McGill Law Journal Revue de droit de McGill
DECREASING THE DATA DEFICIT: IMPROVING POST-
MARKET SURVEILLANCE IN PHARMACEUTICAL
REGULATION
Trudo Lemmens & Shannon Gibson*
The drug regulatory system is currently largely based on
market-entry review of safety and efficacy data and involves only
very limited post-market review. Failures in the industry-controlled
production of pre-market data and the lack of solid post-market sur-
veillance contribute significantly to highly problematic drug prescrip-
tion and consumption practices, which have become a very serious
public health concern. In this paper, we first discuss how historically
grown drug regulations have contributed to the development of in-
dustry control over clinical trials, which is one of the key factors be-
hind the limits of pre-market evidence. We then explore some prob-
lematic aspects related to the fixation of the drug approval system on
pre-market activities, including the lack of good real-world evidence
on drug safety; the lack of comparative evidence on patient benefit
between different drugs; the lack of evidence of the safety and efficacy
of off-label prescribed drugs; and the inadequate reporting of adverse
drug reactions (ADRs). We argue that a more rigorous and intense
post-market surveillance system could counterbalance, at least in
part, the distorted situation created by the regulatory reliance on pre-
market, industry-produced clinical trials data. In particular, we ad-
vocate for improvements to the current ADR reporting system, more
explicit requirements for both comparative effectiveness studies and
post-market clinical research in real-world settings, the promotion of
transparency of pharmaceutical data, and insulating clinical research
from industry control.
Le systme rglementaire des mdicaments se base actuel-
lement en grande partie sur ltude des donnes de scurit et
defficacit au point dentre sur le march, mais ne comporte quune
tude trs restreinte aprs lentre sur le march. Des dfauts dans la
production de donnes avant lentre sur le march, entreprise par
lindustrie mme, et labsence de surveillance apprciable aprs
lentre sur le march contribuent de faon importante des pra-
tiques trs problmatiques de prescription et de consommation des
mdicaments. Ces pratiques sont devenues un problme de sant
publique trs grave. Dans cet article, nous tudions dabord comment
la
rglementation des mdicaments dveloppe partir
dantcdents historiques a contribu au dveloppement du contrle
par lindustrie pharmaceutique des essais cliniques. Ceci est lun des
facteurs essentiels qui explique les limites des donnes prleves
avant lentre sur le march. Nous abordons ensuite les aspects pro-
blmatiques de la fixation du systme dautorisation des mdica-
ments sur les activits avant lentre sur le march, y compris :
labsence de donnes relles sur la scurit des mdicaments;
labsence de donnes comparatives par rapport aux avantages de dif-
frents mdicaments pour les patients; labsence de donnes sur la
scurit et lefficacit des mdicaments prescrits pour des usages non
autoriss; et le rapport inadquat des effets indsirables des mdica-
ments (EIM). Nous soutenons quun systme de surveillance aprs
lentre sur le march plus rigoureux et intense pourrait compenser,
du moins en partie, la situation dforme qua engendre la dpen-
dance rglementaire sur les donnes bases sur des essais cliniques
conduits par lindustrie avant lentre sur le march. En particulier,
nous prnons lamlioration du systme actuel de rapport des EIM,
des exigences plus explicites pour les tudes sur lefficacit compara-
tive et pour la recherche clinique aprs lentre sur le march dans un
contexte rel, la promotion de la transparence dans les donnes de re-
cherche, et le cloisonnement de la recherche clinique du contrle par
lindustrie.
* Trudo Lemmens is Associate Professor and Scholl Chair in Health Law and Policy, and
Shannon Gibson is Research Associate, Faculty of Law, University of Toronto. A Ge-
nome Canada grant on Ethical and Legal Issues of Cancer Initiating Stem Cell Re-
search, and a Social Sciences and Humanities Research Grant, Promoting Integrity of
Medical Research: The Janus Face of Regulation provided financial support. During
part of the writing of this paper, TL was a Faculty Member, Center for Transnational
Legal Studies; Academic Visitor, HeLEX Centre & Faculty of Law, University of Ox-
ford; and Plumer Visiting Fellow, St. Annes College. The authors are grateful to the or-
ganizers and participants of the workshop on Innovations technologiques, incertitude
et droit de la responsabilit for fruitful discussions and feedback that contributed to
the writing of this paper. A short version of this paper was presented at a 2012 hearing
of the Canadian Senate Committee on Social Affairs, Science and Technology. Kelly Tai
did excellent background research and work on the references. Insightful comments by
two anonymous reviewers helped to improve the paper.
Trudo Lemmens & Shannon Gibson 2014
Citation: (2014) 59:4 McGill LJ 943 Rfrence : (2014) 59 : 4 RD McGill 943
945
947
947
950
953
955
958
961
963
967
969
969
973
973
976
977
979
986
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
944
Introduction
I.
II.
III.
Challenges in Pre-Market Evidence Development
A. Historical Developments in Requirements for Market
Entry
B. Issues in the Production of Clinical Trials Data
C. Understanding the Interests of Industry
D. The Limitations of Pre-Market Clinical Trials Evidence
Specific Challenges in Post-Market Surveillance
A. Adverse Drug Reaction Reporting
B. Issues Around Off-Label Prescribing
C. Improving Information Sharing with Patients and
Consumers
Tackling the Data Deficit
A. The Need for Independent Research
B. Promoting Transparency and Data Access
1. The Raison dtre of Transparency
2. The International Move Toward Data Access
3. Canadas Hesitant Steps Toward Transparency
4. The Saga of Pharmaceutical Data Transparency
in Europe
Conclusion
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 945
Introduction
The drug approval system in industrialized countries is generally di-
vided into two distinct phases: a pre-market phase and a post-market au-
thorization phase. The pre-market phase is characterized by substantial
investments in clinical trials aimed at providing data demonstrating the
safety and efficacy of a product. Subsequently, in the post-market phase,
drugs are generally perceived by both the medical community and the
public at large to be safe and effective. In this phase, drug companies
rarely undertake additional studies to assess the safety and efficacy of the
products in the real-life context in which these drugs are then consumed.1
To date, the focus of drug regulation, certainly in Canada, has primar-
ily been on pre-market activities and on ensuring that products entering
the market fulfill certain basic requirements with respect to safety and ef-
ficacy. Unfortunately, lay peoplebut also to a large degree those in-
volved in the regulation, prescription, use, and funding of pharmaceutical
products (including health care professionals)ignore the limitations of
the data produced in pre-market clinical trials. These trials are conducted
under the control of an industry that has a vested and significant finan-
cial interest in showing that the products tested in these trials work. Yet
the approval of drug products by regulatory agencies such as Health Can-
ada is easily taken as a firm confirmation of their safety and value.
As we explain below, the drug regulatory system that has developed
throughout the twentieth century, while obviously intended to improve
the safety and efficacy of pharmaceutical products, has also had perverse
side effects. This system has strengthened the dominant position of indus-
try over the creation of clinical trials data; it has created a false assurance
of safety and efficacy by the very fact of approving the marketing of these
drugs on the basis of the submitted data; and finally, it has enabled in-
dustry, which is largely in control of the production and publication of the
results of clinical trials, to use this false assurance as a marketing tool. At
the same time, the pharmaceutical industry has at this point little inher-
ent interest in critically evaluating or questioning whether the data pro-
duced in pre-market testing for the purpose of drug approval is sufficient
or provides the most meaningful information. In fact, as we will argue,
the opposite is often true.
We will not discuss in detail in this article the various components of
industry control over pharmaceutical research, product development, and
distribution of information. Instead, we want to explore, with an emphasis
on Canadian developments, how a shift toward better post-marketing
1 See H-G Eichler et al, Adaptive Licensing: Taking the Next Step in the Evolution of
Drug Approval (2012) 91:3 Clinical Pharmacology & Therapeutics 426.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
946
surveillance could counterbalance, at least in part, the distorted situation
created by our reliance on pre-market clinical trials data produced under
the control of industry.
The importance of correcting the knowledge deficit resulting from in-
dustry control over data and from the lack of post-market data cannot be
overestimated. Bad information leads to bad health care practices, often
resulting in serious injuries and death. The wide-ranging health risks as-
sociated with industry manipulation and misrepresentation of data, com-
bined with aggressive marketing practices, have been highlighted, partic-
ularly in the last decade, in a panoply of media reports and publications.2
As these publications and reports highlight, a remarkably high number of
serious injuries or deaths are linked to these practices, which makes it in-
creasingly hard to understand the so far very timid response by regulators
and governmental decision makers. The extent of the control of industry
over scientific data also has another important consequence that negative-
ly affects potential legal responses (for example, tort law) to this problem:
by avoiding the collection of evidence about adverse effects of pharmaceu-
tical products that are currently on the market, or by directly manipulat-
ing the presentation of data in the scientific literature, industry can shield
itself more easily from potential liability for deficiencies associated with
its products. The powerful influence, if not control, of industry over the
production of scientific evidence thus also impacts on legal tools that could
otherwise be used to deter or provide compensation for injury resulting
from defective products.3 The legal community should therefore also start
2 For recent detailed discussions, see e.g. Peter C Gtzsche, Deadly Medicines and Organ-
ised Crime: How Big Pharma Has Corrupted Healthcare (London: Radcliffe, 2013); Ben
Goldacre, Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients
(London: Fourth Estate, 2012); Ray Moynihan & Allan Cassels, Selling Sickness: How
the Worlds Biggest Pharmaceutical Companies Are Turning Us All into Patients (New
York: Nation Books, 2005); Terence H Young, Death by Prescription: A Father Takes On
His Daughters Killer: The Multi-Billion-Dollar Pharmaceutical Industry (Toronto: Key
Porter Books, 2009) (with a focus on Canada). See also David Healy, Pharmageddon
(Berkeley: University of California Press, 2012); Carl Elliott, White Coat Black Hat: Ad-
ventures on the Dark Side of Medicine (Boston: Beacon Press, 2010); Trudo Lemmens &
Ron A Bouchard, Regulation of Pharmaceuticals in Canada in Jocelyn Downie, Timo-
thy Caulfield & Colleen M Flood, eds, Canadian Health Law and Policy, 3d ed (Mark-
ham, Ont: LexisNexis Canada, 2007) 311 at 31114; Trudo Lemmens, Leopards in the
Temple: Restoring Scientific Integrity to the Commercialized Research Scene (2004)
32:4 JL Med & Ethics 641 [Lemmens, Leopards]; Trudo Lemmens, Pharmaceutical
Knowledge Governance: A Human Rights Perspective (2013) 41:1 JL Med & Ethics
163 [Lemmens, Pharmaceutical Knowledge Governance]; Trudo Lemmens & Candice
Telfer, Access to Information and the Right to Health: The Human Rights Case for
Clinical Trials Transparency (2012) 38:1 Am J L & Med 63.
3 See Gary Edmond, Supersizing Daubert Science for Litigation and Its Implications for
Legal Practice and Scientific Research (2007) 52:4 Vill L Rev 857 (Edmond is one of the
only authors who has explicitly warned about the impact of industry control over sci-
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 947
to take much more seriously the consequences of the gaps in the scientific
data production processes, as they compromise the integrity of the legal
system itself.
This paper provides recommendations to improve, within the contours
of the current Canadian drug regulatory system, the ways in which data
is being produced and distributed. We begin by describing the historical
development of industry control over clinical trials, followed by a brief
overview of the limitations of pre-market clinical trials. Subsequently, we
argue that the inadequacy of the safety and efficacy data created at the
pre-market stage only increases the importance of enhanced surveillance
activities during the post-market phase. Expanded post-market surveil-
lance is necessary in order to combat a host of problems, including the
lack of good evidence on long-term safety and comparative effectiveness;
the need for improved adverse event reporting; and the phenomenon of
off-label prescribing. We will briefly highlight these issues and discuss
some avenues for improvement. Finally, we highlight the importance of
promoting the transparency of clinical trials, which constitutes a crucial
condition for correcting the failure in data production at both the pre-
market and post-market stages.
I. Challenges in Pre-Market Evidence Development
A. Historical Developments in Requirements for Market Entry
There is a historical reason for the focus of the drug regulatory system
on the pre-market stage and the more limited attention to what happens
after drugs have entered the market. Pharmaceutical regulations were
first introduced in the late nineteenth century, a time of untrammelled
and rapidly expanding market capitalism. These regulations aimed to es-
tablish a regulatory barrier against the mass marketing of often highly
toxic products of the so-called patent medicines industryan industry of
home-brew quackeries that was rapidly expanding at that time, partly as
a result of new mass media technology and improved transportation in-
frastructure.4 As Philip J. Hilts describes in his history of the US Food
and Drug Administration (FDA), this patent medicines industry had
ence for tort law and litigation). See also Simon Stern & Trudo Lemmens, Legal Rem-
edies for Medical Ghostwriting: Imposing Fraud Liability on Guest Authors of Ghost-
written Articles (2011) 8:8 PLoS Medicine 1 [Stern & Lemmens, Legal Remedies].
4 See Lemmens & Bouchard, supra note 2 at 31415. For a good historical overview of
the development of drug regulation in the US, see Philip J Hilts, Protecting Americas
Health: The FDA, Business, and One Hundred Years of Regulation (New York: Albert A
Knopf, 2003); Daniel Carpenter, Reputation and Power: Organizational Image and
Pharmaceutical Regulation at the FDA (Princeton: Princeton University Press, 2010).
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
948
gained a reputation for aggressively promoting often highly toxic or seri-
ously addictive products for the treatment of a variety of ailments, includ-
ing serious diseases.5 Numerous deaths, serious injuries, and high rates of
addiction were associated with several of these products. The goal of the
first regulations was thus to provide some, albeit fairly limited, govern-
mental control at market entry to ensure that uninformed consumers
were protected against manifestly dangerous products. State control over
the sale of such products was an exceptional step at the margins of the
market, aimed at preventing market excesses that threatened public safe-
ty.
Later controversies throughout the twentieth century resulted in
piecemeal extensions of regulatory review and additional regulatory re-
quirements. These controversies took place also in the context of signifi-
cant changes to the pharmaceutical development and production process
and to the overall nature of the pharmaceutical industry, particularly
since the middle of the twentieth century. It is fair to state that the more
science-based pharmaceutical industry that began its gradual develop-
ment in the 1940s was quite distinct from the patent medicines industry
of the earlier era. Daniel Carpenter describes how various interacting po-
litical, scientific, medical, and market changes contributed to a gradual
shift in pharmaceutical policy in the United States, which in turn influ-
enced pharmaceutical policies in other countries. Pharmaceutical policy
became more sophisticated, based on detailed interactions and discussions
between regulatory agencies and pharmaceutical companies. Products
were increasingly submitted to more detailed procedural scrutiny.6 This
culminated in the formal legislative introduction of a general efficacy re-
quirement in the 1960s, in the wake of the Thalidomide disaster.7
5 During the Prohibition era, it is interesting to note that many medicines were pur-
chased for their alcohol or heroin content. See Hilts, supra note 4 at 2327.
6 See Carpenter, supra note 4.
7 The drug Thalidomide, which was marketed in forty-six countries around the world,
was touted as a wonder drug for the treatment of morning sickness in pregnant women.
The drug had, however, catastrophic side effects. It caused phocomelia (or seal limbs,
a serious malformation of the limbs) in an estimated ten thousand newborns. Even
though the tragedy revealed, in the first place, serious problems with drug safety re-
view, it also significantly augmented overall public support for more rigorous review of
both drug safety and efficacy.
Carpenter argues that many historical accounts of the development of drug regula-
tions fail to appreciate that the formal introduction of an efficacy requirement was not
such a radical shift, but was the culmination of a gradual process that had already
started more informally in the 1940s (ibid at 11920, 15051). For a brief overview of
the Canadian history, see Health Canada, Brief History of Drug Regulation in Cana-
da (11 April 2007), online: Health Canada
History of Drug Regulation].
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 949
In the United States, the introduction of the efficacy requirement with
the Kefauver Harris amendments to the FDA was accompanied by the
Drug Efficacy Study (DES), a historically unique and massive regulatory
initiative to evaluate, particularly from an efficacy perspective, the phar-
maceutical products already on the market at the time.8 Between 1966
and 1969, the FDA organized the evaluation of approximately four thou-
sand marketed products, including many top-selling drugs, with support
from many academic specialists in clinical pharmacology and medicine.9
The outcome of the DES constitutes probably one of the most impressive
debunkings of the market-efficiency myth in the context of the market-
ing of pharmaceutical products. Jeremy A. Greene and Scott H. Podolsky
indicate that by the 1970s, the evaluation program had deemed six hun-
dred of the drugs to be ineffective.10 This clearly suggests that market
forces alone ensure neither that only effective products remain on the
market nor that reliable information on the most effective products be-
comes easily available as a result of market competition. Yet the very im-
portant lessons to be heeded from this study appear to be lost on many
commentators.
Notwithstanding the DES and the worldwide introduction of efficacy
requirements in drug regulation, it is fair to state that this did not result
in a more sophisticated assessment of the safety and efficacy of drugs once
they entered the market. Neither did it result in a systematic comparative
review of the merits of new drugs as part of the drug approval process. To
this day, pharmaceutical regulations in both the US and Canada basically
still only require drug manufacturers to provide limited evidence that a
drug has some efficacy, based on a few so-called pivotal trials.11 The divi-
sion of labour within the drug regulatory system has also remained large-
ly the same. Notwithstanding increased scrutiny and growing interaction
8 In his impressive analysis of the historical development of the regulatory power of the
US FDA, Daniel Carpenter points out that even before the formal introduction of this
efficacy requirement, FDA practice had developed in such a way that the regulator did
evaluate efficacy in the context of its safety assessment, rightly arguing, for example,
that an ineffective drug inherently raises safety concerns (Carpenter, supra note 4 at
14956, 18894).
9 Ibid at 34557.
10 Reform, Regulation, and Pharmaceuticals: The Kefauver-Harris Amendments at 50
(2012) 367:16 New Eng J Med 1481 at 1482. See generally The Drug Efficacy Study of
the National Research Councils Division of Medical Sciences, 19661969, online: Na-
tional Academy of Sciences
study). According to Hilts, the 1969 FDA final report concluded that about 7 percent of
the drugs reviewed were completely ineffective for every claim made. Another 50 per
cent of drugs were judged effective to some degree on some claims and ineffective on
others (supra note 4 at 176).
11 See Lemmens & Bouchard, supra note 2 at 32025.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
950
between regulatory agencies and producers, the latter have remained in
control of both the production of the pharmaceutical product and the ac-
companying primarily safety-related data. Regulatory agencies have exer-
cised a marginal level of control based on information submitted by the
industry.
B. Issues in the Production of Clinical Trials Data
The gradually increasing demand for data by regulatory agencies was
obviously intended to enable a more informed regulatory decision at mar-
ket entry. But it has also had an interesting side effect: it has strength-
ened the dominant position of industry over the overall production of sci-
entific data. In the long run, an increasingly demanding regulatory envi-
ronment and the mounting level of data required from pharmaceutical
producers has contributed to a decline in the independence of data gather-
ing and the rise of clinical research led by those with a vested interest in
the approval and promotion of new drug products. The blockbuster model
of pharmaceutical development that has evolved since the 1960swith of-
ten phenomenal profits associated with the sale of a product during the
life of its patenthas increased the financial pressure to collect the data
required for regulatory approval as fast as possible, as any delay under
this model results in a potential loss of thousands, if not millions, of dol-
lars in sales revenue each day. This growing request for data, combined
with the pressure for expedient and efficient data gathering, has led to
the development of a specialized service industry, the clinical trials indus-
try, which has as its exclusive mandate to assist the producers of pharma-
ceutical products with the collection of data needed for regulatory approv-
al.
Clinical trials increasingly became the nearly exclusive business of
specialized contract research organizations (CROs), the key players with-
in this clinical trials industry. The demand for increasingly detailed clini-
cal trials data sets needed for regulatory approval resulted in more indus-
trial data production processes, organized by CROs with direct contractu-
al commitments to pharmaceutical companies. CROs could conduct these
trials more efficiently and collect data more quickly than academic medi-
cal centres, which had originally played a more important role in the con-
duct of clinical drug trials, and which arguably were originally more inde-
pendent from industry. Particularly from the 1980s onward, clinical trials
gradually moved away from academic medical centres to these specialized
CROs,12 which offer the pharmaceutical industry assistance with all the
12 See Robert Steinbrook, Gag Clauses in Clinical-Trial Agreements (2005) 352:21 New
Eng J Med 2160 at 2161; Philip Mirowski & Robert Van Horn, The Contract Research
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 951
services surrounding the increasingly industrialized production of data:
designing the clinical trials, obtaining research ethics approval,13 recruit-
ing human research subjects, collecting and analyzing data, and, often
with the assistance of other specialized entities such as medical communi-
cations agencies,14 writing up the trial results and strategically publishing
and presenting the data. Academic investigators and academic clinical re-
search centres remain to some degree involved in the conduct, analysis,
and publication of clinical trials, but they clearly no longer play a leading
role. In addition, to the extent that they are still involved in clinical drug
trials, academic medical centres have become much more dependent on
corporate funding,15 with various governmental funding agencies also ag-
gressively promoting industry co-funding for various forms of research.16
Academic investigators often function as expert advisors, or are recruited
as key opinion leaders, sometimes primarily to provide lectures touting
the benefits of new products. Others contribute to the recruitment of pa-
tients in multi-centre clinical trials without having any real control over
the final analysis of the full data. There is also a significant body of evi-
dence suggesting that academic authors are frequently only involved as
guest authors after all the research has been conducted and the results
Organization and the Commercialization of Scientific Research (2005) 35:4 Social
Studies of Science 503 at 505.
13 Note that in many jurisdictions, research ethics review itself has gradually been inte-
grated as part of the new research service industry. See the debate about the desirabil-
ity of research ethics as a commercial service in Ezekiel J Emanuel, Trudo Lemmens &
Carl Elliot, Should Society Allow Research Ethics Boards to Be Run As For-Profit En-
terprises? (2006) 3:7 PLoS Medicine 941; Trudo Lemmens & Benjamin Freedman,
Ethics Review for Sale? Conflict of Interest and Commercial Research Review Boards
(2000) 78:4 The Milbank Quarterly 547.
14 See Sergio Sismondo, Ghosts in the Machine: Publication Planning in the Medical Sci-
ences (2009) 39:2 Social Studies of Science 171.
15 Industry funding ofand consequent influence overresearch conducted in academic
centres has increased significantly in recent decades. Sheldon Krimsky reports that in
the US, corporate contributions to research and development in academic institutions
increased by 875 per cent between 1980 and 2000 (Science in the Private Interest: Has
the Lure of Profits Corrupted Biomedical Research? (Lanham: Rowman & Littlefield,
2003) at 7981).
16 See ibid. For more discussion of Canadian developments, see Janet Atkinson-Grosjean,
Public Science, Private Interests: Culture and Commerce in Canadas Networks of Cen-
tres of Excellence (Toronto: University of Toronto Press, 2006); Jocelyn Downie & Mat-
thew Herder, Reflections on the Commercialization of Research Conducted in Public
Institutions in Canada (2007) 1:1 McGill JL & Health 23; Trudo Lemmens, Commer-
cialized Medical Research and the Need for Regulatory Reform in Colleen M Flood, ed,
Just Medicare: Whats In, Whats Out, How We Decide (Toronto: University of Toronto
Press, 2006) 396.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
952
have been prepared for publication by scientific writers working directly
for industry.17
In this new clinical trials environment, trials not only provide the da-
ta necessary for regulatory approval, they have also become the near-
exclusive source of knowledge about new products. Clinical trials have
thus gradually come to serve a double purpose: enabling regulatory ap-
proval, while at the same time providing data that could be used for the
further promotion of an approved product. Strategically placed in leading
medical journals, the dataor at least the data useful for promotional
purposescollected, organized, and written up in close collaboration be-
tween the pharmaceutical industry, CROs, and medical communications
agencies, bestows upon pharmaceutical products the necessary creden-
tials to boost prescriptions and sales.
Since the 1980s, other factors have contributed to the growing im-
portance of clinical trials data within the context of medicine and health
care practice.18 Among the most significant is the growing emphasis on ev-
idence-based medicine (EBM). The EBM paradigm gained in strength in
parallel with the further development of more sophisticated statistical
clinical trials methodologies. A detailed discussion of these developments
exceeds the scope of this paper, but it is worth emphasizing that in the
context of EBM, the evidence produced in clinical trials increased in im-
portance to the detriment of other sources of evidence, such as reports of
individual experiences by clinicians in medical practice.19 The published
results of clinical trials conducted by CROs for pharmaceutical companies
became the most important source of information influencing physician
prescription behaviour, either directly through their reporting in the sci-
entific literature, or indirectly through the influence of these reports on
clinical practice guidelines developed by experts in the field based in part
on a detailed analysis of the scientific literature and available meta-
analyses.
There are, however, serious limitations associated with the publica-
tion of clinical trial results, limitations that undermine their overall relia-
bility. There is little control, other than the peer-review systemwhich is
increasingly criticized for its failure to prevent the publication of flawed
17 See Lemmens, Leopards, supra note 2; Stern & Lemmens, Legal Remedies, supra
note 3.
18 For a more detailed discussion of these factors, see Lemmens, Pharmaceutical
Knowledge Governance, supra note 2 at 17374.
19 See Healy, supra note 2 at 12994.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 953
research20over how these data are presented in the medical literature,
which, as several authors have discussed in detail, has often become a
marketing conduit for pharmaceutical products.21
C. Understanding the Interests of Industry
The pharmaceutical industry has little inherent interest in evaluating
or questioning whether the production of pre-approval data is sufficient or
provides the most meaningful information; this is not what the industry is
asked to do, and there is no incentive for industry to do so once a drug re-
ceives market approval. In fact, the contrary is true: drug producers often
benefit from the deficit of data on drug safety and efficacy caused by, first,
a lack of transparency and of rigorous control during the pre-market
phase, and subsequently, a lack of post-market surveillance. This is not to
say that there are no circumstances under which the pharmaceutical in-
dustry may have significant incentives in evaluating its product in the
post-market phase. There may be concerns that a failure to follow-up spe-
cific preliminary indicators could result in future lawsuits. Companies
may want to expand the approved indications in drug labelling. Addition-
al post-market studies may also be demanded by payers before reimburs-
ing a drug product. More negatively, pharmaceutical companies may also
use large phase III and phase IV trials as marketing tools. These trials
can boost the status of products and accustom physicians and patients to
the new drugs even before they hit the market.22
Yet there are various reasons to be skeptical about the reliability of
post-market data produced by drug manufacturers. The most important
one is that contrary to dominant market ideology, market actors do not
always have an interest in exposing the inadequacy of competitors prod-
ucts, particularly not when the industry as a whole benefits from artifi-
cially created needs. No pharmaceutical company will, for example, en-
thusiastically expose that competitors anti-depressants are too widely
used for often very minor conditions, that antibiotics are overprescribed,
and that the use of cholesterol-lowering drugs may not be the best or the
only way to reduce the risk of heart disease. Even if companies have, in
theory, an interest in showing that their own products are better than
20 See Richard Smith, Peer Review: A Flawed Process at the Heart of Science and Jour-
nals (2006) 99:4 Journal of the Royal Society of Medicine 178. See also John P A Ioan-
nidis, Why Most Published Research Findings Are False (2005) 2:8 PLoS Medicine
696.
21 See e.g. Sismondo, supra note 14.
22 See Lemmens, Leopards, supra note 2 at 645; Carl Elliott, The Deadly Corruption of
Clinical Trials, Mother Jones (September/October 2010), online: Mother Jones
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
954
competing products, they have a shared interest in luring patients into
over-consumption. In fact, particularly when a class of medication is not
or is only marginallyeffective, producers of these products have a joint
interest in avoiding critical analyses. Companies may also sometimes pre-
fer a situation of uncertainty, when it is not so clear which product could
turn out to be better in a head-on comparison. They may want to avoid
engaging in potentially damaging comparative trials that could affect the
status of their own products, since marketing techniques may provide
more efficient and more easily controlled means to promote consumption.
Moreover, illness tends to render people vulnerable. Particularly when
patients are suffering from serious or even life-threatening diseases, they
are much more easily manipulated by misinformation and pushed toward
overconsumption of pharmaceuticals.23
It is becoming increasingly clear that the reliance on industry-
controlled pre-market data combined with the absence of incentives to
conduct reliable post-market surveillance is resulting in serious conse-
quences. In recent decades, the nature of post-market risk has shifted due
to the rise of blockbuster drugs that are so widely prescribedand often
for the long-term treatment of chronic conditionsthat even a minor
change in relative risk can result in thousands of adverse reactions.24 As
various controversies involving massively prescribed pharmaceuticals
have shown, deficiencies during the pre-market phase can create a host of
serious problems that escalate rapidly when a drug is widely promoted af-
ter it enters the market.25 This may in part explain why serious adverse
events appear to have increased in recent decades.26
23 See the more detailed discussion in Lemmens, Pharmaceutical Knowledge Govern-
ance, supra note 2 at 16465. An interesting example of the harm caused by increased
prescription of pharmaceuticals is the threefold increase in deaths resulting from pre-
scription opioids in the US between 1999 and 2007, which is associated with an in-
crease in opioid prescription rates for chronic pain with causes other than cancer: see
Irfan A Dhalla, Navindra Persaud & David N Juurlink, Facing Up to the Prescription
Opioid Crisis (2011) 343:7823 British Medical Journal 569.
24 Between 1998 and 2005, there was an increase in serious adverse drug reactions
(ADRs) in the US from 34,966 to 89,842. During the same period, fatal ADRs increased
2.7-fold, a rate four times faster than outpatient prescriptions, placing ADRs among the
top ten leading causes of death in the US: see Thomas J Moore, Michael R Cohen &
Curt D Furberg, Serious Adverse Drug Events Reported to the Food and Drug Admin-
istration, 19982005 (2007) 167:16 Archives of Internal Medicine 1752, cited in Mary
Wiktorowicz, Joel Lexchin & Kathy Moscou, Pharmacovigilance in Europe and North
America: Divergent Approaches (2012) 75:1 Social Science and Medicine 165
[Wiktorowicz et al].
25 See, for example, the controversy surrounding the widely prescribed pain relief medica-
tion Vioxx (see Topol, infra note 30).
26 See Wiktorowicz et al, supra note 24.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 955
D. The Limitations of Pre-Market Clinical Trials Evidence
Clinical trials primarily detect common and frequently occurring ad-
verse drug reactions; they often miss important reactions that take a long
time to develop or that only occur infrequently.27 Importantly, they are
conducted under controlled conditions that do not necessarily reflect how
a drug will be used in the real world.28 Research subjects in clinical trials
receive the drug under direct medical supervision, are not necessarily ex-
posed to other drug products or suffering from underlying diseases,29 and
are usually younger, healthier, and less diverse than patients in the real
world.30 Further, as will be discussed below, drugs are often prescribed
off-label to patient and disease groups that were never assessed in clini-
cal trials.31
27 Adverse Reaction Information (2012), online: Health Canada
[Health Canada, Adverse Reaction Information].
28 For example, clinical trials for new drugs are of short duration and are conducted in
populations that number from a few hundred to several thousand [patients]; … most
trials exclude the elderly, children, pregnant women, patients with multiple diseases,
and those on medications suspected of interaction with the study drug (Syed
Rizwanuddin Ahmad, Adverse Drug Event Monitoring at the Food and Drug Admin-
istration: Your Report Can Make a Difference (2003) 18:1 Journal of General Internal
Medicine 57 at 57).
29 See Health Canada, Adverse Reaction Information, supra note 27.
30 See Richard Gliklich & Christina DeFilippo Mack, Comparative Effectiveness Re-
search in the Real World (2009) Q3 Next Generation Pharmaceutical 94 at 94. The risk
that important safety data only becomes apparent after many people start using the
medication is highlighted by the highly publicized market withdrawal of Vioxx. The
now infamous drug has been associated with major adverse events, including myocar-
dial infarctions or strokes in tens of thousands of patients (see Eric J Topol, Failing the
Public Health: Rofecoxib, Merck, and the FDA (2004) 351:17 New Eng J Med 1707;
David J Graham, COX-2 Inhibitors, Other NSAIDs, and Cardiovascular Risk: The Se-
duction of Common Sense (2006) 296:13 Journal of the American Medical Association
1653). Another major controversy was the promotion of hormone replacement therapy
(see Jacques E Rossouw et al, Risks and Benefits of Estrogen Plus Progestin in
Healthy Postmenopausal Women: Principal Results from the Womens Health Initia-
tive Randomized Controlled Trial (2002) 288:3 Journal of the American Medical Asso-
ciation 321). In addition, the off-label promotion of the arthritis drug valdecoxib (Bex-
tra), the epilepsy drug gabapentin (Neurontin), and the schizophrenia drug olanzapine
(Zyprexa) resulted in criminal prosecutions, fines and settlements of hundreds of mil-
lions of dollars, and a record-setting settlement of US$2.3 billion in the case of valdecox-
ib (see Gardiner Harris, Pfizer Pays $2.3 Billion to Settle Marketing Case, The New
York Times (2 September 2009), online: The New York Times
31 See Wiktorowicz et al, supra note 24 at 165. Franoise Baylis has pointed out, for ex-
ample, how pre-market clinical trials routinely exclude pregnant women, who subse-
quently receive prescription medication without sufficient evidence base (see Franoise
Baylis, Pregnant Women Deserve Better (2010) 464:7299 Nature 689; Franoise Bay-
lis & Chris Kaposy, Wanted: Inclusive Guidelines for Research Involving Pregnant
Women (2010) 32:5 Journal of Obstetrics and Gynaecology Canada 473). The authors
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
956
There remains a surprising lack of research, publicly funded or other-
wise, on the long-term safety and effectiveness of post-market drugs, de-
spite the fact that regulators, policy-makers, health care providers, and
consumers need this information to make well-informed decisions.32 A re-
lated problem is that clinical trials conducted for market authorization
rarely investigate whether the drug is actually more effective than other
alternative treatments already on the market.33 Additional post-market
studies on both long-term safety and efficacy and comparative effective-
ness are therefore an essential supplement to pre-market clinical trials;34
drug regulatory reforms should include a more explicit requirement for
such studies.35 Unfortunately, in Canada, inferior drugs simply tend to
remain on the market unless a product is found to be unsafe; inferiority to
existing products in itself is typically insufficient to ground market with-
drawal.
Other important limitations relate to the design, conduct, and analysis
of clinical trials. Sponsoring pharmaceutical companies are responsible for
funding the majority of clinical trials, but various studies have shown that
industry-sponsored research is more likely to be biased and yield positive
outcomes than research with alternate sources of sponsorship.36 Various
recommend more systematic requirements for research involving pregnant women, but
this could certainly also extend to the post-market phase.
32 See Lorraine E Ferris & Trudo Lemmens, Governance of Conflicts of Interest in Post-
marketing Surveillance Research and the Canadian Drug Effectiveness Network
(2010) 4:2 Open Medicine E123 at E123.
33 Another related problem worth mentioning is that the evidence required at the point of
drug regulation and approval (occurring at the federal level) is different from that re-
quired at the point of public funding (at the provincial level)this is referred to by
Flood and Dyke as the data divide. In particular, they argue that the evidence of min-
imal efficacy generated for the purposes of regulatory approval does cast light on im-
portant questions around relative effectiveness or cost-effectiveness of drugs
information that is critical to making informed funding and treatment decisions (Col-
leen M Flood & Patrick Dyke, The Data Divide: Managing the Misalignment in Cana-
das Evidentiary Requirements for Drug Regulation and Funding (2012) 45:2 UBC L
Rev 283 at 300).
34 See Gliklich & Mack, supra note 30 at 94.
35 In other countries, regulatory agencies appear to intervene more firmly in this context.
For example, in India, the Intellectual Property Appellate Board recently revoked a pa-
tent held by the pharmaceutical company Roche because of the drugs unproven superi-
ority to other products already on the market (see Kaustubh Kulkarni & Ben Hirschler,
India Revokes Roche Patent in New Blow for Big Pharma, Reuters (2 November 2012)
online: Reuters International
36 According to Lexchin, biases may be introduced through a variety of measures, includ-
ing the choice of comparator agents, multiple publication of positive trials and non-
publication of negative trials, reinterpreting data submitted to regulatory agencies, dis-
cordance between results and conclusions, conflict-of-interest leading to more positive
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 957
reasons can account for this trend, but there is sufficient evidence that
carefully crafted research design, choices made in the context of statistical
analyses, and exclusion of negative findings or over-inclusion of positive
findings contribute to this phenomenon.37
conclusions, ghostwriting and the use of seeding trials (Joel Lexchin, Those Who
Have the Gold Make the Evidence: How the Pharmaceutical Industry Biases the Out-
comes of Clinical Trials of Medications (2012) 18:2 Science and Engineering Ethics 247
at 247 [Lexchin, Those Who Have the Gold]). For an empirical study of the phenome-
non, see also Andreas Lundh et al, Industry Sponsorship and Research Outcome (Re-
view) (2012) 12 Cochrane Database of Systematic Reviews, Art No MR000033.
37 There are a multitude of examples of pharmaceutical companies intentionally withhold-
ing and misrepresenting the results of unfavourable clinical trials in order to advance
drug sales. Two high-profile examples involved the blockbuster drugs Paxil and Cele-
brex. In 2004, the State Attorney General of New York launched a fraud lawsuit
against GlaxoSmithKline (GSK) for concealing clinical trial data about the safety and
efficacy of Paxil. The suit alleged that GSK had conducted at least five studies on the
use of Paxil in children and adolescents, but only published and disseminated one of the
studies (suppressing the negative results of the other four studies) (see Press Release,
Major Pharmaceutical Firm Concealed Drug Information (2 June 2004), online: New
York State Office of the Attorney General
analysis of the Paxil studies in Jon N Jureidini, Leemon B McHenry & Peter R Mans-
field, Clinical Trials and Drug Promotion: Selective Reporting of Study 329 (2008)
20:1/2 International Journal of Risk & Safety in Medicine 73.
Pfizer has similarly been implicated for, among other things, its promotion of
Celebrex: results of a 1999 clinical trial found that elderly patients taking
Celebrex were at a much higher risk of suffering heart problems than patients taking a
placebo. However, Pfizer chose not to publish the study and it was not submitted to the
FDA until June of 2001 (see Alex Berenson & Gardiner Harris, Pfizer Says 1999 Trials
Revealed Risks With Celebrex, The New York Times (1 February 2005), online: New
York Times
Pharmaceutical companies may also use their influence to try to suppress the publi-
cation of negative research results produced by independent researchers. For example,
Nissen reports that GlaxoSmithKline attempted to stop the publication of their meta-
analysis that showed that the diabetes drug Avandia was associated with a significant-
ly increased risk of myocardial infarction (Steven E Nissen, Setting the RECORD
Straight (2010) 303:12 Journal of the American Medical Association 1194).
A host of other products have been associated with equally problematic practices,
including Oxycontin, Neurontin, Zyprexa, Fen-Phen, Prempro, Prepulsid, Depakote,
Actimmune, Avandia, and Risperdal (see e.g. David Evans, Big Pharmas Crime
Spree, Bloomberg Markets (December 2009) 73, online: Canadian Health Coalition
Columbia University Press, 2010); Healy, supra note 2; Moynihan & Cassels, supra
note 2; Young, supra note 2). See also the report by the public interest organization
Public Citizen: Sammy Almashat et al, Rapidly Increasing Criminal and Civil
Monetary Penalties Against the Pharmaceutical Industry: 1991 to 2010 (2010), online:
Public Citizen
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
958
II. Specific Challenges in Post-Market Surveillance
Health Canada itself admits that once a drug is approved, as long as
the drug causes no adverse reactions or the manufacturer does not need
to make changes to the drug, it may never be subject to review by Health
Canada again.38 Yet the drug manufacturer is in charge of collecting and
forwarding adverse events to the regulator and therefore plays a key role
in determining whether changes are needed. It has no financial inter-
ests in diligently pursuing the proactive verification of the safety and effi-
cacy profile of an already marketed drug. It is therefore more accurate to
state that as long as no evidence is being collected and forwarded by the
manufacturer to Health Canada, a drug will rarely be reviewed again.
Moreover, even a totally inferior drug can remain perpetually on the mar-
ket as long as no strong evidence of serious side effects is reported.
Although, as we have emphasized, the current drug regulatory system
focuses on pre-market activities, there are a number of obligations that
manufacturers are required to adhere to once their drug enters the mar-
ket. These include informing Health Canada of any reported serious ad-
verse drug reactions (ADRs) to their product; complying with advertising
restrictions set out in the Food and Drugs Act and regulations; updating
safety information pertaining to their products; maintaining the quality of
their drug to the appropriate standard; and applying for further authori-
zation from Health Canada for significant changes to their product.39
Unfortunately, Health Canada has only a limited ability to ensure
continued compliance with regulations once a drug enters the market.40
Health Canada traditionally appeared to lack the power to compel manu-
facturers to conduct new efficacy or safety studies or studies on therapeu-
tic effectiveness once a drug product hits the market.41 Although drug
manufacturers are encouraged to conduct post-market studies and to
comply with post-market commitments, the current evidence suggests
that they often fail to do so and that the regulatory agencies do not suffi-
38 Health Canada, Brief History of Drug Regulation, supra note 7.
39 See Health Canada, Guidance Document: Submission of Pharmacogenomic Information
(2008), online: Health Canada
40 Some of the few measures available are inspecting manufacturing compliance and re-
moving a drug from the market for safety reasons (ibid).
41 Health Canada believes it lacks the regulatory authority to explicitly require post-
market studies as a condition for further sales of a pharmaceutical product (see Mary E
Wiktorowicz et al, Keeping an Eye on Prescription Drugs, Keeping Canadians Safe:
Active Monitoring Systems for Drug Safety and Effectiveness in Canada and Interna-
tionally ((Health Council of Canada, November 2010) at 45, online: Government of
Canada
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 959
ciently monitor and enforce post-market commitments.42 As such, there is
a need for ongoing assessment of pharmaceutical industry compliance
with post-market commitments and for appropriate sanctions for a failure
to meet these commitments.43 Since 1998, Health Canada has offered a
special type of authorization, a Notice of Compliance with conditions
(NOC/c), which allows the sponsor to market a drug in Canada on the
condition that they undertake additional studies to confirm the clinical
benefit of the product.44 The NOC/c policy allows for earlier market access
to potentially life-saving drugs, while the conditions tied to the authoriza-
tion enable Health Canada to monitor the drug through enhanced post-
market surveillance. Yet there are questions about Health Canadas con-
trol of the fulfillment of these conditions. In 2003, the original policy was
modified following complaints that it was not being consistently applied
and that there was a greater need to disseminate materials to accompany
products that have been issued a NOC/c.45 Joel Lexchin suggests there
continues to be little oversight for drugs once they are approved under the
NOC/c policy.46
Other jurisdictions have strengthened more significantly the power of
regulatory agencies to impose post-market studies under regular approval
procedures. In the US, for example, the Food and Drug Administration
Amendments Act of 2007 expanded the FDAs authority to require manu-
facturers to conduct post-market studies (including clinical trials and ob-
servational studies) after market approval if new safety information
42 See Ferris & Lemmens, supra note 32 at E123.
43 As recommended in a recent Health Council of Canada discussion paper, [a]ny fines
that are imposed for failure to complete Phase IV trials must be significant enough to
achieve their objective; fines that are too small will not have any value (Wiktorowicz et
al, Health Council Paper, supra note 41 at 41). Unfortunately, Health Canada has re-
ceived significant criticism in recent years for its lack of enforcement activities. Of par-
ticular concern is the fact that the drug approval system is largely funded by the phar-
maceutical industry through a user-fee system. Through these financial contributions,
the industry has gained significant influence within the drug regulatory system and us-
es this influence to request a disproportionate distribution of funding towards fast drug
approval, rather than long-term safety monitoring (see generally Tracey Epps, Regula-
tion of Health Care Professionals in Downie, Caulfield & Flood, supra note 2 at 100).
44 Health Canada, Guidance Document: Notice of Compliance with conditions (NOC/c),
(2011), online: Health Canada
45 See Joel Lexchin, Notice of Compliance with Conditions: A Policy in Limbo (2007) 2:4
Healthcare Policy 114 at 116.
46 Ibid. See also Wiktorowicz et al, Health Council Paper, supra note 41 at 12 (discussing
examples of cases in which Health Canada appears to have dropped the ball in terms of
a follow-up after expedited approval).
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
960
comes to light.47 The FDA can mandate that these studies be completed on
a definite timetable. Following the passage of new pharmacovigilance
legislation by the European Union in 2010, European regulatory authori-
ties now also have the authority to require manufacturers to conduct post-
market studies within the framework of a Risk Management Plan. Reg-
ulatory authorities may mandate both safety and efficacy studies … and
the obligation may be imposed at the time of authorisation, or [after ap-
proval] if an important safety concern emerges.48
Fortunately, the federal government finally appears to be ready to
take concrete steps toward enhancing regulatory oversight during the
post-market phase. In December 2013, it tabled new legislation, Bill C-17
(Vanessas Law)49, that includes provisions granting Health Canada the
authority to compel drug companies to compile information, conduct new
tests or studies, and monitor the use of drug products after market entry.
An important complement to these amendments is a significant increase
in penalties for violations of the Food and Drugs Act and its associated
regulations: while the current maximum fine is $5,000, the new maxi-
mum on indictment is $5 million per day that the contravention is contin-
ued.50 Moreover, where a person knowingly or recklessly causes a serious
risk of injury to human health,51 there is no maximum fine set out, giving
the court the discretion to impose even more substantial fines.
Overall, the expansion of Health Canadas authority during the post-
market phase provided for in Bill C-17, including the ability to impose
further testing, represents an important step forward in strengthening
the integrity of the drug regulatory system, particularly when backed by
the threat of much more substantial penalty provisions. The bill undoubt-
edly represents an improvement over the status quo. Yet there are a
number of important omissions in the legislation. As Herder and col-
leagues also point out, the bill does nothing to improve the transparency
of clinical trials or Health Canadas decision-making process around drug
47 See the discussion in Barbara J Evans, Seven Pillars of a New Evidentiary Paradigm:
The Food, Drug, and Cosmetic Act Enters the Genomic Era (2010) 85:2 Notre Dame L
Rev 419.
48 Patrick Waller, Getting to Grips with the New European Union Pharmacovigilance
Legislation (2011) 20 Pharmacoepidemiology and Drug Safety 544 at 545.
49 Bill C-17, An Act to Amend the Food and Drugs Act, 2nd Sess, 41st Parl, 2013.
50 Ibid, ss 21.31, 21.32, 31.2. Where an offence is committed or continued over multiple
days, the maximum fine is multiplied by the number of days, significantly increasing
the potential upper limit of the penalty. The maximum fines for a summary conviction
are increased to $250,000 for a first offence and $500,000 for subsequent offences.
51 Ibid, s 31.4.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 961
approval52; the central importance of such transparency provisions is dis-
cussed in more detail below. Further, the ultimate success of these provi-
sions will depend on Health Canada having both the resources and politi-
cal will to police and enforce the new provisions. Bill C-17 commenced
second reading in April 2014 and will hopefully be passed later in 2014.
A. Adverse Drug Reaction Reporting
In Canada, the most common means of monitoring the ongoing safety
and effectiveness of pharmaceuticals is through voluntary reporting of
(actual or suspected) ADRs by health professionals and consumers. Ac-
cording to Health Canada, [t]hese individual reports may be the only
source of information concerning previously undetected [ADRs] or chang-
es in product safety and effectiveness profiles.53 Although pharmaceutical
manufacturers are required to report ADRs they become aware of to
Health Canada, physicianswho are on the front lines in encountering
ADRs in patients and therefore in a better position to gather this infor-
mationonly need to submit ADR reports on a voluntary basis. Reporting
is therefore conditional on doctors having the time and inclination to do
the additional paperwork.54
Mandatory physician reporting of ADRs is one measure that had been
suggested to strengthen post-market surveillance.55 Unfortunately, physi-
cian groups are resistant to mandatory reporting because the requirement
would be time and labour intensive. Moreover, trying to enforce such a
requirement would likely be very complex and time consuming for regula-
tory authorities. At the very least, physicians would want to be compen-
sated for that time. Indeed, even in the absence of mandated reporting,
simply remunerating physicians to report ADRs could provide a sufficient
incentive to substantially increase the amount and quality of reports re-
ceived. In addition, the expanded use of electronic medical records has
been suggested as an important means of improving the quality and
52 Matthew Herder et al, Regulating Prescription Drugs for Patient Safety: Does Bill C-
17 Go Far Enough? (2014) 186:8 Canadian Medical Association Journal E287.
53 Health Canada, Adverse Reaction Information, supra note 27.
54 Lauren Vogel & Tomek Sysak, Physicians Should Be Paid to Report Adverse Drug Re-
actions, Experts Say (2012) 184:8 Canadian Medical Association Journal E409.
55 At a minimum, some commentators suggest that Canada should follow New Zealands
lead and require doctors to report all instances in which one of their patients suffer[s]
an adverse effect from the use of a drug or a combination of drugs (ibid).
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
962
quantity of ADR reports56and one that could streamline the reporting
process and reduce the time commitment required of physicians.
Any measures to increase ADR reporting need to be coupled with im-
provements in the systems for collecting, storing, and analyzing ADR da-
ta;57 the problem is not simply a lack of data, but a failure to structure da-
ta in an interpretable manner. Presently, the data collected by Health
Canadas adverse drug reaction reporting tool, MedEffect, is sporadic and
often incomplete. Moreover, the MedEffect tool itself has been described
as poorly structured and as compiling information that is often unin-
terpretable.58 Many physicians are sceptical of the value of simply collect-
ing more data since it is unclear how Health Canada uses the infor-
mation. One way Health Canada could counter this is through being more
open about how ADR reports inform the post-surveillance process. Health
Canada should also receive more explicit powers to impose more stringent
and independent ADR reporting requirements as part of the approval pro-
cess.
Finally, it is worth mentioning here that privately developed data-
bases are being developed, partly in response to the limited adverse event
data collection of governments.59 An interesting initiative to improve both
the reporting of adverse events and the availability of the resulting data is
the website RxISK.org. The website is described as the first free website
(that is, not sponsored by the pharmaceutical industry or by advertising)
to provide a means to easily report side effects to assist in individual pa-
tient care and to help other patients by identifying problems and possible
solutions earlier than is currently happening.60 RxISK.org is funded by
selling subscriptions to the anonymous, aggregated data collected through
the site. The initial basis for the RxISK adverse events database is cur-
56 See Standing Senate Committee on Social Affairs, Science and Technology, Prescription
Pharmaceuticals in Canada: Post-Approval Monitoring of Safety and Effectiveness
(March 2013) at 10, online: Parliament of Canada
Pharmaceuticals].
57 Members of the medical community have argued that the process for making an ADR
report should be easier and should incorporate sufficient detail to draw conclusions on
cause and effect (ibid). A common way of recording data should also be established (see
Health Council Paper, supra note 41 at 40).
58 For example, clinically important information, such as the time between a patient
starting a drug and experiencing an adverse effect, must be dumped in free text entry
fields with a jumble of other information (see Vogel & Sysak, supra note 54 at E409).
59 See Shannon Gibson, Direct-to-Consumer Advertising in the Digital Age: The Impact of
the Internet and Social Media in the Promotion of Prescription Drugs in Canada (LLM
Thesis, University of Toronto, 2012) at 71 [unpublished, archived online at University
of Toronto
60 RxISK, About Us, online: RxISK
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 963
rent data supplied by the US FDA, with new adverse events data to be
contributed by patients from around the worldthis information is made
anonymous and added to the database in real time.61 As another exam-
ple, AdverseEvents, Inc. (AEI)62 also uses data taken from the FDA to
create a database on side effects associated with FDA-approved prescrip-
tion medications.63 Users can pay to gain access to detailed information,
such as the relationships between adverse events and patient de-
mographics or prescription regimens, or outcomes such as hospitalization
or death. These tools can be used to track potential trends and problems
in the pharmaceutical industry.
The success of both RxISK and AEI depend in particular on consum-
ers to directly report ADRs to these databases. Given the reluctance of
many physicians to be burdened with the requirement to increase ADR
reporting, consumers themselves will likely play an increasingly im-
portant role in reporting ADRs. Many consumers already share their ex-
periences with drugs through online patient communities and would like-
ly be eager to participate in initiatives like RxISK and AEI. While these
online resources are certainly not a replacement for government-run re-
porting systems and rigorous scientific studies, they could nonetheless
play an important role in raising warning flags about potentially serious
ADRs earlier than might otherwise be possible.64
B. Issues Around Off-Label Prescribing
As in many countries, off-label prescribing (that is, prescribing for us-
es for which drugs have not been officially approved) is legal in Canada
based largely on the premise that regulators lack the authority to control
the practice of medicine.65 Many physicians are in fact unaware that they
61 Ibid.
62 See AdverseEvents, About AdverseEvents, online: AdverseEvents
63 Users can search the database for specific drugs to find the percentage of respondents
that reported a specific adverse effect or the percentage that found a drug to be ineffec-
tive. According to their press material, AEI has created a unique set of online tools that
is optimized to provide unparalleled access to adverse event information on over four
thousand drugs, in an easy to understand and navigate format (AdverseEvents, Ad-
verseEvents Launches Innovative Comparative Drug Side Effect Reporting System at
Health 2.0 Conference, online: AdverseEvents
64 See Gibson, supra note 59 at 73.
65 However, it is worth noting that Health Canada prohibits the promotion of off-label use
by the drug manufacturer. Specifically, section 9(1) of the Food and Drugs Act prohibits
false, misleading or deceptive advertising. According to Health Canada, messages dis-
cussing off-label use of a drug product could be considered a contravention of section
9(1) (Health Canada, Regulation of Health Products Advertising in Canada: Overview
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
964
are prescribing drugs for off-label use.66 Given the complexity of medical
practice, as Patrick OMalley argues, it would be an enormous task to
support labeling for every possible potential use.67 A certain amount of
off-label prescribing can arguably therefore be acceptable, to allow physi-
cians some flexibility in offering therapeutic options. Off-label prescribing
can contribute to innovation in clinical practice, particularly when ap-
proved treatments have failed[;] [i]t offers patients and physicians earlier
access to potentially valuable medications based on emerging evidence;
and it may even provide the only available treatments for orphan condi-
tions.68 However, as Joseph Emmerich, Nathalie Dumarcet, and Annie
Lorence rightly emphasize, such prescriptions must remain the exception
to the rule and should be scrutinized and controlled by regulatory agen-
cies using well-defined frameworks.69 To the extent that off-label pre-
scription is acceptable, it has to be integrated in a framework that pro-
motes evidence gathering.
Pharmaceutical companies are responsible for submitting the neces-
sary safety and efficacy data to Health Canada to support the use of a
drug for a specific indication. Additional indications may be added to a
drugs label through a supplemental new drug submission. Generating
this data requires substantial resources and financial risk.70 Where a
for Physicians (2011), online: Health Canada
the FDA also prohibits the promotion of off-label prescribing, a recent federal appeals
court decision threw out the conviction of a sales representative who sold a drug for us-
es not approved by the [FDA] [on the grounds that] the ban on off-label marketing vio-
lated the representatives freedom of speech (Katie Thomas, Ruling Is Victory for
Drug Companies in Promoting Medicine for Other Uses, The New York Times (3 De-
cember 2012), online: The New York Times
concerns about the enforcement of restrictions on off-label prescription promotion, at
least in the US. However, the right to freedom of expression is construed much more
broadly in the US legal system than in Canada, so it is highly unlikely that a Canadian
court would reach the same conclusion.
66 [S]ome off-label prescribing may be driven by doctors mistaken belief that certain
drugs are actually approved … for such use. In a 2009 survey of 1,200 psychiatrists
and primary care doctors … [t]he average doctor was wrong about the FDA approval
status of 45 percent of the drugs he or she was asked about. In the study, doctors were
asked to identify whether 14 common drugs were approved by the FDA for treatment
of various illnesses (Monifa Thomas, Many Docs Off Mark on Off-Label Scripts, Chi-
cago Sun-Times (23 August 2009) 11A).
67 Patrick G OMalley, What Does Off-Label Prescribing Really Mean? (2012) 172:10 Ar-
chives of Internal Medicine 759 at 759.
68 Randall S Stafford, Regulating Off-Label Drug Use: Rethinking the Role of the FDA
(2008) 358:14 New Eng J Med 1427 at 1427.
69 Frances New Framework for Regulating Off-Label Drug Use (2012) 367:14 New Eng
J Med 1279 at 1280.
70 See OMalley, supra note 67.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 965
brand-name drug is already widely used off-label, manufacturers have lit-
tle incentive to conduct costly clinical trials to support a new indication,
especially since such trials could potentially yield non-supportive evi-
dence.71 Randall Stafford warns that off-label prescribing undermines the
incentives for manufacturers to perform rigorous studies.72 Many of the
serious controversies in the last couple of years relate to the subtle and
not so subtle promotion by industry of off-label prescription.73
Drug labels provide information about the identified benefits and risks
of drug products to help guide physicians and patients in drug treatment
decisions. By disregarding the drugs labelling, off-label prescribing un-
dercuts expectations that drug safety and efficacy have been fully evalu-
ated [and] may discourage evidence-based practice.74 A 2012 study re-
vealed that around 11 per cent of prescription drugs in Canada are pre-
scribed for off-label use, and of these, 79 per cent lacked strong scientific
evidence for the off-label use.75 A similar study conducted in the US in
2006 found that 21 per cent of prescriptions were for off-label uses, and of
that number, 73 per cent had little or no scientific support.76
It is difficult to judge the scope or severity of the risk presented by off-
label prescribing without knowing the corresponding clinical outcomes.
Off-label prescribing is just one aspect of the broader issue of inappropri-
ate use of medications.77 The prevalence of off-label prescribing reflects
the fact that there is inadequate monitoring of the use of medications dur-
ing the post-market phase, particularly with regard to whether medica-
71 See Stafford, supra note 68.
72 Ibid at 142728. Stafford further states that off-label prescribing subtly encourages
[drug companies] to game the system by seeking approval for secondary indications for
which clinical trials are less complicated and less expensive (ibid).
73 See e.g. Adriane Fugh-Berman & Douglas Melnick, Off-Label Promotion, On-Target
Sales (2008) 5:10 PLoS Medicine 1432.
74 Stafford, supra note 68 at 142728.
75 Off-label use was highest for central nervous system drugs (26.3 per cent), including an-
ticonvulsants (66.6 per cent), antipsychotics (43.8 per cent), and antidepressants (33.4
per cent) (see Tewodros Eguale et al, Drug, Patient, and Physician Characteristics As-
sociated with Off-Label Prescribing in Primary Care (2012) 172:10 Archives of Internal
Medicine 781).
76 David C Radley, Stan N Finkelstein & Randall S Stafford, Off-Label Prescribing
Among Office-Based Physicians (2006) 166 Archives of Internal Medicine 1021 at 1024.
Note that one difference between this American study and the Canadian Eguale study
(supra note 75) that may account for the higher reported prevalence of off-label pre-
scribing in the former is that it included prescribing to children, whereas the Canadian
study did not.
77 As stated by OMalley, off-label use may actually be only a very crude marker of inap-
propriate use (supra note 67 at 759).
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
966
tions are being prescribed appropriately. Therefore, reform efforts need to
focus on moving toward better measurement and assessment of drug use
based on clinical outcomes and on explicit and stringent requirements to
conduct post-market clinical research to obtain reliable evidence of the ef-
fects of drugs in real-world settings.78
Effective monitoring of off-label drug use requires the co-operation of
the medical and scientific communities to identify and collect information
about emerging off-label uses. Physician buy-in, in particular, would be
crucial to implementing successful systems for the monitoring of off-label
use since such systems would likely rely on physicians to share infor-
mation about side effects and patient outcomes observed during clinical
practice. However, as with mandatory ADR reporting, physician groups
would likely be resistant to any increased monitoring obligations because
the requirements could potentially be time and labour intensive. Both
Lexchin and OMalley note that the movement toward electronic health
records would be an important development in improving our ability to ef-
fectively track both labelling indication and patient outcomes more sys-
tematically;79 electronic health records may thus facilitate the collection of
this information without requiring a significant reporting burden on phy-
sicians.
An example of a new regulatory initiative to control off-label use is
Frances recently launched Temporary Recommendations for Use (TRU)
mechanism, which provides for temporary supervision of the prescribing
of drugs for indications for which they are not yet approved. The TRU
process establishes an observation window of up to three years to assess
the benefits and risks of a marketed drug for an unlicensed indication and
to collect scientific information to ensure its safe use.80 One potential
drawback in the design of the TRU process is that pharmaceutical firms
are given the responsibility for controlling off-label prescribing: they
must monitor prescriptions adherence to marketing authorizations and
must not market their drugs for unlicensed indications.81 As has been
78 In particular, OMalley advocates an expanded comparative effectiveness research
agenda [that focuses on] ways to more accurately link and monitor use with indication
and disease identification with efficacious interventions (ibid at 75960).
79 Comments by Joel Lexchin in Kate Lunau, Off-Label Drugs Are Off the Charts in
Canada, Macleans (29 May 2012), online: Macleans
supra note 67.
80 Emmerich, Dumarcet & Lorence, supra note 69 at 1280. Several factors must be con-
sidered and carefully balanced by an expert committee before a TRU can be issued: (1)
the quality of the scientific evidence; (2) the drugs safety profile; (3) the prognosis asso-
ciated with the given disease; and (4) the frequency of the diseases occurrence.
81 Ibid.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 967
demonstrated with the significant bias that arises in industry-controlled
clinical trials, it is questionable whether the pharmaceutical industry
should be given the responsibility for monitoring off-label prescribing,
since they tend to benefit from the practice. While the success of the TRU
approach has yet to be proven, it nonetheless represents a possible means
to encourage the development of new uses for marketed drugs while still
monitoring the benefit-risk ratios for new indications.
C. Improving Information Sharing with Patients and Consumers
The goal of improved post-market surveillance is to ameliorate patient
and consumer protectionnot only by strengthening the knowledge base
of regulators and health care decision makers, but also by improving pa-
tient and consumer knowledge and understanding. They are the ultimate
beneficiaries of the entire drug regulatory system. This is not unique to
post-market surveillance. Since this paper focuses on how processes can
be developed to create tools that underlie better knowledge production in
the post-market phase, it exceeds the scope of this paper to discuss all the
possible ways in which patients and consumers can be better informed or
protected against misinformation. But it is worth emphasizing, briefly,
the crucial role of improved consumer information in the context of the
surveillance system itself.
We already indicated that the assessment of pharmaceutical product
safety ought not to be a one-dimensional process based on a one-point-in-
time assessment by the regulator of data submitted by industry. It is not
a process where patients and consumers ought to be purely passive re-
ceivers of information. Not only should they be involved in the decision-
making process surrounding their care, they should also play a role in
feeding back essential information on pharmaceutical products for the
benefit of other patients and consumers. For example, in the context of
off-label prescribing, it is essential that they are fully informed of the lim-
ited information available on the safety and efficacy of a particular prod-
uct (which has not been tested for their specific condition or for their spe-
cific patient population), so that they are alerted to the need to be particu-
larly cautious, and are thereby empowered to identify possible problems
faster. As noted above, patients are expected to play an increasingly im-
portant role in the context of ADR reporting, whether it is reporting for
off-label prescribing or not. But in order to do so, patients and consumers
have to be adequately informed in the first place.
There is a growing realization that good knowledge translation is es-
sential in the context of health care. Yet experts have pointed out how the
risk information about marketed products provided by regulatory agen-
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
968
cies often fails to reach busy health care providers.82 It is even less likely
to adequately inform patients. For example, while Health Canada re-
quires that the product monograph that accompanies approved pharma-
ceutical products also contain information in lay language, most patients
are not aware that they can access this information on the regulators
website.83 This exacerbates a more fundamental problem: the reliability
and completeness of the information being transmitted, which is largely
based on data produced by industry.
Clearly, improving post-market surveillance will require improving
communication with patients and health care providers. Some interesting
initiatives in this context are worth mentioning here briefly as examples
of what can be done. In Canada, the Minister of Health announced in
June 2013 a new Plain Language Labelling initiative to improve the
basic information provided to patients when they receive pharmaceutical
products.84 Those who have been critical of the existing drug regulatory
system, such as parliamentary member Terence Young, who has been ac-
tively pushing for better patient protection in response to the death of his
daughter due to the drug product Prepulsid, have been arguing for years
that the basic information provided to consumers in Canada at the time of
purchase of pharmaceutical products was problematic, and less clear than
in other countries such as the United States.85 A coroners inquest into the
death of Vanessa Young recommended in 2001 that the information pro-
vided in the labelling of drug products should be improved.86 The Plain
Language Labelling initiative aims to make relevant information more
accessible though standardization of forms, provision of a basic table of
facts about the drug (a so-called drug facts table), and the use of accessi-
ble language. Interestingly, it also explicitly includes plans to promote
feedback from consumers by introducing mandatory contact information
that patients can use to report adverse reaction reporting. This laudable
initiative thus recognizes the connection between communication of risk
information and ongoing interaction between patients, health care provid-
ers, pharmaceutical producers, and the regulator.
82 See Prescription Pharmaceuticals, supra note 56 at 11. See also Lisa M Schwartz &
Steven Woloshin, Lost in Transmission: FDA Drug Information That Never Reaches
Clinicians (2009) 361:18 New Eng J Med 1717.
83 See Prescription Pharmaceuticals, supra note 56 at 12.
84 See Health Canada, News Release, Harper Government Launches Plain Language
Labelling Initiative to Improve Drug Safety for Canadians: Proposed Changes Will
Make It Easier for Canadians to Read Drug Labels (14 June 2013), online: Health
Canada
85 See e.g. Young, supra note 2 at 139.
86 Ibid at 31213 (a discussion of these recommendations and others regarding improved
communications).
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 969
In the United States, the FDA already has more consumer-oriented
communication tools at its disposal. This includes a Black Box warning
system,87 used to emphasize serious risks associated with specific prod-
ucts, and a system of Medication Guides to be handed out to consumers
for specific pharmaceutical products when the FDA deems that certain
information is necessary to prevent serious adverse effects or when spe-
cific instructions are important to ensure the effective use of a product.88
Still, pharmaceutical policy experts Lisa Schwartz and colleagues argue
that patients in the US are not sufficiently or accurately informed with
the information they receive. In particular, they point out that there is lit-
tle information about how well a product works, and that patients tend to
be overwhelmed by a laundry list of side effects without clear emphasis on
what they may most commonly experience. They have therefore recom-
mended the introduction of a very basic Drug Facts Box, a single page
with essential information on drug products. They have successfully test-
ed this system in two randomized trials.89 When implementing its new
Plain Language Labelling system, Health Canada may take some les-
sons from these findings.
III. Tackling the Data Deficit
A. The Need for Independent Research
While communication is essential, its success depends on the reliabil-
ity of the knowledge to be transmitted. Given the evidence that clinical
trials conducted by industry for the purpose of approval of drug products
are more likely to be biased and yield positive results, the same problems
would likely arise in industry-sponsored post-market research.90 The ideal
87 A Black Box warning is a prominently displayed box that the FDA may require to be
placed on the label of a pharmaceutical product with appropriate warning of specific
problems, and particularly those that may lead to death or serious injury (21 CFR
201.57(c) (2013)). For an analysis of the FDAs use of Black Box warnings, see Judith E
Beach et al, Black Box Warnings in Prescription Drug Labeling: Results of a Survey of
206 Drugs (1998) 53:3 Food & Drug LJ 403.
88 US Food and Drug Administration, Medication Guides, online: FDA
89 Lisa M Schwartz, Steven Woloshin & H Gilbert Welch, Using a Drug Facts Box to
Communicate Drug Benefits and Harms: Two Randomized Trials (2009) 150:8 Annals
of Internal Medicine 516.
90 Various studies have revealed a statistical correlation between study outcomes and
funding source, reports of misleading selection of trial designs, and the exposure of in-
stances of data suppression, data misrepresentation, ghost authorship or research arti-
cles by industry-funded writers and other related practices (Ferris & Lemmens, supra
note 32 at E12324). See also Lemmens, Leopards, supra note 2 at 653; Wayne A Ray
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
970
solution to this fundamental problem would be to remove industry control
over the conduct of all clinical trials and improve government oversight of
clinical trials.91 This need not rise to the level of state-run clinical trials
centres;92 simply removing the direct relation between the clinical trials
industry and the pharmaceutical companies would already be an im-
portant step forward in improving the quality of clinical trials. Some ju-
risdictions have already implemented a system of governmentally com-
missioned, publicly funded post-market studies. In New Zealand, for ex-
ample, the Medicines and Medical Devices Safety Authority (Medsafe)
oversees a limited number of post-market studies through its contracted
research centre, the National Pharmacovigilance Centre at the University
of Otago. As a matter of policy, Medsafe prefers to conduct the studies
through its research centre rather than have the clinical trials industry
involved. Moreover, the organization has no legal authority to request
studies from drug sponsors.93
& C Michael Stein, Reform of Drug Regulation: Beyond an Independent Drug-Safety
Board (2006) 354:2 New Eng J Med 194.
91 See recommendation in Lemmens, Leopards, supra note 2 at 653 (see additional ref-
erences there to others who have also made this recommendation). See also Jerome H
Reichman, The Eleventh Annual Honorable Helen Wilson Nies Memorial Lecture in
Intellectual Property: Rethinking the Role of Clinical Trial Data in International Intel-
lectual Property Law: The Case for a Public Goods Approach (2009) 13:1 Marq Intell
Prop L Rev 1 at 5864; Marc A Rodwin, Independent Clinical Trials to Test Drugs: The
Neglected Reform (2012) 6 Saint Louis University Journal of Health Law & Policy 113.
92 For more detailed discussions of the idea of independent clinical trials, see Reichman,
supra note 91 at 5864; Lexchin, Those Who Have the Gold, supra note 36 at 25758;
Rodwin, supra note 91.
93 See Health Council Paper, supra note 41 at 30. The commissioning of independent
clinical trials is categorically different from some other interesting initiatives, which
aim to ensure more rigorous independent assessment of data. See, for example, the Yale
University Open Data Access (YODA) Project, which enables industry sponsors to
commission independent analyses of their clinical trials by two independent research
teams, with strict conditions of full data access (including access to all patient-level clin-
ical data) and full transparency and independent publication of the results (Yale School
of Medicine, Yale University Open Data Access (YODA) Project: A New Approach to
Evaluation and Transparency, online: Yale School of Medicine
See also Harlan M Krumholz & Joseph S Ross, A Model for Dissemination and Inde-
pendent Analysis of Industry Data (2011) 306:14 Journal of the American Medical As-
sociation 1593). Two independent research teams recently published the first independ-
ent analyses conducted through the YODA Project. The analyses were funded by the
company Medtronic and focused on Medtronics recombinant human bone morphoge-
netic protein-2, an orthobiologic agent used to promote bone growth in spinal surgery.
The research team analyzed all the publicly available and company held data, and also
re-evaluated the published literature on the product. Not surprisingly, the teams con-
cluded that the early publications of the product underreported adverse events and em-
phasized results favourable to the product. The product was deemed not to offer any ef-
ficacy advantage and was associated with increased risks, even if the product was still
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 971
The more drastic reform of fully removing the control over the conduct
of clinical trials from the pharmaceutical producers is necessary, in our
view. It is unlikely to happen anytime soon, however, considering the
aversion of current governments toward firmer governmental oversight
over industry practices and also, in Canada in particular, the recent sig-
nificant reductions in the federal bureaucracy. Even if such a reform
would not necessarily require a huge new bureaucracy, it would certainly
require some higher level of governmental control and more active in-
volvement of regulatory agencies in the organization of clinical trials. It is
therefore worth mentioning one partially positive initiative in Canada in
this area: the establishment by the federal government of the Drug Safety
and Effectiveness Network (DSEN). This organization supports inde-
pendent and scientifically rigorous real-world studies on the safety and ef-
fectiveness of post-market drugs in Canada and will connect to the re-
search being conducted through a virtual network of Canadian centres of
excellence in postmarketing pharmaceutical research.94 The DSEN can
be an important source of funding for independent post-market research
on the safety and effectiveness of drugs after they enter the market
research that serves as an important supplement to mandatory ADR re-
ports and any post-market studies conducted by industry.95 Yet the DSEN
has two important limitations: its budget pales in comparison to the mar-
keting and clinical trials budgets of large pharmaceutical companies, and
its independence is weakened by the fact that it is housed under the Ca-
nadian Institutes of Health Research (CIHR), which itself is now aggres-
sively promoting close collaboration with industry, particularly in the or-
ganization and conduct of clinical trials. In the last couple of years, the
CIHR has also appointed two pharmaceutical industry executives to its
Governing Council, appointments that highlight the closer ties with in-
dustry and create a perception of lack of independence.96 Considering the
seen as being a valuable treatment option in specific circumstances (see Mark C Sim-
monds et al, Safety and Effectiveness of Recombinant Human Bone Morphogenetic
Protein-2 for Spinal Fusion: A Meta-Analysis of Individual-Participant Data (2013)
158:12 Annals of Internal Medicine 877; Rongwei Fu et al, Effectiveness and Harms of
Recombinant Human Bone Morphogenetic Protein-2 in Spine Fusion: A Systematic Re-
view and Meta-Analysis (2013) 158:12 Annals of Internal Medicine 890; Harlan M
Krumholz et al, A Historic Moment for Open Science: The Yale University Open Data
Access Project and Medtronic (2013) 158:12 Annals of Internal Medicine 910).
94 Ferris & Lemmens, supra note 32 at E123.
95 Ibid at E124.
96 See Ann Silversides, Appointment of Pfizer Executive to CIHR Stirs Controversy
(2009) 181:11 Can Med Assoc J E256; William Ghali, Claire Kendall & Anita Palepu,
Pharmaceutical Industry Representation on CIHRs Governing Council (2010) 1:1
Open Medicine E26; Steven Lewis, Neoliberalism, Conflict of Interest, and the Gov-
ernance of Health Research in Canada (2010) 1:1 Open Medicine E28; Franoise Bay-
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
972
overwhelming evidence of problems with industry-controlled clinical tri-
als, these developments are worrisome and undermine the potentially
positive role the DSEN could play in establishing trust in the clinical tri-
als system and promoting independent post-market research.
In a 2013 report on post-approval surveillance, the Canadian Standing
Senate Committee on Social Affairs, Science and Technology acknowl-
edged these concerns, which were raised both at its hearings on pharma-
ceutical regulation and in a memorandum we submitted to the Committee
in the context of these proceedings. In response, the Senate Committee
called for a further independent assessment of the DSENs ability to work
independently from CIHR and Health Canada97 but it declined to make
any specific recommendations about its reporting structure, noting that
CIHR is currently viewed by Canadians as a trustworthy and well-
regarded organization.98 In our view, the high public esteem of an organi-
zation is of little relevance to a relatively straightforward institutional
conflict of interest analysis, but the Senates prudent recommendation
opens the door for further reflection on the DSENs structure. The Senate
Committee also recommended that the Minister of Health provide assur-
ances of the DSENs financial sustainability, and that a publicly account-
able oversight mechanism be established that would regularly evaluate
its activities.99 Finally, the Senate Committee suggested that the DSEN
ought to develop an active post-market surveillance program focusing on
adverse drug reactions.100 If implemented, those recommendations should
strengthen the independence and public accountability of the DSEN and
improve post-market surveillance in Canada. But clearly, much will de-
pend on the financial means that are put at the disposal of the organiza-
tion and the concrete implementation of measures to strengthen its inde-
pendence from industry.
lis, An Intractable Conflict of Interest, The Mark News (3 December 2009), online: The
Mark News
Standing Committee on Health, Evidence, 40th Parl, 2nd Sess, No 47 (30 November
2009); Canada, House of Commons, Standing Committee on Health, Evidence, 40th
Parl, 2nd Sess, No 49 (7 December 2009).
97 See Prescription Pharmaceuticals, supra note 56 at 20.
98 Ibid at 19.
99 Ibid at 20.
100 Ibid at 21.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 973
B. Promoting Transparency and Data Access
1. The Raison dtre of Transparency
An essential component of improving the quality of clinical trials data
is transparency, an issue that has received significant attention in recent
years, particularly with respect to clinical trials data. Yet transparency is
clearly a crucial condition for the improvement of all forms of data gather-
ing, both pre- and post-market entry. Mandatory clinical trials registra-
tion, publication of research results, and increased access to clinical trials
data and other relevant clinical data related to pharmaceuticals are now
widely viewed as crucial means to prevent industry from manipulating
and obfuscating unfavourable clinical trial results.
Greater transparency and access to data have been recommended
more generally, and many of the examples discussed above deal with pre-
market clinical trials data. Yet transparency is as, or perhaps even more,
crucial for post-market studies. Drug companies often benefit from the
scarcity of post-market research, since studies may reveal the inferiority
of their drug to existing treatments, or reveal the inappropriateness of the
drug for indications for which it is already widely prescribed off-label;
drug companies may be equally motivated to hide unfavourable study re-
sults in the post-market phase. Since post-market research is essential for
the ongoing evaluation of drug safety and efficacy, and for improving clin-
ical guidance on both comparative effectiveness and off-label use, any ex-
panded transparency and access to data requirements should be extended
to all clinical trials no matter at what stage of the product lifecycle they
occur.
As has been discussed elsewhere in detail, data transparency gained
momentum in the wake of controversies that involved pharmaceutical
companies hiding and misrepresenting data in the scientific literature
and subsequently using misleading publications to promote off-label pre-
scriptions.101 More recently, the driving force behind an invigorated de-
bate over data access has been the issue of wasteful spending of public
funding. In 2012, Peter Doshi, Tom Jefferson, and Chris Del Mar pub-
lished a paper in which they laid out the failure of the company to produce
solid data that could undergird the World Health Organizations and vari-
ous public health agencies recommendations to stockpile the drug
Tamiflu as part of pandemic preparedness.102 In their review, they docu-
101 See the more detailed discussion in Lemmens & Telfer, supra note 2.
102 Peter Doshi, Tom Jefferson & Chris Del Mar, The Imperative to Share Clinical Study
Reports: Recommendations from the Tamiflu Experience (2012) 9:4 PLoS Medicine 1
[Doshi et al].
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
974
ment the hurdles put up by the drugs manufacturer, Roche, in their quest
for access to the data allegedly supporting this recommendation, the re-
markable lack of diligence by public health agencies, and the official agen-
cies blind confidence in data hidden behind corporate walls. Doshi and
colleagues note that while systematic reviews of published randomized
clinical trials (RCTs) are considered the gold standard source of synthe-
sized evidence for interventions, … their conclusions are vulnerable to dis-
tortion when trial sponsors have strong interests that might benefit from
suppressing or promoting selected data.103 Supported by their report of
the billions of dollars in waste associated with the unsubstantiated stock-
piling of Tamiflu, a drug they argued was not more effective than aspirin
for the treatment of influenza, Doshi and his colleagues argue that re-
searchers need access to clinical study reports (standardized documents
submitted to drug regulators that contain the most complete record of the
planning, execution, and results of clinical trials) in order to conduct the
most reliable evidence synthesis.104
Unfortunately, both industry and regulators tend to treat clinical
study reports as confidential documents, thereby preventing the data
from being studied by independent researchers. For example, when
Health Canada approves a drug, they only make available a Summary
Basis of Decision, which explains the scientific and benefit-risk infor-
mation that the department considered in making their approval deci-
sion.105 According to Lexchin, with only this limited information base, it is
virtually impossible to independently assess the safety and efficacy of
new products.106
Beyond clinical study reports, providing access to raw clinical trials
data would significantly reduce the incentive for industry to misrepresent
data, since, as stated by Peter Gtzsche, it would be a risky affair when
others can check the methods and calculations against the trial protocol
103 Ibid at 2.
104 Doshi et al show in the article how the claims of actual effectiveness of Tamiflu in re-
ducing hospital admissions and complications from influenza are questionable. In par-
ticular, significant evidence has emerged that Roche, the manufacturer of Tamiflu,
failed to disclose large amounts of data from the clinical trials on the drug, and contin-
ues to sidestep requests from Cochrane Collaboration researchers to make this data
available for review (ibid).
105 Lexchin notes that [t]hese documents lack information about the study protocol, the
baseline characteristics of trial participants, the number of participants who withdrew
and reasons for their withdrawal, primary and secondary efficacy outcomes, and fatal
and nonfatal serious ADRs, by treatment arm (Joel Lexchin, Harmony in Drug Regu-
lation, But Whos Calling the Tune?: An Examination of Regulatory Harmonization in
Health Canada (2012) 42:1 International Journal of Health Services 119 at 129).
106 Ibid at 129.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 975
and the raw data.107 Other potential benefits of granting access to raw
data include improving the efficiency of health research by allowing the
use of existing data for new studies, rather than requiring patients and
researchers to go through the expensive and time-consuming process of
collecting new data; and making meta-analyses of trials studying similar
interventions and patient groups much more reliable than if based on
published summary data.108 Some medical journals are already moving
toward full access to the data as a publication requirement. For instance,
in direct response to the Tamiflu controversy, the British Medical Journal
announced in October 2012 that, as of January 2013, the journal will only
publish studies where the sponsor commits to making the relevant anon-
ymized patient level data available on reasonable request.109
The importance of access to the EMA dataand, in particular, the
way in which it allows public interestoriented scientists to strengthen
the reliability of pharmaceutical information already out thereis high-
lighted by a recent proposal by Peter Doshi and colleagues, which aims at
Restoring Invisible and Abandoned Trials (RIAT).110 With this bold
one could also say provocativeRIAT concept, Doshi and colleagues want
to help correct the scientific literature on a host of blockbuster drugs. As a
result of data disclosure in the context of litigation in the United States,
as well as data made accessible as part of EMAs access to information
policy, which we will discuss further,111 they have obtained access to
around 178,000 pages of previously confidential company research docu-
ments. Preliminary analysis of this data has revealed possible instances
of failure to publish relevant data and misrepresentation of clinical trials
data, issues that Doshi and colleagues identified in their proposal. They
are pushing for a correction of the distorted scientific reports and for a
publication of analyses of the unpublished trials, inviting the research
funders (mostly pharmaceutical companies) and investigators involved in
these trials to declare their interest in doing so within the next year. Fail-
ure to do so, they argue, should enable independent investigators to take
up this task and to either correct distorted publications, or to publish new
analyses of unpublished data. With this eminent proposal, the authors are
putting the companies, which obviously defend the reliable scientific basis
107 Peter C Gtzsche, Why We Need Easy Access to All Data from All Clinical Trials and
How to Accomplish It (2011) 12:249 Trials 1 at 6.
108 Ibid.
109 See Fiona Godlee, Clinical Trial Data for All Drugs in Current Use (2012) 345 Brit
Med J 7.
110 Peter Doshi et al, Restoring Invisible and Abandoned Trials: A Call for People to Pub-
lish the Findings (2013) 346 Brit Med J f2865.
111 See Subsection III-B-4, below.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
976
of their products, between a rock and a hard place: ignoring their invita-
tion opens up the door to public criticism by external researchers and re-
traction or serious correction of the research that has been used to pro-
mote their products. Accepting the invitation means a potential public
admission of serious flaws in the representation of the safety and efficacy
data on their products. Ignoring the invitation will thus be impossible,
even if it can be expected that companies will try to use legal means to
challenge the initiative. The public nature of the data and the support the
proposal has already received by editors of two leading medical journals112
bodes well for this initiative. The initiative and its strong support also
highlight the crucial role of access to data for the scientific community.
2. The International Move Toward Data Access
Internationally, the debate seems no longer to be about whether trial
registration and results reporting are needed, but whether and how fuller
access to data, often including raw data, supporting submissions to regu-
latory agencies and publications can be ensured. Already in 2004, a Min-
isterial Summit organized in Mexico by the World Health Organization
(WHO) called for the establishment of a clinical trials registry.113 The
WHO responded swiftly by developing an international clinical trials reg-
istry platform and by promoting the development of regional registries.114
In the context of this global momentum toward trial registration, several
countries implemented trial registration and results reporting obligations.
In the United States, Congress introduced in 2007 various changes to the
FDA Amendments Act that include strict trial registration and results re-
porting requirements, accompanied by significant financial penalties for
non-compliance.115 The Pan American Health Organization provided sig-
nificant support for the establishment of regional clinical trial registries
in Brazil and Cuba.116 Other organizations contributed to the promotion
112 See Elizabeth Loder et al, Restoring the Integrity of the Clinical Trial Evidence Base
(2013) 346 Brit Med J f3601.
113 The Mexico Statement on Health Research, Knowledge for Better Health: Strengthen-
ing Health Systems (Joint statement from the Ministerial Summit on Health Research
in Mexico City, 1620 November 2004) at 12628, online: WHO
Ministerial Summit on Health Research, WHA58/2005/REC/1 (2005), 58th Sess, at 126,
online: WHO
114 For a detailed overview of international and national initiatives related to transparen-
cy, with a particular focus on the Americas, see Karmela Krlea-Jeri et al, Prospective
Registration and Results Disclosure of Clinical Trials in the Americas: A Roadmap To-
ward Transparency (2011) 30:1 Pan American Journal of Public Health 87.
115 Food and Drug Administration Amendments Act of 2007, Pub L No 110-85, 801, 121
Stat 823 at 90422 [FDA Amendments Act].
116 For more details, see Krlea-Jeri et al, supra note 114.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 977
of trial registration and other transparency measures. In 2008, the World
Medical Association included for the first time in its Declaration of Hel-
sinki, which can be considered the most influential international research
ethics standard, the requirement to prospectively register every clinical
trial and to report the results of research.117 Around the same time, medi-
cal journal editors announced that prospective registration of clinical tri-
als prior to recruiting human research subjects would henceforth be a
condition of the subsequent publication of the results of studies.118 As
mentioned above, some medical journals demand full access to the data as
a publication requirement.119
3. Canadas Hesitant Steps Toward Transparency
While at the international level various steps have been taken, Cana-
dian progress remains embarrassingly slow.120 In Canada, both an Audi-
tor General121 report and a report of the Standing Senate Committee on
Social Affairs, Science and Technology122 have recommended in recent
years that Health Canada improve the transparency of clinical trials data
used in its decision-making process. Yet these recommendations have not
resulted in an unambiguous and firm regulatory response. Trial registra-
tion and results reporting requirements have been introduced in the re-
vised Tri-Council Policy Statement by the federal funding agencies in
2011.123 However, the provisions remain vague and can only be enforced
117 World Medical Association, Declaration of Helsinki: Ethical Principles for Medical Re-
search Involving Human Subjects (Declaration, amended at the 64th WMA General As-
sembly, Fortaleza, Brazil in 2013) at 3435, online: WMA
118 For example, in 2005, the International Committee of Medical Journal Editors imple-
mented a policy that requires all clinically directive trials to be registered in an ap-
proved clinical trials registry in order for the study results to be published in any of the
member journals. An expanded definition of clinical trials requiring registration was
subsequently adopted in 2007. See International Committee of Medical Journal Edi-
tors, Frequently Asked Questions: Clinical Trials Registration, online: IMCJE
119 See Godlee, supra note 109.
120 See the detailed discussion in Lemmens & Telfer, supra note 2.
121 Auditor General of Canada, Chapter 4: Regulating Pharmaceutical DrugsHealth
Canada in Report of the Auditor General of Canada to the House of Commons (Ottawa:
Office of the Auditor General of Canada, 2011) at 12, online: OAG
122 Senate, Standing Senate Committee on Social Affairs, Science and Technology, Cana-
das Clinical Trial Infrastructure: A Prescription for Improved Access to New Medicines
(November 2012) (Chair: Kelvin K Ogilvie).
123 Canadian Institutes of Health Research, Natural Sciences and Engineering Council of
Canada & Social Sciences and Humanities Research Council of Canada, Tri-Council
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
978
through withdrawal of funding by the funding agencies. Most important-
ly, they only apply to research conducted in federally funded institutions
and thus not to the majority of industry-controlled clinical trials used in
the drug approval process. These trials are largely conducted by private
contract research organizations, whose compliance with Canadian re-
search ethics standards and good clinical practices is in several provinces
being verified by private, commercial REBs.124
Although Health Canada announced back in 2007 that it was looking
into the development of trial registration and results reporting, no such
regulatory requirement has yet been introduced. Most recently, in a ra-
ther ambiguous move, the federal Minister of Health announced the es-
tablishment of a Health Canada registry of clinical trials. However, it is
clear from the announcement and subsequent discussions that the regis-
try simply aims at providing information to potential research subjects on
the type of trials that are ongoing in Canada and that have been the ob-
ject of a clinical trials submission with the agency.125 This initiative ap-
pears to do little to diminish the problem of hiding relevant clinical trials
results and aims rather at the promotion of clinical trials per se. Pressure
to take firm action to promote transparency increased in April 2014 in the
wake of media reports about the problems with respect to access to im-
portant safety data of some popular approved drugs.126 Health Canadas
responsethe publication of a summary report of data about the contro-
versial acne pill Diane-35 (six months after it announced it would do so)
does little to reassure Canadians that we are really catching up with oth-
er countries.127 Surprisingly, the recent proposal to improve Health Can-
adas post-marketing powers (Vanessas Law),128 which contains several
Policy Statement: Ethical Conduct for Research Involving Humans (2010), online: Gov-
ernment of Canada Panel on Research Ethics
124 See Trudo Lemmens, Federal Regulation of REBs Review of Clinical Trials: A Modest
but Easy Step Towards an Accountable REB Review Structure in Canada (2005)
13:2&3 Health Law Review 39.
125 See Health Canada, Update: Registration and Disclosure of Clinical Trial Information
(19 October 2012), online: Health Canada
taries in Macleans magazine by numerous industry experts on this development: Julia
Belluz, Dear Leona Aglukkaq: Researchers Write the Health Minister on Clinical Tri-
als, Macleans (30 October 2012), online: Macleans
luz, Why is Canada Such a Laggard on Clinical Trials Regulation?, Macleans (5 No-
vember 2012), online: Macleans
126 See Diana Zlomislic, Health Canada Releases Summary Report of Controversial Drug
Diane-35, Toronto Star (9 April 2014), online: Toronto Star
127 Health Canada, Summary Safety Review: DIANE 35 (Cyproterone Acetate and Ethi-
nyl Estradiol)Venous Thromboembolism (Blood Clot) (2014), online: Health Canada
128 See discussion at 96365, above.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 979
interesting improvements to post-market surveillance, also does not cur-
rently contain any specific reference about clinical trials transparency.
4. The Saga of Pharmaceutical Data Transparency in Europe
It is not clear why Canada trails so far behind other countries with re-
gard to transparency and why the government seems so reluctant to move
forward. But some recent European developments may be revealing in
that context. Europe, which used to take a similar approach to Canada,
has taken major steps over the last four years to improve data transpar-
ency and is now facing strong opposition by industry.
The European Medicines Agency (EMA) was first pushed to change its
data access policy in response to recommendations in the context of a pe-
diatric cancer drug trial inquiry, in which the European Ombudsman crit-
icized the EMA for not providing access to all relevant data and urged the
agency to change course.129 Criticism by leading health researchers of the
Cochrane Collaboration, such as Peter Gtzsche, Peter Doshi, and Tom
Jefferson, who had sought access to data for independent meta-
analyses,130 also added to the pressure on the EMA. In response, the EMA
introduced in 2010 a new policy on access to documents related to medici-
nal products for human and veterinary use.131 Rather than starting from
the premise that that data is confidential and that researchers have to
justify their requests to access the data, the EMA embraced a presump-
tion of accessibility: companies had to justify their requests that specific
data sets should not be disclosed. Under this new policy, the EMA has re-
leased more than 1.9 million pages in response to such requests for non-
clinical and clinical data related to EMA-approved products.132
Then, in November 2012, during a public consultation process focusing
on the implementation of transparency measures for clinical trials da-
ta,133 the EMA announced that it would introduce a prospective data re-
129 European Ombudsman, Decision of the European Ombudsman closing his inquiry into
complaint 2560/2007/BEH against the European Medicines Agency (Strasbourg: Euro-
pean Ombudsman, 2010) at paras 83, 88, online: European Ombudsman
130 Gtzsche, supra note 2 at 13942.
131 European Medicines Agency, Policy on Access to Documents (Related to Medicinal
Products for Human and Veterinary Use): Policy 0043, EMA/110196/2006, online: EMA
132 Court Orders EU Medicines Agency to Withhold Clinical Trial Results, EurActiv (2
May 2013), online: EurActiv
133 One of the authors (TL) participated in this process. See European Medicines Agency,
Press Release, Workshop on Access to Clinical-Trial Data and Transparency Kicks Off
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
980
lease policy in 2013.134 This prospective data release policy would result in
the public sharing of data related to EMA-approved drugs on a publicly
accessible website even without explicit request. At the start of this No-
vember 2012 consultation meeting, the EMA executive director bluntly
stated that the question was no longer if access to the data had to be
provided, but how.135
The EMAs plans were clearly placing Europe ahead of other jurisdic-
tions with respect to data transparency, but they met strongalbeit not
very openresistance from industry. The ongoing saga around the EMAs
data policy may be indicative of the larger political and economic battles
that are now emerging over data access, and it is therefore worth men-
tioning here in more detail.
While industry organizations participated in consultation meetings on
the new data release plans, presenting themselves as constructive part-
ners willing to support at least the concept of transparency and data shar-
ing, 136 they also helped to mount legal opposition. They first did so
through supporting the legal offensive of two US-based pharmaceutical
companies, AbbViea spin-off consisting of much of the former medicines
division of Abbottand InterMune. In the winter of 2013, these compa-
nies launched a procedure before the General Court of the European Un-
ion requesting the annulment of the EMAs decision to grant access to
clinical and non-clinical information related to their drugs Humira and
Esbriet. The EMA had informed the companies that it would release the
clinical trials reports related to these drugs to researchers, and for one of
the drugs also to a competitor who had requested access under the 2011
policy.
The Federation of Pharmaceutical Industries and Associations and the
Pharmaceutical Research and Manufacturers of America applied to obtain
intervener status in support of AbbVie, while the European Confederation
of Pharmaceutical Entrepreneurs intervened on behalf of InterMune; the
European Ombudsman and the British Medical Journal did the same in
Process Towards Proactive Publication of Data (23 November 2012), online: EMA
134 European Medicines Agency, Publication and Access to Clinical-Trial Data: Policy 0070,
EMA/240810/2013, online: EMA
135 European Medicines Agency, Access to Clinical-Trial Data and Transparency: Work-
shop Report (2012), online: EMA
136 See e.g. European Medicines Agency, Overview of Comments Received on HMA/EMA
Guidance Document on the Identification of Commercially Confidential Information and
Protection of Personal Data Within the Structure of the Marketing Authorisation (MA)
Dossier: Release of Information After Granting of a Marketing Authorisation (2012)
EMA/330149/2012, online: EMA
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 981
support of the EMA.137 The support of major trade organizations for the
court challenge is particularly noteworthy in light of the troubling recent
history of both companies with respect to misrepresentation of data. As
Jim Murray reports,
[i]n May 2012, Abbott … pleaded guilty to misbranding a medicine
(Depakote) for dementia and schizophrenia [and] acknowledged …
that they had for years delayed disclosing the full results of trials
showing the medicine was no better than a placebo. The total fines
and penalties amounted to [US]$1.5 billion.138
InterMune, for its part, has been in trouble with US authorities for failing
to make full and timely disclosure of clinical trial results. Very recently,
the Chief Executive of InterMune was criminally convicted for dissemina-
tion of false and misleading information about a clinical trial.139 In light of
this, the industrys frequent expressions of support for improved data
sharing and transparency140 lose much of their credibility.
The companies most prominent arguments against the data sharing
policy were based on the allegedly confidential nature of various compo-
nents of the clinical trial reports: access decisions would violate the access
to information regulations of the European Council141 and the companies
fundamental right to protection of confidential information as guaranteed
by the Charter of Fundamental Rights of the European Union142 and by
137 InterMune UK Ltd v European Medicines Agency, Order of 25 April 2013, T-73/13 at pa-
ra 19, online: Curia
pean Medicines Agency, Order of 25 April 2013, T-44/13 at para 27, online: Curia
138 Fighting for Transparency on Clinical Trials, EuropeanVoice (7 May 2013), online:
EuropeanVoice
139 Ibid. See United States v Harkonen, 510 Fed Appx 633 (Cal 9th Cir Ct 2013).
140 See the numerous expressions of support for transparency in EMA, Overview of Com-
ments, supra note 136. See also Ronald L Krall & Frank Rockhold, Trial Registration:
Putting Principles into Practice, online comment in response to Karmela Krlea-Jeri
et al, Principles for International Registration of Protocol Information and Results
from Human Trials of Health Related Interventions: Ottawa Statement (Part 1) (2005)
330:7497 Brit Med J 956.
141 More specifically, EC, Council Regulation (EC) 1049/2001 of 30 May 2001 regarding
public access to European Parliament, Council and Commission documents, [2001] OJ L
145/43, art 4. This article specifies when European institutions have to refuse access to
specific documents, including when privacy rights are at stake (ibid, art 4(1)(b)), and
when the protection of commercial interests, including intellectual property, will be un-
dermined (ibid, art 4(2)). In the latter case, disclosure is still warranted if there is an
overriding public interest. The EMA policy is based on this notion of public interest in
disclosure.
142 EC, Charter of Fundamental Rights of the European Union, [2000] OJ C 364/1, arts 7, 8,
17.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
982
the European Convention on the Protection of Human Rights and Funda-
mental Freedoms. 143 They also claimed that access would violate the
EMAs TRIPS-based obligations to keep data submitted to the regulator
confidential.144
The companies also requested an interim suspension of the data ac-
cess decision, awaiting the final outcome of the case. Under European
law, such a suspension can be granted if there is a reasonable basis in fact
and in law for the claim (and thus a reasonable chance of success) and if
the regulatory decision is likely to create harms that would be irreparable
and irrevocable. Lawyers for the EMA had reasonably argued that if the
court would conclude in the end that the EMAs disclosure had violated
the companies rights, the court could order financial compensation for the
harm suffered.145 Surprisingly, the president of the General Court sided
with the companies, ruling that there was a reasonable basis for the claim
that the companies fundamental right to the protection of private and
family life would be violated, and that the fundamental rights nature of
these interests would make it impossible to undo the harms done if the fi-
nal decisions would favour the claimants.146
The presidents suspension of data access was a clear victory for in-
dustry. It immediately affected data access decisions by the EMA. The
agency likely became concerned that any company faced with a data ac-
cess decision would now challenge the decision in court and obtain an in-
terim injunction. Industrys support of the legal challenges made this out-
come more likely.
143 Convention for the Protection of Human Rights and Fundamental Freedoms, 4 Novem-
ber 1950, 213 UNTS 221, art 8, Eur TS 5.
144 For a discussion (and rejection) of the TRIPS argument in the context of data disclo-
sure, see Lemmens & Telfer, supra note 2.
145 AbbVie (25 April), supra note 137 at para 51. See also InterMune, supra note 137 at pa-
ra 40.
146 AbbVie (25 April), supra note 137 at para 52; InterMune, supra note 137 at para 41. The
interim ruling based on a privacy-rights characterization of clinical trials data secrecy
seems extraordinary in light of the case law from the European Court of Human Rights
that supports exactly the opposite claim: that access to important health- and safety-
related information is an essential component of, among others, the right to the protec-
tion of private and family life and the right to life. See the discussion of primarily Euro-
pean case law in Lemmens, Pharmaceutical Knowledge Governance, supra note 2 at
16672. For other discussions of the human rights dimensions of data access, see Aaron
A Dhir, Corporate Selective Reporting of Clinical Drug Trial Results as a Violation of
the Right to Health in Marcia H Rioux, Lee Ann Basser & Melinda Jones, eds, Critical
Perspectives on Human Rights and Disability Law (Leiden: Martinus Nijhoff, 2011) 341
at 35664; Lemmens & Telfer, supra note 2 at 99110.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 983
But industrys victory was short-lived. The EMA appealed the decision
to the European Court. In two prudent appeal decisions, the European
Courts vice-president stuck to a careful narrow analysis of the legal ar-
guments of the presidents interim decision.147 Without entering into a
discussion of the reasonableness of the companies claims and the nature
of the rights at stake, he simply affirmed that even if companies would
have a fundamental privacy right over clinical trials data, this type of pri-
vacy right would not be of such nature that disclosure would inevitably
create irreparable harm. The companies would have to show that it would
be impossible to calculate the level of financial harm they would suffer as
a result of disclosure in order to have the potential harm qualified as ir-
reparable.148
This reversal of the suspension put the EMA again in the driver seat,
particularly after one of the companies, AbbVie, withdrew its challenge of
the data access decision related to Humira. While this legal battle was un-
folding, the European Parliament also provided strong support for data
transparency in Europe when it voted in favour of a new clinical trials
regulation, whichwith the amendments provided in the European Par-
liamentcontains strict requirements for data access and reporting.149
Nothing seemed to stand in the way of the EMAs further implementa-
tion of a more proactive data release policy. It therefore came as a major
surprise when, in May 2014, the EMA distributed draft documents to par-
ticipants of its Clinical Trials Data Policy Advisory group,150 documents
that appeared to significantly strengthen pharmaceutical companies con-
trol over data and to make prospective data access a technical obstacle
course. Scientists severely criticized, for example, the EMAs proposed
prohibition on downloading data, which would require them to look at
clinical trials reports, sometimes consisting of tens of thousands of pages,
on screen and on a page-by-page basis. The Terms of Use and Redac-
tion Principles also made it clear that companies would submit two dif-
ferent files to the EMA, one with the full data, and one redacted for public
147 European Medicines Agency v AbbVie Inc and AbbVie Ltd, Order of 28 November 2013,
C-389/13 P(R), online: Eur-Lex
Mune, supra note 137. For a more detailed commentary on the developments in the
cases, see Trudo Lemmens, Access to Pharmaceutical Data, Not Data Secrecy, is an
Essential Component of Human Rights (8 April 2014), online: University of Toronto
Faculty of Law Blog
148 AbbVie (28 November), supra note 147 at para 46.
149 EC, Regulation (EU) No 536/2014 of the European Parliament and of the Council of 16
April 2014 on clinical trials on medicinal products for human use, and repealing Di-
rective 2001/20/EC, [2014] OJ L 158/1, art 81 (in particular).
150 One of the authors (TL) is a member of this advisory group and received the draft doc-
uments in that capacity.
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
984
use,151 thus undermining the idea of full data sharing. Most troubling,
however, are the potential legal implications of the request that research-
ers sign Terms of Use (TOU) and Redaction Principles, which appear
to trap the signer into a particular interpretation of the law. By signing
the TOU, a data user would acknowledge that the Information is protect-
ed by copyright and proprietary rights … and can be considered commer-
cially valuable.152 Brand name pharmaceutical companies increasingly
insist that clinical trials data is protected by copyright and proprietary
rights, and is considered confidential commercial information. But wheth-
er and to what extent data is protected by such rights remains hotly con-
tested, and is not firmly established in law. Many have convincingly ar-
gued that this data is a public good153 and that case law supports access to
data on the basis of health-related human rights.154 Remarkably, the
EMA would now ask scientists to recognize these rights while it is still in-
volved in litigation where the EMA has an interest in arguing for the
most restrictive recognition of these rights.
Perhaps even more troublesome is that researchers would also con-
tractually agree that they could be sued under UK law in accordance
with the provisions of the Contracts (Rights of Third Parties) Act 1999.155
As a result, pharmaceutical companies will be able to challenge research-
ers directly in court for violation of the EMAs TOU. This puts them in a
very comfortable legal position, particularly when this right to sue is con-
nected with the recognition of proprietary rights, copyrights, the commer-
cial and confidential nature of some information, and the severe technical
restrictions on data use, which make good faith violations of the TOU
more likely.
This could have a significant chilling effect. The mere risk or threat of
litigation could seriously hamper open scientific debate. Researchers may
think twice about publicly challenging a companys interpretation and
151 European Medicines Agency, Redaction Principles (Draft) (5 May 2014), Appendix 1
[unpublished, on file with authors, Redaction Principles].
152 European Medicines Agency, Terms of Use (Draft) (5 May 2014) [unpublished, on file
with authors, TOU]. One of the authors (TL) raised these concerns in a letter to
EMAs executive director, Guido Rasi. For a link to the letter, and a further discussion
of the legal implications see Trudo Lemmens, EMAs Proposed Data Release Policy:
Promoting Transparency or Expanding Pharma Control over Data? (30 May 2014),
online: PLoS Blogs: Speaking of Medicine
153 See Reichman, supra note 91.
154 See supra note 36.
155 TOU, supra note 152. The Contracts (Rights of Third Parties) Act 1999 ((UK), c 31)
aims at overcoming the common law doctrine of privity, and allows third parties to en-
force contractual provisionsfor example, if the contract contains a specific provision to
that effect.
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 985
representation of data or correct the published record on the basis of
EMA-held data, as is being undertaken in the context of the RIAT pro-
posal.
This troubling development is not entirely surprising. Even if the
transparency movement has had some major victories, including the
adoption of transparency requirements in the European Clinical Trials
Regulation, opposition has been mounting. Other regulatory initiatives
may be used by industry to fight transparency. The European commission
recently released a draft directive156 aimed at streamlining and strength-
ening trade secret protection in Europe. The draft directive does not ex-
plicitly refer to clinical trials, but the European Federation of Pharmaceu-
tical Industries and Associations (EFPIA) has already jumped enthusias-
tically at the occasion, hinting at the need to protect the proprietary
know-how of drug development, including in the clinical trials phase.157
In the context of ongoing and largely secret transatlantic trade negotia-
tions between Europe and the United States and Canada, the pharmaceu-
tical industry has also been lobbying to strengthen data and IP protec-
tion.158
Could the same reasons be behind Canadas apparent reluctance to
move toward full transparency of data? We cannot know for sure. But it is
clear that significant financial interests are at stake and that the phar-
maceutical industry is mobilizing worldwide to counter the push by the
scientific community and civil society to promote data transparency. It is
interesting to note that a recent Canadian initiative to promote clinical
trials in the country, involving closer collaboration between the CIHR,
Health Canada, and Rx&D, does not contain any reference to transparen-
cy, but mentions in contrast the strengthening of IP rights in clinical tri-
als as an explicit component of the new initiative.159
156 EC, Commission, Proposal for a Directive of the European Parliament and of the Coun-
cil on the Protection of Undisclosed Know-How and Business Information (Trade Se-
crets) Against Their Unlawful Acquisition, Use and Disclosure (Brussels: EC, 2013).
157 European Federation of Pharmaceutical Industries and Associations, News Release,
EFPIA Welcomes the Commissions Proposal on the Protection of Undisclosed Know-
How and Business Information (Trade Secrets) (28 November 2013), online: EFPIA
158 See e.g. Biotechnology Industry Organization, Transatlantic Trade and Investment
Partnership: Comments (Docket No USTR-2013-0019), online:
159 Government of Canada, News Release, Strengthening Clinical Trials: New Coordinat-
ing Centre to Help Attract Research Investment in Canada (24 April 2014), online:
Government of Canada
cal Companies, Canadian Institutes of Health Research & Association of Canadian Ac-
ademic Healthcare Organizations, To Your Health and Prosperity: An Action Plan to
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
986
Conclusion
At the pre-market phase, decreasing the deficit of reliable safety and
efficacy data depends on measures to demystify industry practice in the
conduct of clinical trials and to improve access to the resulting data, in-
cluding making it available for systemic review by independent research-
ers. At the post-market phase, decreasing the deficit of data on real-world
use of drugs requires ongoing studies on the long-term safety and efficacy
of drugs, including on comparative efficacy and off-label use. Concurrent-
ly, improved ADR reporting by industry, physicians, and consumers, cou-
pled with better systems for collecting, storing, and analyzing reported
data, is critical for early detection of potential problems with drugs. Effec-
tive ADR reporting can signal the need for more research into the safety
and efficacy of a drug and lead to faster response by regulators and the
medical profession in placing limitations on, or in extreme cases suspend-
ing, the use of a drug product.
Despite the historical fixation on the pre-market phase of drug ap-
proval, post-market surveillance is becoming increasingly important in
the evolving drug regulatory environment.160 In particular, Health Cana-
da has been working since 2005 on a progressive licensing framework161
that, among other things, would enhance regulatory oversight of the post-
market phase and give the government authority to require drug manu-
facturers to conduct post-market studies and to submit the resulting data
for review. The progressive licensing framework embraces what is re-
ferred to as the life-cycle approach to drug approval, referencing the fact
Help Attract More Clinical Trials to Canada (Ottawa: Association of Canadian Academ-
ic Healthcare Organizations, 2012), online: ACAHO
160 One jurisdiction that has already made the move toward greater emphasis on the post-
market phase is Japan. In 2002, Japan reformed its Pharmaceutical Affairs Law to fo-
cus more attention on post-market duties (rather than on pre-market and manufactur-
ing activities) in an effort to ensure safety and efficacy of medical products. In particu-
lar, Japan more closely regulates post-market activities through a focus on precaution-
ary phases and early intervention, as well as the reexamination/reevaluation periods to
ensure the drugs continuing safety and efficacy (Vanessa Eng, Drug Safety: Its a
Learning Process (2009) 24:1 St. Johns Journal of Legal Commentary 159 at 188).
161 In 2005, Health Canada launched the Progressive Licensing Project to explore reform
proposals for the entire federal drug licensing system, from early clinical trials to mar-
ket authorization, and the monitoring of drugs once they are on the market (see Health
Canada, Progressive Licensing: Background, online: Health Canada
to evolve in response to a number of important medical and social trends. In particular,
Health Canada pointed to changes in public expectation around, inter alia, growing
demands for compassionate access to drugs, for early access to promising new thera-
pies, and for greater patient autonomy in drug treatment decisions, including determin-
ing acceptable levels of risk (see Health Canada, Life-Cycle Management, online:
Health Canada
POST-MARKET SURVEILLANCE IN PHARMACEUTICAL REGULATION 987
that drug regulatory decisions are based not simply on pre-market evi-
dence, but on the entire body of evidence that accumulates throughout the
life cycle of the drug.162 As such, an important element of this model is
the initial and ongoing collection, evaluation, and communication about
drug information throughout the product life-cycle.163 Although the pro-
gressive licensing framework offers a number of improvements to the Ca-
nadian drug regulatory system,164 one limitation is that the framework
still relies primarily on drug manufacturers to conduct post-market stud-
ies of their own products.165 Concurrently, an important trend on the drug
development side has been the growing emphasis on personalized medi-
cines and drugs for niche marketsmarkets in which the problem of the
lack of solid evidence on drug safety and efficacy becomes even more pro-
nounced because of the inherently limited patient population. Drugs for
niche markets, such as pharmacogenomic products,166 drugs that treat ra-
162 The progressive licensing model recognizes that knowledge about how a drug works,
both positively and negatively, increases throughout th[e] life-cycle as more and more
people are exposed to the drug (Health Canada, The Progressive Licensing Framework:
Concept Paper for Discussion (Ottawa: Health Canada, 2006) at 11).
163 Ibid at 12. Interestingly, models similar to the life-cycle approach have also been sug-
gested in the context of drug funding decisions under the title of Coverage with Evi-
dence Development (CED). As summarized by Flood and Dyke, [r]ather than demand-
ing that evidence of effectiveness and cost-effectiveness be presented at the time of the
listing decision, a new drug may be publicly funded on the condition that further evi-
dence is developed about the relative effectiveness of the drug (supra note 33 at 313).
While making reimbursement conditional on expanded post-market evidence genera-
tion represents a step in the right direction, significant challenges remain with respect
to the effective design and governance of CED initiatives (see e.g. Danielle Bishop & Jo-
el Lexchin, Politics and Its Intersection with Coverage with Evidence Development: A
Qualitative Analysis from Expert Interviews online: (2013) 13 BMC Health Services
Research 88
164 For a good discussion of the potential effects of the progressive licensing framework on
post-market safety, see Ron A Bouchard & Monika Sawicka, The Mud and The Blood
and The Beer: Canadas Progressive Licensing Framework for Drug Approval (2009)
3:1 McGill JL & Health 49 at 66.
165 See Ferris & Lemmens, supra note 32.
166 Pharmacogenomics involves the development of pharmaceutical agents in combination
with diagnostic tests that are able to provide information about the possible effective-
ness and toxicity of a drug based on the genetic profile of individuals. Some new phar-
macogenomic drugs have received rapid market approval on the basis of targeted clini-
cal trials in genetically selected patient populations (see Shannon Gibson & Trudo
Lemmens, Niche Markets and Evidence Assessment in Transition: A Critical Review
of Proposed Drug Reforms (2014) 22:2 Medical L Rev 200 at 20911). Proponents of the
rapid approval of pharmacogenomic drugs claim that post-market research can clarify
the safety and effectiveness of the drug and include testing the drug in other genetically
selected populations. A case study by Raziee, Lemmens, and Kimmelman on the devel-
opment of the pharmacogenomic cancer drug cetuximab raises some questions about
the integration of biomarker evidence in drug development (Hamid Raziee, Trudo
Lemmens & Jonathan Kimmelman, Scientific Evidence and Targeted Phar-
(2014) 59:4 MCGILL LAW JOURNAL REVUE DE DROIT DE MCGILL
988
re diseases, or breakthrough drugs for the treatment of life-threatening
diseases are sometimes given special consideration during the drug ap-
proval process. While there is a case for being more flexible about the
amount of data required to support the approval of these drugs, particu-
larly when no other treatment is currently available,167 the limited ap-
proval basis makes it even more critical to assess the benefits and risks of
these drugs through the ongoing collection and analysis of data after they
enter the market.168 If such post-market studies are not properly conduct-
ed, the uncertainty regarding the safety and efficacy of the niche product
is not effectively counterbalanced.
Ultimately, while there have been some promising new initiatives in
various jurisdictions and at the international level to combat the data def-
icit that arises during both the pre- and post-market phases of drug de-
velopment, Canada continues to lag on a number of fronts. Health Cana-
das lack of progress in mandating clinical trials registration, result re-
porting, and overall transparency of clinical trials data, in particular, has
been disappointing. While the creation of the DSEN and the enhanced
post-market commitments contemplated under the proposed progressive
licensing framework offer some hope of restoring public confidence in the
independence of the drug regulatory system, greater emphasis needs to be
placed on insulating clinical research from the influence of the pharma-
ceutical industry at all stages of drug development.
macogenomics Drug Approval (Draft Paper delivered at the conference Using and
Abusing Evidence in Science and Health Policy, Banff, Alberta, 1 June 2012) [un-
published]). The study suggests that a failure to validate molecular hypotheses early in
drug development can result in inefficiencies in research and healthcare delivery and
can evoke concerns about the equitable nature of more narrowly tailored drug develop-
ment.
167 See Health Canada, Life-Cycle Management, supra note 161.
168 Ibid.
