key: cord-0847790-ixd6mwul
authors: Morreim, E H
title: Corporations, high-stakes biomedical research, and research misconduct: yes they can (and sometimes do)
date: 2021-07-01
journal: J Law Biosci
DOI: 10.1093/jlb/lsab014
sha: 8f143804846a846a0119f3044801d6dfcd7e015e
doc_id: 847790
cord_uid: ixd6mwul

Science has long been vulnerable to research misconduct (RM). Biomedical sciences, with vast financial stakes, carry heightened temptations. However, RM is standardly seen as an undertaking of individual scientists, not as something that could be committed by an organization such as a corporation or university. Rather, organizations are generally regarded merely as supervisors to encourage scientific integrity and investigate suspected RM. Indeed, federal regulations expressly embrace this perspective, and the federal Office of Research Integrity has never deemed an organization guilty of committing RM. This article aims to rewrite this corner of research integrity: organizations can directly commit RM and should be held accountable as such. Although the conclusions apply to organizations such as universities and government agencies, the focus here is on corporations in the biomedical sciences. After defining ‘research misconduct’ in Part II, Part III describes corporate-level RM and distinguishes it from individuals’ misconduct. Part IV provides five case studies exemplifying corporate RM, while Part V discusses implications, describes ways in which federal regulations could already encompass organization-level RM, and identifies some needed legal and regulatory adjustments.

ward individual scientists . . . who somehow managed the considerable collaboration that would be needed to produce bogus research that is persuasive enough to gain regulatory approval and marketing panache.

And yet that argument has never been made. That is, it has never been argued that research misconduct-by the definition-has been or could be committed by an organization. 7 Historically, misconduct in science focuses exclusively on wayward individuals, from Galileo, to Newton, to Mendel, and beyond. 8 In the same vein, regulations governing RM for federally funded research expressly regard organizations just as supervisors whose job is to encourage scientific integrity and to investigate when RM is suspected. 9 Moreover, throughout its history the federal Office of Research Integrity (ORI), which addresses scientific integrity affecting a substantial portion of federally funded biomedical research, 10 has investigated hundreds of allegations of RM. All of them focus on individual persons, never on organizations. 11 7 Seeinfra n. 167 for discussion of a somewhat parallel concept, 'institutional misconduct' , and the limitations of that concept. 8 See sources cited supra note 1. 9 See infra Part II. In federally funded research, for instance, the PHS expects institutions receiving federal grants to supervise their scientists and to 'be fully accountable for the appropriate use of any funds awarded' . Grant-receiving organizations must 'take all reasonable and practical steps to foster research integrity' , 'ensur[e] that it is conducting its NIH-funded project in accordance with the approved application and budget and the terms and conditions of the award,' and 'report promptly to ORI any decision to initiate an investigation of research misconduct. U.S. Dep't Health & Human Serv's, NIH Grants Policy Statement §4.1.27 (2019), https://grants.nih.gov/grants/policy/nihgps/nihgps.pdf [hereinafter NIH Policy Statement].The applicant organization is responsible for verifying conformity with the most current guidelines for all administrative, fiscal, and scientific information in the application, . . . . [and] certifies that the applicant organization has the ability to provide appropriate administrative and scientific oversight of the project. Id. at §2.3.6 (I-48 to I-49); see also id. at §2.3.10 (I-63). 42 CFR 93.412 provides:

(a) Institutions must foster a research environment that discourages misconduct in all research and that deals forthrightly with possible misconduct associated with PHS-supported research. (b) ORI may decide that an institution is not compliant with this part if the institution shows a disregard for, or inability or unwillingness to implement and follow the requirements of this part and its assurance . . . . Organizations have of course been held accountable in the setting of RM, but never for directly committing RM. This was the case, e.g., when in 2019 Duke University agreed to pay the federal government $112.5 million to settle False Claims Act (FCA) charges in connection with RM. Duke was held responsible for its employee's actions, not because Duke was said to have committed misconduct, but because the university agreed to be financially responsible as a condition of receiving federal research money, and because the researcher was acting within the scope of her employment. 12 Admittedly, detecting and proving RM by an organization, orchestrated within the highest levels of management, can be difficult without substantial legal action and prodigious discovery to unearth the in-house emails and other communications that would reveal organization-level RM. Yet, as discussed below, 13 this is precisely what becomes available where corporate misdeeds spark intensive litigation.

This article aims to rewrite this corner of research integrity, arguing that organizations can directly commit research misconduct-by the definition-and that organizations can and should be held accountable not just for failures of oversight, but also for their own direct actions to corrupt science. Although these conclusions apply broadly, including to universities and government agencies, this article focuses on corporations engaged in biomedical research.

After defining 'research misconduct' in Part II, we will consider in Part III what it would look like for a corporation to commit RM, and how we might distinguish between corporate versus individual actors. Part IV first reframes several historic lawsuits to show how corporate misdeeds almost certainly amounted to RM even though RM was never expressly alleged. Part IV then details a new instance, also exposed through litigation, in which arguably a corporation directly committed RM. Finally, feld, one retrospective review of ORI guilty verdicts opined that misconduct was largely attributable to 'individual psychological traits and the circumstances in which the researchers found themselves' , whereas another review found that '[a]pproximately one-third of the respondents (the accused) were support staff, one-third were postdoctoral fellows and graduate students, and one-third were faculty. Accusations of fabrication represented 45 per cent of the offenses, falsification 66 per cent, and plagiarism 12 per cent. The first two offenses frequently occurred together. Approximately three-quarters of the respondents admitted their guilt or did not provide a defense' . Id Part V explores some implications of this novel finding. Although this article does not purport to provide an exhaustive exploration, federal regulations illustrate the necessary adjustments in how to identify and sanction research misconduct.

To make the case that an organization can commit research misconduct-by the definition-we must first define 'research misconduct' . The concept is variably defined, 14 yet two elements are universally included-fabrication and falsification. Both are included in the federal definition, which serves as the template for definitions worldwide 15 (b) The misconduct be committed intentionally, knowingly, or recklessly; and (c) The allegation be proven by a preponderance of the evidence. 17 Importantly, this definition emphasizes the research record. A solitary scientist inventing numbers in a lab is not a serious threat to science until she/he actually shares those fictitions with a claim that they represent proper science. Federal regulations governing research funded by the Public Health Service (PHS) are instructive:

Research record means the record of data or results that embody the facts resulting from scientific inquiry, including but not limited to, research proposals, laboratory records, both physical and electronic, progress reports, abstracts, theses, oral presentations, internal reports, journal articles, and any documents and materials provided to Department of Health and Human Services (HHS) or an institutional official by a respondent in the course of the research misconduct proceeding. 18 In other words, the research record essentially is any assertion purporting to describe the underlying scientific work and its findings.

This article proposes distinguishing RM from a broader concept: scientific misconduct. Research misconduct focuses on individual research projects and violations of proper design, execution, analysis, and reporting, whereas scientific misconduct emphasizes damage to the broader scientific literature, where the problem may not specifically reside within any particular study/studies.

To illustrate the difference, suppose a pharmaceutical company seeking Food and Drug Administration (FDA) approval for a new drug provides just one study showing the drug to be safe and effective, while withholding nine studies concluding the contrary. This would not be research misconduct if all 10 studies feature high-quality methodology, data gathering, analysis, and reporting. It might, however, be scientific misconduct if the company purposely mischaracterizes the overall science as supporting the drug, when greater, but buried, evidence suggests otherwise. 19 20 In contrast, if the company were "cooking" or selectively deleting data to produce desired conclusions in a given study, this would be direct, organization-level RM. Research misconduct is thus a subset of scientific misconduct.

Plagiarism, often listed as a core genre of RM, 21 exemplifies the distinction. Inasmuch as the plagiarist falsely states that she/he, rather than the true author, created the research concept, design, execution, etc., then she/he arguably commits RM, in essence falsifying the report of authorship. 22 It is akin to another well-recognized genre of falsification, namely, stating that a research project has secured institutional review board (IRB) approval when it has not. Ghostwriting, another form of authorship falsification, is discussed below. 23 In contrast, plagiarism can also be a form of scientific misconduct, e.g., where it amounts to duplication. Duplicate publication damages the scientific literature as a whole by incorrectly augmenting the number of studies that reach a particular conclusion, 24 even where that study (and thereby its duplications) has internal scientific integrity.

The cornerstone of the distinction is that, for scientific misconduct: if the underlying individual research projects and reports are faithful to the truth, we can at least potentially wade through the mess and find our way to correct it. However, for research misconduct: if the design of a study is contoured to avoid truth, or if the data are falsified/fabricated or the report is otherwise untruthful, we have nowhere else to Parrish does not distinguish between RM and scientific misconduct and, per the definitions presented in this article, actually conflates the two. His discussion, however, mainly falls within the bounds of RM. 23 Admittedly, team science and the large numbers of authors identified in much of contemporary biomedical research pose challenges. The challenges, however, concern not the basic concepts of RM versus SM, but rather the question of how to allocate responsibility where participation in the research is so diffuse. Some approaches to authorship expect that everyone claiming to be an author should be responsible where RM has occurred, whereas others propose that accountability should be more narrowly focused. See turn. This article thus focuses on research misconduct because, as it violates science's integrity at the most basic level-the individual study-it is a particularly pernicious form of scientific misconduct. We next consider the questions of what RM would look like when undertaken by an organization and, in any given instance, how one would determine whether the actor is the organization versus the individual scientist(s).

Organizations can commit fabrication or falsification via the same avenues as individuals: in design, execution, analysis, and reporting of the research. Although data manipulation comprises the great majority of instances in which, e.g., ORI has found RM, 25 myriad other avenues are available for producing a research record that does not accurately reflect underlying realities.

To ensure a product will shine regardless of realities, potential avenues by which design could implicate RM include:

* selecting for research only those questions whose results will likely favor the corporation; 26 * unrepresentative patient selection, e.g., formulating the study's inclusion/exclusion criteria to enroll only those individuals likely to benefit and unlikely to experience adverse effects; 27 * selective use of surrogate endpoints; * failure to define endpoints clearly; * follow-up intervals that maximize the likelihood of favorable outcomes and minimize the tally of adverse events; * trials that are too small to show differences from competitor drugs; * multiple endpoints in the trial, selecting for publication only those with favorable results; * testing a drug against competing treatment(s) known to be inferior; * testing a drug against too low a dose of a competitor drug (making one's own drug appear more effective), or too high a dose (making one's drug appear less toxic). 28 25 See Case Summaries, supra notes 11-12. 26 "Large pharmaceutical companies have the financial means to conduct wide-ranging trials, but may be selective in the type of questions they want answered. Once a drug is approved, there is considerable commercial pressure not to inquire further into potential side-effects, and this absence of statistical evidence is then used as a shield against criticism." Lemmens, supra note 20, at 652. This is not to say that each of the foregoing is RM per se. Every research project must circumscribe its attention, and focusing on whatever is most likely to succeed could exhibit bias shy of misconduct. However, a line is crossed when research is consciously designed to produce a result that is, or likely is, contrary to underlying realities.

Once a clinical trial is designed, companies must select investigators to recruit human participants, provide research interventions according to protocol, gather data, and submit results. Ideally a company will select highly qualified investigators known for meticulous work. Instead, however, it might select and financially incentivize investigators known to recruit subjects and complete trials with lightning speed, potentially glossing over eligibility criteria, steering people selectively into trial groups rather than honoring randomization, or 'tuning' data to whatever is needed. 29 Contract research organizations (CROs) often play a large role in conducting clinical trials, planning logistics and running trials. '[P]harma companies sponsor some 70% of all clinical trials, and 70-75% of these are run by CROs' 30 that typically recruit the physicians who then recruit the trial subjects and collect the data. Those physicians, in turn, may or may not adhere scrupulously to trial protocols. 31 Unlike academicians or private practice physicians, CROs generally 'have no interest in publishing the results under their own names-they produce data that is wholly owned by their sponsors. As a result, pharma companies have complete control over an enormous trove of clinical trial data' 32 they can then potentially use to their best advantage. 33 Data manipulation can occur during data gathering or at the ensuing stage of data analysis (discussed just below). Investigators could, e.g., concoct, shade, delete, or workers who recruit them, has been discussed in the literature. Commentators have also pointed out that financial interests may negatively impact researchers' dealings with research subjects during a trial. When huge profits lure, and pressure mounts to bring novel drugs or therapies quickly to the market, potential risks may be perceived somewhat more lightly, and inclusion or exclusion criteria may become more flexible.' Lemmens. supra note 20, at 645. 30 Sismondo, supra note 27, at 68. 31 '[P]erhaps the simplest way to boost recruitment is to establish good rapport with your patients. "I'd say that about 20 percent of the volunteers we recruit are our own patients, who enroll because they like the close attention they get when they are in a study" says Elinolf. Another 30 to 40 percent are also patients, who sign up on the strength of our relationship with them. simply not record unfavorable data. 34 Sometimes an investigator might be sloppy, failing to conform to protocol, yet the company may use the data anyway. 35 Overall, if an investigator or a company 'ma[kes] up data or results and record[s] or report[s] them,' that is fabrication. 36 If it 'manipulat[es] research materials, equipment, or processes, or chang[es] or omit[s] data or results such that the research is not accurately represented in the research record,' 37 then it commits falsification. Importantly, one act (e.g., fabricating a dataset) can corrupt multiple trials that rely on that dataset. 38 Finally, if all this is done with the intent, encouragement or tacit knowledge of the company, it could be corporate-level RM.

Once data are gathered they must be analyzed. Data can 'cooked', a concept 'Charles Babbage in 1830 defined as "an art of various forms, the object of which is to give to ordinary observations the appearance and character of those of the highest degree of accuracy"; it can be "mined" to find a statistically significant relationship that is then presented as the original target of the study' . 39 Data points can be dropped to conform to expectations, or design or methodology descriptions can be changed, post hoc, to produce favorable results. 40 In one instance, a packaging problem permitted nine subjects to see which medication they were receiving-a violation of the trial's blinded protocol. Initially the pharmaceutical company 'ran the numbers without those nine, and the result fell just short of statistical significance, at which point eight of the subjects were added back into the analysis, improving the statistics enough to make the results significant-the problem was later hidden in the submission to the FDA . . . ' 41 .

In another genre of data manipulation, 'many published results of randomised trials are biased by selective reporting of trials with positive results, and by researchers . . . selectively reporting nominally statistically significant outcomes. This latter practice has been called P-hacking' . 42 It can involve 'monitoring data accrual and stopping a trial if an analysis yields a significant P-value or using different statistical analyses, data 34 eligibility criteria, outcomes and treatment groups before deciding which to report post-analysis' . 43 Perhaps worse, in nonregistered clinical trials, postulated outcomes may be silently switched once analysis shows where statistical significance lies. 44 In a particularly troubling development, drug makers or their CROs can commit covert data manipulation in bioequivalence trials. 45 Essentially, the practice involves 'manipulat[ing] bioequivalence trial data for non-approvable formulations by performing an interim analysis followed by re-analysis of pharmacokinetic [PK] profiles under new subject aliases, with a switch of Test and Reference and/or dilutions', 'performing an undocumented interim statistical analysis after a portion of the subject PK-data has become available' . 46 Federal regulations identify myriad kinds of behavior that can trigger warnings or sanctions at the execution and analysis stages, including 47 : fabrication of data and their recording and reporting; manipulation of data so that data no longer accurately reflect what was observed; repeated and systematic deviation from the established protocol; deviation from investigational plan; failure to maintain adequate/accurate source documentation; failure to personally supervise the study; submission of false information to the FDA and sponsor. 

Even if a trial is properly designed, its procedures followed, and data gathered and analyzed appropriately, RM can still occur if write-ups do not accurately reflect reality. As noted above, the research record is essentially any assertion that purports to describe the underlying scientific work and its findings. 53 Reports can misrepresent the findings and, if underlying details of design, data, and analysis are not made available, misrepresentation can be difficult to detect. 54 In one example discussed below, Merck pharmaceutical corporation is said to have 'omitted 43 key incidences of heart troubles, creating "misleading" conclusions about [the drug Vioxx's] safety' . 55 Ghostwriting is a particularly insidious form of RM. 56 As early as 1994, editors of prominent medical journals lamented:

There are ghosts as well as guests lurking in the bylines-shadowy figures who, increasingly, are in fact the actual writers . . . . These ghosts might be clinical research associates at pharmaceutical companies who are bound by employment contracts not to be listed as authors or public relations officers writing for government officials or organization executives . . . and prominent physicians have then been hired to allow their names to be attached as 'authors' before the reviews were submitted to learned journals. 57

Ghostwriting has been defined as 'the failure to designate an individual (as an author) who has made a substantial contribution to the research or writing of a manuscript' . 58 The treachery is then completed by inducing-usually paying-a well-known academic thought leader(s) to serve as the first author(s) of the article. This latter is a form of guest/honorary authorship, defined as 'the designation of an individual who does not meet authorship criteria as an author' . 59 It is clear, albeit under-recognized, that ghost and guest authorship are a form of research misconduct. Analogous to plagiarism, 62 they falsely state who bears responsibility (not just credit) for the work in question. Hence, the research record thus does not correctly reflect the underlying research realities. More importantly, they conceal what may be ulterior agendas embedded in the work, thereby rendering more difficult the task of discerning its credibility.

In biomedical science ghostwriting as RM is, by its very nature, almost uniquely attributable to organizations rather than to individual scientists. In the case of pharmaceutical studies, the corporation as an entity designs the research, hires the CRO or local investigators, analyzes the data and does the write-up. Only then does it hire the lead 'authors' . The correlative guest authorship RM, conversely, is attributable to the individuals who allow themselves to be listed as authors when they know they have done no real authoring.

Where these research offenses are committed intentionally, knowingly, or recklessly, and constitute a significant departure from accepted research practices, 63 research misconduct has arguably occurred. 64 As explored in Part IV, discerning corporation-level intentionality can be difficult-until litigation reveals the in-house communications that evidence actors' underlying agendas. Several historic examples showing just such intent/knowledge will be explored, as well as a newer case of corporate RM.

It is one thing to suggest that an organization can commit RM, but quite another to discern, in a given instance, whether the organization or individual actors are responsible. The answer begins by considering how and by whom the project was conceived, and under whose direction it was carried out. Concepts of agency and of authorship help us identify that 'moving force' .

US law has long recognized that a corporation is a 'person' capable of setting and accomplishing objectives, mainly by directing employees. 65 The Court has recognized that First Amendment protection extends to corporations . . . . ('The identity of the speaker is not decisive in determining whether speech is protected. Corporations and other associations, like individuals, contribute to the "discussion, debate, and the dissemination of information and ideas" that the First Amendment seeks to foster' . . . ) The Court has thus rejected the argument that political speech of corporations or other associations should be treated differently under the First Amendment simply because such associations are not 'natural persons' . vicariously liable for employees' conduct, they can also be directly liable for their own actions. Hence, whereas respondeat superior penalizes a company when its employee carelessly causes a car accident, that same company can be directly liable for negligent entrustment if it knew or should have known the employee was a chronic drunk with a lengthy crash record. 66 Guidance regarding corporation-as-actor comes from the Restatement 3d of Agency: 'an employee is an agent whose principal controls or has the right to control the manner and means of the agent's performance of work . . . ' . 67 Here, where the corporation acts as principal steering the employee's action, then we look to the corporation to assess potential RM. Thus, a pharmaceutical firm typically decides which areas of research will be pursued or abandoned, which employees will work on which projects, etc.

Corporate persons are also capable of acting knowingly, intentionally or recklessly. Since 1863, for instance, the FCA has been invoked to hold both corporations and individual persons responsible for defrauding the federal government. 68 A key criterion of that statute is that the false claim be 'knowing', defined to include deliberate ignorance and reckless disregard of the truth. 69 Accordingly, so long as a research project is undertaken by an organization controlling its agents' performance and designed to serve the organization's purposes, RM can ordinarily be attributed to the organization. In contrast, research grants illustrate initiatives by individual scientists. The scientists conceive of an idea, write up a grant An employer who is liable solely because of respondeat superior is not necessarily negligent; instead, it is liable because we have made a 'public policy determination that liability for acts committed within the scope of employment should be allocated to the employer as a cost of engaging in that business' . In situations of liability for negligent entrustment, on the other hand, the negligence of the employer is likely the focus of the claim. In other words, '[d]irect liability is liability for breach of one's own duty of care, while vicarious liability . . . is liability for breach of another's duty of care. A negligent entrustment claim is therefore an instance of direct liability' , as one of the key components of the tort is that the employer breached its own duty of care. proposal and seek funding. An organization such as a university might then, as envisioned in federal grant regulations, 70 manage the funds to ensure the work complies with funding rules. Here, the individual scientists would most likely be the perpetrators if RM occurs. That said, although this article focuses on corporations and RM, we note here that a university could also be the entity to directly commit RM, for instance through the Bayh-Dole Act that permits not-for-profit organizations such as academic institutions to patent products they develop from federally sponsored research. 71 University officials could identify promising projects, then direct and incentivize investigators to design and execute those studies most likely produce the desired outcomes regardless of their veracity. It has been suggested that incentives in academia have become increasingly perverse, becoming a system that 'selectively weeds out ethical and altruistic actors, while selecting for academics who are more comfortable and responsive to perverse incentives from the point of entry' . 72

Authorship likewise illuminates the distinction between corporations versus individuals in RM. To be an 'author' is to be the creator, source, or originator of a work. 73 If the organization authored the study, then it should be the entity directly accountable. 74 International Committee of Medical Journal Editors (ICMJE) guidelines are followed by myriad medical journals:

The ICMJE recommends that authorship be based on the following 4 criteria: The first two criteria are particularly relevant: [1] substantial contribution to conception/design or data acquisition/analysis/interpretation and [2] drafting/critically revising the work. Regarding [1]: as noted above, organizations can and do undertake activities. They need not literally 'hold the pen' with which bogus data are created. It is sufficient that they direct someone to. Indeed, in one of the ORI's earliest findings of research misconduct, Dr. Roger Poisson was found guilty of fabrication and/or falsification because '[i]nterviews with the project staff revealed that the actual data changes had been made by the data management staff at the direction of the Principal Investigator, Dr. Roger Poisson' . 76 Broadly, much of medical researchers' authoring is quite indirect. In large clinical trials the clinician-investigator often is not the one who actually screens potential subjects, provides the intervention, and records the data. She/he may do some of these things, but much of the work may be done by a research nurse coordinator, or outsourced to a CRO. 77 Nurse coordinators and CROs generally are not on the list of authors even though they do much or most of the actual data gathering. The one who hired them will ordinarily be deemed author.

In 'bench research', again the named investigators may not be the ones to pour chemicals into beakers, handle lab mice, or record numbers in notebooks. Lab techs often do much or all of this. 78 The investigator is author because she/he is the one who initiated such work and engaged the lab techs, or who agreed with colleagues to participate in the collaborative effort to gather such data according to the agreed-on methodology for the agreed-on purpose.

Regarding [2]: similarly, the process of drafting, critiquing, and editing manuscripts is often not entirely done by named authors. Many universities, for instance, provide services for scientific writing. 79 While such an office might require a first draft capturing the overall gist of the research, much of the further drafting and editorial refinement may be done by such services. 80 Scientists, after all, are not always the most lucid writers.

Moreover, medical science often requires numerous collaborators across many sites, hence numerous authors. 81 The upshot: if individual persons can be authors by hiring others to do the actual work or by affirming what others wrote or revised, then organizations are capable of the same kinds of authorship. Hence, in the end we must look to see which entity is the driving force conceiving the project and directing its execution.

It is one thing to show that organizations such as corporations can directly commit RM and quite another to show they actually have. As discussed in Part II, the most common offenses are fabrication or falsification 'such that the research is not accurately represented in the research record', 83 done knowingly, intentionally, or recklessly. 84 Part IV describes several cases in which a corporation arguably committed RM. At the corporate level the greater challenge is not so much to show, e.g., data were manufactured or falsified, as to prove knowledge, intent, or recklessness. 85 This latter is not ordinarily exposed unless litigation and extensive discovery provide access to communication among upper-level management. Here we discuss five litigated cases: two pharmaceutical seeding trials, concealment of adverse events in a drug trial, ghostwriting, and finally a case from the fitness industry. 86 Legal claims in these cases included False Claims Act violations, off-label marketing, anti-competitive behavior, The trial recruited 772 investigators who enrolled 2759 patients, about four patients per investigator. All were told the study was 'designed to assess the safety and tolerability of doses of Neurontin (gabapentin) from 900 to 3600 mg daily whose partial seizures are not completely controlled by other drugs' . 90 The company 'recruited site investigators with little or no clinical trial experience, provided insufficient training, and did not audit study sites prior to the beginning of the trial, which led to poor trial data quality' . 91 Per an in-house April 1996 memo: Seeding trials are designed to appear as if they answer a scientific question but primarily fulfill marketing objectives. Kessler and colleagues portrayed seeding trials as 'attempts to entice doctors to prescribe a new drug being marketed by the company' while the company puts its product in the hands of practicing physicians, hoping that the experience of treating patients with the study drug and a pleasant, even profitable, interaction with the company will result in more loyal physicians who prescribe the drug. Thus, management knew, perhaps intended that, the trial was poorly designed. Independent sources questioned the trial's scientific validity before it was initiated. The Johns Hopkins University IRB, for instance, rejected the application for the trial, both initially and on appeal, stating 'the board in its deliberation, voted to disapprove the protocol, since we believe that the entry criteria and outcome measures are too vague to allow any scientific conclusions to be reach [sic]' . 93 Moreover, following the trial's completion, some of the corporate customer business units-'autonomous, regionally focused branches of Warner-Lambert that planned and implemented marketing strategies [-] conceded that the study design was not rigorous enough for dissemination' . 94 Statistical analysis also was problematic. Site investigators did not adhere to the follow-up protocol, so that fewer than 25% of patients were assessed during the requisite time-frame. 'There were no mentions of data irregularity in either the internal research report or the published articles.' 95 Even from this brief summary, it appears that the Neurontin seeding trial clearly departed from established scientific norms in its design, execution, data analysis, and research report, arguably committing falsification. Additionally, the company's knowledge of these departures is evident from in-house communications. RM could reasonably be ascribed.

Merck's seeding trial for Vioxx likewise is best explored through confidential internal documents exposed by litigation, particularly Cona v Merck and Co., Inc., and McDarby v Merck and Co., Inc. and made directly available to litigation consultants. 96 Even prior to FDA approval for Vioxx, Merck launched a trial ostensibly comparing Vioxx (rofecoxib) with naproxin for people with osteoarthritis. 97 Merck recruited 600 investigators and over 5500 patients-just a few patients per investigator-in this three-month trial. 98 As with the Neurontin seeding trial, the objective was not science, but marketing: introduce as many physicians and patients to the drug as possible, prior to its FDA approval.

Although ADVANTAGE was called a seeding trial in many internal documents, the marketing objectives of the trial were not described on the informed consent form. objectives were to provide [a] product trial among a key physician group to accelerate uptake of VIOXX as the second entrant in a highly competitive new class and gather data important to this customer group.' 99 However, that aim was actively concealed in the publication's stated objectives and conclusion:

Objective: To assess the tolerability of rofecoxib compared with naproxen for treatment of osteoarthritis. Conclusions: In patients with osteoarthritis treated for 12 weeks, rofecoxib, 25 mg/d, was as effective as naproxen, 500 mg twice daily, but had statistically significantly superior GI tolerability and led to less use of concomitant gastrointestinal (GI) medications. Benefits of rofecoxib in subgroup analyses were consistent with findings in the overall sample. 100

This falsification of the research record is accompanied by evidence of knowledge and intent. Per a marketing division in-house email: 'It may be a seeding study, but let's not call it that in our internal documents' . 101 Hence, RM could reasonably be ascribed.

Merck's considerably more consequential act of likely RM involved concealing heart attack data from its main study to secure FDA approval for Vioxx: the VIGOR trial. 102 As with the other examples in Part IV, litigation provided the otherwise-unavailable in-house documents necessary to complete the case for RM, documenting knowledge and/or intent. Among various accounts, 103 perhaps the clearest is a narrative from a litigation consultant. 104 Early in Vioxx's development, Merck's scientists 'were concerned that the drug might adversely affect the cardiovascular (CV) system by altering the ratio of prostacyclin to thromboxane, which act in opposition, balancing blood flow and clotting.' 105 Their investigations continued, ultimately helping clarify the pathways by which the drug causes CV events. Despite knowing the drug could increase thrombus formation, Merck included none of this information in its application to the FDA. 106 The VIGOR trial, designed to measure upper gastrointestinal toxicity of rofecoxib versus naproxin in patients with rheumatoid arthritis, 'was designed to continue until a predetermined number of confirmed uncomplicated or complicated gastric perforations, ulcers, or bleeds had occurred.' 107 Despite Merck scientists' concern for potential CV events, the trial's Data Safety Monitoring Board (DSMB) had no cardiologist, and the trial had no standard procedure for collecting information on CV events. 108 In its second safety analysis, the DSMB found elevated CV risk in one group and recommended adding an analysis plan to capture that data. Noting this increased risk of myocardial infarction (MI), Merck's chief scientist emailed colleagues that this increased risk apparently had a 'mechanism based as we worried it was', namely the prostacyclin findings. 109 Nevertheless, published reports obscured the trial's adverse MI data. Although the cut-off date for enumerating adverse GI events was March 9, 2000, the cut-off date chosen for adverse CV events was set a month earlier, at February 10, 2000-thereby excluding three MIs that would otherwise have been counted. 110 Moreover, per the editors of New England Journal of Medicine, litigation documents showed that 'at least two of the authors knew about the three additional myocardial infarctions at least two weeks before the authors submitted the first of two revisions and 4 1/2 months before publication of the article' . 111 Issues associated with the VIGOR trial and the two seeding trials are considerably more complex than these brief summaries suggest. 112 Here, suffice it to say that the underlying research processes and data were evidently manipulated so that the 'research was not accurately represented in the research record', 113 hence falsified, and that corporate management knew this to be the case, thus supporting an inference of RM.

Part III discussed ghostwriting as a fairly common, distinctively corporate form of RM. Several Vioxx trials exemplify. As above, litigation revealed the in-house documents attesting to Merck's commission of ghostwriting, and its managerial knowledge and intent.

'When publishing their own clinical trials (designed, conducted, and sponsored by Merck), documents were found describing Merck scientists often working to prepare manuscripts and subsequently recruiting external, academically affiliated investiga-tors to collaborate on the manuscript as guest authors.' 114 'Documents were found describing Merck employees contracting with medical publishing companies to ghostwrite review manuscripts focused on rofecoxib and subsequently recruiting external, academically affiliated investigators to be guest authors.' 115 Indeed, the lead 'author' of Merck's seeding trial 116 expressly acknowledged that he played no role in data collection or analysis, stating:

Merck designed the trial, paid for the trial, ran the trial. Merck came to me after the study was completed and said, 'We want your help to work on the paper' . The initial paper was written at Merck, and then it was sent to me for editing. 117 Inasmuch as bogus ascriptions and deletions of authorship are falsifications of the research record-much like plagiarism is a falsification of authorship attributionand given that Merck executives clearly acted intentionally, RM can reasonably be concluded.

The final case study, also in bioscience, comes from the fitness industry. Here too, details became available only through years of litigation for which this author served as a paid consultant. The story adds value because it exposes management machinations in great detail, and because it describes somewhat less obvious forms of RM, aside from common data manipulation. Here, because the corporation owned the scientific journal publishing fabricated data, its 'corrections' of the research record actually re-entrenched the earlier fabrications.

The saga begins with an apparently simple story about graduate students run amok and the defendant organization's failure properly to address allegations of RM. The realities in CrossFit, Inc. v National Strength and Conditioning Association (NSCA) 118 eventually exposed multiple acts of RM undertaken at the highest levels of the corporation. 119 The NSCA is a research-based nonprofit corporation providing educational services, trainer certifications, and publications. 120 Of its several peer-reviewed scientific journals, premier is the Journal of Strength and Conditioning Research (JSCR). 121 CrossFit, Inc., a privately owned company, uses a somewhat unconventional approach to fitness, including high-intensity interval training, weightlifting, gymnastics, and calisthenics with an emphasis on everyday functional movement. In an intensely competitive market, CrossFit became very popular internationally, in military as well as civilian markets, posing a significant challenge for NSCA as both offer similar products-training seminars, coaching, certifications, etc. 122 -in the same markets.

In 2011 three Ohio State graduate students studying exercise physiology linked up with a local CrossFit gym as it was about to launch a 10-week fitness challenge. The students proposed taking before/after measurements of such parameters as body fat and VO 2 Max. The gym owner and participants agreed.

Finding that everyone who completed the program showed strong improvement regardless of age, gender, or prior fitness, the students wrote up their findings. With faculty advisor Dr. Steven Devor as last author, they submitted their manuscript to NSCA's JSCR (the 'Devor article'). Although the initial manuscript said nothing about injuries or risk, after multiple rounds of revision the authors stated in the final publication that, of 54 people who began the program, nine cited injury or overuse as their reason for not completing it. The authors emphasized that the benefits of CrossFit may not be worth the risk. 123 When the article appeared online in February 2013 the gym owner, startled to see injury data because to his knowledge nobody had been hurt, contacted CrossFit headquarters. CrossFit sought out the nonfinishing participants and, after contacting four of them, published online that none of the four had been injured or spoke with anyone about why they didn't complete the study. 124 CrossFit shared this finding with NSCA's Board of Directors, adding that the study coordinator believed the injury data were fabricated. Neither the NSCA nor the editor of JSCR responded.

In the fitness industry, a purportedly scientific claim that a particular workout approach has a high rate of injury can be economically devastating. Hence, with evidence the injury claims were fraudulent and with no response from NSCA, CrossFit sued in federal court in May 2014. 125 The following January, 10 of the 11 nonfinishers provided sworn declarations. Not one had been injured by the workouts, and not one had told anyone why they did not finish. 126 Although those declarations clearly showed the injury data were false, NSCA did nothing until September 2015, when JSCR published a brief Erratum. 127 At this point, the story depicts graduate students fabricating data and NSCA responding poorly to an allegation of RM. 128 The story soon became much more complicated.

Amidst both sides' preparations for federal trial, NSCA separately sued CrossFit in state court for defamation. 129 Discovery for that case revealed that NSCA had withheld numerous documents and committed perjury in depositions for the federal case. 130 The federal court granted significant sanctions against NSCA in May 2017, including a requirement that NSCA hire a neutral forensic evaluator to find whatever evidence had been withheld or destroyed. 131 That effort unearthed vast discovery abuse. Among many other offenses, NSCA had withheld nearly 280,000 relevant documents and lost or destroyed over 200 devices and 16 separate servers. 132 Granting in December 2019 the severest sanction-termination of the suit in Cross-Fit's favor-the Southern District of California was not amused. '[I]n twenty-five years on the bench, this is the first case that the Court has ever had that has gotten to this point . . . . [T] he severity and frequency of defendant's bad faith misconduct is as egregious as anything this Court has ever seen or read in any of the cases.' 133 More important here, final tranches of evidence arguably reveal research misconduct-by the definition-orchestrated at the highest levels of NSCA management: fabrication of the injury data, falsification in the Erratum, and falsification in the article's eventual Retraction.

Well before the Devor manuscript was submitted to JSCR, NSCA management discussed the need for science to combat CrossFit's competition. 'Crossfit is growing at an astronomical rate and is one of the hottest trends; if the NSCA is to be viewed as an industry leader information needs to be developed and released assessing the strengths of the programs, strategies to capitalize on what Crossfit does well, and arming NSCA-CPT's with scientific rationale as to the weaknesses of Crossfit.' 134 However, at the time no science whatever existed regarding CrossFit; hence, there could not possibly have been science showing its purported weaknesses.

In June 2012, the Devor authors submitted their manuscript, drawing robustly favorable conclusions regarding CrossFit and saying nothing about injuries because the authors had gathered no injury data.

Nevertheless, as JSCR's Editor-in-Chief (EiC) sent the manuscript to the Senior Articles Editor who would shepherd the piece through peer review, he made his expectations clear: 'so we put it in your good hands to get it ok, so a lot of context is needed for this, fit but at what cost etc. . . . . get some good reviewers to take a close look at this as catabolism will break you down but while fit are you in a catabolic state etc.' 135 As subsequent emails made clear, 'context' was the EiC's code-word for CrossFit's alleged propensity to cause injury ('fit but at what cost').

Two months later, although the peer reviewers said nothing about injury, the EiC warned authors the piece could not be published until review comments were taken into account: phones and "lost" more than one NSCA-owned computer (with NSCA's knowledge); NSCA was busily destroying documents and devices even while the forensic neutral was scampering to find them. CrossFit You also need to caution readers as to the context of your findings due to the fact many people do get injured doing these types of workouts. Typically a lack of general preparation is seen or people do to [sic] much to [sic] quickly and get hurt so how this was dealth [sic] with is of particular importance. 136 The revised manuscript still said nothing about injury. The EiC again emphasized the article could not be published unless it incorporates review comments, noting CrossFit supervision is 'not too well done in most CrossFit gyms for exercise technique, Y-Tube is full of this' . 137 The authors' third version stated that, of the 54 subjects who began, nine cited injury or overuse for quitting. The EiC demanded more. 'This [injury rate] is very important and needs to be emphasized,' because overuse injuries come from 'too much too soon' and participants must be 'properly screen [ed] .' 138 He added an entirely new criticism: 'CrossFit certification is not a certification that has met certification standards in Washington DC as American College of Sports Medicine (ACSM) and NSCA certifications and this needs again to be established as to the qualifications of the trainers in more detail as this is important.' 139 The final publication added a strong paragraph emphasizing that CrossFit's benefits may not be worth the risks. 140 NSCA senior management, already looking for science discrediting CrossFit, welcomed the article. A unique concern with any high-intensity training program such as HIPT or other similar programs is the risk of overuse injury. Despite a deliberate periodization and supervision of our Crossfit-based training program by certified fitness professionals, a notable percentage of our subjects (16 per cent) did not complete the training program and return for follow-up testing. Although peer-reviewed evidence of injury rates pertaining to high-intensity training programs is sparse, there are emerging reports of increased rates of musculoskeletal and metabolic injury in these programs (1). This may call into question the risk-benefit ratio for such extreme training programs, as the relatively small aerobic fitness and body composition improvements observed among individuals who are already considered to be 'above average' and 'well-above average' may not be worth the risk of injury and lost training time. Further work in this area is needed to explore how to best realize improvements to health without increasing risk above background levels associated with participation in any nonhigh intensity-based fitness regimen. This richer story suggests fabrication, directly committed by NSCA: 'making up data or results and recording or reporting them,' 142 at least knowingly if not outright intentionally. 143 Admittedly this was not literal gun-to-the-head coercion, and the student authors who invented injury data arguably likewise committed fabrication. But coercion appears real 144 and, as noted above, 145 data fabrication can arise via directing others, not just by making up numbers oneself. Students are vulnerable to Academia's wellknown 'publish or perish' mandate, and perhaps also naïvely saw guidance from longexperienced mentors. Every review piled on more: the students must 'caution readers [that] many people do get injured doing these types of workouts'; 'lack of general preparation'; supervision is 'not too well done . . . for exercise technique, Y-Tube is full of this'; the belated injury data are 'very important and need[] to be emphasized'; overuse injuries come from 'too much too soon'; and CrossFit trainers' certification is dubious. In the end it appears that NSCA, seeking 'science' with which to discredit its competitor and acting through its premier journal's EiC, pushed until it got what it wanted.

The 

Nine months after 10 of the 11 nonfinishers swore under oath they had neither been injured by CrossFit workouts nor told anyone their reasons for not finishing 148 NSCA, which owns JSCR, published an Erratum. not finishing, with only 2 mentioning injury or health conditions that prevented them from completing follow-up testing. In light of this information, injury rate should not be considered a factor in this study. This change does not affect the overall conclusion of the article. 149 This erratum falsifies the research record, first, by falsely reducing the injury data question to a squabble between gym owner and student-authors-'He said, they said.' Second, by stating that two people mentioned injury, NSCA falsely implies that the CrossFit workouts had caused their injuries-which they did not. 150 Third, it leaves intact the article's overall conclusion warning readers that CrossFit's benefits may not be worth its risks of injury, thereby tacitly reinforcing the injury claims rather than correcting the research record.

The erratum emerged from the highest levels of management. When allegations of data fabrication first emerged, NSCA management collectively agreed not to respond. 151 Later amidst litigation, NSCA's Publications Director created a two-person panel 152 that recommended the erratum. 153 NSCA knew the Erratum was misleading but chose not to amend it: 'we did not clarify that the injury and medical condition were not associated with the their workouts at the club[;] people are assuming that they were' . 154 . 23, 2015) . the public and harmed CrossFit'; 'the NSCA was aware of the misleading nature of the Erratum' . 155 Falsification is 'changing or omitting data or results such that the research is not accurately represented in the research record' . 156 Here, instead of plainly renouncing the fabricated injury data, the company's move to 'correct' the research record in its wholly owned scientific journal actually perpetuated the conclusion that CrossFit's risks may outweigh its benefits. Per in-house communications, senior management's conduct quite clearly appears intentional. RM can reasonably be inferred.

NSCA arguably also committed falsification in finally retracting the article. Over one year following the Erratum and nearly two years after sworn declarations proved no one had been injured, NSCA's Publications Director finally asked Ohio State University's Office of Responsible Research Practices (ORRP) 157 to investigate the Devor article. Focusing solely on the Devor authors' failure to obtain proper IRB oversight, the letter expressly presupposed the bogus injury data were actually truthful.

Of those [54 initial participants], . . . nine subjects cit[ed] overuse or injury for failing to complete the program and finish follow-up testing. It is our understanding that the overuse and injury published in the Article was an 'unanticipated event.' However, to the NSCA's knowledge, the Authors did not alert the participants of the article to the possibility of overuse or injury when obtaining their informed consent. 158 Unsurprisingly, the ensuing OSU investigation focused on the Devor authors' false statement of IRB approval, 159 having been expressly steered away from data fabrication.

NSCA's subsequent Retraction Notice 160 misled and obfuscated. The entire first paragraph emphasized IRB problems and left injury claims on the discussed in Part III, key criteria for determining who committed RM-individuals versus an organization-concern who initiated the research, whose purposes are being served, who has primary control over choice of project, research design, data collection, analysis, and reporting. Although this article does not purport to provide an exhaustive analysis, several major implications can be highlighted.

Initially, we should amend the current 'bad apples' view of RM, as something committed only by individual persons, and recognize a system-level perspective. Over 30 years ago health policy analysts began to recognize that a focus on individual miscreants could not adequately explain how errors happen in healthcare. Rather, we needed also to recognize the role of organizational culture, and structural factors such as 'latent errors' that can render adverse outcomes more likely. 170 Here, however, we add a dimension. Whereas errors in healthcare are ordinarily unintentional and a systems-approach looks for multifactorial explanations and solutions, 171 RM is indeed intentional, knowing or reckless. So we do still look for bad actors. But the foregoing says we must elevate the search for RM to encompass organizations, not just wayward individuals.

As adequately pursuing and penalizing organizations committing RM will require adjustments throughout civil and criminal law, a useful beginning comes from federal regulations regarding RM.

Current regulations governing federally funded research regard organizations exclusively as funding conduits and as supervisors to foster scientific integrity and investigate RM where indicated, not as entities capable of committing RM. 172 This must change, for several reasons.

First, the temptation to cheat has grown stronger with evolution of FDA approval processes. 173 In response to corporate and public demand for speedier approval of new products, numerous new tracks have been developed, including accelerated approval, breakthrough therapies, expanded access, fast-track, and priority review. 174 Particularly, in 1997 Congress permitted the FDA to approve a drug on the basis of just one research trial rather than the usual two or more. By 2015-17, nearly half of all new drug approvals were granted on the basis of just one pivotal trial. 175 Those trials, in turn, are now increasingly based on surrogate endpoints (e.g., particular lab values or tumor shrinkage) rather than actual clinical outcomes. 176 Although surrogate endpoints can be valuable and sometimes unavoidable, they do not always correlate with clinical benefit. 177 The results are concerning.

Scientists at biotech pioneer Amgen, for instance, reported in 2011 that they could confirm only 6 of 53 landmark studies in cancer biology. Researchers at pharma giant Bayer announced in 2012 that only 14 of 67 attempts to confirm claims in oncology, women's health and cardiovascular disease succeeded. Officials at Novartis and AstraZeneca told a recent cancer meeting that they encountered the same problem. 178 Similarly, although the FDA often requires postmarketing research as a condition of speedier approval, enforcement is limited. '[A] study of 614 postapproval requirements and commitments imposed in 2009 and 2010 found that by the end of 2015, 20% had not been started, 25% were delayed or ongoing, and only 54% were completed.' 179 Additionally, although the FDA Amendments Act of 2007 180 requires investigators to register their clinical trials and post data on a federal website within a year of a trial's completion, historically the latter provision has not been well-enforced. 181 As a result, even if hypothetically a manufacturer tests its new drug in 10 trials, of which nine produce dismal results, the company has been able to report only the lone favorable trial and keep quiet the overall negative results. That is now set to change, as a federal court ruled in February 2020 that the Department of Health and Human Services must begin enforcement. 182 However, time will tell how vigorously that is undertaken.

In the end, corporations face significant temptation to cut corners, given that: (i) a new drug or device might be approved on the basis of just one study; (ii) that lone study may not need any control group; (iii) the outcomes measured may not be actual improvements in human health and function; (iv) the requirement to post negative trial results may not be vigorously enforced; (v) FDA inspectiondetected RM may never be pursued; and (vi) any negative results or reviews might simply be blended into an overall document that may tend to mask doubts or scientific challenges. 183 Nevertheless, some ingredients for federal recognition of corporate RM are already in place. In suspected RM with PHS funds, for instance, a 'charge letter' 184 is sent to the 'respondent', defined as a 'person' . 185 Likewise, ORI can 'propose administrative actions against any person to the HHS' . 186 Both potentially apply to a corporate person, given that corporations as well as individuals are legally deemed persons.

Similarly, since 1977 the FDA has had authority to inspect investigators and sites doing research subject to FDA oversight, to ensure regulatory and ethical compliance. 187 Those inspections encompass organizations' as well as individuals' actions and can flag certain situations as 'Official Action Indicated'-essentially, RM. 188 Relevant offenses include: deviation from investigational plan; failure to maintain adequate/accurate source documentation; violations related to investigational product; failure to personally supervise the study; submission of false information to the FDA and sponsor; failure to communicate with sponsor. 189 These and other offenses could be more expressly labeled as potential research misconduct.

That said, clearer guidance is needed regarding who/what entity should pursue allegations or suspicions of RM. For federally funded research, the standard presumption is that the funding recipient will undertake RM investigations. 190 Plainly this is untenable where the organization itself is the likely perpetrator. Even so, ORI already in fact has authority to undertake direct investigations; 191 hence, it needs only to exercise that authority where an organization, rather than individual persons, is the target of suspicion. But that authority needs to be supplemented with specific provisions for recognizing and pursuing organization-level RM.

For FDA-governed research, better guidance for pursuing corporate-level RM, e.g., during site inspections, must likewise be developed. Although FDA visits several hundred research sites every year and periodically detects misconduct, often little or nothing is done other than sometimes to delete the corrupt data from analysis. RM is rarely pursued, nor are such findings noted in the resultant scientific literature. 192 This is seriously problematic, because knowingly permitting corrupt data to remain in a research record can more broadly corrupt the scientific literature-scientific misconduct. action, the HHS debarring official issues a notice of proposed debarment or suspension to the respondent as part of the charge letter. The charge letter includes the ORI findings of research misconduct and the basis for them and any HHS administrative actions. The letter also advises the respondent of the opportunity to contest the findings and administrative actions under Subpart E of this part. up fraud enforcement actions in this area. Enforcement is anticipated to focus, inter alia, on Covid-19 research. 198 FDA penalties for improper research conduct already apply to organizationsalbeit not under the express heading of 'RM' . They include warning letters, disqualifications, 199 rejections of data, injunctions, civil money penalties, and the like. 200 Analogous actions are likewise applied to CROs for such violations as failure to select qualified clinical investigators, failure to adequately monitor the study, failure to bring noncompliant investigators into compliance, and the like. 201 Where research misconduct is applicable, these failings should be expressly labeled, and penalized, as forms of RM.

In the end we may ask why it matters what we call yet another episode of fast-andloose-in-science-RM, scientific misconduct, or just bad science-and why does it matter whether we pin RM on organizations as well as on individuals.

Classification matters because individual research projects, and the data and methodology on which they are founded, are the bedrock of science. They are what we use to prove or disprove broader factual claims or to show that the scientific literature has been distorted. These individual research projects are the domain of research misconduct. If a high-quality research project is completed but withheld from publication for wrongful reasons we may have scientific misconduct, but there is still a chance it will be published one day. 202 But if the underlying research has itself been corrupted at the core, we have nowhere else to turn. 203 Now, more than ever, we need to be able to trust science. But now, more than ever, corruption of science is a problem that must be addressed forthrightly and precisely.

Reporting Mortality Findings in Trials of Rofecoxib for Alzheimer Disease or Cognitive Impairment: A Case Study Based on Documents from Rofecoxib Litigation

Report of Specific Cardiovascular Outcomes of the ADVANTAGE Trial, 144 Anns

Comparison of Upper Gastrointestinal Toxicity of Rofecoxib and Naproxen in Patients with Rheumatoid Arthritis, 353 New Eng

Response to Expression of Concern Regarding VIGOR Study, 354 New Eng

Response to Expression of Concern Regarding VIGOR Study, 354 New Eng

Expression of Concern Reaffirmed, 354 N. Eng

What Have We Learnt From Vioxx?

at 253 (quoting statement by Jeffrey Lisse in the New York Times (A. Berenson, Evidence in Vioxx Suits Shows Intervention by Merck Officials

14cv1191-JLS(KSC)

Order Granting in Part and Denying in Part Defendant's Motion to Amend the Scheduling Order to Allow Additional Expert Discovery, Crossfit, Inc. v. Nat'l Strength and Conditioning Ass'n, No.: 14cv1191-JLS(KSC)

This journal wishes to promote the publication of peer-reviewed manuscripts that add to our understanding of conditioning and sport through applied exercise and sport science

Crossfit-Based High Intensity Power Training Improves Maximal Aerobic Fitness and Body Composition

CrossFit Study' Fraud?

Nat'l Strength and Conditioning Ass'n, No.: 14cv1191 JLS (KSC)

Two people experienced injuries during the relevant time-period

See infra note 150 and accompanying text

Among other problems, NSCA and JSCR failed to follow internationally accepted standards for addressing an allegation of fabrication in a published manuscript

See also Discoverability of Identity of Peer Reviewers, Nat'l Strength and Conditioning Ass'n v. Glassman

Reflecting an earlier distinction between corporations' direct versus indirect liability, see supra note 66 and accompanying text, in these cases the corporation did not merely suborn perjury; arguably it directly committed perjury, as the persons in question were acting under the direction and in accordance with the overall plan of top management

Order Granting in Part and Denying in Part Motion for Sanctions, Crossfit, Inc. v. Nat'l Strength and Conditioning Ass'n No

NSCA's senior management had deleted numerous documents (e.g., the Executive director deleted 1378 documents, the Publications Director 1251); the EiC of JSCR had wiped several NSCA-owned cell

he authors have stated that the reasons for participants not completing follow-up testing, as reported in the article, were provided to the authors by the club owner. The club owner has denied that he provided this information

See supra note 76 and accompanying text

Order Granting in Part and Denying in Part Motion for Sanctions, Crossfit, Inc. v. Nat'l Strength and Conditioning Ass'n

) (order of statements is somewhat rearranged). The court encompassed related fraudulent 'science' in NSCA's other publications

Support of Plaintiff 's Motion for Partial Summary Judgment on the Element of Falsity, Exhibits D-N, Crossfit, Inc. v. Nat'l Strength and Conditioning Ass'n

Testimony to the U.S. Cong. Committee on Oversight and Government Reform on Examining Federal Administration of the Safe Drinking Water Act in Flint

Lucian Leape pointed out in 1994 that health system errors caused some 180,000 deaths per year, or 'the equivalent of three jumbo-jet crashes every [two] days' . Lucian L. Leape, Error in Medicine

The American Health Care System. The Movement for Improved Quality in Health Care, 340 New Eng

Is Health Care Ready for Six Sigma Quality

The Case for Using Industrial Quality Management Science in Health Care Organizations, 262 JAMA 2869

The Quality of Care: How Can it Be Assessed? 260 JAMA 1743

It looks at outliers as if they were significant when in fact they are most often highly unusual and sometimes random events. The strategy of quality improvement system design, by contrast, is to recognize that errors occur, to recognize that people work within systems, and to design the systems to do two things: (1) to make it difficult for individuals to make errors and (2) to make the whole system capable of 'absorbing' individuals' errors when they occur by identifying and correcting errors before they can be harmful. Even when a doctor has committed an error of judgment or skill

Adapting Mediation to Link Resolution of Medical Malpractice Disputes with Health Care Quality Improvement, 60 L. & Contemp

see also Policies-Regulations Q&A: Question and Answers-42 CFR Part 93

FDA Approval and Regulation of Pharmaceuticals

The proportion of new drugs supported by at least 2 pivotal trials decreased from 80.6 per cent in 1995-1997 to 52

When these measures do reliably predict how a patient feels, functions, or survives, their use to approve new drugs can accelerate the availability of useful medications. By contrast, a surrogate measure that does not predict actual patient benefit may accelerate approval of a drug

In their related article in Nature, Collins and Tabak note that irreproducibility is particularly a problem in preclinical studies

Approved in 2011 under accelerated approval regulations, the drug became best seller for its manufacturer, AMAG Pharmaceuticals. The required postparketing studies, not completed until 2019, showed the drug to be no more effective than placebo. Sixteen FDA panelists voted unanimously that Makena did not provide verifiable benefits for neonatal outcomes, and 13 of those 16 agreed that Makena did not provide substantial evidence of effectiveness. Nevertheless, Makena remains on the market. R. Rubin, Confirmatory Trial for Drug to Prevent Preterm Birth Finds No Benefit, So Why Is It Still Prescribed?

As of that date, these habitual violators had failed to report any results for 67% of their trials and averaged 268 days late for those and all trials that missed their deadlines . . . . In all 4768 trials Science checked, sponsors violated the reporting law more than 55% of the time. And in hundreds of cases where the sponsors got credit for reporting trial results

Transparency on Trial: Many Clinical Trial Results Aren't Posted Publicly, as U.S. Law Requiresand a Promised Crackdown has Fizzled

Important questions, cautions, and contrary information may be lost, to the potential detriment of public health. Id. The authors provide several examples to illustrate the potential loss of important

General statement of ORI authority. (c) HHS administrative actions. (1) In response to a research misconduct proceeding, ORI may propose administrative actions against any person to the HHS and, upon HHS approval and final action in accordance with this part, implement the actions. (2) ORI may propose to the HHS debarring official that a person be suspended or debarred from receiving Federal funds and may propose to

The FDA uses the same definitions, rules and regulations regarding research misconduct as other federal agencies; see U.S. Food & Drug Admin., 09-10-0020 FDA Records Related to Research Misconduct Proceedings

See sources cited supra note 172

Research Misconduct Identified by the US Food and Drug Administration Out of Sight, Out of Mind, Out of the Peer-Reviewed Literature

See also Department of Justice. Medical Doctor and Study Coordinator Sentenced to Prison in Scheme to Falsify Clinical Trial Data

Disqualification' , or A Notice of Initiation of Disqualification Proceedings and Opportunity to Explain (NIDPOE), for instance, refers to 'a clinical investigator has: a. Repeatedly or deliberately failed to comply with the requirements of 21 CFR 312, 511, and/or 812, as appropriate, and/or 21 CFR 50 or 56 or b. Repeatedly or deliberately submitted false information to FDA or to the sponsor in any required report

Compliance Program 7348

Office Regulatory Affairs, ORA-WIDE PROCEDURE

Compliance Program 7348

The study was completed and accepted for publication in The Journal of the American Medical Association. However, because the science showed the brand name product to be no better, the company blocked the paper from publication. Several years later the study-and the block-came to light and publication followed

$100m Payout After Drug Data Withheld. 388

We need to know as exactly as possible how and to what that statute applies, so that it will hit and not miss its intended mark. Indeed, the federal definition of

, 2017 by one of the authors that the study was not conducted under an IRB approved protocol for that study, as was stated in the article. Because the study was performed without proper IRB approval, the article has been retracted.This retraction follows the October 2015 erratum which stated that 'after the article was published, 10 of the 11 participants who did not complete the study have provided their reasons for not finishing, with only ACKNOWLEDGEMENTS The author is deeply grateful for the very helpful comments provided on earlier drafts by David Resnik, PhD, JD; Jon Merz, PhD, JD, MBA; Robert Jerry, JD; Terrence Ackerman, PhD; Peter Jacobson, JD, MPH; James Jones, PhD, and two anonymous reviewers for the Journal of Law and the Biosciences. The author also wishes to acknowledge the landmark legal results obtained by the litigation team led by Justin Nahama, Esq., lead counsel for CrossFit in CrossFit, Inc. v National Strength and Conditioning Association (discussed in Part IV), for which the author served as expert witness.

The author filed one report and one declaration as an expert on behalf of Plaintiff CrossFit, Inc. in Part IV's final case study.

Corporations, high-stakes biomedical research, and research misconduct • 31 . . . was [sic] false', 161 implying the injury data might still be true notwithstanding such a ruling. 162 The retraction thus abjures any straight statement that the injury data were completely fabricated and that no one was harmed by the workouts. Hence, we see falsification: 'changing or omitting data or results such that the research is not accurately represented in the research record' . 163 NSCA's retraction added to the research record and, by failing to state clearly that injury data were fabricated, it further entrenched discrepancies between the research and the record.Ample evidence attests that the concealment was committed 'knowingly, intentionally or recklessly' . 164 Indeed, NSCA's Publications director admitted in deposition that he knew the 2016 letter failed to notify OSU that NSCA possessed clear evidence the injury data were likely fraudulent. 165 Clearly, such dishonesty also represented a 'significant departure from accepted practices of the research community' . 166

Although research misconduct has not been attributed to organizations, 167 Penalties for RM must likewise be adjusted to penalize organizations expressly as perpetrators and to do so proportionately to the offense. Currently, the avowed purpose of PHS administrative sanctions for RM is just remedial. 193 And many PHS administrative actions inherently apply mainly to individuals, e.g., letters of reprimand, supervision requirements, exclusion from serving as advisor to PHS. 194 Hence, greater force, clarity, and options are needed for sanctioning corporate-level RM.Options are readily available. FCA 195 actions are already applied in the setting of RM, as noted above. 196 However, FCA claims have not, to date, expressly been tied to organizations as direct perpetrators of RM, a deficit that could be remedied by statute or by regulation, or by expressly alleging RM in complaints and identifying it in court rulings. Additionally, although many RM penalties for federally funded research focus on individuals, some could also apply to organizations, e.g., debarment from future federal funding. 197 Given ORI's history of finding only natural persons guilty of RM, however, sanctions applicable to corporate persons need to be made more explicit. Finally, although historically there have been few direct criminal prosecutions for RM, in late 2020 the Department of Justice's Consumer Protection Branch began stepping Fifty-seven published clinical trials were identified for which an FDA inspection of a trial site had found significant evidence of 1 or more of the following problems: falsification or submission of false information, 22 trials (39%); problems with adverse events reporting, 14 trials (25%); protocol violations, 42 trials (74%); inadequate or inaccurate recordkeeping, 35 trials (61%); failure to protect the safety of patients and/or issues with oversight or informed consent, 30 trials (53%); and violations not otherwise categorized, 20 trials (35%). Only 3 of the 78 publications (4%) that resulted from trials in which the FDA found significant violations mentioned the objectionable conditions or practices found during the inspection. No corrections, retractions, expressions of concern, or other comments acknowledging the key issues identified by the inspection were subsequently published . . . . When the FDA finds significant departures from good clinical practice, those findings are seldom reflected in the peer-reviewed literature, even when there is evidence of data fabrication or other forms of research misconduct.