1. Introduction
As private funding becomes increasingly prevalent in scientific research, questions arise regarding the ethical and epistemic appropriateness of research guided by industrial and commercial interests. Successful cases of democratic inclusion of traditionally marginalized populations in the scientific process, such as the pharmaceutical industry’s alliance with the gay movement in the development of retroviral drugs for the treatment of HIV (Epstein Reference Epstein1995), make the democratization of science a particularly promising strategy for countering the alleged negative consequences emerging from the private funding and performance of research (see, e.g., Grasswick Reference Grasswick2010; Harding Reference Harding2015; Koskinen Reference Koskinen2017).
The article examines the process of niche standardization in medical research as a noteworthy example of the way in which democratizing strategies have been implemented in industry-funded science. Niche standardization resulted from the struggle for inclusion and recognition of difference from health advocacy groups in the late twentieth century and consists in the reconceptualization of the standard human subject for medical research, traditionally the white man, into a variety of the standard research subject representative of different groups: women, African Americans, children, and so on (Epstein Reference Epstein2007).
Despite its well-intended aims, I argue that niche standardization can lead to undesirable epistemic and ethical consequences, if the various goals of research are not properly aligned. To do so, I examine two salient examples of niche standardization: the case of Sarafem, a drug approved for the treatment of premenstrual dysphoric disorder (PMDD) in women, and the case of BiDil, the first drug approved for exclusive use in African Americans for the treatment of congestive heart failure (CHF). I conclude by cautioning against the unqualified support of democratizing strategies in industry-funded science and instead claim that assuring the mutual support of the ethical, epistemic, and commercial aspects of research is needed to avoid undesirable consequences.
2. Niche Standardization in Clinical Research
Following Epstein (Reference Epstein2007), I understand niche standardization as the replacement of the standard human subject in medical research, the white middle-class man, with an intersecting set of standard subtypes, leading to a new way of “transforming human populations into standardized objects available for scientific scrutiny, political administration, marketing, or other purposes that eschews both universalism and individualism and instead standardizes at the level of the social group” (135). Accordingly, and with the aim of achieving the appropriate inclusion of underrepresented groups in medical research, niche standardization favors the creation of a diversity of standards: a standard for blacks, a standard for women, a standard for Asians, and so on.
Niche standardization emerged from efforts to diversify medical research. In the early 1990s, as a result of health advocacy movements, both the Federal Drug Administration (FDA) and the National Institutes of Health (NIH) changed their research guidelines to guarantee the inclusion of marginalized groups in clinical trials. For instance, the 1993 NIH Revitalization Act states that NIH-funded clinical research must ensure that “(a) women are included as subjects in each project of such research; and (b) members of minority groups are included in such research” (NIH 1993, 492Ba1). As recruitment of patients for clinical trials became more diverse, the possibility of analyzing results per niche also arose. Thus, in addition to general and inclusive results, niche standardization allowed the production of results relevant to a specific subtype (racial, gender, age, etc.). Moreover, for the pharmaceutical industry, the shift toward niche standardization has come together with the opportunity for marketing drugs to specific health-care consumers. Two salient examples of such practice have been the cases of Sarafem and BiDil.
2.1. Sarafem and the Rebranding of Old Drugs
In June 1999, concerned with the upcoming expiration of Prozac’s patent, Eli Lilly organized a roundtable to discuss the use of fluoxetine, the active component in Prozac, for the treatment of other diseases or disorders different from depression, which would make the drug eligible for FDA patent extension. Among the participants were a number of members of the subcommittee for the study of PMDD of the Diagnostic and Statistical Manual of Mental Disorders (DSM). Originally named Late Luteal Phase Dysphoric Disorder, PMDD was first included in the DSM-III-R (1987, 3rd ed., rev.) under a new appendix on categories requiring further studying. The development of the new DSM-IV gave Eli Lilly a crucial opportunity. If PMDD could be ‘promoted’ to a real condition, it would be a suitable target for treatment, allowing Eli Lilly to argue for a patent extension (Caplan Reference Caplan and Chrisler2004).
An article asserting the existence of PMDD as a ‘distinct clinical entity’, and recommending selective serotonin reuptake inhibitors such as Prozac as effective treatment, was published by participants in the fluoxetine roundtable (Endicott et al. Reference Endicott1999). The article played a crucial role when the FDA committee met later that year to assess Eli Lilly’s patent extension application and was used to create consensus among committee members on both the existence of PMDD and the effectiveness of treatment with fluoxetine. The article, however, did not present an accurate overview of the relevant scientific research at the time. Surprisingly, many of the authors who also belonged to the DSM PMDD subcommittee had recently evaluated the scientific research on PMDD and concluded that the “research was sparse and methodologically poor” (Caplan Reference Caplan and Chrisler2004, 61; see also Cosgrove and Riddle Reference Cosgrove and Riddle2003, 38).
Chemically identical to Prozac, but now in appealing pink and purple pills, Sarafem hit the market in 2000, only 1 year before the expiration of Eli Lilly’s blockbuster antidepressant. The pharmaceutical giant successfully obtained FDA approval to market Sarafem for the treatment of PMDD. With the rebranding of Prozac for the treatment of PMDD, allegedly a new application of the active component, Sarafem was patent protected until 2007, which represented millions of dollars in revenue for Eli Lilly.
Niche marketing tactics allowed Eli Lilly to appeal to the needs of women patients and doctors, as well as justify the rebranding of such a well-known drug as Prozac. As Laura Miller, a spokesperson from Eli Lilly, stated in an interview: “We asked women and physicians, and they told us that they wanted a treatment with its own identity,” and moreover, “Women do not look at their symptoms as a depression, and PMDD is not depression but a separate clinical entity. Prozac is one of the more famous pharmaceutical trademarks and is closely associated with depression” (Vedantam Reference Vedantam2001). However, the introduction of Sarafem was not received without some skepticism. As Greenlit reports, “the media picked up on the coincidence of Prozac’s patent expiration. Much of the news coverage referred to PMDD as a ‘new’ illness category and not infrequently raised the question of whether PMDD was in actuality created for the marketing of Sarafem” (Reference Greenlit2005, 489).
2.2. BiDil and Race-Specific Drugs
The rebranding of already existing drugs has not been the only way pharmaceutical companies have used niche standardization. Another example is the case of BiDil, the first drug approved for exclusive use in African Americans. The analysis of racial subsets after a failed clinical trial (V-HeFT) suggested that the compound drug hydralazine and isosorbide dinitrate, the active components in BiDil, might have significant positive effect for the treatment of CHF in African Americans (Carson et al. Reference Carson, Ziesche, Johnson and Cohn1999). To get FDA approval, BiDil’s manufacturer NitroMed conducted a confirmatory race-specific study, the African American Heart Failure Trial (A-HeFT), which included 1,050 African American subjects (Taylor et al. Reference Taylor2004), with the support of the Association of Black Cardiologists (ABC) and the Congressional Black Caucus. In V-HeFT, BiDil was directly compared with angiotensin converting enzyme inhibitors (ACE inhibitors), a medication that reduces blood pressure and is currently the standard treatment for CHF. In contrast, A-Heft used ACE inhibitors in both the treatment and the control groups, comparing the use of BiDil plus ACE inhibitors in the treatment group to the use of ACE inhibitors alone in the control group (Bloche Reference Bloche2004). Because of its positive results, the trial was stopped earlier than expected to start treatment with patients in the control group (Henig Reference Henig2004). BiDil obtained FDA approval for the treatment of CHF in 2005.Footnote 1
BiDil became the first ethnic-specific drug to be approved in the United States. Patient ethnicity in the A-HeFT trial was determined through self-identification, the same method normally used in studies analyzing ethnic-specific impacts (Franciosa et al. Reference Franciosa2002, 133). Another retrospective analysis of the V-HeFT trial (Dries et al. Reference Dries1999) was used to argue that socioeconomic factors did not account for the disparity in mortality rates due to CHF between black and whites, by controlling for education and financial history. In the end, A-HeFT supported the hypothesis that adding BiDil to the normal course of treatment with ACE inhibitors was better for the treatment of CHF in self-identified African American patients than the standard treatment alone.
3. Consequences for the Goals of Research
Despite its laudable motivations, an inappropriate implementation of niche standardization can lead to undesirable epistemic and ethical consequences in the production of medical knowledge. By examining the cases of Sarafem and BiDil, this section highlights some of these consequences. As the case of Sarafem illustrates, niche marketing can be used as a tool for extending patent protection through rebranding of old drugs. While some might consider the rebranding of old drugs a positive practice, given that it helps with the treatment of diseases without further financial investment, in the case of Sarafem rebranding implied the reification of an otherwise nonexistent condition. In this sense, one might argue that this rebranding is actually ignorance productive, leading to the social construction of pseudodiseases as a strategy to market drugs instead of searching for drugs to treat existing diseases.
Eli Lilly’s support of the reification of PMDD as a mental disorder that can be successfully treated with Sarafem is a good example of the social construction of a pseudodisease (Chrisler and Caplan Reference Chrisler and Caplan2002; Angell Reference Angell2004; Offman and Kleinplatz Reference Offman and Kleinplatz2004; Greenlit Reference Greenlit2005). According to Angell, “Lilly’s launch of Sarafem made premenstrual symptoms a disease—now called ‘premenstrual dysphoric disorder’ (PMDD)” (Reference Angell2004, 189). PMDD was quickly added to an appendix on “Criteria Sets and Axes Provided for Further Study” in the DSM-IV-TR (2000, the year Sarafem hit the market) and then moved to section 2 on “Diagnostic Criteria and Codes” and into the chapter on “Depressive Disorders” in the most recent DSM-5 (2013).
The construction of pseudodiseases in connection with the rebranding of old drugs raises concerns. A central goal of medical research, both ethical and epistemic, is to find successful treatment for afflicted patients. The goal is ethical because it seeks to alleviate human suffering by treating patients’ symptoms, and it is epistemic because it seeks empirical accuracy through treatment effectiveness. Now, in the case of Sarafem, the goal has been inverted: instead of finding successful treatment for afflicted patients, doctors were seeking new afflicted patients for an already existing treatment. This compromises the original goal: instead of successfully treating sick people, we are now making people ‘sick’ in order to treat them, which is neither ethical nor epistemically sound.Footnote 2
A clear reason behind this goal inversion was the patent extension of Prozac, Eli Lilly’s blockbuster drug. Thus, a strong commercial interest, which steered research away from a central epistemic and social goal, was behind the support of Sarafem for the treatment of PMDD. Niche standardization therefore contributed to obtaining the patent extension, a commercial goal, instead of helping achieve a better understanding of diseases and treatment for women, that is, the ethical and epistemic goals that were put forward by Eli Lilly spokesperson Laura Miller to justify it. In fact, insofar as we have good reasons to believe that PMDD was added to the DSM under questionable circumstances, this implementation of niche standardization can be understood as an obstacle to obtaining better medical knowledge and treatment for women.
The case of BiDil illustrates another limitation of niche standardization. As mentioned before, BiDil was approved to treat CHF in African Americans, and the ethnicity of the patients was determined by self-identification (Franciosa et al. Reference Franciosa2002, 133). Self-identification is used given the absence of biological markers to differentiate human beings by race. As Epstein explains, “Unlike other species, including other primates, humans cannot be disaggregated into clearly defined genetic subspecies—meaning that the eighteenth- and nineteenth-century racist and imperialist conception of humanity as divided into biologically discrete groups simply has no basis in fact” (Reference Epstein2007, 211; see also Fuentes Reference Fuentes2012). In other words, race-specific clinical studies are not tracing causal relations between group-specific genetic markers, diseases, and successful treatment but instead are producing correlations (real or spurious) between self-identified racial group membership and success in treatment. In this sense, studies such as A-HeFT aim to establish, for patients suffering from CHF, a significant correlation between being a member of the African American community and benefiting from adding BiDil to the standard course of treatment.
Here one can distinguish between two different concerns: the ethical concern that race-specific trials such as A-HeFT and race-specific drugs such as BiDil promote biological racial essentialism, strengthening our racial stereotypes and serving to justify racial inequality, and the epistemic concern that A-HeFT does not help us understand the underlying mechanisms that explain why BiDil works specifically for this group of people.
This is a controversial issue. On the one hand, opponents of BiDil argue that race-specific studies such as A-Heft are problematic because they encourage racial essentialism and they do not provide the relevant mechanistic knowledge, highlighting both an ethical and an epistemic concern. As Schwartz claims, “Such research mistakenly assumes an inherent biological difference between black-skinned and white-skinned people. It falls into error by attributing a complex physiological or clinical phenomenon to arbitrary aspects of external appearance. It is implausible that a few genes that account for such outward characteristics could be meaningfully linked to multigenic diseases such as diabetes mellitus or to the intricacies of the therapeutic effect of a drug” (Reference Schwartz2001, 1392). According to this view, niche standardization can encourage the use of nonbiological categories such as race for the development and marketing of drugs for the treatment of symptoms, neglecting a deeper understanding of the mechanisms associated with the disease. Thus, critics of BiDil have an epistemic concern about the depth of the medical knowledge produced.
In addition, opponents also have an ethical concern. As Schwartz highlights, “A racial designation in the context of medical management not only defies everything we have learned from biology, genetics, and history, but also opens the door to inequities in medical care” (Reference Schwartz2001, 1392), inequities such as improper treatment for patients who do not conform to the stereotype, encouraging the discovery of dubious associations between race and medical results, fostering an inadequate understanding of race, and obscuring the real causes of health disparities (Epstein Reference Epstein2007, 219–35). In other words, niche standardization and the racial profiling that it encourages might be importantly obstructing our understanding of the social, cultural, and economic factors tied to health disparities (Valles Reference Valles2012). And this not only diminishes our knowledge of the causes of health differences but also supports an inherently unjust health system.
On the other hand, supporters of BiDil might argue that clinical trials can provide strong evidence that a treatment works, without having to explain the underlying mechanisms that make the treatment successful. As philosopher of epidemiology Alex Broadbent says, “The identification of a mechanism is not necessary for a warranted causal inference” (Reference Broadbent, Illari, Russo and Williamson2011, 60). Specifically, BiDil is just one among many treatments, acetaminophen and penicillin included, that work, which we know work, but we do not know why they work. In this sense, mechanistic reasoning is not necessary for developing a successful treatment, and thus the opponent’s epistemic concern should not be particularly worrisome.
In fact, if developing a successful treatment can relieve patients from suffering, we might argue that there’s a moral duty to develop such a treatment, even when we do not have an appropriate explanation for its success. A similar argument has been made in favor of BiDil, which weakens the opponent’s ethical concern. The ABC provided strong support for the development and marketing of BiDil, acknowledging that race is not a biological category but thinking that race was a useful proxy for finding a successful treatment for a traditionally marginalized group. Former president of ABC, Paul Underwood has made this point clear: “I think there’s no implication at all that genetics has anything to do with BiDil’s mechanism of action. … I believe that race in the United States is a social construct, not really a genetic construct. … This is really just looking at a group of individuals who seem to have a higher propensity toward heart failure and more severe ramifications of heart failure” (Goodman Reference Goodman2005, cited in Pollock Reference Pollock2012, 175).
While the epistemic and ethical concerns with BiDil discussed over the last paragraphs are conceptually distinct, they are significantly entangled in practice. In this sense, one might argue that BiDil’s case is different from the other successful treatments for which we do not have mechanistic explanations, precisely because we are dealing here with a social construct that is politically laden and that has contributed to the marginalization of a group for centuries. In this sense, ‘race’ is not merely a useful proxy for confounding variables that we do not yet understand. Instead, ‘race’ is the social category that actively obscures those other variables (e.g., economic and health disparities, unequal education and job opportunities). Thus, in this case we have a compelling reason for demanding a more detailed explanation of the mechanisms involved, given that such explanation would presumably help us dismantle racial prejudice. In this sense, the opponents’ epistemic concern is bolstered by the ethical concern regarding racial equality.
In other words, given that supporters of BiDil acknowledge that race is not biologically determined and that it has been used here only as a proxy, and given that we know how politically laden and troublesome race in fact is as a proxy, then settling for the success of BiDil as a treatment without mechanistic knowledge is not enough. I consider this to be a strong reason to side up with the critics of the race-specific prescription of BiDil, without contesting that BiDil works for the treatment of CHF in African Americans.
In the same vein, one cannot ignore the subtler epistemic limitations of the research supporting the race-specific prescription of BiDil. First, only one retrospective analysis of V-HeFT tried to account for confounding factors to show that they did not account for disparity in mortality rates (Dries et al. Reference Dries1999). However, the study only focused on education and financial history, while many social factors correlate with racial social categories (see, e.g., Williams Reference Williams1992; Kahn Reference Kahn2004). Second, there was an important difference in study design between V-HeFT and A-HeFT. While V-HeFT directly compared BiDil to the standard ACE inhibitors, A-HeFT used ACE inhibitors in all patients. Given that A-HeFT focused on African Americans, we do not know whether the combination of BiDil plus ACE inhibitors would be a successful therapy for the general population. However, as with Sarafem, patent protection was a strong commercial incentive to market BiDil specifically to African Americans, given that the original patent for use in the general population was filed in 1989, while the race-specific patent was filed in 2002 (Kahn Reference Kahn2004). So here again niche standardization served economic interests, instead of serving the ethical and epistemic goals it was originally conceived to serve.
4. On the Relation among Different Aspects of Research
What is the relation among the ethical, epistemic, and commercial aspects of scientific research, and how do (or should) these aspects affect one another? Lately, philosophers of science have become increasingly concerned with questions of this kind. In particular, a growing body of work rejecting the value-free ideal of science and focusing on the role of ethical and social values in scientific research shows philosophers’ concern with understanding the influence of ethical and social values in science and identifying the appropriate role these values should play in the various stages of research. Underdetermination arguments (Longino Reference Longino1990; Intemann Reference Intemann2005; Elliott Reference Elliott2011; Biddle Reference Biddle2013), inductive risk arguments (Douglas Reference Douglas2000, Reference Douglas2009; Wilholt Reference Wilholt2009; Elliott and Richards Reference Elliott and Richards2017), arguments challenging the lexical priority of evidence (Anderson Reference Anderson2004; Kourany Reference Kourany2010; Brown Reference Brown2013), and more recently, arguments distinguishing different types of epistemic risk (Biddle Reference Biddle2016; Biddle and Kukla Reference Biddle and Kukla2017) have all been used to clarify the character and the appropriate extent of the influence of ethical and social values in scientific research. These arguments, however, focus on the relation between the ethical and the epistemic aspects of research without really considering the commercial interests at play.Footnote 3
I claim that the relation among the different aspects of research is more complex when commercial interests are involved and that, given the current organization of science, which is driven increasingly by the private industry, it is paramount that philosophers also consider such interests (Fernández Pinto Reference Fernández Pinto2014, Reference Fernández Pinto2015). Of course, the role that ethical and social values play (or ought to play) in the process of knowledge production is important, but so is the way in which the process of knowledge production affects human beings and society and the role that commercial interests play in directing this impact. In fact, the ethical, epistemic, and commercial aspects are so intertwined in industry-driven research that changes in one aspect trigger changes in the others, as the cases examined in this article illustrate.
Strategies to democratize medical research, such as niche standardization, look for the inclusion and proper treatment of traditionally marginalized subjects (a social justice goal) and for a more accurate understanding of diseases, symptoms, and treatment for these subjects (an epistemic goal). In this sense, niche standardization was originally conceived to achieve both ethical and epistemic goals. However, when we also consider the commercial goals of industry-driven science, niche standardization does not necessarily contribute to the democratization of medical research. The examples of BiDil and Sarafem illustrate how democratizing strategies such as niche standardization can be steered away from the initial ethical and epistemic commitments to serve commercial goals. Hence, these cases show a misalignment among the different goals of research. The ethical and epistemic goal of acquiring more accurate knowledge regarding health differences of underrepresented groups is not properly aligned with commercial goals, such as procuring patent protection. Consequently, some goals are unjustifiably privileged to the detriment of others. As Intemann and de Melo-Martin explain, “Commercial values are problematic in biomedical research insofar as they arbitrarily promote economic interests to the exclusion of epistemic or social values justifiably held by stakeholders affected by the research” (Reference Intemann, de Melo-Martin, Amoretti and Vasallo2016, 86).
In this sense, the problem with strategies to make the scientific process more democratic and inclusive in industry-funded research can be characterized in terms of a misalignment between the ethical, epistemic, and commercial aspects of research. As the examples illustrate, instead of complementing democratizing strategies, leading to more empirically accurate, socially responsible, and commercially successful medical knowledge, commercial interests have compromised the democratization process, steering research away from its ethical and epistemic targets. Given their close interrelation in industry-funded science, the ethical, epistemic, and commercial aspects of research ought to work in tandem, that is, they should be mutually aligned, or they might end up obstructing, instead of supporting, one another.
5. Conclusion
Strategies to democratize the scientific process in order to make it more diverse and inclusive have been suggested to counteract some of the damages to the process of knowledge production emerging from industry-driven scientific research. In the article, I have expressed some concerns about this strategy, showing that purportedly democratizing strategies such as niche standardization can be used to favor commercial interests in detriment of ethical and epistemic goals. To better understand the influence of commercial interests in scientific research, and to better identify some of its detrimental effects, I characterize the problem as a misalignment of the different goals of research.
As Kourany claims, philosophers of science should aim for an understanding of scientific rationality “that integrates the ethical with the epistemic. The ideal of socially responsible science … maintains that sound social values as well as sound epistemic values must control every aspect of the scientific research process, from the choice of research questions to the communication and application of results” (Reference Kourany2012, 344). In addition, I maintain that any attempt at democratizing industry-funded research must not only preserve the integration of ethical and epistemic interests but also acknowledge and integrate its commercial goals, so that the epistemic, ethical, and commercial aspects of research complement one another. In this sense, I agree with Hicks’s view that philosophers of science ought to pay more attention to the ethical issues raised by commercialized science and, in particular, to “the interactions between the ethical and epistemological aspects of the problem” (Reference Hicks2014, 3292). Also, my analysis supports Harding’s (Reference Harding2015) recent claim that democratic values, such as inclusion and diversity, and epistemic values, such as objectivity, can be mutually supportive. However, instead of showing how this mutually supportive relation has in fact occurred, as Harding does, I am highlighting cases in which it has failed, precisely because of a misalignment, where the mutual support has not been properly ensured.
Accordingly, I conclude by cautioning against the unqualified support of strategies to democratize commercially driven science and instead claim that the mutual support of the different aspects of research is needed to achieve epistemically and ethically sound research results.
