The industrialization of pharmaceutical research and the problem of transparency in the early 2000s

The history of the pharmaceutical industry during the past five decades has been characterized by a “biotechnology revolution” (between 1970s and the 2000s), and a ‘Winter of Discontent?’ (2000–2010) [8]. Strictly speaking, the events related to trial registration correspond to this last era, but the circumstances that led to the implementation of registries must be traced back to the so-called revolution.

In this sense, the biotechnology revolution era does not only refer to an increase in technological development, but also to a redefinition of the geopolitics of the pharmaceutical industry embedded in neoliberal macroeconomic processes. One of these main redefinitions consisted in the shift from Europe to the US in terms of R&D dominance due to changes in the patent regimes and the large investment and strengthening of public funding for health research in the early 1970s, including the National Research Act of 1974 [8], [9]. This period was also characterized in the US by the creation and implementation of ethical regulatory codes for clinical trials with human subjects which, along with FDA establishment of randomized controlled trials (RCT) as the gold standard for assessing safety and efficacy of new therapies by 1970, boosting clinical trials [10].

In 2000, Richard Rettig [7] described the scenario of clinical research as essentially industrialized: “a large, rapidly growing line of business.”, especially in the field of drug development. During this period —partly influenced by the new information technology industry— pharmaceutical companies underwent a process of strong commercialization of scientific research, venture capital, and Intellectual Property Rights (IPR) [8]. The growing demand for new drugs both in high and low- and middle-income countries created large markets, resulting in increasingly high revenues for companies [10]. Investing in R&D became a growing interest for the drug industry throughout the 1980s, displacing public funding by the National Institutes of Health (NIH) by 1992 [10]. The implementation of the “Big Pharma” model of business [11] also consisted in large scale practices of out-sourcing (primarily through Contract Research Organizations) and off-shoring to low- and middle-income countries [10].

However, this rapid financial growth and geopolitical expansion, wasn’t exempt of problems or crisis by the turn of the century. The “Winter of discontent?” that Malerba and Orsenigo refer to when describing the situation of the drug industry during the 2000s, is named so because of “growing doubts about the sustainability of the business model” [8], and a deterioration of public opinion due to scandals like Paxil and heated debates regarding access to healthcare and prices of medications worldwide. Indeed, by the early 2000s, the Big Pharma business model, dominant in the 1990s, was being highly criticized: “The blockbuster business model that underpinned Big Pharma’s success is now irreparably broken. (…) The pharmaceutical industry is a prisoner of its past successes.” [11].

This difficulty to adapt was not only regarding its financial and R&D model, but its role in society. By 2004, the industry was under public scrutiny for what was perceived as a generalized lack of transparency and setting priorities that prepended profit rather than patient safety and access. A growing literature condemning malpractices of political lobby, biased results of new drugs’ effectiveness, and generalized misleading marketing, appeared in scientific and non-scientific outlets, including a book by the former editor of the New England Journal of Medicine, Marcia Angell titled: “The truth about the Drug Companies: How they deceive us and what to do about it” [12].

This was the general background where the debate on trial registration took place. How could registering trials in a specific platform help with these “deceptions”, in Angell’s words? Rettig’s paper concludes with the following reflection:

“Transparency. This concern links all questions in the clinical research domain. Concerns have been raised about suppression of research results by drug firms, bias in interpreting inconclusive research as a function affiliation with nonprofit or for-profit institutions, multiple reporting of the results of a single trial, and ghost authorship of articles, especially among nonacademic investigators for whom publishing is not a strong incentive. Among the remedies suggested, the most significant is the proposal to register all clinical trials at inception, whether governmentally or privately sponsored.” (italics by me) [7].

The connection between transparency and registration was already well established. Nonetheless, it took years for the WHO to step into the debate.

Why registries? why register?

Rettig’s quote shows that there was no single motive to advocate for the implementation of a trial registry. “Transparency concerns” is a broad notion that designates several issues happening in the drug industry at the time. For this reason, it is important to distinguish these problems and observe the arguments used to justify both the existence of a registry, as well as the mandatory registration of trials, as a “remedy” [7]. This distinction is also important to examine how the WHO addressed each of these issues and their arguments.

In one word, during the two decades prior to the launching of the ICTRP, the problem with transparency was framed as a problem of bias. “Not surprisingly, bias is now rampant in drug trials”, Angell wrote in 2004 [12]. In 1993, Kay Dickersin —one of the strongest advocates for registration and an expert in publication bias since the late 1980s— and Y. Min, defined: “Publication bias is any tendency on the parts of investigators or editors to fail to publish study results on the basis of the direction or strength of the study findings.” [13]. In other words, publication bias happens when the trial results do not go as expected, thus publishing the study does not favor one of the parties involved, and it remains unpublished. Bias may, however, also occur in the opposite way: overreporting positive results, and/or publishing them in different journals, “in slightly different forms” [12]. A related practice, selective reporting or misreporting, refers to publication of only part of the data: for example, the publication of subgroups of patients where the drug proved effectiveness (though not on the whole sample), and reporting specific time frames of the trial and not the outcomes of its whole duration according to the initial design [12]. By 2004, it was estimated that only half of existing trials were published [2], [14].

Although the study of publication bias and misreporting can be traced back to the 1950s, it was not until the 1960s that different registries specific to conditions or types of study became to be implemented in the US, according to Dickersin and Rennie [15]. The first mention of a trial registry gathering all areas of medicine can be found in a letter written by Thomas Chalmers in 1977 in the New England Journal of Medicine [16]. The first researcher to demonstrate the impact on publication bias regarding interpretation of clinical trial results was John Simes, in 1986 [17], when he compared the results of two different meta-analyses of chemotherapy in ovarian cancer and showing the divergent results depending on the studies included (only published vs. registered). More interestingly, in that same paper published in the Journal of Clinical Oncology, Simes called for the creation of an international registry of clinical trials, more than a decade before the FDA trial registry platform was launched, and two decades before the WHO’s platform.

Following Simes’ study, the upcoming years and 1990s decade had an active debate on the issue. Prominent figures in the discussion were Thomas C. Chalmers (who held positions at the National Institutes of Health), Sir Iain Chalmers (one of the founders of the Cochrane Collaboration), and Dickersin herself (who was mentored by both Thomas and Iain Chalmers in Boston during the 1980s and later integrated the WHO ICTRP Advisory Group). Several meetings were held in Europe starting in Brussels in 1991 to debate the importance of implementing registries and trial registration:

“A workshop on clinical trial registries, organized by Jean-Pierre Boissel (Lyon) and Kay Dickersin (Baltimore), was held in Brussels on July 12. The participants agreed on the need for all trials to be registered, and that there should be a directory of registries, both national and supranational.” [18].

In their 1993 meta-analysis study, Dickersin & Min demonstrated that “in every case, failure to publish was investigator-based, and not due to editorial decisions.” [13]; and acknowledged that “the identification of planned and ongoing trials will be a continuing challenge until registration is required of investigators (…). Who will take the lead?” [13].

The grounds for tackling publication bias and misreporting in clinical research —and specifically in pharmaceutical sponsored trials— were well documented and grounded. Following I. Chalmers’ claim that “adequate reporting of clinical trials is required for both scientific and ethical reasons” [19], inadequate reporting can be grouped in two categories, to differentiate two dimensions of the problem. While the first line of these arguments can be seen as focusing on the “clinical” aspect, this is, the fact that these studies are made with human participants who volunteer to enroll, the second can be analyzed from the “trial” piece, or in other words, a standardized way of creating scientific knowledge.

Misreporting and publication bias as an unethical practice

Health research, including pharmaceutical sponsored drug trials, relies on the participation of human subjects. Unfortunately, it is no secret that the history of medical experimentation has not been exempt of unethical practices, both in the US and overseas [10], [20]. Notwithstanding several of these infamous experiments took place under totalitarian regimes during the second World War, others correspond to trials conducted in the era of megatrials —multicentric, often multi-country studies—, directly contributing to the decay in public opinion of the drug industry. As the model of industrialization became a standard model of business, the demand for larger samples and cheaper, faster recruiting boosted the offshoring of trials to developing countries with weaker regulations or ethical codes of health research, in a process that has been described as the globalization of clinical trials [21]. This process raised concerns regarding accountability of US based companies when conducting trials overseas, especially in poor populations, with fatal outcomes or questionable use of placebo. How to measure transparency in offshored trials? How to make drug companies accountable for reporting the totality of clinical trials conducted out of the US accurately, where FDA regulations are not legally binding? The issue of global clinical trials prompted a report from the National Bioethics Advisory Commission, titled “Ethical and Policy Issues in International Research Clinical Trials in Developing Countries”, released in 2001 [22]. Aiming to deliberate on the question: “Can a research design that could not be ethically implemented in the sponsoring, developed country be ethically justified in the country in which the research is conducted?”, it provided a series of recommendations intended to reduce the risk of exploiting vulnerable populations.

While the events described above imply an atmosphere of mistrust on industry practices, debates over publication bias highlighted the backfires that these practices had for enrolling voluntary participants whose trust could be irreversibly damaged. Moreover, publication bias and misreporting could undermine the very action of participating in a trial: “Failure to provide adequate, publicly available reports of the results of clinical trials does injustice to the patients who have participated in them.” [19]. This argument of betrayed trust was echoed by several authors and actors in the trial business. Advocates to end bias and selective reporting not only aimed to end unethical practices occurring during and after trials, but also emphasized the importance of reporting trial findings to avoid unnecessary suffering stemming from questionable protocols, futile replication and public spending on ineffective trials already conducted but unpublished, misinformation in decision makers and healthcare providers, and as a form of improving access to patients and families who might be benefited by experimental treatments [10], [14], [19], [23], [24].

Misreporting and publication bias as scientific misconduct

“One of the most common ways to bias trials is to present only part of the data —the part that makes the product look good— and ignore the rest” [12]. Angell highlights that these “suppression of negative results”, or selective reporting, is a conscious commercial strategy used by the industry in order to maximize profit at the expense of data accuracy. Paxil’s “study 329” constitutes a case of selective reporting as the data regarding dangerous effects in children was purposefully disguised under a report of general effectiveness. Nevertheless, this was a widespread reporting practice.

The misconduct that Chalmers argued in 1990 resonated like a motto throughout the decade, as it underscored not only the fact that publications in medical journals were inaccurate because of the studies themselves, but because the bias between what was being published and what wasn’t was eventually affecting his own endeavors: the quality of systematic reviews and meta-analyses of the Cochrane collaboration (created in 1993 in the UK), which aims to provide the highest quality of systematized evidence to inform decisions. While publication bias was not an element included in systematic reviews methodology during the 1990s and early 2000, as early as 1986 this issue was pointed out by Simes in his seminal review [17], who demonstrated that the conclusions could vary dramatically depending on the sources included. Unfortunately, controlling bias in publication was out of the scope of reviewers, putting these scholars also at risk of scientific misconduct [15], [24]. And more importantly, as the Paxil case showed, the industry’s selective reporting practices was putting the population at risk of unsafe medications —and not only vulnerable populations in low- and middle-income countries-.

Furthermore, it also affected credibility of medical journals, and the role of journal editors, as they were accepting for publication manuscripts without certainty of the accuracy of the reported data [12], [19], or even transparent information about the authors —due to practices of ghostwriting common in industry-sponsored articles [23]—. Both evidence from the early 1900 and 2000 s analyzed by Dickersin showed that the practice of publication bias was strongly associated with either the investigator or the sponsor (especially when sponsored by the industry), rather than editorial decisions [13], [23]. Nonetheless, Horton and Smith, both editors of The Lancet and the British Medical Journal by the turn of the century, acknowledged that journal editors were also actors with “a part to play”:

“Editors also have a part to play. During peer review, editors increasingly find themselves requesting copies of the original trial protocol to check against the final submitted report. That “protocol culture” has led one of us to begin (and the other to plan) a protocol registration scheme. Editors are unwilling to fill their journals with promises of what might be, but they can publish these protocols on their web sites, perhaps linking them to a central registry.” (italics by me) [24].

Answering the questions: the rationale for (unified) registration

As reviewed above, the problems associated with pharmaceuticals publication practices were numerous and serious during the first half of the 2000s. Both publication bias and selective reporting were not only difficult to track in an increasingly globalized industry, but their effects were threatening to undermine the very mission of the evidence-based medicine paradigm: on commenting the antidepressants’ crisis, Rennie wrote: “This storm over SSRIs is a good demonstration that no clinician can possibly practice evidence-based medicine if prevented from seeing the evidence.” [2].

Initiatives for trial registries have a history as old as the discussion, with different degrees of success and sustainability over time, and increasing endorsement by key agencies [15]. Indeed, by the time the WHO became an actor in trial registration, ambitious projects such as the FDA clinicaltrials.gov platform —launched in 2001 in the US—, the Cochrane Controlled Register of Trials (CENTRAL) launched in 1996, and the International Standard Randomized Controlled Trial Number (ISRCTN, launched in 2000), were already in force. However, registration initiatives were as scattered as the problem, and each database functioned with their own standard of registered data and had different scopes. Nevertheless, there was consensus between the authors about rationale regarding the role of registries and publication bias: for a registry to be effective in counteracting bias, it had to register trials at their inception, that is, before patients were recruited. This would avoid publication bias as it would be possible to detect those trials, ongoing, interrupted or completed, despite their later results, and despite their subsequent publication or not. Depending on the data required by the registry, it would be possible to avoid selective reporting by knowing beforehand details from the protocol and expected outcomes. With the massification of the internet, public status of trials would make it possible for patients, physicians, and other interested parties to identify trials they might be interested in participating or referring patients to, increasing transparency and ensuring access to the information in a timely manner [15], [24], [25].

One month before the Paxil lawsuit, The Lancet published a commentary by Timothy Evans, Metin Gülmezoglu and Tikki Pang, claiming that the WHO had an essential role in the task of trial registration, after a meeting held in London a few days earlier [25]. In this Commentary, Evans and colleagues provide a concise synthesis of the previous debate on the issue. Encompassing the major problems and arguments regarding transparency and registration, they underscore the potential of the WHO in the matter of registration, as “there is no registry that has comprehensive international coverage. There is a need for coordinated international collaboration to either build a single register or to link together all that exist.” [25]. Indeed, by 2004 the WHO was eager to step in the discussion.