccine potency. Claim 7 of the application does claim a strain that came from the

Flu Network, but only insofar as it includes the modified HA molecules. A subse-

quent and similar patent application by St. Jude, WO 2008/033105 A8, claims

very potent HA molecules that can induce the production of antibodies that are to

be used for diagnosis and therapy. See Hammond Study, supra note 371, at 15.

Note that patent applications may not have been granted or may have been

narrowed if granted.

376

In the second category, U.S. Patent No. 6,951,754 (filed Apr. 27, 2001)

(assigned to St. Jude), claims broadly and generally reverse genetics techniques to

produce compositions of various influenza genes and viral components, which is

an efficient method for producing vaccine viruses.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 84 29-SEP-17 12:59

1622 CORNELL LAW REVIEW [Vol. 102:1539

cess known as reverse genetics.

377

These patents are general

in nature, covering processes important to new techniques for

building viruses from scratch, and in this sense are not directly

related to influenza or the work of the Network.

378

These pat-

ents may well be relevant to the production of pandemic vac-

cines, because they cover the process used by the Network to

create reference strains for pandemic vaccines.

379

In recogni-

tion of the sensitivity of the access to medicine issues posed by

these patents—and perhaps also the fact that some of them are

held by a CC—MedImmune has publicly announced that it will

license these patents to government organizations and develop-

ing countries at no cost to develop vaccines for public health

purposes.

380

With the exception of these reverse genetics patents (which

were not perceived as the result of Network work) the sporadic

patents that Network labs have secured have apparently gener-

ated no revenue for the Network or its labs, nor impinged di-

rectly on its activities.

381

CC participants described a few

377

See IVR R

EPORT

2007, supra note 371, at 18–19; Anatole Krattiger et al.,

Intellectual Property Management Strategies to Accelerate the Development and

Access of Vaccines and Diagnostics: Case Studies on Pandemic Influenza, Malaria

and SARS, 2 I

NNOVATION

TRATEGY

ODAY

67, 94–96 (2006). Influenza reverse ge-

netics is the process by which scientists choose which HA and NA genes they will

express in a particular virus strain, relying on their chosen HA and NA genes to be

expressed from pre-designed plasmids, rather than resorting to natural assort-

ment between different HA and NA strains like in conventional virus production.

See Reverse Genetics, Flu (Influenza), N

ATIONAL

NSTITUTE OF

LLERGY AND

NFECTIOUS

ISEASES

(Jan. 14, 2011), https://web.archive.org/web/20140710042216/

https://www.niaid.nih.gov/topics/Flu/Research/basic/Pages/ReverseGenetics

Illustration.aspx [https://perma.cc/8UY2-T99Q].

378

The technique is, however, critical to Network’s work creating seed strains

for pandemic vaccines. There are many patents beyond those belonging to St.

Jude that are relevant to the process of reverse genetics, and the suite of relevant

patents was assembled by a private company—which licensed the St. Jude’s

patents among others—called MedImmune. Press Release, MedImmune, MedIm-

mune Expands Patent Estate for Reverse Genetics with New Rights from Mount

Sinai School of Medicine (Dec. 7, 2005), available at http://www.prnewswire.com/

news-releases/medimmune-expands-patent-estate-for-reverse-genetics-with-

new-rights-from-mount-sinai-school-of-medicine-55415082.html [http://perma.

cc/WG9A-7WGJ].

379

IVR R

EPORT

2007, supra note 371, at 18.

380

See id. MedImmune requires licensing fees for commercial producers of

vaccines, and recently was acquired by AstraZeneca with uncertain implications

for these policies. Id. at 19.

381

See, e.g., Interview with Michael Shaw, supra note 370 (reporting that the

CDC has no revenue stream coming from the CDC’s diagnostic patents); Interview

with Alan Hay, supra note 370 (stating that “generally throughout the Network,

people aren’t patenting anything that really impinges on the Network”); Interview

with John McCauley, supra note 368 (indicating, vis-`a-vis the emergence of pat-

enting in the viruses and genomics, “I don’t think within the Network it changed

much in practice”). While the reverse genetics patents at St. Jude’s were licensed,

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 85 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1623

patents that could theoretically have interfered with Network

work, for example because they covered techniques related to

flu diagnosis. Why did Network labs seek these patents? A

lead scientist at the CDC, one of the few labs to hold such

patents, described the reasons as “defensive.”

382

More specifi-

cally, he described them as intended to prevent private firms

from taking unfair advantage of the Network’s work.

383

broadly, they may also be intended to preserve freedom to oper-

ate for Network labs,

384

though there is no obvious evidence

that patents, held by public or private entities, have interfered

with the Network’s work to date.

The recent seeking of patents by Network labs reflects—in

attenuated fashion—a broader shift in attitudes toward pat-

ents that occurred in public scientific institutions in the 1980s

and 1990s.

385

Government entities and academic institutions

rarely patented before this period, and many had policies

against patenting. The broader trend has been described as

the result of factors that include the rise of biotechnology, ex-

pansions in the scope of patentability, and legislative shifts

(particularly in the United States, where the Bayh–Dole Act

they are not influenza specific, and the head of the CC there also suggested that

any revenues that came to their lab from this patent would only be indirect,

because the revenues go to St. Jude generally, which then sets the overall budgets

for sub-units such as the influenza lab. See Interview with Richard Webby,

Faculty, St. Jude Children’s Research Hospital and Director, WHO Collaborating

Center for Studies on the Ecology of Influenza in Animals and Birds (May 22,

2014).

382

Interview with Michael Shaw, supra note 370. This is consistent with what

others from some CCs also perceived. See, e.g., Interview with Alan Hay, supra

note 370.

383

Dr. Shaw from the CDC reported, for example, that there had “been situa-

tions where we’ve sort of been caught, . . . [where] we had published some se-

quences of primers that were useful for sequencing of influenza viruses and a

fairly large multinational company started marketing them.” Interview with

Michael Shaw, supra note 370. In response to this, the CDC has occasionally

taken out patents on techniques the CC there has developed; for example, for a

real-time PCR assay used in diagnosing influenza strains. These seem to be U.S.

Patent No. 8,568,981 (filed July 20, 2012) (concerned with methods of detecting

various influenza strains using different PCR primers) and U.S. Patent No.

8,241,853 (filed Aug. 13, 2008) (a continuation-in-part, specifically including

probes that can detect variations of the H5 and H7 types of influenza).

384

See, e.g., Interview with Alan Hay, supra note 370 (reporting that while the

CDC had sought IP related to Network materials, that they had always insisted

that the “reason they took out patents on things was to prevent others doing so.

And so it was to protect the Network.”).

385

See, e.g., David C. Mowrey et al., The Growth of Patenting and Licensing by

U.S. Universities: An Assessment of the Effects of the Bayh–Dole Act of 1980, 30

. P

’

99, 103–04 (2001) (giving evidence of the expansion of patenting by

universities in these decades).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 86 29-SEP-17 12:59

1624 CORNELL LAW REVIEW [Vol. 102:1539

was passed to encourage patenting of the results of federal

research).

386

2. The Broader Context: Patenting in the Public and

Private Sectors Generally

The Network’s approach is clearly distinct from the more

general trend in influenza. As these same patent mappings

also show, patenting in connection with pandemic influenza

viruses has grown dramatically in recent years, particularly

around the re-emergence of H5N1. This explosion of patents

includes applications and granted patents for genetic se-

quences, and those that are directed to primers, probes, diag-

nostic tests, and vaccine technologies.

387

The public and

especially academic sector outside of the Network has actively

participated in this new patent race, with a particularly high

share of sequence patents (though altogether these are rela-

tively few), and a more minor but not trivial share of patents

that relate to vaccine production and manufacture.

388

These mappings are incomplete and do not reveal, in par-

ticular, whether public sector and academic institutions seek

to profit from such patents, or rather also view them as defen-

sive in nature. However, they do make clear that many public

sector institutions have begun to seek patents that could cover

key vaccine components or processes.

In the private sector, patents related to vaccines are com-

mon, reflecting the broad importance of patents in the business

model of the pharmaceutical and biotechnology industries.

Patents related to vaccines may be multiple and fragmented,

386

Id. at 100, 116.

387

See H

AMMOND

TUDY

, supra note 371, at 4 (showing that 36% of patent

applications for influenza vaccines filed between 1983 and September 2008 have

been filed since January 1, 2006); Krattiger et al., supra note 377, at 94–104

(collecting granted patents on sequences and important techniques for vaccine

production, etc.); P

ATENT

ENS

EPORT

, supra note 363 (documenting granted pat-

ents on genetic sequences, reassortment viruses, and amino acid sequences);

WIPO 2011 R

EPORT

, supra note 365, at 3–4, 20, 23 (identifying 73 patent applica-

tions directly or indirectly related to the H5N1 and H1N1 viruses that claim virus

sequences, isolated antigens, and sequences included in patents on novel vectors,

vaccines, and adjuvants); IVR R

EPORT

2007, supra note 371 (documenting 123

patents relevant to the production of pandemic influenza vaccines).

388

For sequence patents, see, for example, U.S. Patent No. 6,685,946 (filed

Sep .19, 2002), U.S. Patent No., U.S. Patent No. 6,287,570 (filed Nov. 28, 1998),

and U.S. Patent No. 6,824,784 (filed May 8, 2003). See also P

ATENT

ENS

EPORT

supra note 363. The other patent mappings do not break out the public sector

specifically, but tallying them reveals that public sector institutions account for

less, and in certain cases significantly less, than a third of the patents identified.

See H

AMMOND

TUDY

, supra note 371; IVR R

EPORT

2007, supra note 371; WIPO

2011 R

EPORT

, supra note 365.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 87 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1625

and relate not to single vaccines but to broader systems that

are used in processes of many different vaccines.

389

This is

broadly reflected in influenza, particularly in recent years.

The most common form of influenza vaccine, grown in

eggs, is more than sixty years old, and therefore unpatented.

390

However, there have been extensive efforts to improve the man-

ufacturing and efficacy of vaccines, resulting in many patents

(mostly—but not only—held by the private sector), that could

potentially be critical to the production of vaccines in a pan-

demic.

391

The clearest example of a key patent stake is the

suite of patents assembled by MedImmune that cover the pro-

cess of reverse genetics, described above.

392

Several therapeu-

tic drugs that target the influenza virus are also covered by

patents, but in forms far less complex than the patent land-

scape for influenza vaccines.

393

As this shows, the Network exists at one end of a spectrum:

private sector entities have patented aggressively around influ-

389

See Hillary Greene, Patent Pooling Behind the Veil of Uncertainty: Antitrust,

Competition Policy, and the Vaccine Industry, 90 B.U. L. R

. 1397, 1404–05

(2010).

390

IVR R

EPORT

2007, supra note 371, at 3.

391

For example, improvements in the conventional process (that, for example,

create higher yield) are under patent, at least in some countries. Id. at 3–5. In a

pandemic setting, the use of adjuvants, which increase immune response and

thus allow for smaller doses, may be critical, and many are under patent. Id. at

10–13 (discussing adjuvant-related patents, most of which are held by private

companies). Cell-based techniques to grow viruses outside of eggs also are the

subject of many patents, particularly—but not only—by private sector companies.

Id. at 5–10. Of the 27 such patents identified in the IVR report, five were held by

the University of California (which is home to a scientist who is a leader in the

science of cell-based manufacture), one by the patenting entity of Wisconsin Uni-

versity, and three by St. Jude. Notably, even in the absence of patents, regulatory

requirements may make it difficult to use cell-based techniques without the coop-

eration of the entity holding the original cell-line. Id. at 6. Live-activated vaccines

(“FluMist” in the United States), which are administered as nasal sprays, are built

on backbones of modified strains that give the vaccine virus its attenuated qual-

ity, and so patents could be important barriers to production of certain such

vaccines. See id. at 13–15.

392

See supra note 377 and accompanying text.

393

For example, the most well-known such drug is oseltamivir phosphate

(Tamiflu), which according to the FDA Orange Book was until recently covered by

two patents: U.S. Patent No. 5,866,601 (filed June 6, 1995) and U.S. Patent No.

5,763,483 (filed Dec. 27, 1996). FDA O

RANGE

OOK OF

PPROVED

RUG

RODUCTS

WITH

HERAPEUTIC

QUIVALENCE

VALUATIONS

, N021087, http://www.accessdata.

fda.gov/scripts/cder/ob/patent_info.cfm?Product_No=003&Appl_No=021087&

Appl_type=N [https://perma.cc/2CQB-65RP]. The Orange Book is not compre-

hensive. (For example, it excludes process patents. 21 C.F.R. § 314.53(b)(1).)

However, the landscape is clearly less complex than the landscape for vaccines,

which may be impacted by patents on nucleotide or amino acid sequences, other

intermediaries, on reverse genetics, on cell-based or other manufacturing

processes, as well as on other techniques to optimize or test their efficacy.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 88 29-SEP-17 12:59

1626 CORNELL LAW REVIEW [Vol. 102:1539

enza, public sector entities outside of the Network have pat-

ented in significant numbers, and Network labs have patented

hardly at all. Notably, the fact that private, and some public

entities, have sought patents does not contradict the point

above that the market-exclusionary mode is inadequate to gen-

erate the information goods we need to address influenza.

Much of the work resulting in patents, in both the private and

public sectors, has received substantial public funding, as de-

scribed above. Patents do create some profit opportunities—for

example, if a pandemic occurs—and companies rationally seek

them for this reason. Nonetheless, we cannot expect patents

alone to provide adequate incentives for the range of informa-

tion goods that we need to address influenza, for the reasons

detailed in Part I.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 89 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1627

APPENDIX C

UBLICATION

ECORD OF

URRENT

EADS OF

ETWORK

ABS

The following reports instances of publications where the

current head of a national lab, regulatory lab, or CC was at

least a co-author. The data is up to date as of November 2,

2016.

The data was compiled by searching for each head of a

national lab, regulatory lab, or CC

394

as an author in the Web

of Science database.

395

Web of Science is a citation index sys-

tem maintained by Thomson Reuters that indexes journals,

books, and conference proceedings from all scientific disci-

plines and many social sciences and arts and humanities disci-

plines. It has over 90 million records of scholarly work.

396

aspires to index the world’s most important research, covering

publications from 80 different countries, and has nearly 1 bil-

lion cited references.

397

The search sought to capture potential authorship (or co-

authorship) in all publications with no past time limits. The

number of publications was found by searching in Web of Sci-

ence Advanced Search for (1) a given author’s last name and

first and middle initials and (2) his or her country of residence.

The search was then repeated with the topic “influenza” to see

how many of those publications included the word in their

titles.

398

For common names that returned an inordinate

number of results, the results were limited according to the

author’s institution or via the Web of Science “Article Groups”

feature, which attempts to identify unique authors via a propri-

394

Global Influenza Surveillance and Response System (GISRS), WHO I

NFLU-

ENZA

, http://www.who.int/influenza/gisrs_laboratory/en [https://perma.cc/

BCA5-RFM3].

395

Web of Science, C

LARIVATE

NALYTICS

, http://clarivate.com/scientific-and-

academic-research/research-discovery/web-of-science/ [https://perma.cc/

GM8E-KLCF].

396

Web of Science, C

LARIVATE

NALYTICS

, http://wokinfo.com/citationconnec

tion/ [https://perma.cc/PAE5-5ED6].

397

Id., Web of Science All Databases Help, https://images.webofknowledge.

com/WOKRS516B3/help/WOK/hp_additional_resources.html [https://

perma.cc/YFU7-25CU].

398

For general publication counts, a search query of the following form was

used: “AU=([author’s last name and initials]*) AND CU=([author’s country])”,

where “AU” indicates the author field and “CU” the country field. For “Influenza

Publication” counts, a search query of the following form was used: “AU=([author’s

last name and initials]*) AND CU=([author’s country]) AND TS=(inluenza)”, where

“TS” indicates the topic field. For “Other Publication” counts, the influenza publi-

cation counts were subtracted from the general publication counts.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 90 29-SEP-17 12:59

1628 CORNELL LAW REVIEW [Vol. 102:1539

etary algorithm.

399

The results are not exact, since they are

inflated by the publication counts of other authors with the

same last name and initials as the authors in question. Be-

cause Web of Science catalogues authors by last name and

initials, this effect is impossible to avoid. Limiting the search

results to the author’s country of residence helps attenuate the

effects.

There are three additional noteworthy limitations to the

data generated. First, the publications selected for indexing in

the Web of Science are likely skewed towards the global North.

Second, the tenure duration and activity level of a NIC head

impact his or her ability to compile a publication record. Third,

many important influenza scientists have been affiliated previ-

ously with the Network but are not the current head of a lab, or

are affiliated with the Network but not the head of a lab. The

results below should be understood as illustrative of the dy-

namics of publication in the Network, and generally as under-

estimates of both individuals’ total publications and of the role

of Network scientists in publishing.

399

Article Groups, T

HOMSON

EUTERS

, http://images.webofknowledge.com/

WOKRS522_2R1/help/WOS/hp_results_tellmemore.html [https://perma.cc/

39QH-4SEX].

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 91 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1629

ABLE

Publications by Country & Head

400

Country Head Name Lab

Influenza

Publications

Other

Publications

Afghanistan

Gulam Eshan

Sharifi

Central Public Health Laboratory 0 0

Albania Silvia Bino Institute of Public Health 0 47

Algeria Fawzi Derrar Institut Pasteur d'Algérie 0 1

Argentina

Elsa Baumeister

Instituto Naciona l de

Enfermedades Infecciosas

Jor

e Camara Instituto de Virolo

ia 0 3

Osvaldo Uez

Intituto Nacional de

Epidemiologia

10 0

Country Total 33 11

Australia

Julian Druce

Victorian Infectious Diseases

Reference Laboratory

12 31

David Smith PathWest Laboratory 24 23

Dominic Dwyer Westmead Hospital 95 209

Ian Barr ˠ

Victorian Infectious Diseases

Reference Laboratory

151 25

Mandvi

Bharadwaj ǂ

Therapeutic Goods

Administration Laboratories

3 53

Country Total 285 341

Austria Franz X. Heinz Medical University of Vienna 20 173

Bahrain Amjad Zaed Public Health Laboratory 0 0

Bangladesh

Mahmudur

Rahman

Institute of Epidemiology 10 35

Belarus Natalia Gribkova

Republican Research & Practical

Center for Epidemiology and

Microbiology

Belgium Isabelle Thomas

Scientific Institute of Public

Health

8 111

Brazil

Wyller Alencar

Mello

Instituto Evandro Chagas 0 0

Marilda Siqueira Instituto Oswaldo Cruz 37 67

Terezinha Maria

de Paiva

Instituto Adolfo Lutz 0 0

Country Total 37 67

Bulgaria Neli Korsun

National Centre of Infectious and

Parasitic Diseases

4 9

Cambodia Philippe Dussart Institut Pasteur in Cambodia 2 5

Cameroon Richard N

ouom Centre Pasteur du Cameroun 9 49

400

¥ denotes heads of WHO Collaborating Centres, with a total of six such

individuals: Barr, Shu, Odagiri, McCauley, Katz, and Webby. ‡ denotes heads of

Essential Regulatory Laboratories, with a total of four such individuals:

Bharadwaj, Odagiri, Engelhardt, and Ye. All others are heads of NICs.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 92 29-SEP-17 12:59

1630 CORNELL LAW REVIEW [Vol. 102:1539

Canada Yan Li

Health Canada Canadian Science

Center for Human and Animal

Health

50 41

Central African

Republic

Emmanuel

Nakouné

Institute Pasteur de Bangui 0 0

Chile Rodrigo Fasce

Instituto de Salud Publica de

Chile

16 10

China

Yuelong Shu ˠ

Chinese National Influenza

Center

125 5

Janice Lo Centre for Health Protection 3 22

Country Total 128 27

Colombia

Juliana Barbosa

Ramirez

Instituto Nacional de Salud de

Colombia

1 2

Costa Rica Jenny Lara

Laboratorio Nacional de

Influenza

3 2

Côte d'Ivoire Hervé Kadjo Institut Pasteur de Côte d'Ivoire 0 0

Croatia

Vladimir

Drazenovic

Croatian Ins titute of Public

Health

8 10

Cuba

Betsy Acosta

Herrera

Instituto de Medicina Tropical 4 2

Czech Republic

Martina

Havlickova

National Institute of Public

Health

19 11

Democratic

People's Republic

of Korea

K. Dong Guy

Central Hygienic Anti-epidemic

Station

0 0

Denmark

Thea Kølsen

Fischer

Statens Serum Institut 7 59

Ecuador

Alfredo Bruno

Caicedo

Instituto Nacional de Higiene y

Medicina Tropical

0 0

Egypt

Marwa

Abdalhamid

Egyptian Holding Company for

Biological Products and Vaccines

0 0

Amel Mohamed

Naguib

Central Public Health Laboratory 1 5

Country Total 1 5

El Salvador

Mónica Jeannette

Barahona de

Gámez

Laboratorio Central Ministerio

de Sa lud Pubilca

0 0

Estonia

Natalja

Kuznetsova

Laboratory for Communicable

Diseases

0 8

Fiji Eric Rafai

Center for Communicable

Disease Control

0 6

Finland Niina Ikonen

National Institute for Health and

Welfare

27 2

France

Bruno Lina

Centre de biologie et de

pathologie

114 121

Martine Valette

Centre de biologie et de

pathologie

53 75

Sylvie Van der

Werf

Institut Pasteur 0 0

Country Total 167 196

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 93 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1631

French Guiana

Dominique

Rousset

Institut Pasteur de la Guyane 0 11

French New

Caledonia

Dominique

Baudon

Pasteur Institute 0 0

Georgia

Ann

Machablis hvili

National Influenza Center 1 1

Germany

Brunhilde

Schwei

Robert Koch-Institute 83 59

Ghana William Ampofo National Influenza Laboratory 18 45

Greece

Andreas Mentis Institut Pasteur Hellénique 7 133

Nikolaos

Malisiovas

National Influenza Center 11 27

Country Total 18 160

Guatemala

Leticia Castillo

Signor

Centro Nacional de Influenza 0 0

Honduras

María Luisa

Matute

Laboratorio Nacional de

Vigilancia de la Salud

2 3

Hungary Istvan Jankovics

B. Johan National Center for

Epidemiology

16 37

Iceland Arthur Löve Landspitali - University Hospital 3 3 5

India

Usha Soren Singh National Influenza Center 0 22

Ranjana

Deshmukh

Acharya Donde Marg 3 3

D T Mourya National Institute of Virolo

y 0 88

Country Total 3 113

Indonesia

Pretty

Multihartina

Center for Biomedical and Basic

Technology of Health

1 0

Iran

Talat Mokhtari-

Azad

Iranian National Influenza

Center

12 32

Iraq Imam M. Aufi National Influenza Centre 0 0

Ireland Cillian De Gascu

National Virus Reference

Laboratory

0 0

Israel

Michal

Mandelboim

Ministry of Health 37 10

Italy

Maria Rita

Castrucci

Istituto Superior e di Sanità 48 1

Jamaica Sandra Jackson University of the West Indies 1 10

Japan Takato Odagiri ˠ ǂ

National Institute of Infectious

Diseases

106 105

Jordan

Aktham

Haddadin

Laboratory Directorate 1 8

Kazakhstan

Gauhar

Nusupbaeva

National Reference Laboratory

for Control of Viral infections

0 0

Kenya Japeth Magana Center for Virus Research 4 2

Kuwait S. Al-Mufti Ministry of Public Health 0 9

Kyrgyzstan

Gulbarchyn

Saparova

National Virology Laboratory 0 0

Laos

Phengta

Vongphrachanh

National Centre for Laboratory

and Epidemiology

8 16

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 94 29-SEP-17 12:59

1632 CORNELL LAW REVIEW [Vol. 102:1539

Latvia

Natalija

Zamjatina

Riga East University Hospital 0 1

Gatis Pakarna Ri

a East University Hospital 0 1

Country Total 0 2

Lebanon

Pierre Zalloua National Influenza Centre 0 76

Nisrine Jamal National Influenza Centre 0 0

Country Total 0 76

Lithuania

Vilnele

Lipnickiene

National Public Health

Surveillance Laboratory

0 3

Luxembour

Matthias Opp Laboratoire National de Santé 5 12

Madagascar

Jean-Michel

Heraud

Institut Pasteur de Madagascar 16 22

Malaysia

Zainah Saat Institute of Medical Research 6 15

Jamal I-Ching

Sam

University of Malaya 0 2

Country Total 6 17

Malta

Christopher

Barbara

Virology Laboratory of Mater Dei

Hospital

0 13

Mauritius

Sanjiv

hooputh

Central Health Laboratory 0 50

Mexico

Irma Lopez

Martinez

Instituto de diag nóstico y

Referencia Epidemiologicos

1 5

Moldova Constantin Spinu National Influenza Center 0 3

Mongolia Y. Buyanjargal

National Center for

Communicable Diseases

0 0

Morocco Amal Barakat Institut National d'Hygiène 3 1

Myanmar Htay Htay Tin National Health Laboratory 1 7

Nepal Geeta Shakya

National Public Health

Laboratory

2 27

Netherlands

Marion

Koopmans

National Influenza Centre 66 139

New Zealand

Margaret C.

Croxson

Auckland City Hospital 0 60

Sue Huang National Influenza Centre 5 68

Country Total 5 128

Nicaragua

Angel Balmaceda

Echeverria

Centro Nacional de Diagnóstico y

Referencia

0 0

Nigeria D. Olaleye College of Medicine Ibadan 4 31

Norway Olav Hungnes

Norwegian Institute of Publich

Health

32 17

Oman Amina Al Jardani Central Public Health Laboratory 0 0

Pakistan

Birjees Mazher

Kazi

National Institute of Health 3 10

Panama Brechla Moreno

Instituto Conmemorativo Gorgas

de Estudios de la Salud

1 21

Papua New

Guinea

Amanda Lang Institute of Medical Research 0 0

Paraguay Cy n t h i a Vazquez

Laboratorio Central de Salud

Publica

3 5

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 95 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1633

Peru

Victoria

Guttierez

Peceros

Instituto Nactional de Salud 0 0

Philippines

Socorro P

Lupisan

Department of Health 2 13

Poland Lidia B. Brydak National Influenza Centre 64 0

Portugal Raquel Guiomar

National Influenza Reference

Laboratory

12 5

Qatar Ajayeb Al-Nabet Hamad Medical Corporation 0 1

Republic of

Korea

Kisoon Kim

Korea National Institute of

Health

0 0

Romania

Costin Cernescu

National Influenza Centre in the

Institute of Virolo

0 31

Viorel

Alexandrescu

Cantacuzino Institute 4 38

C. Apetrei

Lasi University of Medicine &

Pharmacy

0 66

Country Total 4 1 3 5

Russia

Petr Grigorievich

Deryabin

Ministry of Health of the Russia

Federation

8 15

Elena Ivanova

Burtseva

Ministry of Health of the Russia

Federation

16 4

Anna A.

Sominina

Research Institute of Influenza 5 1

Country Total 29 20

Senegal Mbayame Niang Institut Pasteur de Dakar 10 67

Serbia

Jasminka

Nedeljkovic

Institute of Immunology 6 103

Vera Jerant-Patic Institute of Public Health 0 3

Country Total 6 1 0 6

Singapore

Raymond Tzer

Pin Lin

National Public Health

Laboratory

12 60

Slovakia Edita Staroňová

Public Health Authority of the

Slovak Republic

1 0

Slovenia Katarina Prosenc

National Laboratory for Health,

Environment and Food

10 13

South Africa

Diana Hardie University of Cape Town 1 30

Florette

Treurnicht

National Institute for

Communicable Diseases

10 33

Country Total 11 63

Spain

Tomàs Pumarola Universidad de Barcelona 54 132

Inmaculada

Casas

Centro Nacional de

Microbiolo

ía

3 9

Francisco Pozo

Centro Nacional de

Microbiología

2 3

Raúl Ortiz de

arazu

Universidad de Valladolid 0 0

Country Total 59 144

Sri Lanka Jude Jayamaha Medical Research Institute 0 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 96 29-SEP-17 12:59

1634 CORNELL LAW REVIEW [Vol. 102:1539

Sudan

Hayat Salah

Eldin Khogali

Federal Ministry of Health 0 0

Sweden Mia Bryttin

Public Health A

ency Sweden 14 42

Switzerland Samuel Cordey University of Geneva Hospitals 1 24

Syria Hazzaa Al Khalaf Public Health Laboratories 0 1

Tanzania Fausta Mosha National Influenza Laboratory 0 2

Thailand

Malinee

Chittaganpitch

National Institute of Health 41 8

Trinidad and

Tobago

Victoria Moris-

Glasgow

Caribbean Epidemiology Centre 0 0

Tunisia Amine Slim Laboratoire de Microbiologie 6 46

Turkey

Gulay

Korukluoglu

National Influenza Centre 11 46

anda Julius Lutwama U

anda Virus Research Institute 4 44

Ukraine Alla Mironenko

L.V.Gromashevsky Institute of

Epidemiology & Infectious

Diseases

5 3

Pamela

Molyneaux

Aberdeen Royal Infirmary 0 3 5

Peter Coyle Royal Victoria Hospital 3 75

William Carman Gartnavel General Hospital 9 73

Maria Zambon Health Protection Agency 181 35

John McCauley ˠ

WHO Collaborating Centre for

Reference and Research on

Influenza

92 86

Othmar

Engelhardt ǂ

National Institute for Biological

Standards and Control

25 1

Country Total 310 305

H.F. Maassab University of Michi

an 71 30

Jacqueline Katz ˠ CDC 275 27

John Janda

Viral and Rickettsial Disease

Laboratory

0 10

L. Grady

New York State Department of

Health

1 37

Richard Webby ˠ

WHO Collaborating Center for

Studies on the Ecology of

Influenza in Animals

308 22

Zhipin

Ye ǂ FDA 50 6

Country Total 705 132

Uruguay Hector Chiparelli

Departamento de Laboratorio de

Salud Publica

2 12

Venezuela

Esperanza

Briceño

Instituto Nacional de Higiene 0 8

Vietnam

Le Quynh Mai

National Institute of Hygiene and

Epidemiology

22 11

Nguyen Thanh

Long

Pasteur Institute 7 89

Country Total 29 100

Zambia Mwaka Monze University Teachin

Hospital 3 36

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 97 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1635

IGURE

Publication Distribution (influenza-only publications):

100

150

200

250

300

350

Number of Publications

Publication Statistics:

ABLE

Influenza Publications Mean Median

All Network Lab Heads 17.53 3

Head of CCs 176.17 138

Head of Regulatory Labs 46 37.5

Head of National Labs 10.79 2

ABLE

Influenza All

Total Number of Publications by All Heads 2700 6612

Percent Publications by Heads of CCs 39.15% 20.07%

Percent of Heads Without Any Publications 36.36% 18.83%

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 98 29-SEP-17 12:59

1636 CORNELL LAW REVIEW [Vol. 102:1539

ABLE

ETWORK

EAD

’

VERAGE

UMBER OF

NFLUENZA

UBLICATIONS BY

EGION

401

0 20406080

Africa

Asia

Europe

North America

South America

Oceania

401

For this and the next figure, the following countries were grouped as Africa:

Algeria, Cameroon, Central African Republic, Cˆote d’Ivoire, Egypt, Ghana, Kenya,

Madagascar, Mauritius, Morocco, Nigeria, Senegal, South Africa, Sudan,

Tanzania, Tunisia, Uganda, and Zambia.

The following countries were grouped as Asia: Afghanistan, Bahrain,

Bangladesh, Cambodia, China, Democratic People’s Republic of Korea, India,

Indonesia, Iran, Iraq, Israel, Japan, Jordan, Kazakhstan, Kuwait, Kyrgyzstan,

Laos, Lebanon, Malaysia, Mongolia, Myanmar, Nepal, Oman, Pakistan,

Philippines, Qatar, Republic of Korea, Singapore, Sri Lanka, Syria, Thailand,

Turkey, and Vietnam.

The following countries were grouped as Europe: Albania, Austria, Belarus,

Belgium, Bulgaria, Croatia, Czech Republic, Denmark, Estonia, Finland, France,

Georgia, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania,

Luxembourg, Malta, Moldova, Netherlands, Norway, Poland, Portugal, Romania,

Russia, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, UK, and Ukraine.

The following countries were grouped as North America: Canada, Costa Rica,

Cuba, El Salvador, Guatemala, Honduras, Jamaica, Mexico, Nicaragua, Panama,

Trinidad and Tobago, and United States.

The following countries were grouped as Oceania: Australia, Fiji, French

Guiana, French New Caledonia, New Zealand, and Papua New Guinea.

The following countries were grouped as South America: Argentina, Brazil,

Chile, Colombia, Ecuador, Paraguay, Peru, Uruguay, and Venezuela.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 99 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1637

ABLE

GISRS H

EADS

’ T

OTAL

NFLUENZA

UBLICATIONS

HADE BY

EGION

Africa

Asia

15%

Europe

39%

North

America

29%

Oceania

11%

South

America

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 100 29-SEP-17 12:59

1638 CORNELL LAW REVIEW [Vol. 102:1539

APPENDIX D

NFLUENZA SAMPLES COLLECTED AND SEQUENCES SHARED

BY EACH COUNTRY

The following reports the number of influenza samples re-

ceived and collected by Network laboratories worldwide, as re-

ported to the Network through WHO FluNet,

402

as well as

influenza virus sequences shared on GISAID EpiFlu by labora-

tories across the world, which include (but are not limited to)

Network labs.

403

This data is up to date as of October 17,

2016.

The FluNet data was compiled by searching the FluNet

interactive database

404

for the category “Influenza laboratory

surveillance data from any week.” For the date range, the

search used a range from week 1 of 1995 to week 40 of 2016.

This data reports the number of recorded influenza virus sam-

ples collected/received as reported to the Network within each

country, where each instance of collection or receipt is tallied

as a unit.

The GISAID data was compiled by searching the EpiFlu

database

405

for the “collection date” from 1/1/1995 to

10/7/2016, and selecting every “originating laboratory” from

an individual country. This data is the number of influenza

sequences shared on the GISAID network for all users to ac-

cess, where each sequence shared is tallied as a unit.

There are four noteworthy observations in this data.

First, there is extensive sharing of genetic sequences by

influenza scientists, and even more extensive reporting of sam-

ple collection by WHO labs. The number of shared sequences

on GISAID from each WHO region is quite high; even the region

contributing the fewest sequences (South-East Asia—11 coun-

tries) had 1/10 of the number of sequences of the region con-

tributing the most (Europe—53 countries). Furthermore,

Network labs in every WHO region reported collecting hun-

dreds of thousands of samples during the last 20 years.

Second, there is regional differentiation in terms of sample

collection and data sharing, consistent with observations made

402

FluNet Database, WHO, http://apps.who.int/flumart/Default?ReportNo=

12 [https://perma.cc/2ZJQ-BEAK].

403

About GISAID, GISAID I

NITIATIVE

, http://platform.gisaid.org/epi3/frontend

- 17a8de [https://perma.cc/B2E6-FTNU] (click “About GISAID” on the top navi-

gation bar).

404

FluNet Database, supra note 402.

405

EpiFlu Database, GISAID I

NITIATIVE

, http://platform.gisaid.org/epi3/

frontend - 17a8de [https://perma.cc/B2E6-FTNU] (click “EpiFlu

™

DATABASE”).

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 101 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1639

by Network scientists. Generally, there are more samples re-

ported and shared from Europe and US labs, and relatively

fewer samples reported and shared from countries with re-

source-poor labs.

Third, despite this differentiation, even in Africa, Latin

America, and especially Asia, there is very extensive reporting

and sharing of influenza sequence information. In fact, Asia’s

contribution is on par with (and regarding viruses, exceeds)

contributions from Europe and the US, while Africa and Latin

America’s contributions are within the same order of magni-

tude as contributions from Europe and the US.

Fourth, overall, the number of specimens received and col-

lected in FluNet looks consistent with the number of sequences

shared on GISAID for each country. This seems to suggest that

collecting and sharing influenza samples are generally corre-

lated. However, it is noteworthy that Mexico, Japan, Austria,

Ukraine, Germany, Sweden, Ireland, and Moldova all made sig-

nificant contributions to GISAID (>100 samples), but were not

recorded as making any collections in FluNet; this is true for a

few more countries with less significant GISAID contributions,

as well. These countries may be reporting their sample collec-

tion in another database.

Regional Comparisons:

Total Sequences Total Viruses

Region Shared in GISAID Collected in FluNet

Americas 16,726 5,342,118

Western Pacific 13,890 3,523,758

Europe 17,143 901,826

Eastern Mediterranean 1,149 344,982

Africa 2,643 321,490

South-East Asia 1,788 259,968

Percent of countries

with no GISAID Top country’s GISAID

Region contributions contribution percent

Americas 31% 76%

Western Pacific 26% 27%

Europe 11% 19%

Eastern Mediterranean 29% 30%

Africa 53% 14%

South-East Asia 45% 33%

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 102 29-SEP-17 12:59

1640 CORNELL LAW REVIEW [Vol. 102:1539

WHO Africa Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Africa 2,643 321,490

South Africa 363 60,123

Madagascar 288 11,445

Ghana 281 24,621

Kenya 274 29,824

Cameroon 226 15,485

Senegal 208 35,900

Mali 180 20,141

Tanzania 148 14,254

Cˆote d’Ivoire 121 13,725

Zambia 81 8,234

Mozambique 68 2,284

Ethiopia 67 3,420

Mauritius 59 6,211

Algeria 55 9,085

Uganda 47 18,243

Nigeria 45 11,649

Burkina Faso 43 2,441

Congo 40 333

Rwanda 31 8,506

Togo 8 2,462

Central African

7 3,986

Republic

Niger 3 1,051

Democratic Republic of

0 14,942

the Congo

Sierra Leone 0 1,782

Angola 0 1,134

Mauritania 0 209

Benin 0 0

Botswana 0 0

Burundi 0 0

Cape Verde 0 0

Chad 0 0

Comoros 0 0

Equatorial Guinea 0 0

Eritrea 0 0

Gabon 0 0

Gambia 0 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 103 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1641

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Guinea 0 0

Guinea-Bissau 0 0

Lesotho 0 0

Liberia 0 0

Malawi 0 0

Namibia 0 0

Sao Tome and Principe 0 0

Seychelles 0 0

South Sudan 0 0

Swaziland 0 0

Zimbabwe 0 0

WHO Americas Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Americas 16,726 5,342,118

United States 12,653 3,821,974

Brazil 1,302 86,253

Chile 494 155,986

Argentina 437 23,015

Mexico 339 0

Canada 322 1,187,428

Peru 151 10,326

Trinidad and Tobago 147 0

Paraguay 117 2,698

Guatemala 108 3,024

Costa Rica 90 4,914

Colombia 81 0

Dominican Republic 81 0

Bolivia 79 0

El Salvador 60 13,381

Honduras 56 8,794

Panama 52 6,422

Uruguay 47 896

Ecuador 42 4,732

Nicaragua 25 0

Jamaica 15 6,536

Venezuela 14 1,153

Haiti 9 0

Barbados 5 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 104 29-SEP-17 12:59

1642 CORNELL LAW REVIEW [Vol. 102:1539

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Cuba 0 4,565

Suriname 0 21

Antigua and Barbuda 0 0

Bahamas 0 0

Belize 0 0

Dominica 0 0

Grenada 0 0

Guyana 0 0

Saint Kitts and Nevis 0 0

Saint Lucia 0 0

Saint Vincent and the

Grenadines

WHO Western Pacific Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Western Pacific 13,890 3,523,758

China 3,757 2,994,997

Japan 3,603 0

Australia 3,155 160,638

Singapore 860 32,325

Laos 572 15,961

Vietnam 553 23,025

New Zealand 494 50,468

Cambodia 214 24,710

Mongolia 138 33,455

South Korea 122 98,245

Malaysia 113 19,596

Philippines 108 67,194

Fiji 99 1,802

Papua New Guinea 49 1,342

Brunei 17 0

Micronesia 15 0

Nauru 7 0

Solomon Islands 7 0

Palau 4 0

Kiribati 3 0

Cook Islands 0 0

Marshall Islands 0 0

Niue 0 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 105 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1643

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Samoa 0 0

Tonga 0 0

Tuvalu 0 0

Vanuatu 0 0

WHO Europe Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Europe 17,143 901,826

Spain 3,249 14,009

United Kingdom 1,547 13,406

Austria 1,333 0

Norway 1,104 30,558

France 990 343,671

Russia 777 271,225

Portugal 753 0

Ukraine 671 0

Italy 631 56,888

Ireland 543 0

Greece 503 41,414

Germany 449 0

Netherlands 419 4,862

Turkey 323 18,240

Sweden 321 0

Romania 314 23,542

Slovenia 244 5,898

Kazakhstan 242 0

Finland 218 1,271

Belgium 211 5,814

Bulgaria 180 0

Denmark 164 14,930

Iceland 150 0

Moldova 148 0

Serbia 145 952

Latvia 131 7,318

Georgia 128 0

Lithuania 122 1,416

Estonia 120 4,739

Slovakia 111 0

Israel 105 494

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 106 29-SEP-17 12:59

1644 CORNELL LAW REVIEW [Vol. 102:1539

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Hungary 100 0

Luxembourg 99 0

Czech Republic 93 0

Switzerland 87 3,045

Albania 60 0

Kyrgyzstan 57 0

Poland 51 38,134

Cyprus 51 0

Macedonia 35 0

Montenegro 34 0

Malta 32 0

Belarus 31 0

Croatia 23 0

Armenia 19 0

Bosnia and

18 0

Herzegovina

Tajikistan 7 0

Andorra 0 0

Azerbaijan 0 0

Monaco 0 0

San Marino 0 0

Turkmenistan 0 0

Uzbekistan 0 0

WHO South-East Asia Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

South-East Asia 1,788 259,968

Bangladesh 585 28,507

Nepal 378 13,050

Thailand 325 33,889

India 224 120,123

Indonesia 215 30,630

Sri Lanka 61 28,478

Bhutan 0 4,658

Myanmar 0 438

Maldives 0 195

North Korea 0 0

Timor-Leste 0 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 107 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1645

WHO Eastern Mediterranean Region:

Sequences shared on Total Viruses

Country GISAID collected in FluNet

Eastern Mediterranean 1,149 344,982

Egypt 347 93,951

Jordan 231 8,844

Oman 162 30,242

Morocco 98 8,733

Bahrain 88 3,831

Iran 79 87,016

Pakistan 39 14,746

Saudi Arabia 23 0

Iraq 22 15,781

Tunisia 18 11,341

Afghanistan 18 7,120

Qatar 12 61,162

Syria 6 72

United Arab Emirates 5 0

Lebanon 1 1,770

Sudan 0 373

Djibouti 0 0

Kuwait 0 0

Libya 0 0

Somalia 0 0

Yemen 0 0

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 108 29-SEP-17 12:59

1646 CORNELL LAW REVIEW [Vol. 102:1539

APPENDIX E

UBLISHERS IN THE

NFLUENZA

IELD

The following table seeks to identify the 100 most prolific

authors and their number of publications in the influenza field.

The data is up to date as of September 25, 2016.

The data was compiled by searching for the topic of “influ-

enza” in the Web of Science database. The search captured

authorship (and co-authorship) in all publications for pub-

lished articles with no past time limits. The results were then

filtered to return the top 100 authors in terms of publication

output. The results again are subject to important limitations,

mirroring those described in Appendix C. Because of these, the

method likely underestimates actual representation of the Net-

work among these prolific authors.

406

Even without this con-

sideration, 7% of the top-100 prolific researchers are Network

lab heads (bolded in the below).

Top-100 Influenza research publishers and their number of

publications

407

Author Publications

Webster RG 766

Kawaoka Y 502

Palese P 343

Garcia-Sastre A 338

Osterhaus ADME 336

Peiris JSM 288

Suzuki Y 280

Klenk HD 262

Li Y 249

Katz JM 248

Doherty PC 243

Webby RJ 242

Guan Y 235

Kida H 234

Fouchier RAM 231

Rimmelzwaan GF 224

Murphy BR 223

406

For example, Nancy Cox, Richard Webster, and Masato Tashiro have been

key members of the Flu Network for many years, but are not bolded here because

they are not current heads of Network labs.

407

Bolded authors denote heads of WHO Collaborating Centres, Essential

Regulatory Laboratories, and national labs in September 2016.

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 109 29-SEP-17 12:59

2017] OPEN SCIENCE IN INFLUENZA 1647

Author Publications

Swayne DE 222

Chen HL 215

Cox NJ 215

Lamb RA 211

Compans RW 205

Zhang Y 205

Skehel JJ 202

Tumpey TM 196

Wang J 191

Cowling BJ 188

Monto AS 182

Suzuki T 169

Hayden FG 167

Subbarao K 166

Scholtissek C 164

Wang Y 164

Taubenberger JK 158

Finelli L 156

Kim JH 155

Tashiro M 153

Edwards KM 152

Capua I 149

Chen Y 148

Gao GF 148

Li J 146

Viboud C 145

Couch RB 143

Shu YL 142

Laver WG 141

Suarez DL 141

Klimov A 140

Uyeki TM 140

Liu Y 137

Rott R 137

Zambon M 137

Yuen KY 136

Krug RM 135

Zhdanov VM 134

Air GM 133

Kendal AP 133

Kilbourne ED 133

Nichol KL 133

\\jciprod01\productn\C\CRN\102-6\CRN601.txt unknown Seq: 110 29-SEP-17 12:59

1648 CORNELL LAW REVIEW [Vol. 102:1539

Author Publications

Wang W 133

Chan KH 132

Oxford JS 132

Glezen WP 129

Katze MG 126

Ludwig S 124

Govorkova EA 123

Klimov AI 122

Schild GC 122

Shay DK 122

Fry AM 121

Poland GA 119

Brown IH 117

Zhang L 117

Treanor JJ 116

Donis RO 115

Lina B 115

Esposito S 114

Gubareva LV 114

Haller O 114

Kuiken T 114

Nakamura K 114

Zhang J 114

Braciale TJ 113

Chen Z 113

Zimmerman RK 113

Belshe RB 112

Stallknecht DE 111

Donatelli I 109

Odagiri T 109

Simonsen L 109

Wiwanitkit V 109

Bridges CB 108

Potter CW 108

Dimmock NJ 107

McCullers JA 107

Wright PF 107

Perez DR 106

Sakoda Y 106

Li X 105

Alexander DJ 104